British Journal of Surgery

doi: 10.1002/bjs.9717pmid: N/A

Article PDF first page preview Close This content is only available as a PDF. © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd

Niederle, B

doi: 10.1002/bjs.9652pmid: 25296738

Medullary thyroid cancer (MTC) develops from the parafollicular or C cells of the thyroid. In the absence of alternative therapy, surgery is the well accepted treatment of choice for MTC1. Up to 20 per cent of patients with sporadic MTC and nearly all patients (97 per cent) with inherited MTC have bilateral or multifocal thyroid disease1. MTC spreads to the central neck nodes at a very early stage of the disease1. The biological tumour characteristics of MTC are well known. As published in 20092, there are clear recommendations regarding the extent of surgery for localized, regional ‘clinically apparent’ MTC (T1b–4 N0/1) and ‘micro’ MTC (10 mm or smaller; T1a N0/1) as well as for patients with distant metastasis. However, in a retrospective North American cohort study3 of patients with MTC, 41 per cent of patients did not receive appropriate surgical therapy because the malignant disease was unknown before surgery. Patients who received surgery that was not in line with recommendations were shown to have shorter survival than those who had surgery according to recommendations3. MTC has a potential for aggressive behaviour independent of size1. Delayed diagnosis and inadequate initial treatment influence the prognosis4. A raised serum calcitonin (CT) level is a sensitive tumour marker for C-cell disease1. A raised basal CT concentration is confirmed by a second determination after exclusion of potential interfering conditions (such as decreased kidney function)1,5. Diagnostic accuracy is increased by pentagastrin or calcium stimulation, which is necessary in patients with a mild or moderate increase in CT level. There is no absolute threshold value for basal CT that differentiates benign from malignant C-cell disease and rules out ‘paraneoplastic’ CT production by various other neuroendocrine tumours1,5. The interpretation of basal and stimulated CT levels should be done with caution owing to differences in sex and in the various available CT assays5. The extent of experience is considerable after applying pentagastrin stimulation1. However, pentagastrin is no longer available for clinical practice. Up to now, experience with use of calcium for stimulation has been limited5. However, as soon as stimulated CT peaks of over 100 pg/ml are documented, total thyroidectomy and bilateral central neck (level VI) dissection should be discussed with the patient. Should patients refuse surgery, close follow-up examinations are mandatory. The available evidence for biochemical screening programmes and their clinical and therapeutic consequences is still limited. Several prospective studies in Europe have recommended routine CT screening of patients in the initial diagnostic evaluation of unselected thyroid nodular disease. These investigations have demonstrated that routine measurement of CT allows the detection of unsuspected MTC with a frequency of one in 200–300 thyroid nodules, and that routine CT screening improves outcome. American authors argue against the European recommendations. As a compromise, the current joint American and European guidelines6 give recommendations neither for nor against routine CT screening. Instead the guidelines state that measurement of basal CT level may be a useful test in the initial evaluation of thyroid nodules and before thyroid surgery for nodular goitre. However, CT measurement is mandatory in patients with a family history or clinical suspicion of MTC or multiple endocrine neoplasia type 2. Arguments against CT screening programmes include the overall low prevalence of occult C-cell malignancy (average 0·14 per cent7), and the low proportion of MTCs detected after surgery in unselected individuals with thyroid nodules and raised CT level (average 0·34 per cent8). Further arguments against CT screening are cost, the difficulty in acquiring pentagastrin, and the current limited experience with calcium stimulation testing5. The cost of CT screening in patients with thyroid nodules independent of size has to be balanced against that of multiple surgical procedures, and arguably worthwhile, but nonetheless frequently used, radiotherapy, chemotherapy and targeted therapy in patients with metastatic disease. The psychological effects of a chronic disease on the patient also need to be considered. The cost-effectiveness of CT screening has been documented recently9. A strong argument for CT screening is that approximately 40 per cent of MTCs detected by surgery in unselected individuals with thyroid nodules and raised (basal or stimulated) level of CT are classified as pT1a or microcarcinoma10. Systematic lymph node dissection reveals positive nodes in up to 12 per cent of patients1. If the underlying malignant C-cell disease is identified before operation, patients with mildly raised CT levels can be cured by meticulous central neck dissection alone1. In CT screening of the thyroid, women are more commonly affected by MTC than men. Therefore, a more liberal indication for surgery is recommended in women with mildly raised CT levels1. Lateral modified neck dissection is recommended only in selected patients with increased basal and stimulated CT levels and with desmoplasia-positive tumours1. Desmoplasia is verified by intraoperative frozen-section analysis1. It has been shown that CT-guided surgery also reduces the extent of operation and surgical morbidity in patients with sporadic C-cell disease1. Routine CT measurements show a higher sensitivity than ultrasound imaging and fine-needle aspiration cytology, especially in diagnosing microcarcinoma5. Even the modification of measuring CT in wash-out fluids from fine-needle aspirates may not improve the initial diagnosis of MTC during the evaluation of thyroid nodules. In microcarcinoma, frozen-section analysis has been shown to diagnose MTC in only 35 per cent of patients1. An early diagnosis of MTC is of crucial importance for prognosis. To apply biochemical CT screening before operation is a decision to be made by the endocrine surgeon. Knowledge of CT levels may help in careful planning of the extent of surgery. Only an adequate surgical procedure in the early phase of the disease can offer the chance of cure. Disclosure The author declares no conflict of interest. References 1 Scheuba C , Kaserer K, Moritz A, Drosten R, Vierhapper H, Bieglmayer C et al. Sporadic hypercalcitoninemia: clinical and therapeutic consequences . Endocr Relat Cancer 2009 ; 16 : 243 – 253 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Kloos RT , Eng C, Evans DB, Francis GL, Gagel RF, Gharib H et al. ; American Thyroid Association Guidelines Task Force. Medullary thyroid cancer: management guidelines of the American Thyroid Association . Thyroid 2009 ; 19 : 565 – 612 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Panigrahi B , Roman SA, Sosa JA. Medullary thyroid cancer: are practice patterns in the United States discordant from American Thyroid Association guidelines? Ann Surg Oncol 2010 ; 17 : 1490 – 1498 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Karga H , Giagourta I, Papaioannou G, Doumouchtsis K, Polymeris A, Thanou S et al. Changes in risk factors and tumor node metastasis stage of sporadic medullary thyroid carcinoma over 41 years, before and after the routine measurements of serum calcitonin . Metabolism 2011 ; 60 : 604 – 608 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Trimboli P , Giovanella L, Crescenzi A, Romanelli F, Valabrega S, Spriano G et al. Medullary thyroid cancer diagnosis: an appraisal . Head Neck 2014 ; 36 : 1216 – 1223 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Gharib H , Papini E, Paschke R, Duick DS, Valcavi R, Hegedüs L et al. ; AACE/AMA/ETA Task Force on Thyroid Nodules . American Association of Clinical Endocrinologists, Associazione Medici Endocrinologi, and European Thyroid Association medical guidelines for clinical practice for the diagnosis and management of thyroid nodules: executive summary of recommendations . Endocr Pract 2010 ; 16 : 468 – 475 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Valle LA , Kloos RT. The prevalence of occult medullary thyroid carcinoma at autopsy . J Clin Endocrinol Metab 2011 ; 96 : E109 – E113 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Ahmed SR , Ball DW. Clinical review: incidentally discovered medullary thyroid cancer: diagnostic strategies and treatment . J Clin Endocrinol Metab 2011 ; 96 : 1237 – 1245 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Cheung K , Roman SA, Wang TS, Walker HD, Sosa JA. Calcitonin measurement in the evaluation of thyroid nodules in the United States: a cost-effectiveness and decision analysis . J Clin Endocrinol Metab 2008 ; 93 : 2173 – 2180 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Daniels GH . Screening for medullary thyroid carcinoma with serum calcitonin measurements in patients with thyroid nodules in the United States and Canada . Thyroid 2011 ; 21 : 1199 – 1207 . Google Scholar Crossref Search ADS PubMed WorldCat © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd

Glass, G E; Murphy, G F; Esmaeili, A; Lai, L-M; Nanchahal, J

doi: 10.1002/bjs.9636pmid: 25294112

Abstract Background Negative-pressure wound therapy (NPWT) promotes angiogenesis and granulation, in part by strain-induced production of growth factors and cytokines. As their expression profiles are being unravelled, it is pertinent to consider the mode of action of NPWT at the molecular level. Methods MEDLINE (January 1997 to present), Embase (January 1997 to present), PubMed (no time limit), the Cochrane Database of Systematic Reviews and the Cochrane Controlled Trials Register were searched for articles that evaluated the influence of NPWT on growth factor expression quantitatively. Results Sixteen studies met the inclusion criteria. Tumour necrosis factor expression was reduced in acute and chronic wounds, whereas expression of interleukin (IL) 1β was reduced in acute wounds only. Systemic IL-10 and local IL-8 expression were increased by NPWT. Expression of vascular endothelial growth factor, fibroblast growth factor 2, transforming growth factor β and platelet-derived growth factor was increased, consistent with mechanoreceptor and chemoreceptor transduction in response to stress and hypoxia. Matrix metalloproteinase-1, -2, -9 and -13 expression was reduced but there was no effect on their enzymatic inhibitor, tissue inhibitor of metalloproteinase 1. Conclusion Cytokine and growth factor expression profiles under NPWT suggest that promotion of wound healing occurs by modulation of cytokines to an anti-inflammatory profile, and mechanoreceptor and chemoreceptor-mediated cell signalling, culminating in angiogenesis, extracellular matrix remodelling and deposition of granulation tissue. This provides a molecular basis for understanding NPWT. Introduction Although negative-pressure wound therapy (NPWT) continues to find new therapeutic applications, the physiological effect of topical negative pressure (TNP) on the surrounding tissues remains the subject of debate. The removal of protease-rich wound fluid and reversal of interstitial oedema, with the consequent reduction in capillary afterload, has been appreciated for some time1. However, it has also been postulated that microdeformational forces2,3 and/or a hypoxic gradient within the microcirculatory environment4 may promote cell recruitment, proliferation and differentiation, culminating in a final common pathway of neovascularization and granulation tissue formation5. These processes are governed by differential gene expression at the wound bed under NPWT6–9, with resulting downstream growth factor, cytokine and chemokine production5,9–18. They may be manipulated by modifying clinical variables such as pressure magnitude and cyclical application. New tissue growth is desirable only in as much as it aids wound healing. Excessive, friable granulation tissue remains the unwanted consequence of NPWT in many therapeutic scenarios as it results in thick, poorly compliant scar tissue19,20. Although the molecular mechanism of action of NPWT remains undefined, refinements to achieve the final desired outcome have been determined largely by empirical use and observation. By summarizing the currently available data, potential avenues for further research may be identified with the aim of defining NPWT at a molecular level. It may thus be possible to improve on a powerful technique that, at present, yields variable results21,22. The aim of this review was to establish how growth factor, cytokine and matrix metalloproteinase (MMP) expression is altered by NPWT compared with conventional dressings. Methods Search strategy PubMed (no date limit), OVID MEDLINE (January 1997 to November 2013), Embase (January 1997 to November 2013), the Cochrane Database of Systematic Reviews and the Cochrane Controlled Trials Register (searched 29 November 2013) were searched using medical subject heading (MeSH) terms and free-text terms. Articles were searched from 1997 as this year corresponds to the first description of NPWT as used in contemporary practice. The search strategy and terms used are shown in Table S1 (supporting information). Articles were cross-referenced to identify others of interest. Additionally, the online trials registers ClinicalTrials.gov and the national research register were scrutinized for completed, discontinued and ongoing trials relating to NPWT and growth factor or cytokine use or expression. The search strategy was performed in accordance with the Cochrane Highly Sensitive Search Strategy guideline in the Cochrane Handbook for Systematic Reviews of Interventions23. The review is reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement24. Inclusion criteria Studies that quantitatively or semiquantitatively evaluated the influence of NPWT on growth factor, cytokine or MMP expression by subjacent cells or tissues and incorporated appropriate control groups were included. In vivo and in vitro studies were included, as were animal models and human trials. Studies were not limited by language or publication status. The methodological quality of included studies was assessed using the PRISMA criteria. Exclusion criteria Publications before the contemporary description of NPWT (pre-1997), descriptive studies relying purely on qualitative data and those lacking a control intervention were excluded. Outcome measures The primary outcome measures were growth factor, cytokine and MMP expression profiles under the influence of NPWT relative to controls. Secondary outcome measures included evidence for differential expression profiles of any other molecule not classed as a cytokine, growth factor or MMP. Study selection Two authors independently performed an abstract review. Full-text review was undertaken for studies that met the inclusion criteria and additional studies that possibly met the inclusion criteria. Discrepancies were resolved by consensus. Data extraction Data were retrieved from all full-text articles accepted for final analysis by two authors. Five Chinese-language and one German paper were translated. A standard pro forma was used to record data variables including the model (human clinical study, animal wound model or in vitro cell study), species, type of wound (acute, chronic or surgical), sample size, magnitude of negative pressure, control intervention, method of analysis, time points, duration of intervention and quantitative outcomes. Owing to diversity in clinical and methodological aspects of the included studies, variations in outcome measures and the addition of semiquantitative outcome data, it was not possible to perform a meta-analysis. Results A total of 16 papers5,9–18,25–29 were accepted for full analysis and formed the basis of the review (Fig. 1). Of these, six10,13–14,17,25–26 reported the influence of NPWT on proinflammatory and anti-inflammatory cytokine expression, 115,9,12,14–15,18,25–29 reported growth factor expression and three11,16–17 MMP expression. Thirteen studies5,9–16,18,25–26,28 used NPWT at –125 mmHg (–16·625 kPa) continuously during the experiment, whereas single studies used –15·96 kPa (–120 mmHg) continuously27 and –80 mmHg continuously29. Two studies5,18 evaluated a cyclical regimen in addition to continuous therapy. One study17 did not specify the pressure employed during TNP. The wound filler material was polyurethane foam in all cases. The methods used to quantify results varied but reverse transcription (RT)–PCR, enzyme-linked immunosorbent assay (ELISA) and immunohistochemistry-aided cell count methods were commonly used. Key characteristics of included studies are summarized in Table 1 (see Table S2, supporting information, for risk of bias assessment). Fig. 1 Open in new tabDownload slide Literature search and study selection. NPWT, negative-pressure wound therapy Table 1 Characteristics of included studies Reference . Participants . Method . Intervention . Outcomes . Notes . 17 Adult humans; grade III–IV pressure sores (n = 8) Fluid aspirate from NPWT tested for cytokines and growth factors on day 0, 1, 3 and 7 NPWT only Effluent concentrations of TNF-α, IL-1β, MMP-2, MMP-3 and TIMP-1 measured by ELISA Magnitude of TNP unclear 10 Adult humans; free muscle transfer for extremity defects (n = 30) Tissue biopsy day 5 NPWT versus petroleum gauze dressings (n = 15 each) Expression of TNF-α and IL-1β in muscle measured by RT–PCR Methodologically robust (randomized and blinded) 25 Murine chronic wound model (n = 40) 1 × 1-cm dorsal wounds colonized with strain of Pseudomonas. Tissue sample at 24 h Dressings ± NPWT for 24 h RT–PCR for IL-1β and VEGF Also reported reduction in bacterial count (via fluorescence) at 24 h and significant increase in vascularity in experimental group 14 Adult humans with traumatic wounds (n = 32) Serial analysis of wound effluent and serum at dressing changes NPWT versus Epigard™ (n = 16 each) IL-6, IL-8, VEGF and FGF-2 expression quantified by ELISA Not randomized or blinded 13 Porcine wound model (n = 12) Serial analysis of serum at 0, 1, 2, 3 and 4 h NPWT versus saline-soaked gauze TGF-β1, IL-6, IL-8 and IL-10 expression quantified by ELISA Experiment terminated at 4 h 5 Murine wound model (n = 50) Tissue biopsies of wound bed at 6 h, 3 and 7 days Continuous versus cyclical NPWT versus Tegaderm™ dressing VEGF dimers and HIF-1α at wound bed Correlated results with wound healing and angiogenesis 12 Rat wound model (n =?) Killed on day 3, 5 and 7 postop. NPWT versus dressings only VEGF and FGF-2 expression quantified by western blot 9 In vitro model of murine fibroblasts Cells harvested at 24, 48 and 72 h Cells cultured under PU foam ± NPWT FGF-2, TGF-β1, α-SMA and type 1 collagen α-1 expression quantified by RT–PCR 15 In vitro model of human fibroblasts in 3D fibrin matrix culture Cell supernatant harvested after 48 h Cells cultured under occlusive dressing ± PU foam and NPWT PDGF expression quantified by Bio-Plex® cytokine bead assay and TGF-β quantified by ELISA Lacked control for PU foam dressing only 18 (Diabetic) murine wound model (n = 40) Killed on day 7 and 14 postop. Continuous versus cyclical NPWT versus PU foam only versus occlusive dressings CGRP, substance P and NGF expression quantified by RT-PCR Fully controlled Only two time points observed 11 Human chronic wounds (n = 3) Samples at day 0, week 1 and week 2 Wound + NPWT ± PU foam filler MMP-9 and MMP-2 semiquantified by gel electrophoresis MMP-9/NGAL and MMP-2 latent also assessed 16 Human chronic wounds (n = 5) Samples at day 0 and at day 1, 4, 7 NPWT only MMP-1, MMP-2 and MMP-13 expression quantified by RT–PCR 26 Adult humans with traumatic wounds (n = 21) Serial analysis of wound effluent and serum at dressing changes NPWT (n = 8) versus Epigard™ (n = 13) IL-6, IL-8, IL-10, VEGF and FGF-2 expression quantified by ELISA Heterogeneous wounds and first measurement taken 2 days after first debridement, hence no early data 27 Adult humans with chronic wounds (failure to heal at 1 month) (n = 5; 4 women, 1 man) Tissue sample × 1 from edge of wounds before and on day 1, 3, 5 and 7 after intervention NPWT only VEGF expression by IHC quantified by counting method Wound aetiology unknown. Wide variation in size and patient age. One biopsy, used 5 sections, and average of 5 sections 28 Rabbit wound model (n = 12) 5 × 5-cm dorsal wounds. Tissue samples at 0·5, 1, 6 and 12 h + 1, 3, 5 and 7 days Dressings ± NPWT HIF-1α measured by RT–PCR, and VEGF by ELISA of tissue homogenate 29 Rat wound model (n = 80) 2 × 2-cm dorsal wounds. Tissue samples at day 1, 3, 6, 9, 12 and 15 Dressings ± NPWT VEGF expression by IHC quantified by counting method Spinal injury groups (innervation or denervation) Reference . Participants . Method . Intervention . Outcomes . Notes . 17 Adult humans; grade III–IV pressure sores (n = 8) Fluid aspirate from NPWT tested for cytokines and growth factors on day 0, 1, 3 and 7 NPWT only Effluent concentrations of TNF-α, IL-1β, MMP-2, MMP-3 and TIMP-1 measured by ELISA Magnitude of TNP unclear 10 Adult humans; free muscle transfer for extremity defects (n = 30) Tissue biopsy day 5 NPWT versus petroleum gauze dressings (n = 15 each) Expression of TNF-α and IL-1β in muscle measured by RT–PCR Methodologically robust (randomized and blinded) 25 Murine chronic wound model (n = 40) 1 × 1-cm dorsal wounds colonized with strain of Pseudomonas. Tissue sample at 24 h Dressings ± NPWT for 24 h RT–PCR for IL-1β and VEGF Also reported reduction in bacterial count (via fluorescence) at 24 h and significant increase in vascularity in experimental group 14 Adult humans with traumatic wounds (n = 32) Serial analysis of wound effluent and serum at dressing changes NPWT versus Epigard™ (n = 16 each) IL-6, IL-8, VEGF and FGF-2 expression quantified by ELISA Not randomized or blinded 13 Porcine wound model (n = 12) Serial analysis of serum at 0, 1, 2, 3 and 4 h NPWT versus saline-soaked gauze TGF-β1, IL-6, IL-8 and IL-10 expression quantified by ELISA Experiment terminated at 4 h 5 Murine wound model (n = 50) Tissue biopsies of wound bed at 6 h, 3 and 7 days Continuous versus cyclical NPWT versus Tegaderm™ dressing VEGF dimers and HIF-1α at wound bed Correlated results with wound healing and angiogenesis 12 Rat wound model (n =?) Killed on day 3, 5 and 7 postop. NPWT versus dressings only VEGF and FGF-2 expression quantified by western blot 9 In vitro model of murine fibroblasts Cells harvested at 24, 48 and 72 h Cells cultured under PU foam ± NPWT FGF-2, TGF-β1, α-SMA and type 1 collagen α-1 expression quantified by RT–PCR 15 In vitro model of human fibroblasts in 3D fibrin matrix culture Cell supernatant harvested after 48 h Cells cultured under occlusive dressing ± PU foam and NPWT PDGF expression quantified by Bio-Plex® cytokine bead assay and TGF-β quantified by ELISA Lacked control for PU foam dressing only 18 (Diabetic) murine wound model (n = 40) Killed on day 7 and 14 postop. Continuous versus cyclical NPWT versus PU foam only versus occlusive dressings CGRP, substance P and NGF expression quantified by RT-PCR Fully controlled Only two time points observed 11 Human chronic wounds (n = 3) Samples at day 0, week 1 and week 2 Wound + NPWT ± PU foam filler MMP-9 and MMP-2 semiquantified by gel electrophoresis MMP-9/NGAL and MMP-2 latent also assessed 16 Human chronic wounds (n = 5) Samples at day 0 and at day 1, 4, 7 NPWT only MMP-1, MMP-2 and MMP-13 expression quantified by RT–PCR 26 Adult humans with traumatic wounds (n = 21) Serial analysis of wound effluent and serum at dressing changes NPWT (n = 8) versus Epigard™ (n = 13) IL-6, IL-8, IL-10, VEGF and FGF-2 expression quantified by ELISA Heterogeneous wounds and first measurement taken 2 days after first debridement, hence no early data 27 Adult humans with chronic wounds (failure to heal at 1 month) (n = 5; 4 women, 1 man) Tissue sample × 1 from edge of wounds before and on day 1, 3, 5 and 7 after intervention NPWT only VEGF expression by IHC quantified by counting method Wound aetiology unknown. Wide variation in size and patient age. One biopsy, used 5 sections, and average of 5 sections 28 Rabbit wound model (n = 12) 5 × 5-cm dorsal wounds. Tissue samples at 0·5, 1, 6 and 12 h + 1, 3, 5 and 7 days Dressings ± NPWT HIF-1α measured by RT–PCR, and VEGF by ELISA of tissue homogenate 29 Rat wound model (n = 80) 2 × 2-cm dorsal wounds. Tissue samples at day 1, 3, 6, 9, 12 and 15 Dressings ± NPWT VEGF expression by IHC quantified by counting method Spinal injury groups (innervation or denervation) NPWT, negative-pressure wound therapy; TNF, tumour necrosis factor; IL, interleukin; MMP, matrix metalloproteinase; TIMP, tissue inhibitor of metalloproteinase; ELISA, enzyme-linked immunosorbent assay; TNP, topical negative pressure; RT, reverse transcription; VEGF, vascular endothelial growth factor; FGF, fibroblast growth factor; TGF, transforming growth factor; HIF, hypoxia-inducible factor; PU, polyurethane; SMA, smooth muscle actin; 3D, three-dimensional; PDGF, platelet-derived growth factor; CGRP, calcitonin gene-related peptide; NGF, nerve growth factor; NGAL, neutrophil gelatinase-associated lipocalin; IHC, immunohistochemistry. Epigard™ (CONSEPT, Eschbach, Germany); Tegaderm™ (3 M, St Paul, Minnesota, USA); Bio-Plex® (Bio-Rad, Hercules, California, USA). Open in new tab Table 1 Characteristics of included studies Reference . Participants . Method . Intervention . Outcomes . Notes . 17 Adult humans; grade III–IV pressure sores (n = 8) Fluid aspirate from NPWT tested for cytokines and growth factors on day 0, 1, 3 and 7 NPWT only Effluent concentrations of TNF-α, IL-1β, MMP-2, MMP-3 and TIMP-1 measured by ELISA Magnitude of TNP unclear 10 Adult humans; free muscle transfer for extremity defects (n = 30) Tissue biopsy day 5 NPWT versus petroleum gauze dressings (n = 15 each) Expression of TNF-α and IL-1β in muscle measured by RT–PCR Methodologically robust (randomized and blinded) 25 Murine chronic wound model (n = 40) 1 × 1-cm dorsal wounds colonized with strain of Pseudomonas. Tissue sample at 24 h Dressings ± NPWT for 24 h RT–PCR for IL-1β and VEGF Also reported reduction in bacterial count (via fluorescence) at 24 h and significant increase in vascularity in experimental group 14 Adult humans with traumatic wounds (n = 32) Serial analysis of wound effluent and serum at dressing changes NPWT versus Epigard™ (n = 16 each) IL-6, IL-8, VEGF and FGF-2 expression quantified by ELISA Not randomized or blinded 13 Porcine wound model (n = 12) Serial analysis of serum at 0, 1, 2, 3 and 4 h NPWT versus saline-soaked gauze TGF-β1, IL-6, IL-8 and IL-10 expression quantified by ELISA Experiment terminated at 4 h 5 Murine wound model (n = 50) Tissue biopsies of wound bed at 6 h, 3 and 7 days Continuous versus cyclical NPWT versus Tegaderm™ dressing VEGF dimers and HIF-1α at wound bed Correlated results with wound healing and angiogenesis 12 Rat wound model (n =?) Killed on day 3, 5 and 7 postop. NPWT versus dressings only VEGF and FGF-2 expression quantified by western blot 9 In vitro model of murine fibroblasts Cells harvested at 24, 48 and 72 h Cells cultured under PU foam ± NPWT FGF-2, TGF-β1, α-SMA and type 1 collagen α-1 expression quantified by RT–PCR 15 In vitro model of human fibroblasts in 3D fibrin matrix culture Cell supernatant harvested after 48 h Cells cultured under occlusive dressing ± PU foam and NPWT PDGF expression quantified by Bio-Plex® cytokine bead assay and TGF-β quantified by ELISA Lacked control for PU foam dressing only 18 (Diabetic) murine wound model (n = 40) Killed on day 7 and 14 postop. Continuous versus cyclical NPWT versus PU foam only versus occlusive dressings CGRP, substance P and NGF expression quantified by RT-PCR Fully controlled Only two time points observed 11 Human chronic wounds (n = 3) Samples at day 0, week 1 and week 2 Wound + NPWT ± PU foam filler MMP-9 and MMP-2 semiquantified by gel electrophoresis MMP-9/NGAL and MMP-2 latent also assessed 16 Human chronic wounds (n = 5) Samples at day 0 and at day 1, 4, 7 NPWT only MMP-1, MMP-2 and MMP-13 expression quantified by RT–PCR 26 Adult humans with traumatic wounds (n = 21) Serial analysis of wound effluent and serum at dressing changes NPWT (n = 8) versus Epigard™ (n = 13) IL-6, IL-8, IL-10, VEGF and FGF-2 expression quantified by ELISA Heterogeneous wounds and first measurement taken 2 days after first debridement, hence no early data 27 Adult humans with chronic wounds (failure to heal at 1 month) (n = 5; 4 women, 1 man) Tissue sample × 1 from edge of wounds before and on day 1, 3, 5 and 7 after intervention NPWT only VEGF expression by IHC quantified by counting method Wound aetiology unknown. Wide variation in size and patient age. One biopsy, used 5 sections, and average of 5 sections 28 Rabbit wound model (n = 12) 5 × 5-cm dorsal wounds. Tissue samples at 0·5, 1, 6 and 12 h + 1, 3, 5 and 7 days Dressings ± NPWT HIF-1α measured by RT–PCR, and VEGF by ELISA of tissue homogenate 29 Rat wound model (n = 80) 2 × 2-cm dorsal wounds. Tissue samples at day 1, 3, 6, 9, 12 and 15 Dressings ± NPWT VEGF expression by IHC quantified by counting method Spinal injury groups (innervation or denervation) Reference . Participants . Method . Intervention . Outcomes . Notes . 17 Adult humans; grade III–IV pressure sores (n = 8) Fluid aspirate from NPWT tested for cytokines and growth factors on day 0, 1, 3 and 7 NPWT only Effluent concentrations of TNF-α, IL-1β, MMP-2, MMP-3 and TIMP-1 measured by ELISA Magnitude of TNP unclear 10 Adult humans; free muscle transfer for extremity defects (n = 30) Tissue biopsy day 5 NPWT versus petroleum gauze dressings (n = 15 each) Expression of TNF-α and IL-1β in muscle measured by RT–PCR Methodologically robust (randomized and blinded) 25 Murine chronic wound model (n = 40) 1 × 1-cm dorsal wounds colonized with strain of Pseudomonas. Tissue sample at 24 h Dressings ± NPWT for 24 h RT–PCR for IL-1β and VEGF Also reported reduction in bacterial count (via fluorescence) at 24 h and significant increase in vascularity in experimental group 14 Adult humans with traumatic wounds (n = 32) Serial analysis of wound effluent and serum at dressing changes NPWT versus Epigard™ (n = 16 each) IL-6, IL-8, VEGF and FGF-2 expression quantified by ELISA Not randomized or blinded 13 Porcine wound model (n = 12) Serial analysis of serum at 0, 1, 2, 3 and 4 h NPWT versus saline-soaked gauze TGF-β1, IL-6, IL-8 and IL-10 expression quantified by ELISA Experiment terminated at 4 h 5 Murine wound model (n = 50) Tissue biopsies of wound bed at 6 h, 3 and 7 days Continuous versus cyclical NPWT versus Tegaderm™ dressing VEGF dimers and HIF-1α at wound bed Correlated results with wound healing and angiogenesis 12 Rat wound model (n =?) Killed on day 3, 5 and 7 postop. NPWT versus dressings only VEGF and FGF-2 expression quantified by western blot 9 In vitro model of murine fibroblasts Cells harvested at 24, 48 and 72 h Cells cultured under PU foam ± NPWT FGF-2, TGF-β1, α-SMA and type 1 collagen α-1 expression quantified by RT–PCR 15 In vitro model of human fibroblasts in 3D fibrin matrix culture Cell supernatant harvested after 48 h Cells cultured under occlusive dressing ± PU foam and NPWT PDGF expression quantified by Bio-Plex® cytokine bead assay and TGF-β quantified by ELISA Lacked control for PU foam dressing only 18 (Diabetic) murine wound model (n = 40) Killed on day 7 and 14 postop. Continuous versus cyclical NPWT versus PU foam only versus occlusive dressings CGRP, substance P and NGF expression quantified by RT-PCR Fully controlled Only two time points observed 11 Human chronic wounds (n = 3) Samples at day 0, week 1 and week 2 Wound + NPWT ± PU foam filler MMP-9 and MMP-2 semiquantified by gel electrophoresis MMP-9/NGAL and MMP-2 latent also assessed 16 Human chronic wounds (n = 5) Samples at day 0 and at day 1, 4, 7 NPWT only MMP-1, MMP-2 and MMP-13 expression quantified by RT–PCR 26 Adult humans with traumatic wounds (n = 21) Serial analysis of wound effluent and serum at dressing changes NPWT (n = 8) versus Epigard™ (n = 13) IL-6, IL-8, IL-10, VEGF and FGF-2 expression quantified by ELISA Heterogeneous wounds and first measurement taken 2 days after first debridement, hence no early data 27 Adult humans with chronic wounds (failure to heal at 1 month) (n = 5; 4 women, 1 man) Tissue sample × 1 from edge of wounds before and on day 1, 3, 5 and 7 after intervention NPWT only VEGF expression by IHC quantified by counting method Wound aetiology unknown. Wide variation in size and patient age. One biopsy, used 5 sections, and average of 5 sections 28 Rabbit wound model (n = 12) 5 × 5-cm dorsal wounds. Tissue samples at 0·5, 1, 6 and 12 h + 1, 3, 5 and 7 days Dressings ± NPWT HIF-1α measured by RT–PCR, and VEGF by ELISA of tissue homogenate 29 Rat wound model (n = 80) 2 × 2-cm dorsal wounds. Tissue samples at day 1, 3, 6, 9, 12 and 15 Dressings ± NPWT VEGF expression by IHC quantified by counting method Spinal injury groups (innervation or denervation) NPWT, negative-pressure wound therapy; TNF, tumour necrosis factor; IL, interleukin; MMP, matrix metalloproteinase; TIMP, tissue inhibitor of metalloproteinase; ELISA, enzyme-linked immunosorbent assay; TNP, topical negative pressure; RT, reverse transcription; VEGF, vascular endothelial growth factor; FGF, fibroblast growth factor; TGF, transforming growth factor; HIF, hypoxia-inducible factor; PU, polyurethane; SMA, smooth muscle actin; 3D, three-dimensional; PDGF, platelet-derived growth factor; CGRP, calcitonin gene-related peptide; NGF, nerve growth factor; NGAL, neutrophil gelatinase-associated lipocalin; IHC, immunohistochemistry. Epigard™ (CONSEPT, Eschbach, Germany); Tegaderm™ (3 M, St Paul, Minnesota, USA); Bio-Plex® (Bio-Rad, Hercules, California, USA). Open in new tab Cytokine expression Two independent studies10,17 of different wound types, chronic (present for more than 1 month) and acute (surgical wounds), using two different methods of analysis, both observed significant reductions in early local tumour necrosis factor (TNF) expression following the application of NPWT. By contrast, a significant reduction in interleukin (IL) 1β expression was observed in acute surgical wounds10, whereas in chronic wounds there was no effect17. The influence of NPWT on IL-6 was examined in three studies13–14,26. In the larger of two studies of wound effluent, Labler and colleagues14 observed no difference in IL-6 expression with the application of NPWT, in contrast to their earlier findings26. An additional study13 using a porcine model observed no difference in serum IL-6 level over the first 4 h following injury. However, there was a significant rise in circulating (systemic) levels of the anti-inflammatory cytokine IL-10 within 4 h of application of NPWT13. It is unclear whether this rise was sustained as the study protocol terminated at 4 h. These changes in cytokine profile were consistent with a reduction in CD68-positive macrophages in reperfused skeletal muscle on application of NPWT10. By contrast, a non-randomized, non-blinded trial26 of 21 human traumatic wounds dressed with either Epigard™ (CONSEPT, Eschbach, Germany) or NPWT failed to yield any observable difference in IL-10 expression at the wound bed, although the first samples were taken on day 2 when the experimental dressings were first applied and subsequent samples taken even later. Hence, these data are not comparable with those from the porcine model described above. Studies evaluating the influence of NPWT on cytokine expression profiles are summarized in Table S3 (supporting information). Chemokine expression IL-8 is a potent chemokine and proangiogenic factor that also plays an important role in regulating neutrophil and macrophage migration30,31. In a study14 of human traumatic wounds, there was a significant rise in local, but not systemic, IL-8 expression following application of NPWT, although the temporal expression profile was not delineated as samples comprised the cumulative wound effluent in the collecting canister, evaluated at each dressing change at intervals of 48–96 h. Growth factor expression Vascular endothelial growth factor (VEGF) was the most frequently studied growth factor in relation to NPWT, with seven papers5,14,25–29 identified. Human data were reported in three studies14,26–27, two14,26 by the same author. Application of NPWT resulted in a significant rise in local VEGF expression in both acute (traumatic)14,26 and chronic27 wounds. The remaining data were derived from animal models of full-thickness wounds, two murine5,25, one rat29 and one rabbit28. Using different methods of quantification, one murine25 and one rabbit28 model demonstrated increased local expression of VEGF with NPWT. Erba and colleagues5 reported a concentration gradient, with the highest level expressed by tissue at the wound surface, adjacent to the filler material. Although the magnitude of expression was greater on days 3 and 7 in the control group, the gradient of expression in the NPWT group was more physiologically appropriate and resulted in better vascularization and enhanced wound healing. Another study29, however, reported no influence on VEGF expression by TNP. This study used semiquantitative immunohistochemistry to evaluate VEGF expression. Fibroblast growth factor (FGF) 2 (b-FGF), a potent mitogen and inducer of an angiogenic phenotype, was evaluated in three studies9,12,26. In an in vitro model of murine fibroblasts, increased FGF-2 expression was observed 48 h after application of TNP9. In a rat wound model, FGF-2 expression was noted to exhibit a biphasic response, with a significant increase on day 5 followed by a significant decrease on day 712. By contrast, a human study26 of 21 traumatic wounds found no difference in FGF-2 expression using NPWT or Epigard™ dressings. Transforming growth factor (TGF) β1 and platelet-derived growth factor (PDGF) were evaluated in three9,13,15 and one15 study respectively. Application of TNP resulted in increased expression of both PDGF and TGF-β1 in a human in vitro fibroblast model15. Murine fibroblasts exposed to TNP also demonstrated increased expression of TGF-β19. However, there was no change in systemic TGF-β1 expression in a porcine wound model13. Individual studies have also sought to observe the relationship between NPWT and the expression profiles of a number of other growth factors. Using a murine (diabetic) wound model, Younan and co-workers18 observed an increase in calcitonin gene-related peptide (CGRP) and substance P (together with an increase in nerve fibre density) at the wound bed following the application of TNP. The authors suggested that improved wound healing may be a direct consequence of the release of the sensory neuropeptide, although a causal link was not established. Finally, a single study28 using a full-thickness wound model in rabbits showed that application of NPWT resulted in increased expression of hypoxia-inducible factor (HIF) 1α, a key regulator of angiogenesis to maintain oxygen homeostasis. By contrast, a full-thickness murine model demonstrated a significant reduction in expression of HIF-1α at the wound bed under NPWT at the later time points (days 3 and 7). Studies evaluating the influence of NPWT on growth factor expression profiles are summarized in Table S4 (supporting information). Matrix metalloproteinase expression MMP expression was reported in three small series11,16–17 of chronic human wounds (defined as pressure ulcers by Stechmiller et al.17 and Shi and colleagues16, and wounds of 4–10 weeks' duration by Greene and co-workers11). Reductions in MMP-9/neutrophil gelatinase-associated lipocalin complex and latent MMP-2 expression were observed in the wound bed under NPWT11. A reduction in MMP-1 and MMP-13, with a tendency towards reduced MMP-2, was also reported16. However, a study17 of aspirates from eight chronic wounds subjected to TNP found no significant difference in expression of MMP-2, MMP-3 and tissue inhibitor of metalloproteinase (TIMP) 1 when NPWT was applied. These data indicate that, although there is some evidence to support reduced collagenase (MMPs 1 and 13) and gelatinase (MMP-9) expression, there is insufficient evidence to suggest that NPWT application results in global inhibition of the MMPs. Data are similarly lacking to enable conclusions with regard to MMP inhibitors, including TIMP-1. Studies evaluating the influence of NPWT on MMP expression profiles are summarized in Table S5 (supporting information). Discussion The present systematic review summarizes current understanding of the mechanism of action of NPWT at the molecular level. Human studies support attenuation of the acute inflammatory response locally: angiogenesis (via VEGF), cell recruitment (via IL-8) and reduced metalloproteinase expression. Animal models suggest a systemic anti-inflammatory response (via IL-10); VEGF and FGF-2 levels that vary with time; and increased CGRP and substance P (in diabetic wounds). In vitro models suggest increased granulation tissue synthesis via FGF-2, TGF-β and PDGF, and angiogenesis (PDGF). It is worth noting that cytokine and MMP data are derived from human studies whereas growth factor data are derived from animal and in vitro models. Overall, the data suggest that NPWT aids wound healing by local immune modulation, and mechanoreceptor and hypoxia-mediated signalling. Tissue trauma is accompanied by the release of alarmins, activation of tissue-resident macrophages, and recruitment of neutrophils and subsequently circulating monocytes32. The inflammatory infiltrate mediates further leucocyte recruitment in a paracrine fashion, and mobilizes mesenchymal stromal and endothelial cells by synthesis and release of the proinflammatory cytokines, particularly TNF and IL-633. However, TNF and IL-1β also upregulate MMP expression34, suggesting that persistent inflammation may have deleterious consequences for the tissues. It has been demonstrated previously that the application of TNP influences expression profiles of proinflammatory cytokines and their associated signalling pathways6. The data summarized in this review suggest that NPWT results in a shift towards an anti-inflammatory phenotype, with reduced expression of TNF and macrophage recruitment locally, and increased expression of IL-8 at the wound bed and IL-10 systemically. The localized expression profiles of other chemokines (particularly CCL-2 and CXCL-8) and their associated receptors remain unexplored. It has been observed previously that NPWT promotes subjacent cell proliferation35. However, this too must be controlled at the molecular level. As a result of a volumetric change in the filler material by the application of NPWT, a deformational force is applied to the wound bed, resulting in cell strain36,37. Although the mechanism converting a mechanical stimulus to a chemical signal remains unclear, some facts are beginning to emerge. Cells under stretch proliferate as a consequence of cytoskeletal tension38. Integrins, adhesive contacts within the cell matrix, act as strain gauges, triggering mechanoreceptor signalling pathways39,40, which upregulate expression of proto-oncogenes including myc, c-jun and Bcl-2; these in turn increase cellular mitotic rate41, along with genes coding for extracellular matrix (ECM) synthesis and remodelling7. The transcription factor early growth response (EGR) 1 appears to be an important nuclear mechanoregulator42 that plays a crucial role in wound healing, in part by inducing expression of VEGF, PDGF, TGF-β and FGF-243–46. Other evidence also supports the conclusion that NPWT-induced strain promotes a proangiogenic and promitogenic phenotype. Hsu and colleagues47 observed increased expression of cell division control protein (Cdc42) in human keratinocytes incubated under negative pressure. Human tendon fibroblasts subjected to mechanical stretch exhibited increased expression of PDGF, TGF-β and FGF-248. An in vitro model of vacuum-induced cyclical cell strain demonstrated increased cell proliferation and PDGF expression49, and shear stress resulted in increased angiogenesis in a rabbit ear model50. Application of NPWT results in positive pressure at the wound bed and hence reduced blood flow in the tissue immediately adjacent to the filler material51. Hypoxic fibroblasts have been shown to synthesize HIF-1, which activates genes encoding proteins that influence collagen deposition and alignment52. In this way, hypoxic fibroblasts influence the composition and structure of the ECM, with a resulting effect on wound healing. A recent study53 showed that hypoxic preconditioning of adipose-derived mesenchymal stromal cells enhanced cellular expression of VEGF and FGF-2. Moreover, it has been demonstrated that NPWT promotes a hypoxic gradient, with increased VEGF expression resulting in physiologically appropriate neovascularization, in contrast to the occlusive dressing controls, which yielded tortuous and friable vessels5. In a rat ear model, mechanical stretch resulted in significant upregulation of genes associated with hypoxia8. Interestingly, both VEGF and TGF-β1 are potent inducers of EGR-1 gene expression as well as growth factors induced by EGR-1 gene expression54,55. Moreover, EGR-1 knockout mice demonstrate an attenuated acute inflammatory response to injury and impaired dermal fibrosis, whereas EGR-1 transgenic mice exhibit accelerated wound healing56. EGR-1 null mice exhibit an attenuated chemokine expression profile, impairing inflammatory cell recruitment following ischaemia–reperfusion injury46. These data suggest that EGR-1 is not only crucial to optimal wound healing56 but might serve as a master switch by which NPWT leads to enhanced healing46. This concept is summarized in Fig. 2, with the? indicating that the relationship between EGR-1 and HIF-1α has yet to be established in this context. Fig. 2 Open in new tabDownload slide Proposed mechanisms governing cytokine, growth factor and matrix metalloproteinase (MMP) expression as influenced by negative-pressure wound therapy (NPWT) (simplified by exclusion of additional intracellular signalling proteins). HMGB, high-mobility group box; IL, interleukin; TNF, tumour necrosis factor; HIF, hypoxia-inducible factor; VEGF, vascular endothelial growth factor; TGF, transforming growth factor; PDGF, platelet-derived growth factor; FGF, fibroblast growth factor; EGR, early growth response. MMPs are collagenases, gelatinases and stromelysins that not only remodel the ECM but are now appreciated as inflammatory modulators implicated in a wide range of cell signalling processes of innate and adaptive immunity57. The MMPs also influence cutaneous wound healing58. MMP-2 and MMP-9 expression is increased in chronic wounds59, and delayed healing of traumatic wounds in humans is associated with raised serum levels of MMP-2 and MMP-7 and wound effluent level of MMP-360. To date, studies of the influence of NPWT on MMP expression profiles remain confined to small clinical series of chronic wounds. From the data available, NPWT appears to reduce expression of the collagenases MMP-1 and MMP-13 and the gelatinase MMP-9 and, possibly, also of MMP-2. However, NPWT appears not to reduce expression of MMP-3 or increase expression of the MMP inhibitor TIMP-1. The influence of NPWT on other MMPs is unknown. A systematic evaluation of the influence of NPWT on MMPs and their inhibitors in acute and chronic wounds is required. This systematic review has a number of limitations. Of the 16 studies identified, seven10–11,14,16–17,26–27 were human in vivo studies, whereas seven5,12–13,18,25,28–29 used animal models (3 mouse, 2 rat, 1 rabbit and 1 pig) and two relied on in vitro models, with human15 and murine9 fibroblasts. The studies of most clinical relevance relating to human wounds were opportunistic, poorly designed and controlled, and the data presented were not suitable for meta-analysis. Molecular assays from wound effluent stored at room temperature in the NPWT collecting canister for, typically, 72 h do not take into account cytokine and growth factor degradation, nor can they provide information on the temporal expression characteristics that can be controlled more easily in animal models. Moreover, effluent volume, a function of both plasma oncotic and capillary hydrostatic pressure, in addition to topical pressure, may exert a dilutional influence on measured concentrations. Additionally, pharmacological variables, including the use of corticosteroids and non-steroidal anti-inflammatory drugs, were not considered. Thus, the quality of the human experimental evidence remains poor and it must be interpreted with caution. Animal models offer some solutions to the problems inherent in human clinical studies, but the data are then subject to criticisms regarding the extent to which non-human models are predictive of clinical biological response. In vitro models using human cells circumvent some of these criticisms. However, only one publication15 used a human in vitro system and residual concerns remain regarding the extent to which a cellular response may be predictive of behaviour at tissue or organ level. Importantly, each study presents quantitative data from preselected time points. Although these time points are largely appropriate given the factors studied, nonetheless a comprehensive temporal expression profile of cytokine, growth factor and MMP by tissue under negative pressure remains elusive given the data available. Negative pressure-mediated cell strain is influenced by tissue stiffness and the magnitude of negative pressure. The combination of these factors is likely to influence mechanosensitive gene expression3. Furthermore, the magnitude and cyclical application of negative pressure influences blood flow, hence hypoxia and reperfusion. These variables have been shown to influence the quantity and quality of granulation tissue51. Most studies reviewed here used a pressure of –125 mmHg, and all but one used continuous suction. Therefore, a study of the effect on the cytokine and growth factor profile of these clinically relevant parameters is urgently required. Finally, six different methods of quantitative analysis were used. Of these, four techniques can be regarded as semiquantitative, with the remaining two (ELISA; Bio-Plex™ cytokine assay, Bio-Rad, Hercules, California, USA) being truly quantitative. These findings support the need for standard methods of observation in studies of this nature, with ELISA, RT–PCR and commercially available cytokine assay tests being the favoured options. Disclosure The authors declare no conflict of interest. Snapshot quiz 14/15 References 1 Morykwas MJ , Argenta LC, Shelton-Brown EI, McGuirt W. Vacuum-assisted closure: a new method for wound control and treatment: animal studies and basic foundation . Ann Plast Surg 1997 ; 38 : 553 – 562 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Pietramaggiori G , Liu P, Scherer SS, Kaipainen A, Prsa MJ, Mayer H et al. Tensile forces stimulate vascular remodeling and epidermal cell proliferation in living skin . Ann Surg 2007 ; 246 : 896 – 902 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Saxena V , Hwang CW, Huang S, Eichbaum Q, Ingber D, Orgill DP. Vacuum-assisted closure: microdeformations of wounds and cell proliferation . Plast Reconstr Surg 2004 ; 114 : 1086 – 1096 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 4 Shweiki D , Itin A, Soffer D, Keshet E. Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis . Nature 1992 ; 359 : 843 – 845 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Erba P , Ogawa R, Ackermann M, Adini A, Miele LF, Dastouri P et al. Angiogenesis in wounds treated by microdeformational wound therapy . Ann Surg 2011 ; 253 : 402 – 409 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Derrick KL , Norbury K, Kieswetter K, Skaf J, McNulty AK. Comparative analysis of global gene expression profiles between diabetic rat wounds treated with vacuum-assisted closure therapy, moist wound healing or gauze under suction . Int Wound J 2008 ; 5 : 615 – 624 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Kessler D , Dethlefsen S, Haase I, Plomann M, Hirche F, Krieg T et al. Fibroblasts in mechanically stressed collagen lattices assume a ‘synthetic’ phenotype . J Biol Chem 2001 ; 276 : 36 575 – 36 585 . Google Scholar Crossref Search ADS WorldCat 8 Saxena V , Orgill D, Kohane I. A set of genes previously implicated in the hypoxia response might be an important modulator in the rat ear tissue response to mechanical stretch . BMC Genomics 2007 ; 8 : 430 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Lu F , Ogawa R, Nguyen DT, Chen B, Guo D, Helm DL et al. Microdeformation of three-dimensional cultured fibroblasts induces gene expression and morphological changes . Ann Plast Surg 2011 ; 66 : 296 – 300 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Eisenhardt SU , Schmidt Y, Thiele JR, Iblher N, Penna V, Torio-Padron N et al. Negative pressure wound therapy reduces the ischaemia/reperfusion-associated inflammatory response in free muscle flaps . J Plast Reconstr Aesthet Surg 2012 ; 65 : 640 – 649 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Greene AK , Puder M, Roy R, Arsenault D, Kwei S, Moses MA et al. Microdeformational wound therapy: effects on angiogenesis and matrix metalloproteinases in chronic wounds of 3 debilitated patients . Ann Plast Surg 2006 ; 56 : 418 – 422 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Jacobs S , Simhaee DA, Marsano A, Fomovsky GM, Niedt G, Wu JK. Efficacy and mechanisms of vacuum-assisted closure (VAC) therapy in promoting wound healing: a rodent model . J Plast Reconstr Aesthet Surg 2009 ; 62 : 1331 – 1338 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Kilpadi DV , Bower CE, Reade CC, Robinson PJ, Sun YS, Zeri R et al. Effect of vacuum assisted closure therapy on early systemic cytokine levels in a swine model . Wound Repair Regen 2006 ; 14 : 210 – 215 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Labler L , Rancan M, Mica L, Härter L, Mihic-Probst D, Keel M. Vacuum-assisted closure therapy increases local interleukin-8 and vascular endothelial growth factor levels in traumatic wounds . J Trauma 2009 ; 66 : 749 – 757 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 15 McNulty AK , Schmidt M, Feeley T, Villanueva P, Kieswetter K. Effects of negative pressure wound therapy on cellular energetics in fibroblasts grown in a provisional wound (fibrin) matrix . Wound Repair Regen 2009 ; 17 : 192 – 199 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Shi B , Chen SZ, Zhang P, Li JQ. [ Effects of vacuum-assisted closure (VAC) on the expressions of MMP-1, 2, 13 in human granulation wound .] Zhonghua Zheng Xing Wai Ke Za Zhi 2003 ; 19 : 279 – 281 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 17 Stechmiller JK , Kilpadi DV, Childress B, Schultz GS. Effect of vacuum-assisted closure therapy on the expression of cytokines and proteases in wound fluid of adults with pressure ulcers . Wound Repair Regen 2006 ; 14 : 371 – 374 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Younan G , Ogawa R, Ramirez M, Helm D, Dastouri P, Orgill DP. Analysis of nerve and neuropeptide patterns in vacuum-assisted closure-treated diabetic murine wounds . Plast Reconstr Surg 2010 ; 126 : 87 – 96 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Dastouri P , Helm DL, Scherer SS, Pietramaggiori G, Younan G, Orgill DP. Waveform modulation of negative-pressure wound therapy in the murine model . Plast Reconstr Surg 2011 ; 127 : 1460 – 1466 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Fraccalvieri M , Zingarelli E, Ruka E, Antoniotti U, Coda R, Sarno A et al. Negative pressure wound therapy using gauze and foam: histological, immunohistochemical and ultrasonography morphological analysis of the granulation tissue and scar tissue. Preliminary report of a clinical study . Int Wound J 2011 ; 8 : 355 – 364 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Armstrong DG , Lavery LA; Diabetic Foot Study Consortium. Negative pressure wound therapy after partial diabetic foot amputation: a multicentre, randomised controlled trial . Lancet 2005 ; 366 : 1704 – 1710 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Ubbink DT , Westerbos SJ, Nelson EA, Vermeulen H. A systematic review of topical negative pressure therapy for acute and chronic wounds . Br J Surg 2008 ; 95 : 685 – 692 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Higgins JPT , Green S (eds). Cochrane Handbook for Systematic Reviews of Interventions , version 5.1.0 [updated March 2011]. http://handbook.cochrane.org [accessed 1 December 2013]. 24 Moher D , Liberati A, Tetzlaff J, Altman DG; PRISMA Group . Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement . PLoS Med 2009 ; 6 : e1000097 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Liu Y , Hu DH, Dong ML, Wang YC, Liu JQ, Bai L et al. [ Efficacy of vacuum sealing drainage in mice infected with Pseudomonas aeruginosa and its mechanism .] Zhonghua Shao Shang Za Zhi 2011 ; 27 : 255 – 259 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 26 Labler L , Mica L, Harter L, Trentz O, Keel M. [ Influence of V.A.C.-therapy on cytokines and growth factors in traumatic wounds .] Zentralbl Chir 2006 ; 131 ( Suppl 1 ): S62 – S67 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Cao DY , Chen SZ, Tang SY, Hu ZH, Song M, Lu XX. Effect of vacuum-assisted closure on angiogenesis of human chronic wound healing . Chinese Journal of Clinical Rehabilitation 2004 ; 8 : 264 – 265 . Google Scholar OpenURL Placeholder Text WorldCat 28 Yang F , Hu D, Bai XJ, Zhang K, Li RJ, Xue CC. [ The influence of oxygen partial pressure change and vascularization of rabbit wound through negative pressure wound therapy .] Zhonghua Wai Ke Za Zhi 2012 ; 50 : 650 – 654 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 29 Tang SY , Xu HR, Qi LJ, Guo YL, Ji HR. [ Effect of vacuum-assisted closure on angiogenesis during denervation of wound healing .] Zhongguo Linchuang Kangfu 2005 ; 46 : 91 – 93 . Google Scholar OpenURL Placeholder Text WorldCat 30 Koch AE , Polverini PJ, Kunkel SL, Harlow LA, DiPietro LA, Elner VM et al. Interleukin-8 as a macrophage-derived mediator of angiogenesis . Science 1992 ; 258 : 1798 – 1801 . Google Scholar Crossref Search ADS PubMed WorldCat 31 Mukaida N , Harada A, Matsushima K. Interleukin-8 (IL-8) and monocyte chemotactic and activating factor (MCAF/MCP-1), chemokines essentially involved in inflammatory and immune reactions . Cytokine Growth Factor Rev 1998 ; 9 : 9 – 23 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Chan JK , Roth J, Oppenheim JJ, Tracey KJ, Vogl T, Feldmann M et al. Alarmins: awaiting a clinical response . J Clin Invest 2012 ; 122 : 2711 – 2719 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Glass GE , Chan JK, Freidin A, Feldmann M, Horwood NJ, Nanchahal J. TNF-alpha promotes fracture repair by augmenting the recruitment and differentiation of muscle-derived stromal cells . Proc Natl Acad Sci U S A 2011 ; 108 : 1585 – 1590 . Google Scholar Crossref Search ADS PubMed WorldCat 34 Sarén P , Welgus HG, Kovanen PT. TNF-alpha and IL-1beta selectively induce expression of 92-kDa gelatinase by human macrophages . J Immunol 1996 ; 157 : 4159 – 4165 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 35 Scherer SS , Pietramaggiori G, Mathews JC, Prsa MJ, Huang S, Orgill DP. The mechanism of action of the vacuum-assisted closure device . Plast Reconstr Surg 2008 ; 122 : 786 – 797 . Google Scholar Crossref Search ADS PubMed WorldCat 36 Wilkes R , Zhao Y, Cunningham K, Kieswetter K, Haridas B. 3D strain measurement in soft tissue: demonstration of a novel inverse finite element model algorithm on MicroCT images of a tissue phantom exposed to negative pressure wound therapy . J Mech Behav Biomed Mater 2009 ; 2 : 272 – 287 . Google Scholar Crossref Search ADS PubMed WorldCat 37 Wilkes R , Zhao Y, Kieswetter K, Haridas B. Effects of dressing type on 3D tissue microdeformations during negative pressure wound therapy: a computational study . J Biomech Eng 2009 ; 131 : 031012 . Google Scholar Crossref Search ADS PubMed WorldCat 38 Huang S , Ingber DE. The structural and mechanical complexity of cell-growth control . Nat Cell Biol 1999 ; 1 : E131 – E138 . Google Scholar Crossref Search ADS PubMed WorldCat 39 Galbraith CG , Sheetz MP. Forces on adhesive contacts affect cell function . Curr Opin Cell Biol 1998 ; 10 : 566 – 571 . Google Scholar Crossref Search ADS PubMed WorldCat 40 Shyy JY , Chien S. Role of integrins in cellular responses to mechanical stress and adhesion . Curr Opin Cell Biol 1997 ; 9 : 707 – 713 . Google Scholar Crossref Search ADS PubMed WorldCat 41 Chen SZ , Cao DY, Li JQ, Tang SY. [ Effect of vacuum-assisted closure on the expression of proto-oncogenes and its significance during wound healing .] Zhonghua Zheng Xing Wai Ke Za Zhi 2005 ; 21 : 197 – 200 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 42 Schwachtgen JL , Houston P, Campbell C, Sukhatme V, Braddock M. Fluid shear stress activation of Egr-1 transcription in cultured human endothelial and epithelial cells is mediated via the extracellular signal-related kinase 1/2 mitogen-activated protein kinase pathway . J Clin Invest 1998 ; 101 : 2540 – 2549 . Google Scholar Crossref Search ADS PubMed WorldCat 43 Biesiada E , Razandi M, Levin ER. Egr-1 activates basic fibroblast growth factor transcription. Mechanistic implications for astrocyte proliferation . J Biol Chem 1996 ; 271 : 18 576 – 18 581 . Google Scholar Crossref Search ADS WorldCat 44 Khachigian LM , Lindner V, Williams AJ, Collins T. Egr-1-induced endothelial gene expression: a common theme in vascular injury . Science 1996 ; 271 : 1427 – 1431 . Google Scholar Crossref Search ADS PubMed WorldCat 45 Kim SJ , Jeang KT, Glick AB, Sporn MB, Roberts AB. Promoter sequences of the human transforming growth factor-beta 1 gene responsive to transforming growth factor-beta 1 autoinduction . J Biol Chem 1989 ; 264 : 7041 – 7045 . Google Scholar Crossref Search ADS PubMed WorldCat 46 Yan SF , Fujita T, Lu J, Okada K, Shan Zou Y, Mackman N et al. Egr-1, a master switch coordinating upregulation of divergent gene families underlying ischemic stress . Nat Med 2000 ; 6 : 1355 – 1361 . Google Scholar Crossref Search ADS PubMed WorldCat 47 Hsu CC , Chow SE, Chen CP, Tsai WC, Wang JS, Yu SY et al. Negative pressure accelerated monolayer keratinocyte healing involves Cdc42 mediated cell podia formation . J Dermatol Sci 2013 ; 70 : 196 – 203 . Google Scholar Crossref Search ADS PubMed WorldCat 48 Skutek M , van Griensven M, Zeichen J, Brauer N, Bosch U. Cyclic mechanical stretching modulates secretion pattern of growth factors in human tendon fibroblasts . Eur J Appl Physiol 2001 ; 86 : 48 – 52 . Google Scholar Crossref Search ADS PubMed WorldCat 49 Wilson E , Mai Q, Sudhir K, Weiss RH, Ives HE. Mechanical strain induces growth of vascular smooth muscle cells via autocrine action of PDGF . J Cell Biol 1993 ; 123 : 741 – 747 . Google Scholar Crossref Search ADS PubMed WorldCat 50 Ichioka S , Shibata M, Kosaki K, Sato Y, Harii K, Kamiya A. Effects of shear stress on wound-healing angiogenesis in the rabbit ear chamber . J Surg Res 1997 ; 72 : 29 – 35 . Google Scholar Crossref Search ADS PubMed WorldCat 51 Glass GE , Nanchahal J. The methodology of negative pressure wound therapy: separating fact from fiction . J Plast Reconstr Aesthet Surg 2012 ; 65 : 989 – 1001 . Google Scholar Crossref Search ADS PubMed WorldCat 52 Gilkes DM , Bajpai S, Chaturvedi P, Wirtz D, Semenza GL. Hypoxia-inducible factor 1 (HIF-1) promotes extracellular matrix remodeling under hypoxic conditions by inducing P4HA1, P4HA2, and PLOD2 expression in fibroblasts . J Biol Chem 2013 ; 288 : 10 819 – 10 829 . Google Scholar Crossref Search ADS WorldCat 53 Liu L , Gao J, Yuan Y, Chang Q, Liao Y, Lu F. Hypoxia preconditioned human adipose derived mesenchymal stem cells enhance angiogenic potential via secretion of increased VEGF and bFGF . Cell Biol Int 2013 ; 37 : 551 – 560 . Google Scholar Crossref Search ADS PubMed WorldCat 54 Chen SJ , Ning H, Ishida W, Sodin-Semrl S, Takagawa S, Mori Y et al. The early-immediate gene EGR-1 is induced by transforming growth factor-beta and mediates stimulation of collagen gene expression . J Biol Chem 2006 ; 281 : 21 183 – 21 197 . Google Scholar Crossref Search ADS WorldCat 55 Liu L , Tsai JC, Aird WC. Egr-1 gene is induced by the systemic administration of the vascular endothelial growth factor and the epidermal growth factor . Blood 2000 ; 96 : 1772 – 1781 . Google Scholar Crossref Search ADS PubMed WorldCat 56 Wu M , Melichian DS, de la Garza M, Gruner K, Bhattacharyya S, Barr L et al. Essential roles for early growth response transcription factor Egr-1 in tissue fibrosis and wound healing . Am J Pathol 2009 ; 175 : 1041 – 1055 . Google Scholar Crossref Search ADS PubMed WorldCat 57 Parks WC , Wilson CL, López-Boado YS. Matrix metalloproteinases as modulators of inflammation and innate immunity . Nat Rev Immunol 2004 ; 4 : 617 – 629 . Google Scholar Crossref Search ADS PubMed WorldCat 58 Singer AJ , Clark RA. Cutaneous wound healing . N Engl J Med 1999 ; 341 : 738 – 746 . Google Scholar Crossref Search ADS PubMed WorldCat 59 Wysocki AB , Staiano-Coico L, Grinnell F. Wound fluid from chronic leg ulcers contains elevated levels of metalloproteinases MMP-2 and MMP-9 . J Invest Dermatol 1993 ; 101 : 64 – 68 . Google Scholar Crossref Search ADS PubMed WorldCat 60 Utz ER , Elster EA, Tadaki DK, Gage F, Perdue PW, Forsberg JA et al. Metalloproteinase expression is associated with traumatic wound failure . J Surg Res 2010 ; 159 : 633 – 639 . Google Scholar Crossref Search ADS PubMed WorldCat © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd

Brunckhorst, O; Challacombe, B; Abboudi, H; Khan, M S; Dasgupta, P; Ahmed, K

doi: 10.1002/bjs.9635pmid: 25312488

Abstract Background Live surgical demonstrations are commonly performed for surgical conferences. These live procedures have recently come under scrutiny, in particular with issues pertaining to patient safety. This systematic review aimed to explore the evidence for live surgery as a training tool, and to investigate the safety of live surgical broadcasts. Methods PubMed, Embase, MEDLINE and Cochrane Library databases were searched using a predefined search strategy from January 1980 to October 2013. Specialty Societies and primary Colleges of Surgeons were searched for guidelines or position statements on live surgical teaching. Results Educational value criteria demonstrated for live surgery included feasibility, acceptability, construct and concurrent validity. Complication rates during live procedures were not compromised in the majority of studies. Patient safety, however, may be affected during live procedures as success rates have been found to be lower in some studies. Only Cardiothoracic, Urology and Vascular Surgical Societies currently offer guidelines on conducting live surgical demonstrations. Conclusion Little evidence exists on the safety and educational value of live surgery, with few studies of high quality conducted. Guidance on live procedures is scarce, with only three major surgical specialties offering any advice. More needs to be done to establish and promote evidence for the value of live surgery demonstrations. Introduction Live operating is a practice that can be traced back to the birth of surgery itself, and indeed the name operating theatre demonstrates the origin of these rooms1. As technology and the internet have developed, live surgery is no longer limited to people within the same room and, since the first live surgical broadcast in 19962, this format has established itself as a common educational tool for surgeons. Live endoscopic, laparoscopic and now robotic surgery is popular as these procedures are easily broadcast. However, questions have been raised over its safety3, and some argue that live surgery broadcasts are no more than a promotional tool for surgeons to display their skills4. With such mixed comments, the educational value of live surgery must be considered. This review therefore aims to explore the evidence for the effectiveness of live surgery as a teaching tool in terms of its feasibility, acceptability, validity, educational impact, cost and value; identify current evidence for the safety of live surgical broadcasts; and outline currently available guidelines provided by surgical specialties on live surgical procedures. Methods This study was performed following guidelines defined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement5. Study eligibility criteria Empirical studies describing the educational value, safety, use and development of guidelines for live surgical teaching were included in the review, including original articles and surveys. Guidelines and position statements offered by Surgical Societies were included. Review articles, non-English articles, letters and abstracts from conferences were excluded. Information sources and search A broad search was performed using the PubMed, Embase, MEDLINE and Cochrane Library databases in the date range January 1980 to October 2013. Search terms included ‘live surgical teaching’, ‘live case demonstration’ and ‘live surgical broadcast’. These were combined with the word ‘guidelines’. A reference review was then conducted. The American, British and European Societies for the 11 major surgical specialties alongside the primary Colleges of Surgeons were searched for guidelines or position statements on live surgical teaching. Study selection and data collection Two reviewers independently identified potentially relevant articles. The full text of each article was obtained and further screened for inclusion if it pertained to live surgical demonstrations. Potentially relevant guidelines identified from respective Surgical Societies were analysed for inclusion in the review. Conflicts between reviewers were subsequently discussed until resolution. Data items Relevant data were extracted from each study. To evaluate the effectiveness of live surgery as a teaching tool, objective and subjective data supporting the feasibility, acceptability, validity, educational impact, cost and value were evaluated. Definitions as set out by Ahmed6,7 and Van Der Vleuten8 were used to assess these parameters (Table S1, supporting information). Outcome measures were divided into subjective and objective data. Objective data collected included the number of questions asked during an operative procedure and the successful transmission of a live procedure. Subjective data were collected from surveys and questionnaires answered by students and experts. These included Likert-scale evaluation of the quality of the topics covered during an operative procedure, educational value and appropriate content of live transmission. Furthermore, the percentage of those believing live surgery to be ethical was examined. Patient safety during live procedures was assessed using key intraoperative and postoperative outcome measures. Primary measures, present in all articles, included total success and complication rates. Secondary measures included duration of operation, delay to procedure, use of general anaesthesia, 30-day mortality, estimated blood loss and length of hospital stay, according to complications. Finally, guidelines identified were analysed for method and date of development, content and regulatory bodies overseeing the practice of live surgery. The quality of educational impact and assessment studies was analysed using the framework proposed by Van Der Vleuten8. These studies and those analysing the safety of live surgery were assessed using the Oxford Centre for Evidence-based Medicine levels of evidence9. Results Study selection In total, 1104 potentially relevant articles were identified; 1056 were excluded following abstract review and removal of duplicates. After full-text review, another 24 articles were excluded, but a further three were added after reference review. Thus, 27 articles were finally included in the review (Fig. 1). Fig. 1 Open in new tabDownload slide PRISMA diagram for study selection Study characteristics and results synthesis The selected articles consisted of empirical studies and individual guideline documents set out by Surgical Societies. Information from these studies was categorized into three headings: educational value, patient safety and surgical specialties' guidelines. Educational value of live surgery as a teaching tool In total six studies2,10–14 were identified that investigated the educational value of live surgery (Table S2, supporting information). Three2,10–11 were conducted by Gandsas and colleagues to assess the feasibility of surgical broadcasting. In 1996, three separate laparoscopic procedures were transmitted successfully from an institution in California to Buenos Aires via the internet during a medical meeting; the authors claimed the system was therefore feasible2. Additionally, in 2002 the same authors10 developed a system that could transmit live and prerecorded surgery via the internet; successful transmission of a live paraoesophageal hernia repair was reported, with good video quality. The feasibility of live broadcasting to a handheld device was confirmed by successfully broadcasting a laparoscopic splenectomy to eight separate handheld devices11. Only one study assessed the validity of live surgical teaching objectively. McIntyre and co-workers12 compared the responses of students in the operating room (OR) and those watching live transmissions; they analysed the questions asked and those left unanswered during and after the sessions. In the 23 sessions observed, it was found that students asked four times as many questions and were left with fewer unanswered questions during live transmissions compared with OR teaching sessions. Although few objective data exist, other studies have used subjective measures to assess the educational value of live surgery. A large survey of 253 urologists attending conferences was conducted by Elsamra and colleagues13. This showed that 93 per cent of respondents rated live demonstrations a minimum of 4 of 5 in helpfulness as an educational tool. Furthermore, 78 per cent of respondents also identified live demonstrations as an ethical and, therefore, acceptable method of teaching. Gul et al.14 achieved similar results with 46 students, who rated live surgical broadcasting a median of 9 of 10 as an educational tool compared with 5 of 10 for traditional surgical theatre teaching. Patient safety A total of eight studies15–22 were identified that analysed patient safety during live surgical demonstrations (Table 1). Schmit and colleagues15 analysed data from 336 patients undergoing endoscopic retrograde cholangiopancreatography (ERCP): 168 underwent ERCP as a live demonstration as part of a workshop and 168 acted as controls. No difference in success, complication rates or duration of hospital stay was found between the workshop and control patients. The authors concluded that live demonstrations were safe. Two further studies analysed the effects of live demonstrations on ERCP safety. Liao and co-workers16 analysed 406 patients who underwent live demonstrations in 36 conferences between 2002 and 2007; they were compared with 406 corresponding controls from various endoscopy centres. The authors found no significant difference in complication rates between cohorts. Similarly, Ridtitid et al.17 analysed 82 patients who underwent live demonstrations compared with patients with the same indications for admission; this study also concluded that complication rates did not differ between the two cohorts. Despite these findings, both studies found a significantly lower success rate for live procedures, especially those considered to be more difficult. Table 1 Overview of patient safety studies Reference . Year . Procedure analysed . No. of live procedures . No. of control procedures . Outcome . Schmit et al.15 2005 ERCP 168 168 Duration of procedure: no significant difference Complications: no significant difference Success rate: no significant difference Delay to procedure: present in 10·7% of procedures Use of general anaesthesia: significantly increased for live procedures (87·5 versus 44%) Liao et al.16 2009 ERCP 406 406 Complications: no significant difference Success rate: significantly lower for live procedure (85·5 versus 92·1%) Ridtitid et al.17 2012 ERCP 82 246 Complications: no significant difference Success rates: significantly lower for live procedure in more difficult patients (73 versus 90%) Franke et al.18 2009 Carotid stenting 186 – Complications: no different from published data Success rate: no different from published data Eliyahu et al.19 2012 Interventional cardiology 101 66 Complications: no significant difference Success rate: no significant difference Seeburger et al.20 2011 Cardiac surgery 250 – Complications: no different from expected standards Success rates: no different from expected standards 30-day mortality: no different from expected standards Mullins et al.21 2012 Robotic partial nephrectomies 39 847 Complications: no significant difference Duration of operation: no significant difference Estimated blood loss: no significant difference Length of hospital stay: no significant difference Chatelain et al.22 1992 Coronary angioplasty 104 – Complications: live surgery inferior to published data Success rate: live surgery inferior to published data Reference . Year . Procedure analysed . No. of live procedures . No. of control procedures . Outcome . Schmit et al.15 2005 ERCP 168 168 Duration of procedure: no significant difference Complications: no significant difference Success rate: no significant difference Delay to procedure: present in 10·7% of procedures Use of general anaesthesia: significantly increased for live procedures (87·5 versus 44%) Liao et al.16 2009 ERCP 406 406 Complications: no significant difference Success rate: significantly lower for live procedure (85·5 versus 92·1%) Ridtitid et al.17 2012 ERCP 82 246 Complications: no significant difference Success rates: significantly lower for live procedure in more difficult patients (73 versus 90%) Franke et al.18 2009 Carotid stenting 186 – Complications: no different from published data Success rate: no different from published data Eliyahu et al.19 2012 Interventional cardiology 101 66 Complications: no significant difference Success rate: no significant difference Seeburger et al.20 2011 Cardiac surgery 250 – Complications: no different from expected standards Success rates: no different from expected standards 30-day mortality: no different from expected standards Mullins et al.21 2012 Robotic partial nephrectomies 39 847 Complications: no significant difference Duration of operation: no significant difference Estimated blood loss: no significant difference Length of hospital stay: no significant difference Chatelain et al.22 1992 Coronary angioplasty 104 – Complications: live surgery inferior to published data Success rate: live surgery inferior to published data ERCP, endoscopic retrograde cholangiopancreatography. Open in new tab Table 1 Overview of patient safety studies Reference . Year . Procedure analysed . No. of live procedures . No. of control procedures . Outcome . Schmit et al.15 2005 ERCP 168 168 Duration of procedure: no significant difference Complications: no significant difference Success rate: no significant difference Delay to procedure: present in 10·7% of procedures Use of general anaesthesia: significantly increased for live procedures (87·5 versus 44%) Liao et al.16 2009 ERCP 406 406 Complications: no significant difference Success rate: significantly lower for live procedure (85·5 versus 92·1%) Ridtitid et al.17 2012 ERCP 82 246 Complications: no significant difference Success rates: significantly lower for live procedure in more difficult patients (73 versus 90%) Franke et al.18 2009 Carotid stenting 186 – Complications: no different from published data Success rate: no different from published data Eliyahu et al.19 2012 Interventional cardiology 101 66 Complications: no significant difference Success rate: no significant difference Seeburger et al.20 2011 Cardiac surgery 250 – Complications: no different from expected standards Success rates: no different from expected standards 30-day mortality: no different from expected standards Mullins et al.21 2012 Robotic partial nephrectomies 39 847 Complications: no significant difference Duration of operation: no significant difference Estimated blood loss: no significant difference Length of hospital stay: no significant difference Chatelain et al.22 1992 Coronary angioplasty 104 – Complications: live surgery inferior to published data Success rate: live surgery inferior to published data Reference . Year . Procedure analysed . No. of live procedures . No. of control procedures . Outcome . Schmit et al.15 2005 ERCP 168 168 Duration of procedure: no significant difference Complications: no significant difference Success rate: no significant difference Delay to procedure: present in 10·7% of procedures Use of general anaesthesia: significantly increased for live procedures (87·5 versus 44%) Liao et al.16 2009 ERCP 406 406 Complications: no significant difference Success rate: significantly lower for live procedure (85·5 versus 92·1%) Ridtitid et al.17 2012 ERCP 82 246 Complications: no significant difference Success rates: significantly lower for live procedure in more difficult patients (73 versus 90%) Franke et al.18 2009 Carotid stenting 186 – Complications: no different from published data Success rate: no different from published data Eliyahu et al.19 2012 Interventional cardiology 101 66 Complications: no significant difference Success rate: no significant difference Seeburger et al.20 2011 Cardiac surgery 250 – Complications: no different from expected standards Success rates: no different from expected standards 30-day mortality: no different from expected standards Mullins et al.21 2012 Robotic partial nephrectomies 39 847 Complications: no significant difference Duration of operation: no significant difference Estimated blood loss: no significant difference Length of hospital stay: no significant difference Chatelain et al.22 1992 Coronary angioplasty 104 – Complications: live surgery inferior to published data Success rate: live surgery inferior to published data ERCP, endoscopic retrograde cholangiopancreatography. Open in new tab Franke and colleagues18 analysed the effect of live transmission on the clinical outcomes of carotid stenting. In a cohort of 186 patients whose procedures were broadcast live to 22 different conferences, there was no difference in success rates or complications compared with a similar contemporary series. This evidence is further strengthened by the findings of Eliyahu and co-workers19, who concluded that interventional cardiology procedures performed in live demonstrations had similar clinical outcomes to controls. In total, 101 patients undergoing various interventional procedures, ranging from coronary and carotid procedures to congenital heart operations, were analysed. There was no significant difference in success and complication rates compared with 66 matched controls. Although several of these studies have confirmed the safety of arguably less invasive procedures, only two studies have analysed the effects of broadcasting major surgical procedures. Seeburger and colleagues20 showed that the mortality, clinical outcome and follow-up after various cardiac surgical procedures were similar to expected standards in a total of 250 procedures. In urology, Mullins and co-workers21 found that there was no difference in several operative parameters and complication rates between live robotic partial nephrectomies in 39 patients and 847 control procedures. Despite evidence supporting the safety of live demonstrations in various surgical procedures, it must be noted that Chatelain et al.22 found that success rates were lower during live demonstrations of coronary angioplasty. In 104 procedures performed during 12 international angioplasty courses, there was an increase in failure rates compared with the literature. Furthermore, although Schmit and co-workers15 found no difference in outcomes in their study, some less favourable parameters were identified: an increased use of anaesthesia during live surgery, and delayed treatment so that procedures could be performed as a live demonstration. Available specialty guidelines Thirteen guidelines, position statements and articles relating to guidelines placed by Surgical Societies were identified (Table 2; Table S3, supporting information)23–35. The first guidelines were developed following the death of a patient in Japan during a live transmission in 200736. This resulted in formation of a joint committee for the establishment of guidelines for the live session of thoracic and cardiovascular surgery. This consisted of the Japanese Society for Cardiovascular Surgery, Japanese Association for Thoracic Surgery and Japanese Society for Vascular Surgery, which drew up the current guidelines23. Several international Cardiothoracic Surgical Societies followed suit. The European Association for Cardio-Thoracic Surgery24 implemented the same guidelines as set out by the three Japanese Societies. In addition, the American Association for Thoracic Surgery jointly set up an Ethics Committee with the Society of Thoracic Surgeons, which identified key ethical considerations for live procedures from which guidelines were developed25. Table 2 Common themes in live surgery guidelines23–33 Patient safety comes above all else and the surgeon must be willing to terminate the live broadcast if this becomes compromised The patient's privacy must be preserved at all times Separate consent should be gained for live surgical broadcast with dedicated forms A moderator between the audience and surgeon should be used to prevent questions distracting the surgeon at key steps The educational value must exceed that of a prerecorded operation, otherwise this should be used instead Live broadcasts should be performed at a surgeon's home institution where possible Surgeons should not participate in broadcasts when rigid schedules constrain the starting time or duration of the procedures Live procedure demonstrations should not be used for marketing or commercial opportunities for the physician, host institution or equipment used in the procedure Broadcasting to the general public does not provide enough educational benefit and is therefore not ethical Patient safety comes above all else and the surgeon must be willing to terminate the live broadcast if this becomes compromised The patient's privacy must be preserved at all times Separate consent should be gained for live surgical broadcast with dedicated forms A moderator between the audience and surgeon should be used to prevent questions distracting the surgeon at key steps The educational value must exceed that of a prerecorded operation, otherwise this should be used instead Live broadcasts should be performed at a surgeon's home institution where possible Surgeons should not participate in broadcasts when rigid schedules constrain the starting time or duration of the procedures Live procedure demonstrations should not be used for marketing or commercial opportunities for the physician, host institution or equipment used in the procedure Broadcasting to the general public does not provide enough educational benefit and is therefore not ethical Open in new tab Table 2 Common themes in live surgery guidelines23–33 Patient safety comes above all else and the surgeon must be willing to terminate the live broadcast if this becomes compromised The patient's privacy must be preserved at all times Separate consent should be gained for live surgical broadcast with dedicated forms A moderator between the audience and surgeon should be used to prevent questions distracting the surgeon at key steps The educational value must exceed that of a prerecorded operation, otherwise this should be used instead Live broadcasts should be performed at a surgeon's home institution where possible Surgeons should not participate in broadcasts when rigid schedules constrain the starting time or duration of the procedures Live procedure demonstrations should not be used for marketing or commercial opportunities for the physician, host institution or equipment used in the procedure Broadcasting to the general public does not provide enough educational benefit and is therefore not ethical Patient safety comes above all else and the surgeon must be willing to terminate the live broadcast if this becomes compromised The patient's privacy must be preserved at all times Separate consent should be gained for live surgical broadcast with dedicated forms A moderator between the audience and surgeon should be used to prevent questions distracting the surgeon at key steps The educational value must exceed that of a prerecorded operation, otherwise this should be used instead Live broadcasts should be performed at a surgeon's home institution where possible Surgeons should not participate in broadcasts when rigid schedules constrain the starting time or duration of the procedures Live procedure demonstrations should not be used for marketing or commercial opportunities for the physician, host institution or equipment used in the procedure Broadcasting to the general public does not provide enough educational benefit and is therefore not ethical Open in new tab Urology is a specialty that has followed in the footsteps of thoracic surgeons. The European Association of Urology26 has endorsed the use of live surgery if it is carried out according to its strict code of conduct; this was developed following the establishment of an Ethics of Live Surgery Panel which produced a policy for live surgery for the Association. Furthermore, a Live Surgery Committee has been established. The American Urological Association27 has also laid out guidelines for live surgical broadcasting; however, it is not clear how these were developed. In vascular surgery, the Japanese Society for Vascular Surgery was part of the committee that laid out the first guidelines in 2007; however, it is the only Vascular Society to have implemented guidelines. No specific guidelines were identified for any of the other major surgical specialties. Despite this, many of these Societies still use live surgery in their annual meetings and some even encourage its use in the training of procedures. Although many specialty Societies have not commented on the issue, several Colleges of Surgery have released position statements. The Royal College of Surgeons held a seminar in 2010 about live surgery, and subsequently developed a position statement34 which applies to all surgeons in England; this document endorses the use of live surgery to teach clinicians, if carried out safely. The Royal Australasian College of Surgeons28 also added a position statement in 2010 owing to concerns raised over patient safety. This indicates that live surgery is acceptable if criteria are met. In contrast, the American College of Surgeons35 has banned the use of live surgery at its major meetings; the American College of Obstetricians and Gynecologists is also following this path. Outside of surgery several other specialties have produced guidelines for live broadcasting of procedures. Several endoscopic organizations, such as the World Endoscopy Organization29, American Society for Gastrointestinal Endoscopy30, European Society of Gastrointestinal Endoscopy31 and European Society of Gastrointestinal Endoscopy Nurses and Associates31, have set guidelines. In addition, an international committee of six separate Cardiac and Interventional Cardiology Societies set out a joint document on the use of live procedure demonstration at cardiology meetings, with a code of conduct32. In America, an institution that plays an important role in live surgical demonstrations is the Food and Drug Administration (FDA). Devices shown in live procedures should be approved by the FDA for the specific use of the surgery being performed. If not approved, or being used for an off-label purpose, the institution must apply for permission to use the device for the broadcast33. Unfortunately, this applies only to transmissions from America and the FDA has no authority elsewhere. Discussion This systematic review looked at the educational value and safety of live surgical training, along with the available guidelines from Surgical Societies. Currently the evidence for live surgery as an educational tool is scarce. Its feasibility has been proven by the studies conducted by Gandsas and colleagues2,10–11. The study carried out by McIntyre and co-workers12 is the only one to use objective measures: concurrent validity was established by comparison with the current standard of OR teaching. Elsamra and colleagues13 and Gul et al.14 used subjective measures to demonstrate content and concurrent validity and acceptability. However, little else has been done to strengthen the validity of live surgery; its educational impact, cost and value lack any evidence. This has not prevented live transmissions of procedures becoming integrated into some medical school curricula37. Additionally, several courses continue to use live surgery workshops as a key element38,39 and have received good feedback from attendees. There remain concerns over performing live demonstrations at a hospital outside of your own40, which provides a visiting surgeon with added challenges. Common problems have been highlighted by a survey demonstrating how anxiety levels experienced by surgeons at a visiting hospital are much higher than at the home institution41. Khan and co-workers35 delineated further issues, such as lack of specific instruments, language difficulties and jet lag causing problems for a visiting surgeon. With increased stress and distraction having already been shown to affect surgical performance42,43, the safety of live surgery must be questioned. However, current evidence suggests that there is no added danger to the patient during a live demonstration, in terms of complication rates. Other studies, however, reported lower success rates. Few studies analysed surgical procedures, most of the research being concentrated in interventional radiology. Regardless, there remains strong support for live surgery, with many believing there is little difference between live surgery and teaching students in the OR42. It is evident that more must be done to establish the educational value and safety of live surgical teaching. Although randomized studies are not feasible, a multispecialty database of all procedures could be established. This could provide the data to compare the safety of live and standard procedures. Subjective data could also be collected from observers during live demonstrations to establish their educational impact and costs. Without this information there is not enough evidence to prove the validity and safety of live surgery as a method of training. It surprising to find that there is little guidance on how live surgery should be conducted. Although available guidelines are similar in content, they are insufficient and not well enough publicized to ensure that live surgery is conducted safely across the globe. International guidelines would be ideal. There needs to be increased awareness of existing guidelines to ensure that adequate precautions are being taken during live procedures. This article has some limitations. As with any systematic review, some useful references may have been missed; however, multiple databases and broad search terms were used to minimize this risk. The review was conducted only on English-language studies and relevant articles in other languages may have been missed. Furthermore, few of the included studies were considered to have a high level of evidence. This reflects the current paucity of available data, with no randomized controlled trials having been carried out. The small numbers in the live surgery and control groups further limit the patient safety studies. Finally, it was not possible to carry out formal meta-analysis of the results owing to the varying outcomes used across the studies. More research is required to ensure the value of live surgical teaching, and international Surgical Societies should be encouraged to lead on guiding and monitoring the safety of the procedures. Acknowledgements The authors thank K. Walsh for his invaluable contribution and suggestions in the process of writing the paper. Disclosure: P.D. and K.A. acknowledge educational research support from the Urology Foundation, Olympus and the Royal College of Surgeons of England. P.D. acknowledges support from the Vattikuti Foundation, MRC Centre for Transplantation and NIHR Biomedical Research Centre at King's College London. All other authors declare no conflict of interest. Open in new tabDownload slide References 1 Rao AR , Karim O. A benedictory ode to urological live surgery . BJU Int 2013 ; 112 : 11 – 12 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Gandsas A , Pleatman M, Altrud R, Migliarini G, Silva Y. Live broadcast of surgery through the Internet . Lancet 1996 : 348 : 1314 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Kallmes DF , Cloft HJ, Molyneux A, Burger I, Brinjikji W, Murphy KP. Live case demonstrations: patient safety, ethics, consent, and conflicts . Lancet 2011 ; 377 : 1539 – 1541 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Darbari A , Tandon S. Live surgery telecast – surgery teaching tool or sensationalism? Indian J Surg 2007 ; 69 : 171 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Moher D , Liberati A, Tetzlaff J, Altman DG; PRISMA Group . Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement . PLoS Med 2009 ; 6 : e1000097 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Ahmed K , Miskovic D, Darzi A, Athanasiou T, Hanna GB. Observational tools for assessment of procedural skills: a systematic review . Am J Surg 2011 ; 202 : 469 – 480 .e6. Google Scholar Crossref Search ADS PubMed WorldCat 7 Ahmed K , Ashrafian H, Hanna GB, Darzi A, Athanasiou T. Assessment of specialists in cardiovascular practice . Nat Rev Cardiol 2009 ; 6 : 659 – 667 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Van Der Vleuten CP . The assessment of professional competence: developments, research and practical implications . Adv Health Sci Educ Theory Pract 1996 ; 1 : 41 – 67 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Centre for Evidence-based Medicine . The Oxford Levels of Evidence 1 . http://www.cebm.net/?o=1025 [accessed 26 September 2013]. 10 Gandsas A , McIntire K, Palli G, Park A. Live streaming video for medical education: a laboratory model . J Laparoendosc Adv Surg Tech A 2002 ; 12 : 377 – 382 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Gandsas A , McIntire K, Park A. Live broadcast of laparoscopic surgery to handheld computers . Surg Endosc 2004 ; 18 : 997 – 1000 . Google Scholar Crossref Search ADS PubMed WorldCat 12 McIntyre TP , Monahan TS, Villegas L, Doyle J, Jones DB. Teleconferencing surgery enhances effective communication and enriches medical education . Surg Laparosc Endosc Percutan Tech 2008 ; 18 : 45 – 48 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Elsamra SE , Fakhoury M, Motato H, Friedlander JI, Moreira DM, Hillelsohn J et al. The surgical spectacle: a survey of urologists viewing live case demonstrations . BJU Int 2014 ; 113 : 674 – 678 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Gul YA , Wan AC, Darzi A. Undergraduate surgical teaching utilizing telemedicine . Med Educ 1999 ; 33 : 596 – 599 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Schmit A , Lazaraki G, Hittelet A, Cremer M, Le Moine O, Devière J. Complications of endoscopic retrograde cholangiopancreatography during live endoscopy workshop demonstrations . Endoscopy 2005 ; 37 : 695 – 699 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Liao Z , Li ZS, Leung JW, Zhang X, Zhang ST, Ji M et al. China ERCP Live Demonstration and Education Study Group . How safe and successful are live demonstrations of therapeutic ERCP? A large multicenter study . Am J Gastroenterol 2009 ; 104 : 47 – 52 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Ridtitid W , Rerknimitr R, Treeprasertsuk S, Kongkam P, Khor CJ, Kullavanijaya P. Outcome of endoscopic retrograde cholangiopancreatography during live endoscopy demonstrations . Surg Endosc 2012 ; 26 : 1931 – 1938 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Franke J , Reimers B, Scarpa M, Span S, Thieme M, Wunderlich N et al. Complications of carotid stenting during live transmissions . JACC Cardiovasc Interv 2009 ; 2 : 887 – 891 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Eliyahu S , Roguin A, Kerner A, Boulos M, Lorber A, Halabi M et al. Patient safety and outcomes from live case demonstrations of interventional cardiology procedures . JACC Cardiovasc Interv 2012 ; 5 : 215 – 224 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Seeburger J , Diegeler A, Dossche K, Lange R, Mohr FW, Schreiber C et al. Live broadcasting in cardiac surgery does not increase the operative risk . Eur J Cardiothorac Surg 2011 ; 40 : 367 – 371 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 21 Mullins JK , Borofsky MS, Allaf ME, Bhayani S, Kaouk JH, Rogers CG et al. Live robotic surgery: are outcomes compromised? Urology 2012 ; 80 : 602 – 607 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Chatelain P , Meier B, de la Serna F, Moles V, Pande AK, Verine V et al. Success with coronary angioplasty as seen at demonstrations of procedure . Lancet 1992 ; 340 : 1202 – 1205 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Japanese Society for Cardiovascular Surgery . Guidelines to Live Presentations of Thoracic and Cardiovascular Surgery . http://jscvs.umin.ac.jp/eng/live.html [accessed 11 November 2013]. 24 European Association for Cardio-Thoracic Surgery . Proposed Guidelines to Live Presentations of Thoracic and Cardiovascular Surgery . http://www.eacts.org/media/96969/GUIDELINES-ON-LIVE-SURGERY-BY-ASCVTS_Aug07.pdf [accessed 13 November 2013]. 25 Sade RM ; American Association for Thoracic Surgery Ethics Committee; Society of Thoracic Surgeons Standards and Ethics Committee. Broadcast of surgical procedures as a teaching instrument in cardiothoracic surgery . J Thorac Cardiovasc Surg 2008 ; 136 : 273 – 277 . Google Scholar Crossref Search ADS PubMed WorldCat 26 European Association of Urology . EAU Policy Statement on Live Surgical Events . http://www.uroweb.org/fileadmin/livesurgery/FULL.pdf [accessed 13 November 2013]. 27 American Urological Association . Standard Operating Practices Live Surgery . http://www.auanet.org/common/pdf/about/AUA-Live-Surgery-SOP.pdf [accessed 9 November 2013]. 28 Royal Australasian College of Surgeons . Live Transmission of Surgery . http://www.surgeons.org/media/14504/POS_2010-02-25_Live_Transmission_of_Surgery.pdf [accessed 8 November 2013]. 29 Organisation Mondiale d'Endoscopie Digestive . OMED Recommendations for the Ethical Performance of Live Endoscopy Demonstrations . http://www.worldendo.org/assets/downloads/pdf/guidelines/omed_guideline_ethics.pdf [accessed 12 November 2013]. 30 Loren DE , Azar R, Charles RJ, Dumot JA, Farooq F, Gopal DV et al. ASGE CME Programs Committee . Updated guidelines for live endoscopy demonstrations . Gastrointest Endosc 2010 ; 71 : 1105 – 1107 . Google Scholar Crossref Search ADS PubMed WorldCat 31 Kruse A , Beilenhoff U, Axon AT; European Society of Gastrointestinal Endoscopy; European Society of Gastrointestinal Endoscopy Nurses and Associates . ESGE/ESGENA guideline for live demonstration courses . Endoscopy 2003 ; 35 : 781 – 784 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Dehmer GJ , Douglas JS, Abizaid Al, Berg JW, Day J, Hall R et al. SCAI/ACCF/HRS/ESC/SOLACI/APSIC statement on the use of live case demonstrations at cardiology meetings: assessments of the past and standards for the future . J Am Coll Cardiol 2010 ; 56 : 1267 – 1282 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Farb A , Zuckerman B. Live case demonstration of interventional cardiology procedures: a regulatory perspective . JACC Cardiovasc Interv 2012 ; 5 : 225 – 227 . Google Scholar Crossref Search ADS PubMed WorldCat 34 Royal College of Surgeons of England . Live Surgery Broadcasts Position Statement . http://www.rcseng.ac.uk/publications/docs/live-surgery-broadcasts-position-statement [accessed 12 November 2013]. 35 Khan SA , Chang RT, Ahmed K, Knoll T, Van Velthoven R, Challacombe B et al. Live surgical education: a perspective from the surgeons who perform it . BJU Int 2014 ; 114 : 151 – 158 . Google Scholar Crossref Search ADS PubMed WorldCat 36 Japan Times . Patient in ‘Live’ Surgery Showing Died Soon After . http://www.japantimes.co.jp/news/2007/06/07/news/patient-in-live-surgery-showing-died-soon-after#.U-tDLfldWQ0 [accessed 15 November 2013]. 37 Pereira BM , Calderan TR, Silva MT, Silva AC, Marttos AC, Fraga GP. Initial experience at a university teaching hospital from using telemedicine to promote education through video conferencing . Sao Paulo Med J 2012 ; 130 : 32 – 36 . Google Scholar Crossref Search ADS PubMed WorldCat 38 Roser F , Pfister G, Tatagiba M, Ebner FH. Live surgery in neurosurgical training courses: essential infrastructure and technical set-up . Acta Neurochir (Wien) 2013 ; 155 : 541 – 545 . Google Scholar Crossref Search ADS PubMed WorldCat 39 Altunrende F , Autorino R, Haber GP, Laydner H, White MA, Khanna R et al. Immediate impact of a robotic kidney surgery course on attendees practice patterns . Int J Med Robot 2011 ; 7 : 165 – 169 . Google Scholar Crossref Search ADS PubMed WorldCat 40 Guillonneau B Live surgical demonstration: is it worthwhile? Nat Clin Pract Urol 2007 ; 4 : 59 . Google Scholar Crossref Search ADS PubMed WorldCat 41 Duty B , Okhunov Z, Friedlander J, Okeke Z, Smith A. Live surgical demonstrations: an old, but increasingly controversial practice . Urology 2012 ; 79 : 1185 e1187 – e1111 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 42 Jones JW , McCullough LB. Operative simulcasts: patient's donations to surgeon's education . J Vasc Surg 2008 ; 47 : 476 – 477 . Google Scholar Crossref Search ADS PubMed WorldCat 43 Arora S , Sevdalis N, Nestel D, Woloshynowych M, Darzi A, Kneebone R. The impact of stress on surgical performance: a systematic review of the literature . Surgery 2010 ; 147 : 318 – 330 , 330 e311 – e316 . Google Scholar Crossref Search ADS PubMed WorldCat © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd

Wu, L M; Sankaran, S J; Plank, L D; Windsor, J A; Petrov, M S

doi: 10.1002/bjs.9665pmid: 25334028

Abstract Background The gut is implicated in the pathogenesis of acute pancreatitis but there is discrepancy between individual studies regarding the prevalence of gut barrier dysfunction in patients with acute pancreatitis. The aim of this study was to determine the prevalence of gut barrier dysfunction in acute pancreatitis, the effect of different co-variables, and changes in gut barrier function associated with the use of various therapeutic modalities. Methods A literature search was performed using PRISMA and MOOSE guidelines. Summary estimates were presented as pooled prevalence of gut barrier dysfunction and the associated 95 per cent c.i. Results A total of 44 prospective clinical studies were included in the systematic review, of which 18 studies were subjected to meta-analysis. The pooled prevalence of gut barrier dysfunction was 59 (95 per cent c.i. 48 to 70) per cent; the prevalence was not significantly affected by disease severity, timing of assessment after hospital admission or type of test used, but showed a statistically significant association with age. Overall, nine of 13 randomized clinical trials reported a significant improvement in gut barrier function following intervention compared with the control group, but only three of six studies that used standard enteral nutrition reported a statistically significant improvement in gut barrier function after intervention. Conclusion Gut barrier dysfunction is present in three of five patients with acute pancreatitis, and the prevalence is affected by patient age but not by disease severity. Clinical studies are needed to evaluate the effect of enteral nutrition on gut function in acute pancreatitis. Introduction The gastrointestinal tract has been considered an innocent bystander that is subjected to collateral injury in critical illness. The reflex splanchnic vasoconstriction to preserve the perfusion of vital organs results in ischaemic injury, and resuscitation promotes reperfusion injury1,2. These events may contribute to the systemic inflammatory response and eventual organ dysfunction3. A critical event with the injury to the intestine is the loss of barrier function, increased permeability, and translocation of luminal bacteria and toxins to the portal venous circulation and mesenteric lymph4. Pancreatic infection and organ failure are determinants of severity in acute pancreatitis5. Gut barrier dysfunction and increased bacterial translocation are implicated in the development of secondary infection, sepsis, multiple organ failure, and death in acute pancreatitis6. Studies have shown that microorganisms responsible for sepsis and pancreatic infection are generally enteric in origin2,7. Dysfunction of the gut barrier and the subsequent translocation of toxins, bacteria and enteric microflora into the portal venous and lymphatic systems may cause multiple organ dysfunction syndrome8. Many interventions have been investigated in patients with acute pancreatitis with the aim of improving clinical outcomes, but only enteral nutrition has been shown to produce clear clinical benefits in patients with acute pancreatitis as it reduces the risk of developing both pancreatic infections and multiple organ dysfunction syndrome9–12. Gut barrier dysfunction is characterized by damage to the gut epithelium and intestinal tight junctions, resulting in increased intestinal permeability7,13 and compromise of the protective role of the gut barrier14,15. However, the permeability pathways, the effect of interventions on gut barrier function, and the implications of gut barrier dysfunction in the pathophysiology of acute pancreatitis are complex and not well understood. Loss of gut barrier integrity and increased gut barrier dysfunction have been observed in patients with acute pancreatitis, but there is discrepancy in the prevalence of gut barrier dysfunction in clinical studies6,16–18. Various interventions aimed at preserving gut barrier function and integrity have been evaluated and are centred mainly on optimizing nutritional management in patients. Although the effect of enteral nutrition in reducing gut barrier dysfunction in animal studies has been coherent, there has been discrepancy amongst clinical studies on the effect of enteral nutrition in reducing gut barrier dysfunction and bacterial translocation19–23. Some of the issues with clinical studies that might have contributed to the conflicting evidence include the heterogeneity of tests used to assess gut barrier function, the variability in disease severity, and the variation in timing of assessment used in different clinical studies. The aim of the present study was to conduct a comprehensive systematic literature review of the best available evidence on gut barrier dysfunction in patients with acute pancreatitis, and to determine the prevalence of gut barrier dysfunction in patients with acute pancreatitis, the effect of different co-variables, as well as the impact of various therapeutic modalities. Methods The reporting of this study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement24 and the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) checklist25. Study identification and selection An electronic literature search of three major biomedical databases, PubMed, Embase and Scopus, was conducted from the earliest available date to 1 January 2014, with no geographical restrictions. The search strategy, by database, is provided in Appendix S1 (supporting information). Titles and abstracts were screened for relevance to the study. Full-text articles were obtained for all relevant studies and were appraised for eligibility of inclusion. Study identification, screening and selection were conducted independently by two reviewers, and any discrepancies were resolved by discussion with the senior author. Eligibility criteria for study selection For a study to be included, the following criteria needed to be met: design (prospective observational or interventional study); population (adult patients with acute pancreatitis, aged at least 18 years); outcome (gut barrier function assessed). A study was excluded if the full-text article was not in English or the sample size was fewer than ten patients. Studies were also excluded if they used blood cultures as the only method of assessing gut barrier function. Data extraction Qualitative and quantitative data from all the included studies were extracted independently by two reviewers and entered into a standard data extraction form. The following data were extracted for each study: author(s); year of publication; country; study design; number of patients with acute pancreatitis; severity of acute pancreatitis; criteria for assessment of disease severity; test(s) of gut barrier function used; number of patients undergoing test; number of patients with gut barrier dysfunction; duration of disease onset before hospital admission; timing of assessment of gut barrier function after hospital admission; sex; age; body mass index; aetiology (alcohol, biliary, other); intervention(s); significance of interventions. For extraction of data on gut barrier dysfunction, only data from discrete individuals were included, and data from repeated measures taken from the same patient were excluded. For studies with sequential assessment of gut barrier function, only the timing of the earliest assessment was extracted, unless prevalence results were combined over sequential assessments. Quality assessment Methodological quality of the included studies was assessed independently by two reviewers using the Newcastle–Ottawa scale26. Quality assessment was stratified by study design, and an interventional study was assessed using similar criteria as a prospective cohort study, owing to the nature of the extracted data from the interventional studies. Cohort studies were assessed on the following categories: selection of the cohorts (0–4 points), comparability of the cohorts (0–2 points), and outcome assessment (0–3 points). Case–control studies were assessed on the following categories: selection of the cases and controls (0–4 points), comparability of cases and controls (0–2 points), and ascertainment of the exposure (0–3 points). Each study could be awarded a maximum of nine points, and studies that were awarded seven or more points were rated as having high methodological quality27,28. Data synthesis and statistical analysis The pooled prevalence of gut barrier dysfunction was the primary outcome of interest in the meta-analysis. Data from the included studies were combined to determine the pooled prevalence of gut barrier dysfunction and the associated 95 per cent c.i. The medical statistical software StatsDirect version 3.0.97 was used29. When studies used the same patient population, only the study that reported the most conservative result was included. For studies that used more than one test to assess gut barrier function, or used sequential assessment of gut barrier function, the most conservative result was included in the pooled analysis. For interventional studies, only baseline data (before commencement of the intervention) were included in the pooled analysis of prevalence. The effect of interventions was assessed by the comparison of gut barrier function in the intervention group versus the control group postintervention. Subgroup analyses were performed according to disease severity (mild versus severe), study location and study design (cohort versus case–control), where randomized clinical trials were analysed as cohort studies. Sensitivity analyses constrained to assessment of gut barrier function within 24, 48 and 72 h of hospital admission, constrained to particular criteria for severity assessment, and constrained to assessment of gut barrier dysfunction using a particular test were conducted (if data were available from 2 or more studies). If a study used more than one test of gut barrier function, prevalence data from all the reported tests were included in sensitivity analyses constrained to different tests used. The different tests used to assess gut barrier function were classified into three major groups: assessments of gut epithelial barrier integrity; functional assessments of the gut barrier; and assessments of bacterial translocation14. Assessments of gut epithelial barrier integrity were subclassified into tests of enterocyte damage or paracellular barrier integrity loss14. Functional assessments of the gut barrier were subclassified into tests that utilize active or passive measurements14. Statistical heterogeneity between the studies was assessed using the Cochran's Q for statistical significance30 and the I2 statistic for quantifying heterogeneity30,31. Statistical significance was set at P < 0·100 using Cochran's Q, and low, moderate and high statistical heterogeneity was set at I2 values of 25, 50 and 75 per cent respectively30. The analysis was performed using the DerSimonian–Laird random-effects method to yield the most conservative results32,33. Potential for publication bias was assessed by visual inspection of funnel plot assymetry and by using the Begg34, Egger35 and Harbord36 tests, with statistical significance set at P < 0·05037. Metaregression analyses were performed using the metafor package in R 3.0.2 statistical software (http://www.metafor-project.org/doku.php) to investigate the possible confounding effect of the following factors: sex, age, biliary aetiology and disease severity33,38. Statistical significance in the metaregression analyses was set at P < 0·10039,40. A random-effects model was used for the analyses to yield the most conservative results33. Each co-variable was fitted both univariably into individual models and also combined into a single model, and the results from the two methods were compared33,40. For the metaregression analyses, only the mean age was used and, where data were available, this was estimated using the values of the median, low end and high end of the range41. If sample size data were available, the mean ages of different patient groups within a study were combined to give a single mean for the metaregression analyses. Results A total of 3584 records were identified through the initial literature search and 90 potentially relevant full-text articles were retrieved and assessed for eligibility (Fig. 1). Forty-six articles were excluded at this stage, yielding a total of 44 studies6,12,16–18,21–23,42–77 that were included in the systematic review. There were 13 randomized clinical trials and 31 observational studies, which included both prospective cohort and case–control studies (Table 1). Twenty-seven studies were conducted in Europe, and the remaining 17 studies in Asia. The 44 studies encompassed a total of 2611 patients with acute pancreatitis, of whom 2094 patients underwent assessment of gut barrier function. The baseline characteristics of individuals in the included studies are presented in Table S1 (supporting information). Fig. 1 Open in new tabDownload slide Flow diagram of the study selection process Table 1 Characteristics of included studies Reference . Year . Country . Study design . No. of patients with acute pancreatitis . No. of patients undergoing test . Tests of gut barrier function used . Timing of assessment of gut barrier function . Foulis et al.42 1982 UK Prospective cohort 24 24 Endotoxin* Daily¶ Beger et al.43 1988 Germany Prospective cohort 95 30 Endotoxin Preoperative Exley et al.44 1992 UK Prospective cohort 38 37 Endotoxin ≤ 24 h¶ Curley et al.45 1993 UK Prospective case–control 29 28 EndoCAb, endotoxin ≤ 48 h¶ Windsor et al.46 1993 UK Prospective cohort 33 33 EndoCAb, endotoxin ≤ 24 h¶ Wig et al.47 1998 India Prospective cohort 14 14 Endotoxin Consecutively over 2 or 4 days¶ Windsor et al.12 1998 UK RCT 34 34 EndoCAb ≤ 48 h¶ Ammori et al.48 1999 UK Prospective case–control 85 64 EndoCAb, endotoxin, PEGs ≤ 24 h¶ Soong et al.49 1999 UK Prospective cohort 19 19 EndoCAb, endotoxin ≤ 12 h¶ Buttenschoen et al.50 2000 Germany Prospective case–control 25 25 EndoCAb, endotoxin ≤ 24 h¶ Juvonen et al.16 2000 Finland Prospective case–control 23 23 L/R, multisugar probes ≤ 48 h¶ Powell et al.51 2000 UK RCT 27 22 EndoCAb, L/R ≤ 24 h¶ Zhang et al.52 2001 China Prospective case–control 13 13 DNA/RNA† n.r. Bose et al.53 2002 India Prospective cohort 20 20 EndoCAb, endotoxin ≤ 72 h¶ Martínez et al.54 2002 Spain Prospective cohort 19 19 EndoCAb, endotoxin ≤ 24 h¶ McNaught et al.55 2002 UK Prospective cohort 59 27 L/R ≤ 72 h¶ Ammori et al.56 2003 UK Prospective cohort  72‡ 72 EndoCAb, endotoxin ≤ 24 h¶ Ammori et al.57 2003 UK Prospective cohort  60‡ 52 EndoCAb, endotoxin, PEGs ≤ 24 h¶ Ammori et al.17 2003 UK Prospective case–control 26 26§ EndoCAb, endotoxin, DNA/RNA, PEGs ≤ 24 h¶ Giamarellos-Bourboulis et al.18 2003 Greece Prospective cohort 33 33 Endotoxin ≤ 24 h¶ Gupta et al.22 2003 UK RCT 17 17 EndoCAb ≤ 24 h¶ Rahman et al.58 2003 UK Prospective case–control  61‡ 61 EndoCAb, I-FABP, PEGs ≤ 72 h# Rahman et al.59 2003 UK Prospective case–control  65‡ 65 EndoCAb, NO, PEGs ≤ 48 h# Zhao et al.60 2003 China RCT 96 96 Endotoxin, L/M ≤ 24 h¶ Penalva et al.61 2004 Spain Prospective case–control 68 68 EndoCAb, L/M ≤ 72 h# Rahman et al.62 2004 UK Prospective case–control 320  46 Reduced GSH ≤ 24 h# De Madaria et al.63 2005 Spain Prospective cohort 31 31 DNA/RNA ≤ 48 h¶ Eckerwall et al.21 2006 Sweden RCT 48 40 EndoCAb, PEGs ≤ 24 h¶ Nagpal et al.64 2006 UK Prospective case–control 46 45 L/M Days 1, 4, 7¶ Liu et al.6 2008 China Prospective case–control 62 26 d-Xylose, endotoxin, L/M ≤ 72 h# Qin et al.65 2008 China RCT 74 71 L/R ≤ 24 h¶ Besselink et al.66 2009 Netherlands RCT 141  115  I-FABP, NO, PEGs ≤ 72 h¶ Chen et al.67 2010 China RCT 40 40 Endotoxin, L/M ≤ 24 h¶ Pan et al.68 2010 China Prospective cohort 32 32 I-FABP, L/M ≤ 24 h¶ Zhang et al.69 2010 China RCT 63 63 Endotoxin, F-actin, serum DAO Preintervention Sharma et al.70 2011 India RCT 50 40 EndoCAb, L/M ≤ 24 h¶ Shen et al.71 2011 China Prospective cohort 72 72 Endotoxin ≤ 24 h¶ Koh et al.72 2012 Korea Prospective cohort 87 87 Endotoxin, L/M ≤ 24 h¶ Pynnönen et al.73 2012 Finland Prospective cohort 30 27 Rectal luminal lactate ≤ 24 h¶ Sharma et al.74 2012 India Prospective case–control 31 31 EndoCAb, L/M ≤ 24 h¶ Singh et al.23 2012 India RCT 78 55 EndoCAb, L/M ≤ 48 h¶ Li et al.75 2013 China Prospective case–control 48 48 DNA/RNA Days 4, 5** Wang et al.76 2013 China RCT 183  183  Endotoxin Preintervention Zhao et al.77 2013 China RCT 120  120  Endotoxin, L/M ≤ 24 h¶ Reference . Year . Country . Study design . No. of patients with acute pancreatitis . No. of patients undergoing test . Tests of gut barrier function used . Timing of assessment of gut barrier function . Foulis et al.42 1982 UK Prospective cohort 24 24 Endotoxin* Daily¶ Beger et al.43 1988 Germany Prospective cohort 95 30 Endotoxin Preoperative Exley et al.44 1992 UK Prospective cohort 38 37 Endotoxin ≤ 24 h¶ Curley et al.45 1993 UK Prospective case–control 29 28 EndoCAb, endotoxin ≤ 48 h¶ Windsor et al.46 1993 UK Prospective cohort 33 33 EndoCAb, endotoxin ≤ 24 h¶ Wig et al.47 1998 India Prospective cohort 14 14 Endotoxin Consecutively over 2 or 4 days¶ Windsor et al.12 1998 UK RCT 34 34 EndoCAb ≤ 48 h¶ Ammori et al.48 1999 UK Prospective case–control 85 64 EndoCAb, endotoxin, PEGs ≤ 24 h¶ Soong et al.49 1999 UK Prospective cohort 19 19 EndoCAb, endotoxin ≤ 12 h¶ Buttenschoen et al.50 2000 Germany Prospective case–control 25 25 EndoCAb, endotoxin ≤ 24 h¶ Juvonen et al.16 2000 Finland Prospective case–control 23 23 L/R, multisugar probes ≤ 48 h¶ Powell et al.51 2000 UK RCT 27 22 EndoCAb, L/R ≤ 24 h¶ Zhang et al.52 2001 China Prospective case–control 13 13 DNA/RNA† n.r. Bose et al.53 2002 India Prospective cohort 20 20 EndoCAb, endotoxin ≤ 72 h¶ Martínez et al.54 2002 Spain Prospective cohort 19 19 EndoCAb, endotoxin ≤ 24 h¶ McNaught et al.55 2002 UK Prospective cohort 59 27 L/R ≤ 72 h¶ Ammori et al.56 2003 UK Prospective cohort  72‡ 72 EndoCAb, endotoxin ≤ 24 h¶ Ammori et al.57 2003 UK Prospective cohort  60‡ 52 EndoCAb, endotoxin, PEGs ≤ 24 h¶ Ammori et al.17 2003 UK Prospective case–control 26 26§ EndoCAb, endotoxin, DNA/RNA, PEGs ≤ 24 h¶ Giamarellos-Bourboulis et al.18 2003 Greece Prospective cohort 33 33 Endotoxin ≤ 24 h¶ Gupta et al.22 2003 UK RCT 17 17 EndoCAb ≤ 24 h¶ Rahman et al.58 2003 UK Prospective case–control  61‡ 61 EndoCAb, I-FABP, PEGs ≤ 72 h# Rahman et al.59 2003 UK Prospective case–control  65‡ 65 EndoCAb, NO, PEGs ≤ 48 h# Zhao et al.60 2003 China RCT 96 96 Endotoxin, L/M ≤ 24 h¶ Penalva et al.61 2004 Spain Prospective case–control 68 68 EndoCAb, L/M ≤ 72 h# Rahman et al.62 2004 UK Prospective case–control 320  46 Reduced GSH ≤ 24 h# De Madaria et al.63 2005 Spain Prospective cohort 31 31 DNA/RNA ≤ 48 h¶ Eckerwall et al.21 2006 Sweden RCT 48 40 EndoCAb, PEGs ≤ 24 h¶ Nagpal et al.64 2006 UK Prospective case–control 46 45 L/M Days 1, 4, 7¶ Liu et al.6 2008 China Prospective case–control 62 26 d-Xylose, endotoxin, L/M ≤ 72 h# Qin et al.65 2008 China RCT 74 71 L/R ≤ 24 h¶ Besselink et al.66 2009 Netherlands RCT 141  115  I-FABP, NO, PEGs ≤ 72 h¶ Chen et al.67 2010 China RCT 40 40 Endotoxin, L/M ≤ 24 h¶ Pan et al.68 2010 China Prospective cohort 32 32 I-FABP, L/M ≤ 24 h¶ Zhang et al.69 2010 China RCT 63 63 Endotoxin, F-actin, serum DAO Preintervention Sharma et al.70 2011 India RCT 50 40 EndoCAb, L/M ≤ 24 h¶ Shen et al.71 2011 China Prospective cohort 72 72 Endotoxin ≤ 24 h¶ Koh et al.72 2012 Korea Prospective cohort 87 87 Endotoxin, L/M ≤ 24 h¶ Pynnönen et al.73 2012 Finland Prospective cohort 30 27 Rectal luminal lactate ≤ 24 h¶ Sharma et al.74 2012 India Prospective case–control 31 31 EndoCAb, L/M ≤ 24 h¶ Singh et al.23 2012 India RCT 78 55 EndoCAb, L/M ≤ 48 h¶ Li et al.75 2013 China Prospective case–control 48 48 DNA/RNA Days 4, 5** Wang et al.76 2013 China RCT 183  183  Endotoxin Preintervention Zhao et al.77 2013 China RCT 120  120  Endotoxin, L/M ≤ 24 h¶ * Serum endotoxin; † bacterial DNA or RNA by PCR; ‡ patients drawn from a similar patient group; § disregarded results from PCR owing to substances inhibitory to the amplification of DNA in two-fifths of patients; ¶ after hospital admission; # after onset of acute pancreatitis; ** after diagnosis of acute pancreatitis. EndoCAb, endogenous antiendotoxin core antibodies; RCT, randomized clinical trial; PEGs, polyethylene glycols; L/R, lactulose/rhamnose ratio; n.r., not reported; I-FABP, intestinal fatty acid-binding protein; NO, nitric oxide; L/M, lactulose/mannitol ratio; GSH, reduced glutathione; DAO, diamine oxidase. Open in new tab Table 1 Characteristics of included studies Reference . Year . Country . Study design . No. of patients with acute pancreatitis . No. of patients undergoing test . Tests of gut barrier function used . Timing of assessment of gut barrier function . Foulis et al.42 1982 UK Prospective cohort 24 24 Endotoxin* Daily¶ Beger et al.43 1988 Germany Prospective cohort 95 30 Endotoxin Preoperative Exley et al.44 1992 UK Prospective cohort 38 37 Endotoxin ≤ 24 h¶ Curley et al.45 1993 UK Prospective case–control 29 28 EndoCAb, endotoxin ≤ 48 h¶ Windsor et al.46 1993 UK Prospective cohort 33 33 EndoCAb, endotoxin ≤ 24 h¶ Wig et al.47 1998 India Prospective cohort 14 14 Endotoxin Consecutively over 2 or 4 days¶ Windsor et al.12 1998 UK RCT 34 34 EndoCAb ≤ 48 h¶ Ammori et al.48 1999 UK Prospective case–control 85 64 EndoCAb, endotoxin, PEGs ≤ 24 h¶ Soong et al.49 1999 UK Prospective cohort 19 19 EndoCAb, endotoxin ≤ 12 h¶ Buttenschoen et al.50 2000 Germany Prospective case–control 25 25 EndoCAb, endotoxin ≤ 24 h¶ Juvonen et al.16 2000 Finland Prospective case–control 23 23 L/R, multisugar probes ≤ 48 h¶ Powell et al.51 2000 UK RCT 27 22 EndoCAb, L/R ≤ 24 h¶ Zhang et al.52 2001 China Prospective case–control 13 13 DNA/RNA† n.r. Bose et al.53 2002 India Prospective cohort 20 20 EndoCAb, endotoxin ≤ 72 h¶ Martínez et al.54 2002 Spain Prospective cohort 19 19 EndoCAb, endotoxin ≤ 24 h¶ McNaught et al.55 2002 UK Prospective cohort 59 27 L/R ≤ 72 h¶ Ammori et al.56 2003 UK Prospective cohort  72‡ 72 EndoCAb, endotoxin ≤ 24 h¶ Ammori et al.57 2003 UK Prospective cohort  60‡ 52 EndoCAb, endotoxin, PEGs ≤ 24 h¶ Ammori et al.17 2003 UK Prospective case–control 26 26§ EndoCAb, endotoxin, DNA/RNA, PEGs ≤ 24 h¶ Giamarellos-Bourboulis et al.18 2003 Greece Prospective cohort 33 33 Endotoxin ≤ 24 h¶ Gupta et al.22 2003 UK RCT 17 17 EndoCAb ≤ 24 h¶ Rahman et al.58 2003 UK Prospective case–control  61‡ 61 EndoCAb, I-FABP, PEGs ≤ 72 h# Rahman et al.59 2003 UK Prospective case–control  65‡ 65 EndoCAb, NO, PEGs ≤ 48 h# Zhao et al.60 2003 China RCT 96 96 Endotoxin, L/M ≤ 24 h¶ Penalva et al.61 2004 Spain Prospective case–control 68 68 EndoCAb, L/M ≤ 72 h# Rahman et al.62 2004 UK Prospective case–control 320  46 Reduced GSH ≤ 24 h# De Madaria et al.63 2005 Spain Prospective cohort 31 31 DNA/RNA ≤ 48 h¶ Eckerwall et al.21 2006 Sweden RCT 48 40 EndoCAb, PEGs ≤ 24 h¶ Nagpal et al.64 2006 UK Prospective case–control 46 45 L/M Days 1, 4, 7¶ Liu et al.6 2008 China Prospective case–control 62 26 d-Xylose, endotoxin, L/M ≤ 72 h# Qin et al.65 2008 China RCT 74 71 L/R ≤ 24 h¶ Besselink et al.66 2009 Netherlands RCT 141  115  I-FABP, NO, PEGs ≤ 72 h¶ Chen et al.67 2010 China RCT 40 40 Endotoxin, L/M ≤ 24 h¶ Pan et al.68 2010 China Prospective cohort 32 32 I-FABP, L/M ≤ 24 h¶ Zhang et al.69 2010 China RCT 63 63 Endotoxin, F-actin, serum DAO Preintervention Sharma et al.70 2011 India RCT 50 40 EndoCAb, L/M ≤ 24 h¶ Shen et al.71 2011 China Prospective cohort 72 72 Endotoxin ≤ 24 h¶ Koh et al.72 2012 Korea Prospective cohort 87 87 Endotoxin, L/M ≤ 24 h¶ Pynnönen et al.73 2012 Finland Prospective cohort 30 27 Rectal luminal lactate ≤ 24 h¶ Sharma et al.74 2012 India Prospective case–control 31 31 EndoCAb, L/M ≤ 24 h¶ Singh et al.23 2012 India RCT 78 55 EndoCAb, L/M ≤ 48 h¶ Li et al.75 2013 China Prospective case–control 48 48 DNA/RNA Days 4, 5** Wang et al.76 2013 China RCT 183  183  Endotoxin Preintervention Zhao et al.77 2013 China RCT 120  120  Endotoxin, L/M ≤ 24 h¶ Reference . Year . Country . Study design . No. of patients with acute pancreatitis . No. of patients undergoing test . Tests of gut barrier function used . Timing of assessment of gut barrier function . Foulis et al.42 1982 UK Prospective cohort 24 24 Endotoxin* Daily¶ Beger et al.43 1988 Germany Prospective cohort 95 30 Endotoxin Preoperative Exley et al.44 1992 UK Prospective cohort 38 37 Endotoxin ≤ 24 h¶ Curley et al.45 1993 UK Prospective case–control 29 28 EndoCAb, endotoxin ≤ 48 h¶ Windsor et al.46 1993 UK Prospective cohort 33 33 EndoCAb, endotoxin ≤ 24 h¶ Wig et al.47 1998 India Prospective cohort 14 14 Endotoxin Consecutively over 2 or 4 days¶ Windsor et al.12 1998 UK RCT 34 34 EndoCAb ≤ 48 h¶ Ammori et al.48 1999 UK Prospective case–control 85 64 EndoCAb, endotoxin, PEGs ≤ 24 h¶ Soong et al.49 1999 UK Prospective cohort 19 19 EndoCAb, endotoxin ≤ 12 h¶ Buttenschoen et al.50 2000 Germany Prospective case–control 25 25 EndoCAb, endotoxin ≤ 24 h¶ Juvonen et al.16 2000 Finland Prospective case–control 23 23 L/R, multisugar probes ≤ 48 h¶ Powell et al.51 2000 UK RCT 27 22 EndoCAb, L/R ≤ 24 h¶ Zhang et al.52 2001 China Prospective case–control 13 13 DNA/RNA† n.r. Bose et al.53 2002 India Prospective cohort 20 20 EndoCAb, endotoxin ≤ 72 h¶ Martínez et al.54 2002 Spain Prospective cohort 19 19 EndoCAb, endotoxin ≤ 24 h¶ McNaught et al.55 2002 UK Prospective cohort 59 27 L/R ≤ 72 h¶ Ammori et al.56 2003 UK Prospective cohort  72‡ 72 EndoCAb, endotoxin ≤ 24 h¶ Ammori et al.57 2003 UK Prospective cohort  60‡ 52 EndoCAb, endotoxin, PEGs ≤ 24 h¶ Ammori et al.17 2003 UK Prospective case–control 26 26§ EndoCAb, endotoxin, DNA/RNA, PEGs ≤ 24 h¶ Giamarellos-Bourboulis et al.18 2003 Greece Prospective cohort 33 33 Endotoxin ≤ 24 h¶ Gupta et al.22 2003 UK RCT 17 17 EndoCAb ≤ 24 h¶ Rahman et al.58 2003 UK Prospective case–control  61‡ 61 EndoCAb, I-FABP, PEGs ≤ 72 h# Rahman et al.59 2003 UK Prospective case–control  65‡ 65 EndoCAb, NO, PEGs ≤ 48 h# Zhao et al.60 2003 China RCT 96 96 Endotoxin, L/M ≤ 24 h¶ Penalva et al.61 2004 Spain Prospective case–control 68 68 EndoCAb, L/M ≤ 72 h# Rahman et al.62 2004 UK Prospective case–control 320  46 Reduced GSH ≤ 24 h# De Madaria et al.63 2005 Spain Prospective cohort 31 31 DNA/RNA ≤ 48 h¶ Eckerwall et al.21 2006 Sweden RCT 48 40 EndoCAb, PEGs ≤ 24 h¶ Nagpal et al.64 2006 UK Prospective case–control 46 45 L/M Days 1, 4, 7¶ Liu et al.6 2008 China Prospective case–control 62 26 d-Xylose, endotoxin, L/M ≤ 72 h# Qin et al.65 2008 China RCT 74 71 L/R ≤ 24 h¶ Besselink et al.66 2009 Netherlands RCT 141  115  I-FABP, NO, PEGs ≤ 72 h¶ Chen et al.67 2010 China RCT 40 40 Endotoxin, L/M ≤ 24 h¶ Pan et al.68 2010 China Prospective cohort 32 32 I-FABP, L/M ≤ 24 h¶ Zhang et al.69 2010 China RCT 63 63 Endotoxin, F-actin, serum DAO Preintervention Sharma et al.70 2011 India RCT 50 40 EndoCAb, L/M ≤ 24 h¶ Shen et al.71 2011 China Prospective cohort 72 72 Endotoxin ≤ 24 h¶ Koh et al.72 2012 Korea Prospective cohort 87 87 Endotoxin, L/M ≤ 24 h¶ Pynnönen et al.73 2012 Finland Prospective cohort 30 27 Rectal luminal lactate ≤ 24 h¶ Sharma et al.74 2012 India Prospective case–control 31 31 EndoCAb, L/M ≤ 24 h¶ Singh et al.23 2012 India RCT 78 55 EndoCAb, L/M ≤ 48 h¶ Li et al.75 2013 China Prospective case–control 48 48 DNA/RNA Days 4, 5** Wang et al.76 2013 China RCT 183  183  Endotoxin Preintervention Zhao et al.77 2013 China RCT 120  120  Endotoxin, L/M ≤ 24 h¶ * Serum endotoxin; † bacterial DNA or RNA by PCR; ‡ patients drawn from a similar patient group; § disregarded results from PCR owing to substances inhibitory to the amplification of DNA in two-fifths of patients; ¶ after hospital admission; # after onset of acute pancreatitis; ** after diagnosis of acute pancreatitis. EndoCAb, endogenous antiendotoxin core antibodies; RCT, randomized clinical trial; PEGs, polyethylene glycols; L/R, lactulose/rhamnose ratio; n.r., not reported; I-FABP, intestinal fatty acid-binding protein; NO, nitric oxide; L/M, lactulose/mannitol ratio; GSH, reduced glutathione; DAO, diamine oxidase. Open in new tab Assessment of methodological quality and publication bias Overall, 42 of the 44 studies were of high methodological quality (Table S2, supporting information). The most common reasons for not scoring a point were potential selection biases and inadequate control for possible confounding factors, such as timing of assessment, severity of acute pancreatitis and age. There was no evidence of publication bias by visual inspection of funnel plot assymetry, and by using the Begg (P = 0·822), Egger (P = 0·986) or Harbord (P = 0·220) test. Tests used to assess gut barrier function A total of 14 different tests were used to assess gut barrier function in the included studies (Table 2). Of the 14 tests used, three assessed gut epithelial barrier integrity, seven were functional assessments of the gut barrier and four assessed bacterial translocation. Of the 44 included studies, five assessed gut epithelial barrier integrity, 37 studies utilized functional assessments of the gut barrier and 25 assessed bacterial translocation (Table 1). Twenty-three of the 44 included studies measured endotoxins, 20 measured endogenous antiendotoxin core antibodies and 11 measured the lactulose/mannitol ratio; these were the three most frequently used tests. Each study used between one and four tests to assess gut barrier function, with a median of 2 tests per study. The 44 studies spanned a total of three decades. Fig. 2 depicts the total frequency of tests used within each 5-year interval, as well as the frequency within each of the three major groups of tests. Table 2 Tests used to assess gut barrier function Classification . Subclassification . Tests . Frequency of tests used . Study design* . Pooled prevalence of gut barrier dysfunction (%)† . Heterogeneity . Assessments of gut epithelial barrier integrity Enterocyte damage I-FABP 3 Observational (2) n.e. Interventional (1) Reduced GSH 1 Observational (1) n.e. Paracellular barrier integrity loss F-actin 1 Interventional (1) n.e. Functional assessments of gut barrier Active measurements L/M ratio 11  Observational (6) 79 (61, 92) (2) P = 0·091 Interventional (5) PEGs 7 Observational (5) n.e. Interventional (2) L/R ratio 4 Observational (2) 66 (44, 84) (2) P = 0·116 Interventional (2) d-Xylose 1 Observational (1) n.e. Multisugar probes 1 Observational (1) n.e. Serum DAO 1 Interventional (1) n.e. Passive measurements Endotoxins 23  Observational (18)  57 (43, 71) (11) I2 = 88%, P < 0·001 Interventional (5) Assessments of bacterial translocation EndoCAb 20  Observational (14) 56 (15, 93) (2) P < 0·001 Interventional (6) DNA/RNA‡ 4 Observational (4) 46 (17, 77) (3) I2 = 89%, P < 0·001 NO 2 Observational (1) n.e. Interventional (1) Rectal luminal lactate 1 Observational (1) n.e. Classification . Subclassification . Tests . Frequency of tests used . Study design* . Pooled prevalence of gut barrier dysfunction (%)† . Heterogeneity . Assessments of gut epithelial barrier integrity Enterocyte damage I-FABP 3 Observational (2) n.e. Interventional (1) Reduced GSH 1 Observational (1) n.e. Paracellular barrier integrity loss F-actin 1 Interventional (1) n.e. Functional assessments of gut barrier Active measurements L/M ratio 11  Observational (6) 79 (61, 92) (2) P = 0·091 Interventional (5) PEGs 7 Observational (5) n.e. Interventional (2) L/R ratio 4 Observational (2) 66 (44, 84) (2) P = 0·116 Interventional (2) d-Xylose 1 Observational (1) n.e. Multisugar probes 1 Observational (1) n.e. Serum DAO 1 Interventional (1) n.e. Passive measurements Endotoxins 23  Observational (18)  57 (43, 71) (11) I2 = 88%, P < 0·001 Interventional (5) Assessments of bacterial translocation EndoCAb 20  Observational (14) 56 (15, 93) (2) P < 0·001 Interventional (6) DNA/RNA‡ 4 Observational (4) 46 (17, 77) (3) I2 = 89%, P < 0·001 NO 2 Observational (1) n.e. Interventional (1) Rectal luminal lactate 1 Observational (1) n.e. * Values in parentheses are number of studies; † values in parentheses are 95 per cent c.i. followed by number of studies. ‡ Bacterial DNA or RNA by PCR. I-FABP, intestinal fatty acid-binding protein; n.e., not estimable; GSH, reduced glutathione; L/M, lactulose/mannitol; PEGs, polyethylene glycols; L/R, lactulose/rhamnose; DAO, diamine oxidase; EndoCAb, endogenous antiendotoxin core antibodies; NO, nitric oxide. Open in new tab Table 2 Tests used to assess gut barrier function Classification . Subclassification . Tests . Frequency of tests used . Study design* . Pooled prevalence of gut barrier dysfunction (%)† . Heterogeneity . Assessments of gut epithelial barrier integrity Enterocyte damage I-FABP 3 Observational (2) n.e. Interventional (1) Reduced GSH 1 Observational (1) n.e. Paracellular barrier integrity loss F-actin 1 Interventional (1) n.e. Functional assessments of gut barrier Active measurements L/M ratio 11  Observational (6) 79 (61, 92) (2) P = 0·091 Interventional (5) PEGs 7 Observational (5) n.e. Interventional (2) L/R ratio 4 Observational (2) 66 (44, 84) (2) P = 0·116 Interventional (2) d-Xylose 1 Observational (1) n.e. Multisugar probes 1 Observational (1) n.e. Serum DAO 1 Interventional (1) n.e. Passive measurements Endotoxins 23  Observational (18)  57 (43, 71) (11) I2 = 88%, P < 0·001 Interventional (5) Assessments of bacterial translocation EndoCAb 20  Observational (14) 56 (15, 93) (2) P < 0·001 Interventional (6) DNA/RNA‡ 4 Observational (4) 46 (17, 77) (3) I2 = 89%, P < 0·001 NO 2 Observational (1) n.e. Interventional (1) Rectal luminal lactate 1 Observational (1) n.e. Classification . Subclassification . Tests . Frequency of tests used . Study design* . Pooled prevalence of gut barrier dysfunction (%)† . Heterogeneity . Assessments of gut epithelial barrier integrity Enterocyte damage I-FABP 3 Observational (2) n.e. Interventional (1) Reduced GSH 1 Observational (1) n.e. Paracellular barrier integrity loss F-actin 1 Interventional (1) n.e. Functional assessments of gut barrier Active measurements L/M ratio 11  Observational (6) 79 (61, 92) (2) P = 0·091 Interventional (5) PEGs 7 Observational (5) n.e. Interventional (2) L/R ratio 4 Observational (2) 66 (44, 84) (2) P = 0·116 Interventional (2) d-Xylose 1 Observational (1) n.e. Multisugar probes 1 Observational (1) n.e. Serum DAO 1 Interventional (1) n.e. Passive measurements Endotoxins 23  Observational (18)  57 (43, 71) (11) I2 = 88%, P < 0·001 Interventional (5) Assessments of bacterial translocation EndoCAb 20  Observational (14) 56 (15, 93) (2) P < 0·001 Interventional (6) DNA/RNA‡ 4 Observational (4) 46 (17, 77) (3) I2 = 89%, P < 0·001 NO 2 Observational (1) n.e. Interventional (1) Rectal luminal lactate 1 Observational (1) n.e. * Values in parentheses are number of studies; † values in parentheses are 95 per cent c.i. followed by number of studies. ‡ Bacterial DNA or RNA by PCR. I-FABP, intestinal fatty acid-binding protein; n.e., not estimable; GSH, reduced glutathione; L/M, lactulose/mannitol; PEGs, polyethylene glycols; L/R, lactulose/rhamnose; DAO, diamine oxidase; EndoCAb, endogenous antiendotoxin core antibodies; NO, nitric oxide. Open in new tab Fig. 2 Open in new tabDownload slide Tests of gut barrier dysfunction used in patients with acute pancreatitis over the past three decades Prevalence of gut barrier dysfunction Eighteen6,17–18,42–48,51–53,55,57,63–64,75 of the 44 studies reported data on the prevalence of gut barrier dysfunction in patients with acute pancreatitis, encompassing a total of 743 patients, 573 of whom had assessment of gut barrier function (Table 1). The pooled prevalence of gut barrier dysfunction in patients with acute pancreatitis, using the most conservative test result from each study, was 59 (95 per cent c.i. 48 to 70) per cent (Fig. 3), in the presence of high statistical heterogeneity (I2 = 87 per cent, P < 0·001). Fig. 3 Open in new tabDownload slide Forest plot showing prevalence of gut barrier dysfunction in patients with acute pancreatitis. Proportions with 95 per cent c.i. are shown A number of prespecified subgroup and sensitivity analyses were carried out to investigate the effect of possible confounders. Subgroup analysis of patients with mild acute pancreatitis included a total of 11 studies6,17,45–46,48,53,55,57,63–64,75 comprising 247 patients who underwent assessment of gut barrier function. The pooled prevalence of gut barrier dysfunction in patients with mild acute pancreatitis was 47 (95 per cent c.i. 30 to 64) per cent (Fig. S1A, supporting information), with no reduction in statistical heterogeneity (I2 = 87 per cent, P < 0·001). Twelve studies17,44–48,53,55,57,63–64,75, comprising a total of 204 patients who underwent assessment of gut barrier function, provided data for subgroup analysis of patients with severe acute pancreatitis. The pooled prevalence of gut barrier dysfunction in patients with severe acute pancreatitis was 59 (47 to 70) per cent (Fig. S1B, supporting information), in the presence of moderate statistical heterogeneity (I2 = 66 per cent, P < 0·001). For the studies that used functional assessments of the gut barrier, subgroup analysis of mild acute pancreatitis6,17,45–46,48,53,55,57,64 resulted in a pooled prevalence of gut barrier dysfunction of 51 (32 to 70) per cent, in the presence of high statistical heterogeneity (I2 = 83 per cent, P < 0·001), whereas severe acute pancreatitis17,44–48,53,55,57,64 resulted in a pooled prevalence of gut barrier dysfunction of 60 (47 to 72) per cent, in the presence of moderate statistical heterogeneity (I2 = 52 per cent, P = 0·028). For the studies that used assessments of bacterial translocation, subgroup analysis of mild acute pancreatitis45–46,63,75 resulted in a pooled prevelance of gut barrier dysfunction of 21 (10 to 40) per cent, in the presence of low statistical heterogeneity (I2 = 40 per cent, P = 0·174), whereas severe acute pancreatitis45–46,63,75 resulted in a pooled prevalence of gut barrier dysfunction of 41 (18 to 69) per cent, in the presence of high statistical heterogeneity (I2 = 79 per cent, P = 0·003). Subgroup analysis of cohort studies17–18,42–44,46–47,51,53,55,57,63 resulted in a pooled prevelance of gut barrier dysfunction of 61 (95 per cent c.i. 48 to 73) per cent, with no reduction in statistical heterogeneity (I2 = 84 per cent, P < 0·001). Subgroup analysis of case–control studies6,45,48,52,64,75 resulted in a pooled prevalence of gut barrier dysfunction of 54 (32 to 76) per cent, with no reduction in statistical heterogeneity (I2 = 91 per cent, P < 0·001). Subgroup analysis of studies performed in Europe resulted in a pooled prevalence of gut barrier dysfunction of 53 (41 to 66) per cent, with no reduction in statistical heterogeneity (I2 = 83 per cent, P < 0·001). Subgroup analysis of studies performed in Asia resulted in a pooled prevalence of gut barrier dysfunction of 74 (60 to 84) per cent, in the presence of low statistical heterogeneity (I2 = 46 per cent, P = 0·114). Sensitivity analyses constrained to assessment of gut barrier function within 24 h17–18,44,46,48,51,57, 48 h17–18,44–46,48,51,57,63 and 72 h6,17–18,44–46,48,51,53,55,57,63 of hospital admission resulted in the pooled prevalence of gut barrier dysfunction of 59 (95 per cent c.i. 41 to 76), 49 (32 to 67) and 56 (41 to 71) per cent respectively. Sensitivity analysis excluding studies where the criteria for assessment of disease severity were not reported resulted in a pooled prevalence of gut barrier dysfunction of 53 (41 to 66) per cent, in the presence of high statistical heterogeneity (I2 = 83 per cent, P < 0·001). Sensitivity analysis constrained to studies using the 1992 Atlanta classification78 resulted in the pooled prevalence of gut barrier dysfunction of 60 (43 to 75) per cent. Sensitivity analyses were also carried out to investigate the effect of using different tests to assess gut barrier function. Sensitivity analysis constrained to assessment of gut barrier function by measurement of serum endotoxins included a total of 11 studies17–18,42–48,53,57, comprising 361 patients who underwent assessment of gut barrier function (Table 1). The pooled prevalence of gut barrier dysfunction was 57 (95 per cent c.i. 43 to 71) per cent, in the presence of high statistical heterogeneity (I2 = 88 per cent, P < 0·001) (Table 2). Three studies52,63,75, comprising a total of 92 patients, provided data for sensitivity analysis constrained to assessment of gut barrier function by measurement of bacterial DNA or RNA in blood, as detected by PCR (Table 1). The pooled prevalence of gut barrier dysfunction was 46 (17 to 77) per cent, in the presence of high statistical heterogeneity (I2 = 89 per cent, P < 0·001) (Table 2). Sensitivity analyses constrained to assessment of gut barrier function by measurement of endogenous antiendotoxin core antibodies45,46, lactulose/mannitol ratio6,64 and lactulose/rhamnose ratio51,55 resulted in the pooled prevalence of gut barrier dysfunction of 56 (15 to 93), 79 (61 to 92) and 66 (44 to 84) per cent, with statistical heterogeneity of P < 0·001, P = 0·091 and P = 0·116 respectively (Table 2). Metaregression analyses showed that mean age of the patients had a statistically significant association with the prevalence of gut barrier dysfunction using both methods of model fitting (Table 3). The prevalence of gut barrier dysfunction against the mean age for individual studies is presented in Fig. 4. The remaining co-variables investigated in the metaregression analyses (sex, aetiology and severity of disease) did not have a statistically significant association with the prevalence of gut barrier dysfunction using either method of model fitting (Table 3). Table 3 Results of metaregression analyses . Overall . Sex* . Age* . Biliary aetiology† . Severity‡ . Each co-variable fitted into individual models Estimate n.e. 0·005 (−0·004, 0·015) −0·012 (−0·026, 0·002) −0·000 (−0·007, 0·007) 0·002 (−0·003, 0·007) P n.e. 0·273 0·090* 0·951 0·484 All co-variables fitted into one model Estimate n.e. 0·000 (−0·011, 0·011) −0·016 (−0·033, 0·001) 0·000 (−0·006, 0·007) 0·002 (−0·003, 0·007) P 0·275 1·000 0·058 0·896 0·423 . Overall . Sex* . Age* . Biliary aetiology† . Severity‡ . Each co-variable fitted into individual models Estimate n.e. 0·005 (−0·004, 0·015) −0·012 (−0·026, 0·002) −0·000 (−0·007, 0·007) 0·002 (−0·003, 0·007) P n.e. 0·273 0·090* 0·951 0·484 All co-variables fitted into one model Estimate n.e. 0·000 (−0·011, 0·011) −0·016 (−0·033, 0·001) 0·000 (−0·006, 0·007) 0·002 (−0·003, 0·007) P 0·275 1·000 0·058 0·896 0·423 Values in parentheses are 95 per cent c.i. * All 18 studies included in model fitting; † 17 studies included in model fitting as co-variable not reported in one study; ‡ 15 studies included in model fitting as co-variable not reported in three studies. n.e., Not estimable. Open in new tab Table 3 Results of metaregression analyses . Overall . Sex* . Age* . Biliary aetiology† . Severity‡ . Each co-variable fitted into individual models Estimate n.e. 0·005 (−0·004, 0·015) −0·012 (−0·026, 0·002) −0·000 (−0·007, 0·007) 0·002 (−0·003, 0·007) P n.e. 0·273 0·090* 0·951 0·484 All co-variables fitted into one model Estimate n.e. 0·000 (−0·011, 0·011) −0·016 (−0·033, 0·001) 0·000 (−0·006, 0·007) 0·002 (−0·003, 0·007) P 0·275 1·000 0·058 0·896 0·423 . Overall . Sex* . Age* . Biliary aetiology† . Severity‡ . Each co-variable fitted into individual models Estimate n.e. 0·005 (−0·004, 0·015) −0·012 (−0·026, 0·002) −0·000 (−0·007, 0·007) 0·002 (−0·003, 0·007) P n.e. 0·273 0·090* 0·951 0·484 All co-variables fitted into one model Estimate n.e. 0·000 (−0·011, 0·011) −0·016 (−0·033, 0·001) 0·000 (−0·006, 0·007) 0·002 (−0·003, 0·007) P 0·275 1·000 0·058 0·896 0·423 Values in parentheses are 95 per cent c.i. * All 18 studies included in model fitting; † 17 studies included in model fitting as co-variable not reported in one study; ‡ 15 studies included in model fitting as co-variable not reported in three studies. n.e., Not estimable. Open in new tab Fig. 4 Open in new tabDownload slide Prevalence of gut barrier dysfunction against the mean age for individual studies in patients with acute pancreatitis Effect of interventions on gut barrier function Thirteen12,21–23,51,60,65–67,69–70,76–77 of the 44 included studies were randomized clinical trials. The types of intervention used were standard enteral nutrition (standard enteral nutrition formulas without specific immunomodulating nutrients) (6 studies12,21–23,51,60), advanced enteral nutrition (enteral nutrition in combination with probiotics or ecoimmunonutrition) (5 studies65–66,70,76–77), continuous blood purification treatment (1 study69) and traditional Chinese medicine (1 study67). Overall, nine12,22,60,65–67,69,76–77 of the 13 randomized clinical trials reported a significant improvement in gut barrier function postintervention compared with the control group (P < 0·050). Three12,22,60 of the six studies that used standard enteral nutrition reported a significant improvement in gut barrier function (P < 0·050), as measured by serum endotoxins, endogenous antiendotoxin core antibodies and lactulose/mannitol ratio. However, the remaining three studies21,23,51 that used standard enteral nutrition reported no significant difference in gut barrier function, as measured by endogenous antiendotoxin core antibodies, polyethylene glycols, lactulose/mannitol ratio and lactulose/rhamnose ratio. Four65–66,76–77 of the five studies that used advanced enteral nutrition showed a significant improvement in gut barrier function (P < 0·050), as measured by serum endotoxins, lactulose/mannitol ratio, lactulose/rhamnose ratio and nitric oxide. The studies that used continuous blood purification treatment69 and traditional Chinese medicine67 both showed a significant improvement in gut barrier function (P < 0·050), as measured by serum endotoxins and lactulose/mannitol ratio. Discussion This study demonstrates that gut barrier dysfunction is present in three of five patients with acute pancreatitis, and that the prevalence is not significantly affected by disease severity or timing of assessment after hospital admission, but shows a statistically significant association with age. A total of 14 different tests, conventionally classified into assessments of gut epithelial barrier integrity, functional assessments of the gut barrier and assessments of bacterial translocation, have been used to assess gut barrier function in patients with acute pancreatitis14. This study shows that the prevalence of gut barrier dysfunction is approximately the same over the three groups of tests and not significantly affected by the type of test used. One of the important clinical findings in this study is that gut barrier dysfunction is common in patients with acute pancreatitis and that there is no statistically significant difference between the prevalence of gut barrier dysfunction in patients with mild and severe acute pancreatitis, in both subgroup and metaregression analysis. This suggests that gut barrier dysfunction may be a characteristic of the disease, rather than a condition that develops with disease progression. However, limitations of the 1992 Atlanta classification of acute pancreatitis severity78 may have influenced the results, and it remains to be seen whether the results observed in the present study will also be seen when modern, more granular, classifications of severity79,80 are employed. Further, prespecified sensitivity analyses show that there is no statistically significant difference between the prevalence of gut barrier dysfunction within 24, 48 and 72 h after hospital admission. A possible explanation for this result is the varying duration between disease onset and hospital admission. However, nine6,18,47–48,51,53,57,64,75 of the 18 studies in the meta-analysis reported on the duration of disease at the time of hospital admission, and only one of these studies47 reported delayed hospital admission of more than 72 h after disease onset. Prolonged duration of disease before hospital admission may have resulted in assessment of gut barrier function during the restitution phase of the gut barrier61, leading to an underestimation of the pooled prevalence of gut barrier dysfunction. Intra-abdominal hypertension is also thought to be associated with gut barrier dysfunction, but the literature lacks consensus on the cause–effect relationship of intra-abdominal hypertension and gut barrier dysfunction in human studies81–83. Only five44,46–48,53 of the 44 included studies specifically investigated the association between gut barrier dysfunction and infected pancreatic necrosis and/or mortality. Future studies investigating gut barrier function in acute pancreatitis should include only a short duration of disease (48 h or less) as a criterion for patient recruitment, in an attempt to capture patients with a similar phase of disease progression, as well as investigate the association between gut barrier dysfunction and clinical endpoints other than severity of acute pancreatitis. The metaregression analyses show that mean age of the patient has a statistically significant association with the prevalence of gut barrier dysfunction. The estimate using the combined model shows that, in the presence of the other co-variables, there is a 2 per cent decrease in the prevalence of gut barrier dysfunction for every 1-year increase in patient age. This suggests that patient age may have a significant impact on the prevalence of gut barrier dysfunction and might have been a confounding variable in past clinical studies of gut barrier function. This finding provides justification for future studies to enrol patients within a particular age group or to control for patient age in the analyses. A wide variety of tests of gut barrier function exists. This provides impetus for the development of improved tests and the standardization of existing tests in order to improve the clinical assessment of gut barrier dysfunction and to improve the comparability of studies. The prevalence of gut barrier dysfunction using the PCR technique as a measure of bacterial translocation was comparable to estimates obtained using other tests of gut barrier function. In fact, sensitivity analyses showed that PCR resulted in the lowest prevalence of gut barrier dysfunction compared with the other tests. This is in contrast to the suggestion that use of PCR to detect bacterial DNA in the circulation is too sensitive to be a useful marker of bacterial translocation in acute pancreatitis84. Only half (three of six) of the studies that used standard enteral nutrition reported a statistically significant improvement in gut barrier function postintervention. This contrasts with high-quality evidence indicating that enteral nutrition has clear clinical benefits in patients with acute pancreatitis, which is generally attributed to the improvement of gut barrier function8–12. There are two potential explanations. The first is that standard enteral nutrition exerts its beneficial effect in acute pancreatitis independent of any significant improvement in gut barrier function. The second is that standard enteral nutrition does improve gut barrier function but some technical limitations in the clinical studies and possible confounding variables, as discussed previously, might have overshadowed its beneficial effect. Further, four of the five studies that used advanced enteral nutrition reported a significant improvement in gut barrier function postintervention. In light of the evidence from the present study, there is justification for further research to understand better the role of enteral nutrition in the pathophysiology of acute pancreatitis and to evaluate the benefit of advanced enteral nutrition in the mitigation of gut barrier dysfunction. There are further points to consider when interpreting the results of this study. First, a wide variety of tests have been used, and they each investigate different aspects of gut barrier dysfunction. This heterogeneity of the tests may be a possible limitation for the pooled estimate. However, the present study included only validated tests for measuring the prevalence of gut barrier dysfunction14. For example, tests of bacterial DNA or RNA using blood cultures were excluded, as recent studies have reported this to be an insensitive method of detection63,75. In addition, prespecified sensitivity analyses constrained to the assessment of gut barrier function using different tests showed no statistically significant difference between the prevalence of gut barrier dysfunction. Second, the high statistical heterogeneity between the studies may be a possible limitation for the pooled estimate and the conclusions that can be drawn from the meta-analysis. However, a random-effects model was used in the present analyses to yield the most conservative estimate of the pooled prevalence33,85. In addition, the most conservative result was taken for each study, so the probability of overestimating the prevalence of gut barrier dysfunction is low. Furthermore, an increase in statistical power was achieved through the use of meta-analytical techniques, as this is known to result in a more reliable estimate compared with individual primary studies of low statistical power25,32–33. Third, the present analysis included observational studies, which may have been subject to confounding variables. To counter this, a number of prespecified subgroup, sensitivity and metaregression analyses using two different methods of model fitting were performed to investigate the effect of possible confounding variables on the prevalence of gut barrier dysfunction. Fourth, by including only studies that have full-text articles in English, the possibility of language bias cannot be excluded. However, given that 27 of the 44 studies were conducted in a country where English is not the official language and the diverse geographical distribution of the included studies, it is unlikely that language of publication would have a significant effect on the results presented here. In addition, exclusion of non-English-language publications has been shown to have little effect on the summary estimates, and studies published in English tend to include more participants, have higher methodological quality and produce more conservative estimates86. Finally, the possibility of publication bias cannot be excluded. However, visual inspection of funnel plot symmetry and the use of three different statistical tests revealed no statistical evidence of publication bias. Disclosure The authors declare no conflict of interest. References 1 Oldenburg WA , Lau LL, Rodenberg TJ, Edmonds HJ, Burger CD. Acute mesenteric ischemia: a clinical review . Arch Intern Med 2004 ; 164 : 1054 – 1062 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Capurso G , Zerboni G, Signoretti M, Valente R, Stigliano S, Piciucchi M et al. Role of the gut barrier in acute pancreatitis . J Clin Gastroenterol 2012 ; 46 : S46 – S51 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Petrov MS . Gastric feeding and “gut rousing” in acute pancreatitis . Nutr Clin Pract 2014 : 29 : 287 – 290 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Fanous MY , Phillips AJ, Windsor JA. Mesenteric lymph: the bridge to future management of critical illness . JOP 2007 ; 8 : 374 – 399 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 5 Petrov MS , Shanbhag S, Chakraborty M, Phillips AR, Windsor JA. Organ failure and infection of pancreatic necrosis as determinants of mortality in patients with acute pancreatitis . Gastroenterology 2010 ; 139 : 813 – 820 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Liu H , Li W, Wang X, Li J, Yu W. Early gut mucosal dysfunction in patients with acute pancreatitis . Pancreas 2008 ; 36 : 192 – 196 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Ohland CL , MacNaughton WK. Probiotic bacteria and intestinal epithelial barrier function . Am J Physiol 2010 ; 298 : G807 – G819 . Google Scholar OpenURL Placeholder Text WorldCat 8 Petrov MS , Loveday BP, Pylypchuk RD, McIlroy K, Phillips AR, Windsor JA. Systematic review and meta-analysis of enteral nutrition formulations in acute pancreatitis . Br J Surg 2009 ; 96 : 1243 – 1252 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Petrov MS , van Santvoort HC, Besselink MG, van der Heijden GJ, Windsor JA, Gooszen HG. Enteral nutrition and the risk of mortality and infectious complications in patients with severe acute pancreatitis: a meta-analysis of randomized trials . Arch Surg 2008 ; 143 : 1111 – 1117 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Kalfarentzos F , Kehagias J, Mead N, Kokkinis K, Gogos C. Enteral nutrition is superior to parenteral nutrition in severe acute pancreatitis: results of a randomized prospective trial . Br J Surg 1997 ; 84 : 1665 – 1669 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 11 Hasibeder WR , Torgersen C, Rieger M, Dünser M. Critical care of the patient with acute pancreatitis . Anaesth Intensive Care 2009 ; 37 : 190 – 206 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Windsor AC , Kanwar S, Li AG, Barnes E, Guthrie JA, Spark JI et al. Compared with parenteral nutrition, enteral feeding attenuates the acute phase response and improves disease severity in acute pancreatitis . Gut 1998 ; 42 : 431 – 435 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Groschwitz KR , Hogan SP. Intestinal barrier function: molecular regulation and disease pathogenesis . J Allergy Clin Immunol 2009 ; 124 : 3 – 20 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Grootjans J , Thuijls G, Verdam F, Derikx JP, Lenaerts K, Buurman WA. Non-invasive assessment of barrier integrity and function of the human gut . World J Gastrointest Surg 2010 ; 2 : 61 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Turner JR . Intestinal mucosal barrier function in health and disease . Nat Rev Immunol 2009 ; 9 : 799 – 809 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Juvonen PO , Alhava EM, Takala JA. Gut permeability in patients with acute pancreatitis . Scand J Gastroenterol 2000 ; 35 : 1314 – 1318 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 17 Ammori BJ , Fitzgerald P, Hawkey P, McMahon MJ. The early increase in intestinal permeability and systemic endotoxin exposure in patients with severe acute pancreatitis is not associated with systemic bacterial translocation: molecular investigation of microbial DNA in the blood . Pancreas 2003 ; 26 : 18 – 22 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Giamarellos-Bourboulis EJ , Nikou GC, Matsaggoura M, Toumpanakis C, Grecka P, Giannikopoulos G et al. Alterations of systemic endotoxemia over the course of acute edematous pancreatitis. Correlation to the advent of an infection? Pancreatology 2003 ; 3 : 323 – 328 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Alpers DH . Enteral feeding and gut atrophy . Curr Opin Clin Nutr Metab Care 2002 ; 5 : 679 – 683 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Gatt M , Reddy B, MacFie J. Review article: bacterial translocation in the critically ill – evidence and methods of prevention . Aliment Pharmacol Ther 2007 ; 25 : 741 – 757 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Eckerwall GE , Axelsson JB, Andersson RG. Early nasogastric feeding in predicted severe acute pancreatitis: a clinical, randomized study . Ann Surg 2006 ; 244 : 959 – 965 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Gupta R , Patel K, Calder PC, Yaqoob P, Primrose JN, Johnson CD. A randomised clinical trial to assess the effect of total enteral and total parenteral nutritional support on metabolic, inflammatory and oxidative markers in patients with predicted severe acute pancreatitis (APACHE II > or = 6) . Pancreatology 2003 ; 3 : 406 – 413 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Singh N , Sharma B, Sharma M, Sachdev V, Bhardwaj P, Mani K et al. Evaluation of early enteral feeding through nasogastric and nasojejunal tube in severe acute pancreatitis: a noninferiority randomized controlled trial . Pancreas 2012 ; 41 : 153 – 159 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Moher D , Liberati A, Tetzlaff J, Altman DG; PRISMA Group . Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement . Ann Intern Med 2009 ; 151 : 264 – 269 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Stroup DF , Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group . JAMA 2000 ; 283 : 2008 – 2012 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Wells GA , Shea B, O'Connell D, Peterson J, Welch V, Losos M et al. The Newcastle–Ottawa Scale (NOS) for Assessing the Quality of Nonrandomised Studies in Meta-Analyses . http://www.ohri.ca/programs/clinical_epidemiology/oxford.htm [accessed 10 December 2013]. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 27 Gong Y , Zhou Q, Zhou Y, Lin Q, Zeng B, Chen R et al. Gastrectomy and risk of pancreatic cancer: systematic review and meta-analysis of observational studies . Cancer Causes Control 2012 ; 23 : 1279 – 1288 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Leonardi-Bee J , Pritchard D, Britton J. Asthma and current intestinal parasite infection: systematic review and meta-analysis . Am J Resp Crit Care Med 2006 ; 174 : 514 – 523 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Ltd StatsDirect . StatsDirect Statistical Software . http://www.statsdirect.com [accessed 28 June 2014]. 30 Higgins JP , Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses . BMJ 2003 ; 327 : 557 – 560 . Google Scholar Crossref Search ADS PubMed WorldCat 31 Higgins JP , Thompson SG. Quantifying heterogeneity in a meta-analysis . Stat Med 2002 ; 21 : 1539 – 1558 . Google Scholar Crossref Search ADS PubMed WorldCat 32 DerSimonian R , Laird N. Meta-analysis in clinical trials . Control Clin Trials 1986 ; 7 : 177 – 188 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Chen DG , Peace KE. Applied Meta-analysis with R . Chapman & Hall/CRC : Boca Raton , 2013 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 34 Begg CB , Mazumdar M. Operating characteristics of a rank correlation test for publication bias . Biometrics 1994 ; 50 : 1088 – 1101 . Google Scholar Crossref Search ADS PubMed WorldCat 35 Egger M , Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test . BMJ 1997 ; 315 : 629 – 634 . Google Scholar Crossref Search ADS PubMed WorldCat 36 Harbord RM , Egger M, Sterne JA. A modified test for small-study effects in meta-analyses of controlled trials with binary endpoints . Stat Med 2006 ; 25 : 3443 – 3457 . Google Scholar Crossref Search ADS PubMed WorldCat 37 Dickersin K Publication bias: recognizing the problem, understanding its origins and scope, and preventing harm. In Publication Bias in Meta-Analysis: Prevention, Assessment and Adjustments , Rothstein HR, Sutton AJ, Borenstein M (eds). John Wiley & Sons : Chichester , 2005 ; 11 – 33 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 38 Viechtbauer W Conducting meta-analyses in R with the metafor package . J Stat Softw 2010 ; 36 : 1 – 48 . Google Scholar Crossref Search ADS WorldCat 39 Maneiro JR , Lopez-Canoa N, Salgado E, Gomez-Reino JJ. Maintenance therapy of lupus nephritis with mycophenolate or azathioprine: systematic review and meta-analysis . Rheumatology (Oxford) 2014 ; 53 : 834 – 838 . Google Scholar Crossref Search ADS PubMed WorldCat 40 Houssami N , Macaskill P, Marinovich ML, Dixon JM, Irwig L, Brennan ME et al. Meta-analysis of the impact of surgical margins on local recurrence in women with early-stage invasive breast cancer treated with breast-conserving therapy . Eur J Cancer 2010 ; 46 : 3219 – 3232 . Google Scholar Crossref Search ADS PubMed WorldCat 41 Hozo SP , Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample . BMC Med Res Methodol 2005 ; 5 : 13 . Google Scholar Crossref Search ADS PubMed WorldCat 42 Foulis AK , Murray WR, Galloway D, McCartney AC, Lang E, Veitch J et al. Endotoxaemia and complement activation in acute pancreatitis in man . Gut 1982 ; 23 : 656 – 661 . Google Scholar Crossref Search ADS PubMed WorldCat 43 Beger HG , Büchler M, Bittner R, Block S, Nevalainen T, Roscher R. Necrosectomy and postoperative local lavage in necrotizing pancreatitis . Br J Surg 1988 ; 75 : 207 – 212 . Google Scholar Crossref Search ADS PubMed WorldCat 44 Exley AR , Leese T, Holliday MP, Swann RA, Cohen J. Endotoxaemia and serum tumour necrosis factor as prognostic markers in severe acute pancreatitis . Gut 1992 ; 33 : 1126 – 1128 . Google Scholar Crossref Search ADS PubMed WorldCat 45 Curley PJ , McMahon MJ, Lancaster F, Banks RE, Barclay GR, Shefta J et al. Reduction in circulating levels of CD4-positive lymphocytes in acute pancreatitis: relationship to endotoxin, interleukin 6 and disease severity . Br J Surg 1993 ; 80 : 1312 – 1315 . Google Scholar Crossref Search ADS PubMed WorldCat 46 Windsor JA , Fearon KC, Ross JA, Barclay GR, Smyth E, Poxton I et al. Role of serum endotoxin and antiendotoxin core antibody levels in predicting the development of multiple organ failure in acute pancreatitis . Br J Surg 1993 ; 80 : 1042 – 1046 . Google Scholar Crossref Search ADS PubMed WorldCat 47 Wig JD , Kochhar R, Ray JD, Krishna Rao DV, Gupta NM, Ganguly NK. Endotoxemia predicts outcome in acute pancreatitis . J Clin Gastroenterol 1998 ; 26 : 121 – 124 . Google Scholar Crossref Search ADS PubMed WorldCat 48 Ammori BJ , Leeder PC, King RF, Barclay GR, Martin IG, Larvin M et al. Early increase in intestinal permeability in patients with severe acute pancreatitis: correlation with endotoxemia, organ failure, and mortality . J Gastrointest Surg 1999 ; 3 : 252 – 262 . Google Scholar Crossref Search ADS PubMed WorldCat 49 Soong CV , Lewis HG, Halliday MI, Rowlands BJ. Intramucosal acidosis and the inflammatory response in acute pancreatitis . Am J Gastroenterol 1999 ; 94 : 2423 – 2429 . Google Scholar Crossref Search ADS PubMed WorldCat 50 Buttenschoen K , Berger D, Hiki N, Buttenschoen DC, Vasilescu C, Chikh-Torab F et al. Endotoxin and antiendotoxin antibodies in patients with acute pancreatitis . Eur J Surg 2000 ; 166 : 459 – 466 . Google Scholar Crossref Search ADS PubMed WorldCat 51 Powell JJ , Murchison JT, Fearon KCH, Ross JA, Siriwardena AK. Randomized controlled trial of the effect of early enteral nutrition on markers of the inflammatory response in predicted severe acute pancreatitis . Br J Surg 2000 ; 87 : 1375 – 1381 . Google Scholar Crossref Search ADS PubMed WorldCat 52 Zhang WZ , Han TQ, Tang YQ, Zhang SD. Rapid detection of sepsis complicating acute necrotizing pancreatitis using polymerase chain reaction . World J Gastroenterol 2001 ; 7 : 289 – 292 . Google Scholar Crossref Search ADS PubMed WorldCat 53 Bose SM , Verma GR, Mazumdar A, Giridhar M, Ganguly NK. Significance of serum endotoxin and antiendotoxin antibody levels in predicting the severity of acute pancreatitis . Surg Today 2002 ; 32 : 602 – 607 . Google Scholar Crossref Search ADS PubMed WorldCat 54 Martínez J , Palazón JM, Muñoz C, López M, Sánchez-Payá J, Laveda R et al. [ Endotoxin and anti-endotoxin antibodies in the prognosis of acute pancreatitis .] Rev Esp Enferm Dig 2002 ; 94 : 412 – 416 . Google Scholar OpenURL Placeholder Text WorldCat 55 McNaught CE , Woodcock NP, Mitchell CJ, Rowley G, Johnstone D, MacFie J. Gastric colonisation, intestinal permeability and septic morbidity in acute pancreatitis . Pancreatology 2002 ; 2 : 463 – 468 . Google Scholar Crossref Search ADS PubMed WorldCat 56 Ammori BJ , Barclay GR, Larvin M, McMahon MJ. Hypocalcemia in patients with acute pancreatitis: a putative role for systemic endotoxin exposure . Pancreas 2003 ; 26 : 213 – 217 . Google Scholar Crossref Search ADS PubMed WorldCat 57 Ammori BJ , Becker KL, Kite P, Snider RH, Nylén ES, White JC et al. Calcitonin precursors: early markers of gut barrier dysfunction in patients with acute pancreatitis . Pancreas 2003 ; 27 : 239 – 243 . Google Scholar Crossref Search ADS PubMed WorldCat 58 Rahman SH , Ammori BJ, Holmfield J, Larvin M, McMahon MJ. Intestinal hypoperfusion contributes to gut barrier failure in severe acute pancreatitis . J Gastrointest Surg 2003 ; 7 : 26 – 35 . Google Scholar Crossref Search ADS PubMed WorldCat 59 Rahman SH , Ammori BJ, Larvin M, McMahon MJ. Increased nitric oxide excretion in patients with severe acute pancreatitis: evidence of an endotoxin mediated inflammatory response? Gut 2003 ; 52 : 270 – 274 . Google Scholar Crossref Search ADS PubMed WorldCat 60 Zhao G , Wang CY, Wang F, Xiong JX. Clinical study on nutrition support in patients with severe acute pancreatitis . World J Gastroenterol 2003 ; 9 : 2105 – 2108 . Google Scholar Crossref Search ADS PubMed WorldCat 61 Penalva JC , Martínez J, Laveda R, Esteban A, Muñoz C, Sáez J et al. A study of intestinal permeability in relation to the inflammatory response and plasma EndoCab IgM levels in patients with acute pancreatitis . J Clin Gastroenterol 2004 ; 38 : 512 – 517 . Google Scholar Crossref Search ADS PubMed WorldCat 62 Rahman SH , Ibrahim K, Larvin M, Kingsnorth A, McMahon MJ. Association of antioxidant enzyme gene polymorphisms and glutathione status with severe acute pancreatitis . Gastroenterology 2004 ; 126 : 1312 – 1322 . Google Scholar Crossref Search ADS PubMed WorldCat 63 De Madaria E , Martínez J, Lozano B, Sempere L, Benlloch S, Such J et al. Detection and identification of bacterial DNA in serum from patients with acute pancreatitis . Gut 2005 ; 54 : 1293 – 1297 . Google Scholar Crossref Search ADS PubMed WorldCat 64 Nagpal K , Minocha VR, Agrawal V, Kapur S. Evaluation of intestinal mucosal permeability function in patients with acute pancreatitis . Am J Surg 2006 ; 192 : 24 – 28 . Google Scholar Crossref Search ADS PubMed WorldCat 65 Qin HL , Zheng JJ, Tong DN, Chen WX, Fan XB, Hang XM et al. Effect of Lactobacillus plantarum enteral feeding on the gut permeability and septic complications in the patients with acute pancreatitis . Eur J Clin Nutr 2008 ; 62 : 923 – 930 . Google Scholar Crossref Search ADS PubMed WorldCat 66 Besselink MG , van Santvoort HC, Renooij W, De Smet MB, Boermeester MA, Fischer K et al. Intestinal barrier dysfunction in a randomized trial of a specific probiotic composition in acute pancreatitis . Ann Surg 2009 ; 250 : 712 – 719 . Google Scholar Crossref Search ADS PubMed WorldCat 67 Chen H , Li F, Jia JG, Diao YP, Li ZX, Sun JB. Effects of traditional Chinese medicine on intestinal mucosal permeability in early phase of severe acute pancreatitis . Chin Med J 2010 ; 123 : 1537 – 1542 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 68 Pan L , Wang X, Li W, Li N, Li J. The intestinal fatty acid binding protein diagnosing gut dysfunction in acute pancreatitis: a pilot study . Pancreas 2010 ; 39 : 633 – 638 . Google Scholar Crossref Search ADS PubMed WorldCat 69 Zhang J , Yuan C, Hua G, Tong R, Luo X, Ying Z. Early gut barrier dysfunction in patients with severe acute pancreatitis: attenuated by continuous blood purification treatment . Int J Artif Organs 2010 ; 33 : 706 – 715 . Google Scholar Crossref Search ADS PubMed WorldCat 70 Sharma B , Srivastava S, Singh N, Sachdev V, Kapur S, Saraya A. Role of probiotics on gut permeability and endotoxemia in patients with acute pancreatitis: a double-blind randomized controlled trial . J Clin Gastroenterol 2011 ; 45 : 442 – 448 . Google Scholar Crossref Search ADS PubMed WorldCat 71 Shen Y , Cui N, Miao B, Zhao E. Immune dysregulation in patients with severe acute pancreatitis . Inflammation 2011 ; 34 : 36 – 42 . Google Scholar Crossref Search ADS PubMed WorldCat 72 Koh YY , Jeon WK, Cho YK, Kim HJ, Chung WG, Chon CU et al. The effect of intestinal permeability and endotoxemia on the prognosis of acute pancreatitis . Gut Liver 2012 ; 6 : 505 – 511 . Google Scholar Crossref Search ADS PubMed WorldCat 73 Pynnönen L , Minkkinen M, Räty S, Sand J, Nordback I, Perner A et al. Luminal lactate in acute pancreatitis – validation and relation to disease severity . BMC Gastroenterol 2012 ; 12 : 40 . Google Scholar Crossref Search ADS PubMed WorldCat 74 Sharma M , Sachdev V, Singh N, Bhardwaj P, Pal A, Kapur S et al. Alterations in intestinal permeability and endotoxemia in severe acute pancreatitis . Trop Gastroenterol 2012 ; 33 : 45 – 50 . Google Scholar Crossref Search ADS PubMed WorldCat 75 Li Q , Wang C, Tang C, He Q, Li N, Li J. Bacteremia in patients with acute pancreatitis as revealed by 16S ribosomal RNA gene-based techniques . Crit Care Med 2013 ; 41 : 1938 – 1950 . Google Scholar Crossref Search ADS PubMed WorldCat 76 Wang G , Wen J, Xu L, Zhou S, Gong M, Wen P et al. Effect of enteral nutrition and ecoimmunonutrition on bacterial translocation and cytokine production in patients with severe acute pancreatitis . J Surg Res 2013 ; 183 : 592 – 597 . Google Scholar Crossref Search ADS PubMed WorldCat 77 Zhao G , Zhang JG, Wu HS, Tao J, Qin Q, Deng SC et al. Effects of different resuscitation fluid on severe acute pancreatitis . World J Gastroenterol 2013 ; 19 : 2044 – 2052 . Google Scholar Crossref Search ADS PubMed WorldCat 78 Bradley EL . A clinically based classification system for acute pancreatitis. Summary of the International Symposium on Acute Pancreatitis, Atlanta, Ga, September 11 through 13, 1992 . Arch Surg 1993 ; 128 : 586 – 590 . Google Scholar Crossref Search ADS PubMed WorldCat 79 Banks PA , Bollen TL, Dervenis C, Gooszen HG, Johnson CD, Sarr MG et al. ; Acute Pancreatitis Classification Working Group . Classification of acute pancreatitis – 2012: revision of the Atlanta classification and definitions by international consensus . Gut 2013 ; 62 : 102 – 111 . Google Scholar Crossref Search ADS PubMed WorldCat 80 Dellinger EP , Forsmark CE, Layer P, Lévy P, Maraví-Poma E, Petrov MS et al. ; Pancreatitis Across Nations Clinical Research and Education Alliance (PANCREA). Determinant-based classification of acute pancreatitis severity: an international multidisciplinary consultation . Ann Surg 2012 ; 256 : 875 – 880 . Google Scholar Crossref Search ADS PubMed WorldCat 81 Al-Bahrani AZ , Abid GH, Holt A, McCloy RF, Benson J, Eddleston J et al. Clinical relevance of intra-abdominal hypertension in patients with severe acute pancreatitis . Pancreas 2008 ; 36 : 39 – 43 . Google Scholar Crossref Search ADS PubMed WorldCat 82 Al-Bahrani AZ , Darwish A, Hamza N, Benson J, Eddleston JM, Snider RH et al. Gut barrier dysfunction in critically ill surgical patients with abdominal compartment syndrome . Pancreas 2010 ; 39 : 1064 – 1069 . Google Scholar Crossref Search ADS PubMed WorldCat 83 Sun JK , Li WQ, Ni HB, Ke L, Tong ZH, Li N et al. Modified gastrointestinal failure score for patients with severe acute pancreatitis . Surgery Today 2013 ; 43 : 506 – 513 . Google Scholar Crossref Search ADS PubMed WorldCat 84 Bhutta HY , Ashley SW. Demonstrating infection in severe acute pancreatitis: a role for polymerase chain reaction and gene sequencing? Crit Care Med 2013 ; 41 : 2048 – 2049 . Google Scholar Crossref Search ADS PubMed WorldCat 85 Das SL , Singh PP, Phillips AR, Murphy R, Windsor JA, Petrov MS. Newly diagnosed diabetes mellitus after acute pancreatitis: a systematic review and meta-analysis . Gut 2014 ; 63 : 818 – 831 . Google Scholar Crossref Search ADS PubMed WorldCat 86 Jüni P , Holenstein F, Sterne J, Bartlett C, Egger M. Direction and impact of language bias in meta-analyses of controlled trials: empirical study . Int J Epidemiol 2002 ; 31 : 115 – 123 . Google Scholar Crossref Search ADS PubMed WorldCat © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd

Moossdorff, M; van Roozendaal, L M; Schipper, R-J; Strobbe, L J A; Voogd, A C; Tjan-Heijnen, V C G; Smidt, M L

doi: 10.1002/bjs.9644pmid: 25308345

Abstract Background Results in breast cancer research are reported using study endpoints. Most are composite endpoints (such as locoregional recurrence), consisting of several components (for example local recurrence) that are in turn composed of specific events (such as skin recurrence). Inconsistent endpoint selection and definition might lead to unjustified conclusions when comparing study outcomes. This study aimed to determine which locoregional endpoints are used in breast cancer studies, and how these endpoints and their components are defined. Methods PubMed was searched for breast cancer studies published in nine leading journals in 2011. Articles using endpoints with a local or regional component were included and definitions were compared. Results Twenty-three different endpoints with a local or regional component were extracted from 44 articles. Most frequently used were disease-free survival (25 articles), recurrence-free survival (7), local control (4), locoregional recurrence-free survival (3) and event-free survival (3). Different endpoints were used for similar outcomes. Of 23 endpoints, five were not defined and 18 were defined only partially. Of these, 16 contained a local and 13 a regional component. Included events were not specified in 33 of 57 (local) and 27 of 50 (regional) cases. Definitions of local components inconsistently included carcinoma in situ and skin and chest wall recurrences. Regional components inconsistently included specific nodal sites and skin and chest wall recurrences. Conclusion Breast cancer studies use many different endpoints with a locoregional component. Definitions of endpoints and events are either not provided or vary between trials. To improve transparency, facilitate trial comparison and avoid unjustified conclusions, authors should report detailed definitions of all endpoints. Introduction When comparing results of breast cancer studies, one is confronted with many different study endpoints and unclear definitions. Most studies have composite endpoints, for example locoregional recurrence, that consist of several components. These components consist of specific events, such as recurrence in axillary lymph nodes. Both the selection and definition of study endpoints (that is which specific events are included) may vary between studies. For instance, survival may be reported using a variety of endpoints, including disease-free survival, distant disease-free survival or breast cancer-specific survival. These endpoints do not always include the same components and events, and the paper may not provide the precise definition. The definition of endpoints in breast cancer studies has been a topic of interest among medical researchers for several years. Cuzick1 discussed inconsistent definitions of disease-free survival and noted that inconsistent selection of endpoints may confound the interpretation of study outcomes. Meropol2 advocated using a common language in cancer research outcome measures in general. Some efforts have been made to achieve uniform breast cancer endpoint definitions. Definitions for neoadjuvant and adjuvant trials were proposed by Hudis and colleagues3 in 2007 (Standardized Definitions for Efficacy End Points in adjuvant breast cancer trials, STEEP) and Fumagalli et al.4 in 2012. These definitions, however, have not been adopted universally into research practice. Since its publication in 2007, the STEEP article has been cited by 125 individual publications, according to PubMed Central, Google Scholar, Web of Knowledge and the Journal of Clinical Oncology website. A STEEP endpoint was used in 64 of these publications. Comparing or pooling the results of studies using different endpoints, or the same endpoint with a different definition, may result in the comparison of apples and oranges. Comparing study results or pooling results in meta-analyses may not be justified, and may lead to incorrect conclusions. The aim of this study was to determine the extent of this problem, by providing an overview of local and regional study endpoints used in breast cancer studies, through a limited but representative review of the literature. The study explored which endpoints are being used, whether definitions are provided for the endpoints and their components, and, if so, which specific events are included. Methods Literature search The PubMed database was searched for experimental and observational research investigating breast cancer in humans. The PubMed limits ‘clinical trials’, ‘randomized controlled trials’ and ‘comparative studies’ were used, and the search was limited to research published between 1 January 2011 and 31 December 2011 in nine leading medical, surgical and radiation oncology journals. Journals were selected based on impact factor in order to provide an impression of study endpoints used in good-quality breast cancer research with considerable impact in the field. Search terms were: breast neoplasms (MeSH), breast cancer, breast carcinoma, Annals of Surgery (Journal, NLM Catalog), Annals of Surgical Oncology (Journal), British Journal of Surgery (Journal), Journal of Clinical Oncology (Journal), Journal of the American Medical Association (Journal), Lancet (Journal), Lancet Oncology (Journal), New England Journal of Medicine (Journal) and Radiotherapy & Oncology (Journal). Selection Articles found through this search were assessed for eligibility. Articles were subjected to review if the abstract met the following inclusion criteria: original research paper; observational or therapeutic study; investigation of any type of invasive early breast cancer; and use of a clinical study endpoint. Articles with study endpoints containing a local or regional component were analysed further. Selection of publications and endpoint extraction were performed independently by two authors. Discrepancies were resolved by consulting a third author. Data extraction All endpoints containing a local or regional component were extracted from the publications. Any definitions of the endpoints provided in the original article or appendix were extracted, including specific events included in the local and regional components. Results Selection of articles The PubMed search identified 159 publications, of which 70 met the inclusion criteria. These 70 articles were evaluated for use of a local or regional study endpoint (or a composite endpoint with a local or regional component). This resulted in inclusion of 44 papers. A Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)5 flow chart is presented in Fig. 1. Fig. 1 Open in new tabDownload slide PRISMA flow chart showing selection of articles for review Local and regional study endpoints used in breast cancer trials The 44 articles6–49 contained 23 different endpoints with a local and/or regional component (Table 1). Various study endpoints were used for similar outcomes. Of these 23 endpoints, disease-free survival was used most frequently (25 articles), followed by recurrence-free survival (7), local control (4), locoregional recurrence-free survival (3) and event-free survival (3). Twelve endpoints were used only once each among the 44 publications. Table 1 Choice, frequency, and definitions of local and regional endpoints in 44 publications . No. of articles . Definition provided (at least partial) . Local endpoints (n = 7) 12 6 Ipsilateral breast recurrence 1 1 Ipsilateral breast tumour recurrence 2 1 Ipsilateral breast relapse 1 1 Ipsilateral local tumour relapse 1 1 Local control 4 1 Local recurrence 2 1 Rate of cancer recurrence after mastectomy 1 0 Regional endpoints (n = 3) 4 3 Axillary relapse 1 1 Crude cumulative incidence of axillary recurrence 1 0 Regional recurrence 2 2 Locoregional endpoints (n = 3) 6 4 Local or regional failure 1 0 Locoregional control 2 1 Locoregional recurrence-free survival 3 3 Composite endpoints with local or regional component (n =10) 44 39 Any breast cancer event 1 0 Breast cancer-free interval 1 1 Breast cancer-free survival 1 1 Disease-free survival 25 25 Event-free survival 3 3 Invasive disease-free survival 1 1 Recurrence-free survival 7 6 Relapse-free survival 2 1 Risk of recurrence 1 1 Time to recurrence 2 0 Total (n = 23 separate endpoints) 66 52 . No. of articles . Definition provided (at least partial) . Local endpoints (n = 7) 12 6 Ipsilateral breast recurrence 1 1 Ipsilateral breast tumour recurrence 2 1 Ipsilateral breast relapse 1 1 Ipsilateral local tumour relapse 1 1 Local control 4 1 Local recurrence 2 1 Rate of cancer recurrence after mastectomy 1 0 Regional endpoints (n = 3) 4 3 Axillary relapse 1 1 Crude cumulative incidence of axillary recurrence 1 0 Regional recurrence 2 2 Locoregional endpoints (n = 3) 6 4 Local or regional failure 1 0 Locoregional control 2 1 Locoregional recurrence-free survival 3 3 Composite endpoints with local or regional component (n =10) 44 39 Any breast cancer event 1 0 Breast cancer-free interval 1 1 Breast cancer-free survival 1 1 Disease-free survival 25 25 Event-free survival 3 3 Invasive disease-free survival 1 1 Recurrence-free survival 7 6 Relapse-free survival 2 1 Risk of recurrence 1 1 Time to recurrence 2 0 Total (n = 23 separate endpoints) 66 52 Open in new tab Table 1 Choice, frequency, and definitions of local and regional endpoints in 44 publications . No. of articles . Definition provided (at least partial) . Local endpoints (n = 7) 12 6 Ipsilateral breast recurrence 1 1 Ipsilateral breast tumour recurrence 2 1 Ipsilateral breast relapse 1 1 Ipsilateral local tumour relapse 1 1 Local control 4 1 Local recurrence 2 1 Rate of cancer recurrence after mastectomy 1 0 Regional endpoints (n = 3) 4 3 Axillary relapse 1 1 Crude cumulative incidence of axillary recurrence 1 0 Regional recurrence 2 2 Locoregional endpoints (n = 3) 6 4 Local or regional failure 1 0 Locoregional control 2 1 Locoregional recurrence-free survival 3 3 Composite endpoints with local or regional component (n =10) 44 39 Any breast cancer event 1 0 Breast cancer-free interval 1 1 Breast cancer-free survival 1 1 Disease-free survival 25 25 Event-free survival 3 3 Invasive disease-free survival 1 1 Recurrence-free survival 7 6 Relapse-free survival 2 1 Risk of recurrence 1 1 Time to recurrence 2 0 Total (n = 23 separate endpoints) 66 52 . No. of articles . Definition provided (at least partial) . Local endpoints (n = 7) 12 6 Ipsilateral breast recurrence 1 1 Ipsilateral breast tumour recurrence 2 1 Ipsilateral breast relapse 1 1 Ipsilateral local tumour relapse 1 1 Local control 4 1 Local recurrence 2 1 Rate of cancer recurrence after mastectomy 1 0 Regional endpoints (n = 3) 4 3 Axillary relapse 1 1 Crude cumulative incidence of axillary recurrence 1 0 Regional recurrence 2 2 Locoregional endpoints (n = 3) 6 4 Local or regional failure 1 0 Locoregional control 2 1 Locoregional recurrence-free survival 3 3 Composite endpoints with local or regional component (n =10) 44 39 Any breast cancer event 1 0 Breast cancer-free interval 1 1 Breast cancer-free survival 1 1 Disease-free survival 25 25 Event-free survival 3 3 Invasive disease-free survival 1 1 Recurrence-free survival 7 6 Relapse-free survival 2 1 Risk of recurrence 1 1 Time to recurrence 2 0 Total (n = 23 separate endpoints) 66 52 Open in new tab Definitions of endpoints used Definitions of the endpoints were not provided consistently (Table 1). Five of 23 endpoints were not defined in any of the papers. The other 18 were defined partially at least once, describing either the time interval for a time-to-event endpoint (for example from randomization or from surgery) or describing which local and/or regional events were included, or both. Of the 18 defined endpoints, 16 contained a local component and 13 a regional component. Two endpoints were defined according to the STEEP guidelines3: (invasive) breast cancer-free interval38 and invasive disease-free survival13. Definitions of local components The 16 defined endpoints with a local component were used 57 times in the 44 articles (Table S1, supporting information). The definitions provided for these local components were compared with respect to the inclusion or exclusion of specific events. Events listed in definitions of the local component of endpoints included ipsilateral breast recurrence, in situ carcinomas, recurrence in skin, surgical scar and chest wall, and, in one case, lymph nodes. Over half of the cases (33 of 57) did not mention which specific events were included as a local recurrence. In the remaining 24, at least some included events were listed (Fig. 2). Tumour recurrence in breast was included specifically 15 of 24 times. In contrast, the breast was not mentioned specifically six times. Three studies used an alternative definition for in-breast recurrence: one excluded resectable recurrences after lumpectomy and the other two subdivided breast recurrences as true/marginal or elsewhere in the breast. One of these papers also included recurrences in ‘nodal basins’ as a local event. None of the other papers made a distinction between true recurrences and new ipsilateral primary breast cancer. Carcinoma in situ was excluded as a local event eight times, but in the remaining 16 was neither included explicitly nor excluded. A skin recurrence was included twice as a local event, excluded once (but included as a regional recurrence) and not specified in the remaining 20 articles. One author included ‘ipsilateral breast tissue and overlying skin’; no other author clarified whether the location of the skin recurrence (such as overlying tumour, in biopsy tract, or anywhere on the breast) was important. Recurrences in the surgical scar were included twice and not specified 22 times. Chest wall recurrences were mentioned as a local event in four articles, excluded once (but included as a regional recurrence) and unclear in the remaining 19 articles. Fig. 2 Open in new tabDownload slide Events included as local recurrence in endpoints for which at least a partial definition was provided. *Distinguished true/marginal recurrence versus elsewhere in breast (n = 2) and excluded resectable recurrence after lumpectomy (n = 1). †Ipsilateral breast tissue and overlying skin Definitions of regional components Thirteen endpoints with a regional component were used 50 times in the 44 selected articles (Table S2, supporting information). Events listed under the regional components of these endpoints were skin and chest wall recurrences, as well as the involvement of lymph nodes in general and/or in specific nodal sites. In 27 of 50 cases, the articles did not specify the events that were considered regional recurrences. Fourteen of the remaining 23 cases included recurrences in ‘lymph nodes’ or ‘nodal’ recurrences, and nine described specific nodal sites that were included (Fig. 3). These sites varied in the articles that provided this information; recurrences in axillary lymph nodes were specifically mentioned in nine, infraclavicular lymph nodes in two, supraclavicular lymph nodes in seven and internal mammary lymph nodes in seven. In six of the 23 cases, lymph nodes were not mentioned in the definition. In the remaining three, the endpoints disease-free survival, breast cancer-free survival and recurrence-free survival were said to include ‘local or distant’ recurrences, but did not refer to inclusion or exclusion of lymph node involvement. Fig. 3 Open in new tabDownload slide Events included as regional recurrence in endpoints for which at least a partial definition was provided. *Included ‘local and distant recurrence’, without further mention of lymph nodes in disease-free survival, breast cancer-free survival and recurrence-free survival. Recurrence in nodal basins was included once as a local recurrence Of the 23 cases that listed the included and excluded events for the regional component, skin recurrences were included as a regional event in one, excluded in four and not specified in the remaining 18. Chest wall recurrences were considered regional events in one of the 23 cases, excluded in six and unclear in 16. Discussion This study looked at breast cancer study endpoints with a local or regional component that were used in papers published in nine leading journals in 2011. There were several observations. First, many different endpoints were used for similar outcomes. Second, endpoint definitions were not provided consistently. For one in five endpoints, no definition at all could be deduced from the article; for others, definitions were often incomplete with respect to the specific events included or excluded as local or regional recurrences. Moreover, several inconsistencies in included local and regional events were observed between the definitions of similar endpoints. Only two of 44 papers used a standard definition of the endpoint. Inconsistencies in the selection and definition of endpoints can limit interpretation and mutual comparison of trial results. Differences in study outcomes can be interpreted incorrectly as differences in treatment effects, leading to false conclusions and possible delays in the implementation of important study outcomes in clinical practice. Breast cancer studies are particularly vulnerable. Many new interventions show only small improvements in outcomes, considering the already favourable prognosis of most patients50. When studying small absolute differences, the relative effect of varying endpoint definitions compared with the treatment effect may be even larger. Almost all breast cancer study endpoints are composite endpoints. Composite endpoints have the advantage of increased event rates and, as a result, fewer patients are needed to provide significant results. However, for a composite endpoint to be a valid outcome measure, all included components, and subsequently the events included in these components, should meet predefined criteria51. First, they should be of similar relevance to patients. If patients consider distant metastases and death to be of similar importance, it is not important how a risk reduction is distributed between the two. In contrast, ipsilateral ductal carcinoma in situ (DCIS) is less important to patients than mortality. In such instances, the distribution of risk reduction is important, and is not reflected properly if both are combined in one endpoint such as disease-free survival. Second, components should be influenced to a similar degree by the intervention. If an intervention effectively prevents breast recurrence but not distant metastasis, one endpoint measuring both does not provide specific information on treatment effect and may decrease the discriminative power of the study. The same applies to mortality, particularly in subgroups at high risk of non-breast cancer death, which is not influenced by the intervention to the same degree as breast cancer-specific mortality. Inclusion of all-cause mortality in an endpoint can therefore distort the results52–54. Finally, the incidence of the more and less important components should be comparable. For instance, a high incidence of ‘locoregional recurrence’ could reflect either many lymph node recurrences and few instances of ipsilateral DCIS, or many cases of ipsilateral DCIS and few lymph node recurrences. In that case, the endpoint does not adequately reflect prognosis. Clearly, not every endpoint currently used in breast cancer research meets these criteria. The standard definitions proposed by Hudis et al.3 and Fumagalli and co-workers4 aimed to solve the problem of inconsistent use and definitions of study endpoints in adjuvant and neoadjuvant settings, but they are not used consistently. It is unknown why these definitions have not been adopted universally; possibilities include lack of awareness, the relatively short interval since publication, criticism of definitions, or anticipated problems in comparing new results with previous findings. For instance, the American College of Surgeons Oncology Group (ACOSOG) Z0011 study22 reported disease-free survival rather than protocol-specified distant disease-free survival to facilitate comparison with other studies. Additionally, these proposals focused on traditional adjuvant therapy trials, whereas the multidisciplinary character of breast cancer care requires easy comparison of results from other fields involved in management of breast cancer. An additional consensus-based proposal for standard definitions of endpoints in cancer research, including breast cancer, might be expected from the Definition for the Assessment of Time-to-event Endpoints in CANcer trials (DATECAN) group55. The detrimental effect of inconsistent endpoint definitions on reliable comparison of trial results may be even larger when different events occur in the same patient. In patients with synchronous distant metastasis and axillary recurrence, researchers may only count distant metastasis and ignore the axillary recurrence, or count distant metastasis and include the axillary recurrence separately in an analysis of locoregional control. As the chosen approach either increases or decreases the event rate, differences between trials may contribute to variations in reported study outcomes. The same applies to the question of whether a thorough search for synchronous locoregional events should be conducted once distant metastases have occurred. These issues should be taken into account when interpreting trial results and again stress the need for a standard approach. The articles selected for this review were published in only nine journals over a relatively short time. Furthermore, only endpoints with a local and/or regional component were selected. Therefore, the list of endpoints and variable definitions is probably not exhaustive, which may limit extrapolation of the results. With a longer time frame and additional journals, even more different endpoints and definitions could be encountered. It is striking, therefore, that such a large variety of endpoints was identified even in this limited search and that definitions were not provided consistently. Furthermore, many different definitions were used for similar endpoints. Additionally, it was found that the lack of definition of local and regional events lies at the root of inconsistent endpoint definitions. These inconsistencies suggest that detailed endpoint definitions do not have the full attention of authors and reviewers. It is unlikely that this problem is limited to the selection of journals or time frame of the search. Therefore, despite these restrictions, the results illustrate that the outcomes of major breast cancer studies are not readily comparable as a result of inconsistencies in endpoint selection and definition. To improve transparency, facilitate trial comparison and avoid unjustified conclusions, authors should provide clear and detailed definitions of the endpoints. Preferably, standard endpoint definitions should be used, to facilitate reliable comparison of results. This also applies to definitions of the components included in the endpoints, such as local and regional events. To ensure transparency in endpoint definitions, clinical trial registries, reviewers of research protocols and journals publishing the results should insist on inclusion of detailed definitions of endpoints and their components. These should comprise at least all included (and excluded) events and, for time-to-event endpoints, the starting point (for example from randomization, from surgery). The Consolidated Standards of Reporting Trials (CONSORT) and STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) checklists already contain an item requiring listing and definition of endpoints. Subsequently, journal editors and reviewers should assess whether these are covered sufficiently. Designing standard endpoints for breast cancer trials should start with standard definitions of the specific components of these endpoints, such as local or regional recurrence. Only when the definitions of these components are used consistently can a valid and relevant combination be chosen as a valid and relevant composite endpoint. Currently, a consensus project using the RAND/UCLA Appropriateness Method56, an adjusted version of the Delphi method, aiming to reach consensus on the definitions of local event, second primary breast cancer, regional event and distant event, is being undertaken. An international expert panel was formed for this purpose, consisting of leading breast cancer specialists, epidemiologists, presidents and members of scientific and clinical societies and boards, research groups, and editors and editorial board members of leading cancer journals. The proposed event definitions can be used to improve existing standard endpoint definitions or, if necessary, to build further towards a new proposal. In anticipation of these proposals, authors reporting trial results should improve transparency in two ways. First, definitions of all endpoints and their components must be provided in the paper, so any differences in definitions between trials become evident. Second, authors should report the incidence of all separate events in a supplement, in addition to the incidence of the endpoint. For instance, a trial using ‘locoregional recurrence’ as the primary endpoint should also provide the incidence of all included events, such as ipsilateral recurrence in the breast, skin recurrence and recurrence in a supraclavicular lymph node. This improves transparency even further, and may help interpret conflicting results. As a result of these improvements, more reliable conclusions will become available, serving patients with breast cancer worldwide. Acknowledgements M.M. and L.M.v.R. are joint first authors of this article Disclosure: The authors declare no conflict of interest. References 1 Cuzick J Controversies in design and interpretation of adjuvant clinical trials . Cancer Invest 2010 ; 28 ( Suppl 1 ): 28 – 34 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Meropol NJ . Comparative effectiveness research to inform medical decisions: the need for common language . J Clin Oncol 2012 ; 30 : 4192 – 4193 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Hudis CA , Barlow WE, Costantino JP, Gray RJ, Pritchard KI, Chapman JA et al. Proposal for standardized definitions for efficacy end points in adjuvant breast cancer trials: the STEEP system . J Clin Oncol 2007 ; 25 : 2127 – 2132 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Fumagalli D , Bedard PL, Nahleh Z, Michiels S, Sotiriou C, Loi S et al. A common language in neoadjuvant breast cancer clinical trials: proposals for standard definitions and endpoints . Lancet Oncol 2012 ; 13 : e240 – e248 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Moher D , Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement . PLoS Med 2009 ; 6 : e1000097 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Abdulkarim BS , Cuartero J, Hanson J, Deschênes J, Lesniak D, Sabri S. Increased risk of locoregional recurrence for women with T1–2 N0 triple-negative breast cancer treated with modified radical mastectomy without adjuvant radiation therapy compared with breast-conserving therapy . J Clin Oncol 2011 ; 29 : 2852 – 2858 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Adkins FC , Gonzalez-Angulo AM, Lei X, Hernandez-Aya LF, Mittendorf EA, Litton JK et al. Triple-negative breast cancer is not a contraindication for breast conservation . Ann Surg Oncol 2011 ; 18 : 3164 – 3173 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Agrawal RK , Alhasso A, Barrett-Lee PJ, Bliss JM, Bliss P et al. ; FAST Trialists group. First results of the randomised UK FAST Trial of radiotherapy hypofractionation for treatment of early breast cancer (CRUKE/04/015) . Radiother Oncol 2011 ; 100 : 93 – 100 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 9 Badwe R , Hawaldar R, Parmar V, Nadkarni M, Shet T, Desai S et al. Single-injection depot progesterone before surgery and survival in women with operable breast cancer: a randomized controlled trial . J Clin Oncol 2011 ; 29 : 2845 – 2851 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Bartlett JM , Brookes CL, Robson T, van de Velde CJ, Billingham LJ, Campbell FM et al. Estrogen receptor and progesterone receptor as predictive biomarkers of response to endocrine therapy: a prospectively powered pathology study in the Tamoxifen and Exemestane Adjuvant Multinational trial . J Clin Oncol 2011 ; 29 : 1531 – 1538 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Bliss JM , Kilburn LS, Coleman RE, Forbes JF, Coates AS, Jones SE et al. Disease-related outcomes with long-term follow-up: an updated analysis of the intergroup exemestane study . J Clin Oncol 2012 ; 30 : 709 – 717 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Bollet MA , Belin L, Reyal F, Campana F, Dendale R, Kirova YM et al. Preoperative radio-chemotherapy in early breast cancer patients: long-term results of a phase II trial . Radiother Oncol 2012 ; 102 : 82 – 88 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Coleman RE , Marshall H, Cameron D, Dodwell D, Burkinshaw R, Keane M et al. Breast-cancer adjuvant therapy with zoledronic acid . N Engl J Med 2011 ; 365 : 1396 – 1405 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Colleoni M , Giobbie-Hurder A, Regan MM, Thürlimann B, Mouridsen H, Mauriac L et al. Analyses adjusting for selective crossover show improved overall survival with adjuvant letrozole compared with tamoxifen in the BIG 1-98 study . J Clin Oncol 2011 ; 29 : 1117 – 1124 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Coombes RC , Bliss JM, Espie M, Erdkamp F, Wals J, Tres A et al. Randomized, phase III trial of sequential epirubicin and docetaxel versus epirubicin alone in postmenopausal patients with node-positive breast cancer . J Clin Oncol 2011 ; 29 : 3247 – 3254 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Eiermann W , Pienkowski T, Crown J, Sadeghi S, Martin M, Chan A et al. Phase III study of doxorubicin/cyclophosphamide with concomitant versus sequential docetaxel as adjuvant treatment in patients with human epidermal growth factor receptor 2-normal, node-positive breast cancer: BCIRG-005 trial . J Clin Oncol 2011 ; 29 : 3877 – 3884 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Elkin EB , Klem ML, Gonzales AM, Ishill NM, Hodgson D, Ng AK et al. Characteristics and outcomes of breast cancer in women with and without a history of radiation for Hodgkin's lymphoma: a multi-institutional, matched cohort study . J Clin Oncol 2011 ; 29 : 2466 – 2473 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Ellis GK , Barlow WE, Gralow JR, Hortobagyi GN, Russell CA, Royce ME et al. Phase III comparison of standard doxorubicin and cyclophosphamide versus weekly doxorubicin and daily oral cyclophosphamide plus granulocyte colony-stimulating factor as neoadjuvant therapy for inflammatory and locally advanced breast cancer: SWOG 0012 . J Clin Oncol 2011 ; 29 : 1014 – 1021 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Fortunato L , Mascaro A, Poccia I, Andrich R, Amini M, Costarelli L et al. Lobular breast cancer: same survival and local control compared with ductal cancer, but should both be treated the same way? Analysis of an institutional database over a 10-year period . Ann Surg Oncol 2012 ; 19 : 1107 – 1114 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Gentilini O , Veronesi P, Botteri E, Soggiu F, Trifirò G, Lissidini G et al. Sentinel lymph node biopsy in multicentric breast cancer: five-year results in a large series from a single institution . Ann Surg Oncol 2011 ; 18 : 2879 – 2884 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Gianni L , Dafni U, Gelber RD, Azambuja E, Muehlbauer S, Goldhirsch A et al. Treatment with trastuzumab for 1 year after adjuvant chemotherapy in patients with HER2-positive early breast cancer: a 4-year follow-up of a randomised controlled trial . Lancet Oncol 2011 ; 12 : 236 – 244 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Giuliano AE , Hunt KK, Ballman KV, Beitsch PD, Whitworth PW, Blumencranz PW et al. Axillary dissection vs no axillary dissection in women with invasive breast cancer and sentinel node metastasis: a randomized clinical trial . JAMA 2011 ; 305 : 569 – 575 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Gnant M , Mlineritsch B, Stoeger H, Luschin-Ebengreuth G, Heck D, Menzel C et al. Adjuvant endocrine therapy plus zoledronic acid in premenopausal women with early-stage breast cancer: 62-month follow-up from the ABCSG-12 randomised trial . Lancet Oncol 2011 ; 12 : 631 – 641 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Gobardhan PD , Elias SG, Madsen EV, van Wely B, van den Wildenberg F, Theunissen EB et al. Prognostic value of lymph node micrometastases in breast cancer: a multicenter cohort study . Ann Surg Oncol 2011 ; 18 : 1657 – 1664 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Hackshaw A , Roughton M, Forsyth S, Monson K, Reczko K, Sainsbury R et al. Long-term benefits of 5 years of tamoxifen: 10-year follow-up of a large randomized trial in women at least 50 years of age with early breast cancer . J Clin Oncol 2011 ; 29 : 1657 – 1663 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Isern AE , Manjer J, Malina J, Loman N, Mårtensson T, Bofin A et al. Risk of recurrence following delayed large flap reconstruction after mastectomy for breast cancer . Br J Surg 2011 ; 98 : 659 – 666 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Jin H , Tu D, Zhao N, Shepherd LE, Goss PE. Longer-term outcomes of letrozole versus placebo after 5 years of tamoxifen in the NCIC CTG MA.17 trial: analyses adjusting for treatment crossover . J Clin Oncol 2012 ; 30 : 718 – 721 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Joensuu H , Kellokumpu-Lehtinen PL, Huovinen R, Jukkola-Vuorinen A, Tanner M, Kokko R et al. Adjuvant capecitabine, docetaxel, cyclophosphamide, and epirubicin for early breast cancer: final analysis of the randomized FinXX trial . J Clin Oncol 2012 ; 30 : 11 – 18 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Kauer-Dorner D , Pötter R, Resch A, Handl-Zeller L, Kirchheiner K, Meyer-Schell K et al. Partial breast irradiation for locally recurrent breast cancer within a second breast conserving treatment: alternative to mastectomy? Results from a prospective trial . Radiother Oncol 2012 ; 102 : 96 – 101 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Kimple RJ , Klauber-DeMore N, Kuzmiak CM, Pavic D, Lian J, Livasy CA et al. Local control following single-dose intraoperative radiotherapy prior to surgical excision of early-stage breast cancer . Ann Surg Oncol 2011 ; 18 : 939 – 945 . Google Scholar Crossref Search ADS PubMed WorldCat 31 Li S , Yu KD, Fan L, Hou YF, Shao ZM. Predicting breast cancer recurrence following breast-conserving therapy: a single-institution analysis consisting of 764 Chinese breast cancer cases . Ann Surg Oncol 2011 ; 18 : 2492 – 2499 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Lipton A , Chapman JA, Demers L, Shepherd LE, Han L, Wilson CF et al. Elevated bone turnover predicts for bone metastasis in postmenopausal breast cancer: results of NCIC CTG MA.14 . J Clin Oncol 2011 ; 29 : 3605 – 3610 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Martelli G , Miceli R, Daidone MG, Vetrella G, Cerrotta AM, Piromalli D et al. Axillary dissection versus no axillary dissection in elderly patients with breast cancer and no palpable axillary nodes: results after 15 years of follow-up . Ann Surg Oncol 2011 ; 18 : 125 – 133 . Google Scholar Crossref Search ADS PubMed WorldCat 34 Perez EA , Romond EH, Suman VJ, Jeong JH, Davidson NE, Geyer CE et al. Four-year follow-up of trastuzumab plus adjuvant chemotherapy for operable human epidermal growth factor receptor 2-positive breast cancer: joint analysis of data from NCCTG N9831 and NSABP B-31 . J Clin Oncol 2011 ; 29 : 3366 – 3373 . Google Scholar Crossref Search ADS PubMed WorldCat 35 Perez EA , Suman VJ, Davidson NE, Gralow JR, Kaufman PA, Visscher DW et al. Sequential versus concurrent trastuzumab in adjuvant chemotherapy for breast cancer . J Clin Oncol 2011 ; 29 : 4491 – 4497 . Google Scholar Crossref Search ADS PubMed WorldCat 36 Pfeiler G , Königsberg R, Fesl C, Mlineritsch B, Stoeger H, Singer CF et al. Impact of body mass index on the efficacy of endocrine therapy in premenopausal patients with breast cancer: an analysis of the prospective ABCSG-12 trial . J Clin Oncol 2011 ; 29 : 2653 – 2659 . Google Scholar Crossref Search ADS PubMed WorldCat 37 Pritchard KI , Shepherd LE, Chapman JA, Norris BD, Cantin J, Goss PE et al. Randomized trial of tamoxifen versus combined tamoxifen and octreotide LAR therapy in the adjuvant treatment of early-stage breast cancer in postmenopausal women: NCIC CTG MA.14 . J Clin Oncol 2011 ; 29 : 3869 – 3876 . Google Scholar Crossref Search ADS PubMed WorldCat 38 Regan MM , Neven P, Giobbie-Hurder A, Goldhirsch A, Ejlertsen B, Mauriac L et al. Assessment of letrozole and tamoxifen alone and in sequence for postmenopausal women with steroid hormone receptor-positive breast cancer: the BIG 1-98 randomised clinical trial at 8·1 years median follow-up . Lancet Oncol 2011 ; 12 : 1101 – 1108 . Google Scholar Crossref Search ADS PubMed WorldCat 39 Shah C , Antonucci JV, Wilkinson JB, Wallace M, Ghilezan M, Chen P et al. Twelve-year clinical outcomes and patterns of failure with accelerated partial breast irradiation versus whole-breast irradiation: results of a matched-pair analysis . Radiother Oncol 2011 ; 100 : 210 – 214 . Google Scholar Crossref Search ADS PubMed WorldCat 40 Slamon D , Eiermann W, Robert N, Pienkowski T, Martin M, Press M et al. Adjuvant trastuzumab in HER2-positive breast cancer . N Engl J Med 2011 ; 365 : 1273 – 1283 . Google Scholar Crossref Search ADS PubMed WorldCat 41 Stull TS , Catherine Goodwin M, Gracely EJ, Chernick MR, Carella RJ, Frazier TG et al. A single-institution review of accelerated partial breast irradiation in patients considered “cautionary” by the American Society for Radiation Oncology . Ann Surg Oncol 2012 ; 19 : 553 – 559 . Google Scholar Crossref Search ADS PubMed WorldCat 42 Tausch C , Taucher S, Dubsky P, Seifert M, Reitsamer R, Kwasny W et al. Prognostic value of number of removed lymph nodes, number of involved lymph nodes, and lymph node ratio in 7502 breast cancer patients enrolled onto trials of the Austrian Breast and Colorectal Cancer Study Group (ABCSG) . Ann Surg Oncol 2012 ; 19 : 1808 – 1817 . Google Scholar Crossref Search ADS PubMed WorldCat 43 Untch M , Fasching PA, Konecny GE, Hasmüller S, Lebeau A, Kreienberg R et al. Pathologic complete response after neoadjuvant chemotherapy plus trastuzumab predicts favorable survival in human epidermal growth factor receptor 2-overexpressing breast cancer: results from the TECHNO trial of the AGO and GBG study groups . J Clin Oncol 2011 ; 29 : 3351 – 3357 . Google Scholar Crossref Search ADS PubMed WorldCat 44 van de Velde CJ , Rea D, Seynaeve C, Putter H, Hasenburg A, Vannetzel JM et al. Adjuvant tamoxifen and exemestane in early breast cancer (TEAM): a randomised phase 3 trial . Lancet 2011 ; 377 : 321 – 331 . Google Scholar Crossref Search ADS PubMed WorldCat 45 Wang J , Shi M, Ling R, Xia Y, Luo S, Fu X et al. Adjuvant chemotherapy and radiotherapy in triple-negative breast carcinoma: a prospective randomized controlled multi-center trial . Radiother Oncol 2011 ; 100 : 200 – 204 . Google Scholar Crossref Search ADS PubMed WorldCat 46 Weaver DL , Ashikaga T, Krag DN, Skelly JM, Anderson SJ, Harlow SP et al. Effect of occult metastases on survival in node-negative breast cancer . N Engl J Med 2011 ; 364 : 412 – 421 . Google Scholar Crossref Search ADS PubMed WorldCat 47 Wennemers M , Bussink J, Grebenchtchikov N, Sweep FC, Span PN. TRIB3 protein denotes a good prognosis in breast cancer patients and is associated with hypoxia sensitivity . Radiother Oncol 2011 ; 101 : 198 – 202 . Google Scholar Crossref Search ADS PubMed WorldCat 48 Werkhoven E , Hart G, Tinteren H, Elkhuizen P, Collette L, Poortmans P et al. Nomogram to predict ipsilateral breast relapse based on pathology review from the EORTC 22881–10882 boost versus no boost trial . Radiother Oncol 2011 ; 100 : 101 – 107 . Google Scholar Crossref Search ADS PubMed WorldCat 49 Wink CJ , Woensdregt K, Nieuwenhuijzen GA, van der Sangen MJ, Hutschemaekers S, Roukema JA et al. Hormone treatment without surgery for patients aged 75 years or older with operable breast cancer . Ann Surg Oncol 2012 ; 19 : 1185 – 1191 . Google Scholar Crossref Search ADS PubMed WorldCat 50 Newman LA . Epidemiology of locally advanced breast cancer . Semin Radiat Oncol 2009 ; 19 : 195 – 203 . Google Scholar Crossref Search ADS PubMed WorldCat 51 Montori VM , Permanyer-Miralda G, Ferreira-Gonzalez I, Busse JW, Pacheco-Huergo V, Bryant D et al. Validity of composite end points in clinical trials . BMJ 2005 ; 330 : 594 – 596 . Google Scholar Crossref Search ADS PubMed WorldCat 52 Nout RA , Fiets WE, Struikmans H, Rosendaal FR, Putter H, Nortier JW. The in- or exclusion of non-breast cancer related death and contralateral breast cancer significantly affects estimated outcome probability in early breast cancer . Breast Cancer Res Treat 2008 ; 109 : 567 – 572 . Google Scholar Crossref Search ADS PubMed WorldCat 53 Mell LK , Jeong J, Nichols MA, Polite BN, Weichselbaum RR, Chmura SJ. Predictors of competing mortality in early breast cancer . Cancer 2010 ; 116 : 5365 – 5373 . Google Scholar Crossref Search ADS PubMed WorldCat 54 Mell LK , Jeong JH. Pitfalls of using composite primary end points in the presence of competing risks . J Clin Oncol 2010 ; 28 : 4297 – 4299 . Google Scholar Crossref Search ADS PubMed WorldCat 55 Bellera CA , Pulido M, Gourgou S, Collette L, Doussau A, Kramar A et al. Protocol of the Definition for the Assessment of Time-to-event Endpoints in CANcer trials (DATECAN) project: formal consensus method for the development of guidelines for standardised time-to-event endpoints' definitions in cancer clinical trials . Eur J Cancer 2013 ; 49 : 769 – 781 . Google Scholar Crossref Search ADS PubMed WorldCat 56 Fitch K , Bernstein SJ, Aguilar MS, Burn B, LaCalle JR, Lazaro P et al. The RAND/UCLA Appropriateness Method User's Manual . RAND : Santa Monica , 2001 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Author notes Presented to the American Society of Breast Surgeons 14th Annual Meeting, Chicago, Illinois, USA, May 2013, Conference of the Association of Surgeons of the Netherlands (Chirurgendagen), Veldhoven, The Netherlands, May 2013, 11e Bossche Mammacongres, Vught, The Netherlands, June 2013, and the European Cancer Congress (ECCO), Amsterdam, The Netherlands, September 2013; published in abstract form as Ann Surg Oncol 2013; 1: 115–116 and Eur J Cancer 2013; 49: S438 © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd

Pucher, P H; Aggarwal, R; Qurashi, M; Singh, P; Darzi, A

doi: 10.1002/bjs.9654pmid: 25350855

Abstract Background Complications are a common and accepted risk of surgery. Failure to optimize the management of patients who suffer postoperative morbidity may result in poorer surgical outcomes. This study aimed to evaluate a checklist-based tool to improve and standardize care of postoperative complications. Methods Surgical trainees conducted baseline ward rounds of three patients with common postoperative complications in a high-fidelity simulated ward environment. Subjects were randomized to intervention or control groups, and final ward rounds were conducted with or without the aid of checklists for management of postoperative complications. Adherence to critical care processes was assessed, in addition to technical (Surgical Ward-care Assessment Tool, SWAT) and non-technical (Ward NOn-TECHnical Skills (W-NOTECHS) scale) performance. Subjects completed a feedback questionnaire regarding their perception of the checklists. Results Twenty trainees completed 120 patient assessments. All intervention group subjects opted to use the checklists, resulting in significantly fewer critical errors compared with controls (median (i.q.r.) 0 (0–0) versus 60 (40–73) per cent; P < 0·001). The intervention group demonstrated improved patient management (SWAT-M) (P < 0·001) and non-technical skills (P = 0·043) between baseline and final ward rounds, whereas controls did not (P = 0·571 and P = 0·809 respectively). A small learning effect was seen with improvement in patient assessment (SWAT-A) in both groups (P < 0·001). Intervention group subjects found checklists easy and effective to use, and would want them used for their own care if they were to experience postoperative complications. Conclusion Checklist use resulted in significantly improved standardization, evidence-based management of postoperative complications, and quality of ward rounds. Simulation-based piloting aided appropriate use of checklists and staff engagement. Checklists represent a low-cost intervention to reduce rates of failure to rescue and to improve patient care. Introduction Complications are frequent occurrences after surgery, with reported morbidity rates following major gastrointestinal procedures of 30–50 per cent1–4. The most frequently reported postoperative complications, such as pneumonia or wound infection, are common to patients undergoing gastrointestinal surgery, irrespective of the specific procedure4–6. Despite their relatively common nature, published evidence7,8 suggests a high degree of variability in the management of postoperative complications. Although several associated structural factors, such as nurse staffing levels, have been identified9–11, these fail to account for the majority of variability seen in surgical outcomes. Process-level failures, at the point of interaction between the patient and clinician, have been investigated less thoroughly. Postoperative patient assessment and the surgical ward round are subject to pressures of time, resource limitation and the prioritization of other clinical activities12. Coupled with natural tendencies for human error, the ward-based phase of surgical care is rife with potential for adverse events13. Checklists are seen as one of the most effective means to combat medical error14,15. As mental prompts and physical reminders, checklists serve to standardize process-led care. They have been adopted widely in recent years, with their use becoming the accepted standard of care in many areas of surgical care16–18. Evidence also suggests that checklists may act beyond the included processes, to improve safety culture as a whole, through wider individual involvement and lessening traditional clinical hierarchies19,20. At the same time, the wider uptake of checklists has served to highlight some of the potential pitfalls that may be encountered when attempting to change established process and clinical culture21. The clinical effectiveness of checklists is dependent on more than the content of the checklist alone. Without effective staff involvement and training, implementation of such interventions may be, at best, ineffective22 or, at worst, a source of staff frustration and discord23. To address this issue, interventions have been shown to benefit from initial trialling in simulated environments to assess feasibility and clinical effectiveness, before full clinical implementation24. Such an approach benefits from the known advantages of simulation, with controlled and reliable scenarios within a dedicated assessment space and elimination of potential risks to patients. Patients who suffer complications in the postoperative phase are the group most at risk of adverse events13,25 and vulnerable to variations in their care7. In the management of postoperative morbidity, failures to adhere to principles of best practice place unwell patients at risk of avoidable harm and poorer outcomes (unpublished work, Pucher et al.). Adherence to care process-oriented guidelines, such as expert body recommendations26 or antimicrobial stewardship policies27, may contribute significantly to standardization of care and optimization of patient management. The aim of this study was to assess the clinical effectiveness of checklists in the management of common postoperative complications within a high-fidelity simulated ward environment. Methods General surgical registrars were recruited from a single academic hospital. Registrar-level trainees (with a minimum of 2 completed years of specialist training in general surgery) were selected, to reflect a level at which trainees can be expected to assume responsibility for routine patient assessment during the ward round in typical practice in the UK28. A sample size calculation was conducted, assuming a study powered to 90 per cent and an α level of 0·05. As this represented an exploratory study, with no available previous data for identical outcomes, published rates of surgical ward round error were used25, with the presumed effect size based on data from a previous study24 of surgical process checklists in the operating theatre. This suggested a necessary sample size of only three subjects in each group. Given the surrogate nature of data used, this was increased to ten to maximize data collection and increase the validity of results, based on experience from previous ward simulation-based trials29. Simulated ward rounds took place within a three-bed simulated ward. This was a fully furnished, high-fidelity simulation environment, complete with patient monitoring equipment and access to simulated patient records, radiology and pathology results (Fig. 1). The environment has been validated and described in detail previously29,30. Professional medical actors were employed to simulate patients and exhibit realistic clinical signs. An integrated audiovisual system recorded all activity digitally, with realistic background ‘white noise’ recorded from a real clinical environment. Fig. 1 Open in new tabDownload slide High-fidelity simulated ward environment Trainees were exposed to a series of simulated scenarios in which patients experienced one of the following postoperative complications: pneumonia, anastomotic leak, wound infection, urinary tract infection, postoperative bleeding and intra-abdominal sepsis. These complications were identified as the most common types following major gastrointestinal surgery, with subsequent development of checklists through expert input, best available evidence and end-user input (unpublished work, Pucher et al.). All trainees conducted a baseline ward round of three patients within the simulated ward, according to personal practice (Fig. 2). Ward rounds were supported by a nurse and a junior doctor trained to provide a standardized level of information relating to patients as part of the simulation, if prompted or asked by the trainee. Trainees were blinded to study design and assessment endpoints. To create a realistic sense of time pressure, trainees were told they were required to finish each ward round within 30 min, based on previous data from similar scenarios29. Patient scenarios assigned to each trainee's ward round were determined through a computer-generated randomization process, with all six scenarios being assessed by each trainee over the course of the study. Fig. 2 Open in new tabDownload slide CONSORT flow diagram for the study Following the baseline ward round assessment, each trainee was randomized (using a further computer-generated randomization sequence) to either a control or an intervention group. Trainees in the intervention group received a didactic session, with written and verbal instruction, on use of the previously developed checklists for management of surgical complications (unpublished work, Pucher et al.), then completed their final ward round, for which the checklists were made available for use (Appendix S1, supporting information). Trainees were informed that use of the checklists was not compulsory, and need be used only if the trainee was comfortable with them or felt they would help improve care given. Trainees in the control group conducted a final ward round according to their own personal practice, thus acting as controls for any potential learning effects resulting from repeated exposure to the simulator. Baseline and final ward rounds were conducted in separate sessions on different days. The primary endpoint for the study was the rate of failure to execute critical management steps for each postoperative complication. Each checklist involved five to eight management processes, such as the ordering of blood cultures and administration of antibiotics for sepsis, or arranging a blood transfusion and preparing a patient for return to the operating room in case of postoperative haemorrhage (unpublished work, Pucher et al.). Completion or omission of each checklist item was recorded in a binary manner through direct observation. Secondary endpoints included technical and non-technical performance. Technical performance was assessed using the Surgical Ward care Assessment Tool (SWAT), a checklist-based, validated tool to assess thoroughness of patient assessment (SWAT-A) and management (SWAT-M)25,29. Non-technical performance was assessed using the Ward NOn-TECHnical Skills (W-NOTECHS) scale, assessing five behavioural domains on a Likert scale, resulting in a final score of 5–2529,31. Prescribing errors were recorded. Each ward round was recorded digitally using two integrated ceiling cameras, with 20 per cent of ward rounds rated by a second blinded rater to assess potential for observer bias. Inter-rater reliability was assessed using Cronbach's α. Trainees in the intervention group were surveyed regarding their perceptions of the checklists, with questionnaire responses graded from 1 (strongly disagree) to 5 (strongly agree) on a Likert scale. Statistical analysis Demographic and ward round performance data for failure rate, SWAT, W-NOTECHS and time taken was compared between groups, in addition to within-group comparison of baseline and final ward rounds. Non-parametric tests were used, with Mann–Whitney U and Wilcoxon signed-rank tests for intergroup and intragroup comparisons respectively. The χ2 test was used to compare categorical data. Secondary post hoc analyses were performed to assess potentially confounding factors. To investigate differences in management of different complications, failure rates were stratified by scenario and differences in means assessed using ANOVA. A linear regression model was used to adjust for trainee factors (sex and years of experience) when interpreting the effect of trial group (control versus intervention) on failure rates. Questionnaire responses were collated and reported as mean(s.d.). Responses given as a 4 or 5 on a Likert scale were considered positive. Data analysis was performed using SPSS® version 21 (IBM, Armonk, New York, USA), with P < 0·050 considered statistically significant. All results are reported as median (i.q.r.) unless indicated otherwise. Results Twenty trainees were recruited, resulting in 120 patient assessments in the final analysis. Each trainee completed all six patient scenarios, randomized equally to control and intervention groups. All intervention group trainees elected to use the checklists in their final ward rounds. There were no differences in sex or experience levels of participants between groups (Table 1). Table 1 Demographics of subjects and controls . Control (n = 10) . Intervention (n = 10) . P . Sex ratio (M : F) 9 : 1 7 : 3 0·264† Length of surgical training (years)* 3 (2–4) 3 (2–4) 0·796‡ . Control (n = 10) . Intervention (n = 10) . P . Sex ratio (M : F) 9 : 1 7 : 3 0·264† Length of surgical training (years)* 3 (2–4) 3 (2–4) 0·796‡ * Values are median (i.q.r.). † χ2 test; ‡ Mann–Whitney U test. Open in new tab Table 1 Demographics of subjects and controls . Control (n = 10) . Intervention (n = 10) . P . Sex ratio (M : F) 9 : 1 7 : 3 0·264† Length of surgical training (years)* 3 (2–4) 3 (2–4) 0·796‡ . Control (n = 10) . Intervention (n = 10) . P . Sex ratio (M : F) 9 : 1 7 : 3 0·264† Length of surgical training (years)* 3 (2–4) 3 (2–4) 0·796‡ * Values are median (i.q.r.). † χ2 test; ‡ Mann–Whitney U test. Open in new tab Of all ward rounds, 20 per cent were assessed by a second blinded assessor, with excellent inter-rater reliability (Cronbach α = 0·893). Considering the rates of failure to complete critical processes in the management of postoperative complications, there was no difference between control and intervention groups for baseline assessment (58 (43–75) versus 67 (50–67) per cent respectively; P = 0·988) (Fig. 3). For the final ward round, process failure rates were significantly different, with near-elimination of errors in the intervention group (0 (0–0) per cent versus 60 (40–73) per cent for the control group; P < 0·001). All trainees in the intervention group elected to use the checklists where these were made available, for the final ward round. Fig. 3 Open in new tabDownload slide Failure to adhere to critical processes in the management of postoperative complications. Box plot comparisons of performance between control and intervention groups for a baseline and b final ward rounds. Thick bar denotes median value, box denotes i.q.r. and whiskers show the range; small circles indicate outliers. aP = 0·988, bP < 0·001 (Mann–Whitney U test) There were no statistically significant differences between groups at baseline for SWAT-A, SWAT-M or W-NOTECHS metrics (Table 2). In the final ward round, the intervention group scored significantly higher than the control group in patient management (SWAT-M: control 0·63 (0·50–0·75) versus intervention 0·83 (0·63–0·88); P = 0·005), although not for patient assessment (SWAT-A, P = 0·094) or non-technical skill (W-NOTECHS, P = 0·218). There was no significant difference in the time taken for each ward round between groups or between ward rounds. Table 2 Simulated ward round performance metrics . Baseline assessment . Final assessment . . Control (n = 30) . Intervention (n = 30) . P‡ . Control (no checklist) (n = 30) . Intervention (checklist) (n = 30) . P‡ . SWAT-A 0·62 (0·55–0·81) 0·73 (0·64–0·81) 0·436 0·74 (0·64–0·82) 0·77 (0·73–0·91) 0·094 Intragroup P† < 0·001 < 0·001 SWAT-M 0·63 (0·50–0·75) 0·56 (0·46–0·67) 0·350 0·63 (0·50–0·75) 0·83 (0·63–0·88) 0·005 Intragroup P† 0·571 < 0·001 W-NOTECHS 19·5 (17–22) 17·0 (16–20) 0·190 19·0 (18·5–21) 20·5 (18·5–23) 0·218 Intragroup P† 0·809 0·043 Ward round time (min)* 26·5(5·1) 26·8(5·1) 0·900 23·3(4·2) 22·2(5·7) 0·730 Intragroup P† 0·349 0·111 Prescription errors 10 22 0·019 15 5 0·043 Intragroup P† 0·298 0·020 . Baseline assessment . Final assessment . . Control (n = 30) . Intervention (n = 30) . P‡ . Control (no checklist) (n = 30) . Intervention (checklist) (n = 30) . P‡ . SWAT-A 0·62 (0·55–0·81) 0·73 (0·64–0·81) 0·436 0·74 (0·64–0·82) 0·77 (0·73–0·91) 0·094 Intragroup P† < 0·001 < 0·001 SWAT-M 0·63 (0·50–0·75) 0·56 (0·46–0·67) 0·350 0·63 (0·50–0·75) 0·83 (0·63–0·88) 0·005 Intragroup P† 0·571 < 0·001 W-NOTECHS 19·5 (17–22) 17·0 (16–20) 0·190 19·0 (18·5–21) 20·5 (18·5–23) 0·218 Intragroup P† 0·809 0·043 Ward round time (min)* 26·5(5·1) 26·8(5·1) 0·900 23·3(4·2) 22·2(5·7) 0·730 Intragroup P† 0·349 0·111 Prescription errors 10 22 0·019 15 5 0·043 Intragroup P† 0·298 0·020 Values are median (i.q.r.), except * mean(s.d.). SWAT(-A/M), Surgical Ward-care Assessment Tool (-Assessment/Management); W-NOTECHS, Ward NOn-TECHnical Skills scale. † Final versus baseline assessment (Wilcoxon signed-rank test); ‡ Mann–Whitney U test. Open in new tab Table 2 Simulated ward round performance metrics . Baseline assessment . Final assessment . . Control (n = 30) . Intervention (n = 30) . P‡ . Control (no checklist) (n = 30) . Intervention (checklist) (n = 30) . P‡ . SWAT-A 0·62 (0·55–0·81) 0·73 (0·64–0·81) 0·436 0·74 (0·64–0·82) 0·77 (0·73–0·91) 0·094 Intragroup P† < 0·001 < 0·001 SWAT-M 0·63 (0·50–0·75) 0·56 (0·46–0·67) 0·350 0·63 (0·50–0·75) 0·83 (0·63–0·88) 0·005 Intragroup P† 0·571 < 0·001 W-NOTECHS 19·5 (17–22) 17·0 (16–20) 0·190 19·0 (18·5–21) 20·5 (18·5–23) 0·218 Intragroup P† 0·809 0·043 Ward round time (min)* 26·5(5·1) 26·8(5·1) 0·900 23·3(4·2) 22·2(5·7) 0·730 Intragroup P† 0·349 0·111 Prescription errors 10 22 0·019 15 5 0·043 Intragroup P† 0·298 0·020 . Baseline assessment . Final assessment . . Control (n = 30) . Intervention (n = 30) . P‡ . Control (no checklist) (n = 30) . Intervention (checklist) (n = 30) . P‡ . SWAT-A 0·62 (0·55–0·81) 0·73 (0·64–0·81) 0·436 0·74 (0·64–0·82) 0·77 (0·73–0·91) 0·094 Intragroup P† < 0·001 < 0·001 SWAT-M 0·63 (0·50–0·75) 0·56 (0·46–0·67) 0·350 0·63 (0·50–0·75) 0·83 (0·63–0·88) 0·005 Intragroup P† 0·571 < 0·001 W-NOTECHS 19·5 (17–22) 17·0 (16–20) 0·190 19·0 (18·5–21) 20·5 (18·5–23) 0·218 Intragroup P† 0·809 0·043 Ward round time (min)* 26·5(5·1) 26·8(5·1) 0·900 23·3(4·2) 22·2(5·7) 0·730 Intragroup P† 0·349 0·111 Prescription errors 10 22 0·019 15 5 0·043 Intragroup P† 0·298 0·020 Values are median (i.q.r.), except * mean(s.d.). SWAT(-A/M), Surgical Ward-care Assessment Tool (-Assessment/Management); W-NOTECHS, Ward NOn-TECHnical Skills scale. † Final versus baseline assessment (Wilcoxon signed-rank test); ‡ Mann–Whitney U test. Open in new tab In terms of intragroup comparisons, a small but significant improvement was seen in both groups for SWAT-A (both P < 0·001). Regarding patient management (SWAT-M), improvement was seen only in the intervention group (P < 0·001). Similarly, total W-NOTECHS scores improved only in the intervention group (P = 0·043; control group, P = 0·809). There were no significant differences in median values for individual W-NOTECHS domain scores. Although the intervention group had a higher number of prescription errors recorded in the baseline round (22 versus 10 in the control group; P = 0·019), the use of checklists resulted in a significant reduction in the final round (5 versus 15 errors respectively; P = 0·043) (Table 2). There were no differences in performance between scenarios (ANOVA for baseline ward round, P = 0·796) (Table 3). Controlling for trainee experience and sex through inclusion as independent variables in a regression model had no effect on results, with trial group allocation the only significant predictor of process failure rates (odds ratio 0·54, 95 per cent c.i. 0·46 to 0·62; P < 0·001). Table 3 Baseline ward round performance stratified by clinical scenario . Failure rate (%) . Pneumonia 65(26) Wound infection 58(13) Anastomotic leak 61(13) Sepsis 61(13) Urinary tract infection 70(33) Postoperative haemorrhage 68(18) . Failure rate (%) . Pneumonia 65(26) Wound infection 58(13) Anastomotic leak 61(13) Sepsis 61(13) Urinary tract infection 70(33) Postoperative haemorrhage 68(18) Values are mean(s.d.). P = 0·796 (ANOVA). Open in new tab Table 3 Baseline ward round performance stratified by clinical scenario . Failure rate (%) . Pneumonia 65(26) Wound infection 58(13) Anastomotic leak 61(13) Sepsis 61(13) Urinary tract infection 70(33) Postoperative haemorrhage 68(18) . Failure rate (%) . Pneumonia 65(26) Wound infection 58(13) Anastomotic leak 61(13) Sepsis 61(13) Urinary tract infection 70(33) Postoperative haemorrhage 68(18) Values are mean(s.d.). P = 0·796 (ANOVA). Open in new tab Questionnaire responses All intervention group study participants returned completed questionnaires. Responses were positive, with all believing that checklists were easy to use and would improve clinical practice (Table 4). All participants indicated that they would want the checklists used if they themselves were to experience a postoperative complication. Table 4 Responses to use of checklists for management of postoperative complications Questionnaire statement . Response score* . Positive response (%) . The checklists were easy to use 4·7 (4–5) 100 The checklists cover clinical conditions relevant to my practice 4·8 (4–5) 100 I found this checklist applicable to my practice 4·1 (3–5)  80 The checklists may improve management of complications 4·4 (4–5) 100 I would want this checklist to be used for me if I experienced a postoperative complication 4·3 (4–5) 100 I would consider using the checklists in my routine practice 4·1 (3–5)  80 Questionnaire statement . Response score* . Positive response (%) . The checklists were easy to use 4·7 (4–5) 100 The checklists cover clinical conditions relevant to my practice 4·8 (4–5) 100 I found this checklist applicable to my practice 4·1 (3–5)  80 The checklists may improve management of complications 4·4 (4–5) 100 I would want this checklist to be used for me if I experienced a postoperative complication 4·3 (4–5) 100 I would consider using the checklists in my routine practice 4·1 (3–5)  80 * Values are mean (range). Responses were received from ten of ten subjects in the intervention group. Open in new tab Table 4 Responses to use of checklists for management of postoperative complications Questionnaire statement . Response score* . Positive response (%) . The checklists were easy to use 4·7 (4–5) 100 The checklists cover clinical conditions relevant to my practice 4·8 (4–5) 100 I found this checklist applicable to my practice 4·1 (3–5)  80 The checklists may improve management of complications 4·4 (4–5) 100 I would want this checklist to be used for me if I experienced a postoperative complication 4·3 (4–5) 100 I would consider using the checklists in my routine practice 4·1 (3–5)  80 Questionnaire statement . Response score* . Positive response (%) . The checklists were easy to use 4·7 (4–5) 100 The checklists cover clinical conditions relevant to my practice 4·8 (4–5) 100 I found this checklist applicable to my practice 4·1 (3–5)  80 The checklists may improve management of complications 4·4 (4–5) 100 I would want this checklist to be used for me if I experienced a postoperative complication 4·3 (4–5) 100 I would consider using the checklists in my routine practice 4·1 (3–5)  80 * Values are mean (range). Responses were received from ten of ten subjects in the intervention group. Open in new tab Discussion Despite their common nature, wide variations in the management of postoperative complications exist across different centres2,8,32–35. Checklists are being adopted increasingly in routine surgical care36, but no such clinical aids have been available for the non-routine care of patients suffering postoperative complications. The results of the present study demonstrate not only the significant variability that exists in the management of common complications, but how error and variation in clinical care may be eliminated essentially through a single intervention. The introduction of checklists to guide patient management, once the diagnosis of a postoperative complication had been made and a decision taken to treat, led to the standardization of treatment in line with evidence-based guidelines, and reduction of median error rates from 60 to 0 per cent. Despite their simple nature, or perhaps because of it, implementing checklists in the manner of this study has the potential to reduce the variability of ward-based care25 and may act significantly to reduce failure-to-rescue rates and improve outcomes. The experimental, simulated context in which this study was conducted represents a potential confounder in the interpretation of results, although the overwhelmingly positive effect is in agreement with previously published data16,24. The introduction of checklist-based protocols for antiseptic central venous catheter placement eliminated catheter-associated sepsis over the course of 3 months16. A simulation-based trial assessing the management of operating room crises reported a reduction in error rates of 75 per cent when checklists were used to guide the operating room team's response to rare events such as asystole, malignant hyperthermia or unstable arrhythmia24. Patient management without the aid of checklists placed patients at risk of critical errors such as the failure to prescribe antibiotics for a patient with an anastomotic leak, insufficient intravenous access in a patient suffering from haemorrhage, or failure to obtain blood cultures in a patient with sepsis and pyrexia. Checklist use led to near elimination of these errors, standardizing care without compromising clinicians' decision-making or patient care. It is important to note that the checklists were introduced as aids to guide care only, with clinicians free to use them (or not) as they felt appropriate, dependent upon the suspected diagnosis and intended treatment. Although they represented an additional piece of paperwork, the checklists were well received, with all reporting that they were easy to use, and improved their practice. Checklist use did not prolong the ward rounds, with no significant difference in time taken between groups. Perhaps most significantly, all trainees reported that they would want checklists used in their own care, were they to require care for postoperative complications themselves. Introduction of the checklist also led to improvements in non-technical skill. The ability of checklists to flatten clinical hierarchies and improve communication has been seen previously in improved safety attitudes19 and staff questionnaire responses37, but this is the first time that the effect of a checklist on non-technical performance has been observed directly and measured in a clinical context. The positive feedback seen in this study represents a crucial aspect of checklist implementation, illustrating the benefit of simulation-based trialling of such an intervention before introduction into clinical practice. Centres have reported significant difficulties associated with attempts to change established clinical process21–23. Negative feedback captured in the simulation can therefore be addressed before clinical implementation. This study, intended as an exploratory study of the effectiveness of checklists for postoperative care, must be considered in the context of its limitations. Only the performance of surgical trainees was assessed. Other health professionals, including nursing staff and other grades of medical staff, were not included. As registrars are the group of surgical trainees commonly responsible for routine patient review, and most likely to make the first diagnosis of a postoperative complication and initiate treatment, this is unlikely to represent a major confounder. Interaction with other members of the clinical team, although important, would be secondary to the clinician's diagnosis and decision to treat. This was a single-centre study of a simulation-based trial. Localization and adaptation to local protocols and practices is a critical component of successful checklist implementation18,38, and adapted checklists would be required for individual centres. The study did not evaluate the longevity or retention of changes to practice, which may experience degradation of checklist use over time. Other groups implementing checklist-type interventions have variably reported both deterioration39 and improvement16 of performance over time. Analogous to the acquisition of technical skill through repeated training40, it may be that repeated training sessions should be considered for such non-technical interventions. Further longitudinal assessments of postoperative checklists would be required. Finally, it must be recognized that the management of postoperative complications is a two-stage process, requiring the correct diagnosis first to be made before appropriate management can be instituted. Checklists address the latter, but not the former7,8. Although management of a given condition is amenable to checklist-based care, based upon best evidence and expert guidelines, diagnostic ability is dependent on a clinician's own ability and practice. To improve patient assessment and the recognition of complications, broader interventions, such as educational curricula to improve trainees' conduct of ward rounds in general, can be effective41,42. This study has demonstrated the viability of checklists in the management of postoperative complications. Future work should include a clinical implementation study. As seen in this and other checklist studies17,24, steps to ensure staff participation and ‘buy-in’ are essential to success with the involvement of local ‘champions’, regular audit and performance feedback, as well as simulation-based trialling and staff training to ensure appropriate use. Development of checklists for further complications, such as pulmonary embolism or deep vein thrombosis, might also be considered. Considering the common nature of complications, improving the care of patients with postoperative morbidity has perhaps the greatest potential for impact in reducing failure-to-rescue rates; the use of appropriate checklists may standardize treatment and improve outcomes. Disclosure The authors declare no conflict of interest. References 1 Ito H , Are C, Gonen M, D'Angelica M, Dematteo RP, Kemeny NE et al. Effect of postoperative morbidity on long-term survival after hepatic resection for metastatic colorectal cancer . Ann Surg 2008 ; 247 : 994 – 1002 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Ghaferi AA , Birkmeyer JD, Dimick JB. Hospital volume and failure to rescue with high-risk surgery . Med Care 2011 ; 49 : 1076 – 1081 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Hii MW , Smithers BM, Gotley DC, Thomas JM, Thomson I, Martin I et al. Impact of postoperative morbidity on long-term survival after oesophagectomy . Br J Surg 2013 ; 100 : 95 – 104 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Merkow RP , Bilimoria KY, McCarter MD, Phillips JD, Decamp MM, Sherman KL et al. Short-term outcomes after esophagectomy at 164 American College of Surgeons National Surgical Quality Improvement Program Hospitals: effect of operative approach and hospital-level variation . Arch Surg 2012 ; 147 : 1009 – 1016 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Cone MM , Herzig DO, Diggs BS, Rea JD, Hardiman KM, Lu KC. Effect of surgical approach on 30-day mortality and morbidity after elective colectomy: a NSQIP study . J Gastrointest Surg 2012 ; 16 : 1212 – 1217 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Molena D , Mungo B, Stem M, Feinberg RL, Lidor AO. Outcomes of esophagectomy for esophageal achalasia in the United States . J Gastrointest Surg 2014 ; 18 : 310 – 317 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Ghaferi AA , Birkmeyer JD, Dimick JB. Variation in hospital mortality associated with inpatient surgery . N Engl J Med 2009 ; 361 : 1368 – 1375 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Ghaferi AA , Birkmeyer JD, Dimick JB. Complications, failure to rescue, and mortality with major inpatient surgery in medicare patients . Ann Surg 2009 ; 250 : 1029 – 1034 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Aiken LH , Clarke SP, Sloane DM, Sochalski J, Silber JH. Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction . JAMA 2002 ; 288 : 1987 – 1993 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Friese CR , Earle CC, Silber JH, Aiken LH. Hospital characteristics, clinical severity, and outcomes for surgical oncology patients . Surgery 2010 ; 147 : 602 – 609 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Ghaferi AA , Osborne NH, Birkmeyer JD, Dimick JB. Hospital characteristics associated with failure to rescue from complications after pancreatectomy . J Am Coll Surg 2010 ; 211 : 325 – 330 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Creamer GL , Dahl A, Perumal D, Tan G, Koea JB. Anatomy of the ward round: the time spent in different activities . ANZ J Surg 2010 ; 80 : 930 – 932 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Vincent C , Neale G, Woloshynowych M. Adverse events in British hospitals: preliminary retrospective record review . BMJ 2001 ; 322 : 517 – 519 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Reason J Human error: models and management . BMJ 2000 ; 320 : 768 – 770 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Reason J Combating omission errors through task analysis and good reminders . Qual Saf Health Care 2002 ; 11 : 40 – 44 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Pronovost P , Needham D, Berenholtz S, Sinopoli D, Chu H, Cosgrove S et al. An intervention to decrease catheter-related bloodstream infections in the ICU . N Engl J Med 2006 ; 355 : 2725 – 2732 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Haynes AB , Weiser TG, Berry WR, Lipsitz SR, Breizat AH, Dellinger EP et al. ; Safe Surgery Saves Lives Study Group. A surgical safety checklist to reduce morbidity and mortality in a global population . N Engl J Med 2009 ; 360 : 491 – 499 . Google Scholar Crossref Search ADS PubMed WorldCat 18 de Vries EN , Prins HA, Crolla RM, den Outer AJ, van Andel G, van Helden SH et al. ; SURPASS Collaborative Group . Effect of a comprehensive surgical safety system on patient outcomes . N Engl J Med 2010 ; 363 : 1928 – 1937 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Haynes AB , Weiser TG, Berry WR, Lipsitz SR, Breizat AH, Dellinger EP et al. ; Safe Surgery Saves Lives Study Group . Changes in safety attitude and relationship to decreased postoperative morbidity and mortality following implementation of a checklist-based surgical safety intervention . BMJ Qual Saf 2011 ; 20 : 102 – 107 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Russ S , Rout S, Sevdalis N, Moorthy K, Darzi A, Vincent C. Do safety checklists improve teamwork and communication in the operating room? A systematic review . Ann Surg 2013 ; 258 : 856 – 871 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Borchard A , Schwappach DL, Barbir A, Bezzola P. A systematic review of the effectiveness, compliance, and critical factors for implementation of safety checklists in surgery . Ann Surg 2012 ; 256 : 925 – 933 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Spence J , Goodwin B, Enns C, Dean H. Student-observed surgical safety practices across an urban regional health authority . BMJ Qual Saf 2011 ; 20 : 580 – 586 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Conley DM , Singer SJ, Edmondson L, Berry WR, Gawande AA. Effective surgical safety checklist implementation . J Am Coll Surg 2011 ; 212 : 873 – 879 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Arriaga AF , Bader AM, Wong JM, Lipsitz SR, Berry WR, Ziewacz JE et al. Simulation-based trial of surgical-crisis checklists . N Engl J Med 2013 ; 368 : 246 – 253 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Pucher PH , Aggarwal R, Darzi A. Surgical ward round quality and impact on variable patient outcomes . Ann Surg 2014 ; 259 : 222 – 226 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Hooton TM , Bradley SF, Cardenas DD, Colgan R, Geerlings SE, Rice JC et al. ; Infectious Diseases Society of America. Diagnosis, prevention, and treatment of catheter-associated urinary tract infection in adults: 2009 International Clinical Practice Guidelines from the Infectious Diseases Society of America . Clin Infect Dis 2010 ; 50 : 625 – 663 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Evans RS , Pestotnik SL, Classen DC, Clemmer TP, Weaver LK, Orme JF et al. A computer-assisted management program for antibiotics and other antiinfective agents . N Engl J Med 1998 ; 338 : 232 – 238 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Intercollegiate Surgical Curriculum Programme . General Surgery Specialty Syllabus ; 2013. http://www.iscp.ac.uk [accessed 1 April 2014]. 29 Pucher PH , Aggarwal R, Srisatkunam T, Darzi A. Validation of the simulated ward environment for assessment of ward-based surgical care . Ann Surg 2014 ; 259 : 215 – 221 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Pucher PH , Darzi A, Aggarwal R. Simulation for ward processes of surgical care . Am J Surg 2013 ; 206 : 96 – 102 . Google Scholar Crossref Search ADS PubMed WorldCat 31 Steinemann S , Berg B, Ditullio A, Skinner A, Terada K, Anzelon K et al. Assessing teamwork in the trauma bay: introduction of a modified ‘NOTECHS’ scale for trauma . Am J Surg 2012 ; 203 : 69 – 75 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Glance LG , Dick AW, Meredith JW, Mukamel DB. Variation in hospital complication rates and failure-to-rescue for trauma patients . Ann Surg 2011 ; 253 : 811 – 816 . Google Scholar Crossref Search ADS PubMed WorldCat 33 McHugh MD , Kelly LA, Smith HL, Wu ES, Vanak JM, Aiken LH. Lower mortality in magnet hospitals . Med Care 2013 ; 51 : 382 – 388 . Google Scholar Crossref Search ADS PubMed WorldCat 34 Almoudaris AM , Mamidanna R, Bottle A, Aylin P, Vincent C, Faiz O et al. Failure to rescue patients after reintervention in gastroesophageal cancer surgery in England . JAMA Surg 2013 ; 148 : 272 – 276 . Google Scholar Crossref Search ADS PubMed WorldCat 35 Henneman D , Snijders HS, Fiocco M, van Leersum NJ, Kolfschoten NE, Wiggers T et al. Hospital variation in failure to rescue after colorectal cancer surgery: results of the Dutch Surgical Colorectal Audit . Ann Surg Oncol 2013 ; 20 : 2117 – 2123 . Google Scholar Crossref Search ADS PubMed WorldCat 36 Adamina M , Kehlet H, Tomlinson GA, Senagore AJ, Delaney CP. Enhanced recovery pathways optimize health outcomes and resource utilization: a meta-analysis of randomized controlled trials in colorectal surgery . Surgery 2011 ; 149 : 830 – 840 . Google Scholar Crossref Search ADS PubMed WorldCat 37 Blanco M , Clarke JR, Martindell D. Wrong site surgery near misses and actual occurrences . AORN J 2009 ; 90 : 215–218 , 221 – 212 . Google Scholar OpenURL Placeholder Text WorldCat 38 Sparks EA , Wehbe-Janek H, Johnson RL, Smythe WR, Papaconstantinou HT. Surgical Safety Checklist compliance: a job done poorly! J Am Coll Surg 2013 ; 217 : 867 – 873 . e1 – e3 . Google Scholar Crossref Search ADS PubMed WorldCat 39 Vats A , Vincent CA, Nagpal K, Davies RW, Darzi A, Moorthy K. Practical challenges of introducing WHO surgical checklist: UK pilot experience . BMJ 2010 ; 340 : b5433 . Google Scholar Crossref Search ADS PubMed WorldCat 40 Crochet P , Aggarwal R, Dubb SS, Ziprin P, Rajaretnam N, Grantcharov T et al. Deliberate practice on a virtual reality laparoscopic simulator enhances the quality of surgical technical skills . Ann Surg 2011 ; 253 : 1216 – 1222 . Google Scholar Crossref Search ADS PubMed WorldCat 41 Pucher PH , Aggarwal R, Singh P, Srisatkunam T, Twaij A, Darzi A. Ward simulation to improve surgical ward round performance: a randomised controlled trial of a simulation-based curriculum . Ann Surg 2014 ; [Epub ahead of print]. Google Scholar OpenURL Placeholder Text WorldCat 42 Pucher PH , Darzi A, Aggarwal R. Development of an evidence-based curriculum for training of ward-based surgical care . Am J Surg 2014 ; 207 : 213 – 217 . Google Scholar Crossref Search ADS PubMed WorldCat © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd

van Broekhoven, D L M; Deroose, J P; Bonvalot, S; Gronchi, A; Grünhagen, D J; Eggermont, A M M; Verhoef, C

doi: 10.1002/bjs.9659pmid: 25296896

Abstract Background Aggressive fibromatoses (desmoid tumours) may be locally aggressive, but do not metastasize. Although a conservative approach is advocated for most patients, pain and functional impairment are indications for active treatment. Tumour necrosis factor (TNF) α and melphalan-based isolated limb perfusion (TM-ILP) is a limb-saving treatment modality for soft tissue tumours. This study reports the results of TM-ILP treatment in patients with aggressive fibromatosis. Methods Institutional databases of three European centres were searched. All patients who received TM-ILP treatment for aggressive fibromatosis between 1990 and 2012 were included. Before therapy, the patients were discussed at multidisciplinary tumour board meetings. Results Twenty-five patients received 28 TM-ILP treatments. The median age of patients was 28 (i.q.r. 19–34) years and median hospital stay was 8 (7–12) days. Median follow-up was 84 (34–114) months. A complete response was achieved after two TM-ILP treatments, and a partial response after 17 treatments in 16 patients. Stable disease was reported after eight treatments in seven patients, including a patient with stable disease after the first treatment and progression after the second TM-ILP. Toxicity was modest after most treatments; Wieberdink grade IV (extensive epidermolysis, and threatening or manifest compartment syndrome) was seen after two TM-ILP treatments. Systemic leakage was reported after one treatment, but did not lead to systemic toxicity. Functional outcome was good; 16 patients had no physical limitations, and six patients had some limitations but did not need medical aids. Amputation was prevented in all but three patients. Conclusion TNF-α-based ILP is effective in patients with aggressive fibromatosis. Introduction Aggressive fibromatosis, or desmoid tumour, is a rare benign soft tissue tumour that is localized throughout the body. Aggressive fibromatosis does not metastasize, but locally advanced or recurrent disease is frequently seen in tumours localized in the extremities1–3. Extra-abdominal aggressive fibromatosis is not associated with mortality, unlike intra-abdominal tumours4. Current literature advocates a conservative approach for these benign tumours, owing to compelling evidence of disease stabilization and spontaneous tumour regression in many patients5,6. Despite the benign nature, some tumours behave aggressively, leading to pain and functional impairment. In patients with severely impaired quality of life, a conservative approach may no longer be an option. The choice of treatment depends on many factors, and an algorithm was proposed recently6. Involvement or proximity of vital structures, for instance major nerves and vessels, may lead to mutilating surgery or even amputation. In these patients, a limb-saving strategy that results in relief of symptoms should be preferred. In 1958, isolated limb perfusion (ILP) was introduced as a treatment modality for extremity malignancies, such as locally advanced sarcomas and melanoma in-transit metastases7. In patients with sarcoma this technique appeared ineffective with cytostatics alone, but gained markedly in efficacy when tumour necrosis factor (TNF) α was added8,9. Currently, TNF-α and melphalan-based ILP (TM-ILP) is the standard of care in Europe for patients with limb-threatening sarcomas, and leads to limb salvage in up to 89 per cent of patients10. The present study reports data from three sarcoma centres of the European Organization for Research and Treatment of Cancer (EORTC) relating to patients with locally advanced aggressive fibromatosis treated with TM-ILP to avoid limb amputation. Methods Consecutive patients with aggressive fibromatosis who underwent treatment with TM-ILP in participating EORTC centres between 1990 and 2012 were included in the study. Participating centres were: Gustave Roussy Cancer Campus, Villejuif, Paris, France, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy, and Erasmus Medical Centre Cancer Institute, Rotterdam, The Netherlands. Owing to the benign tumour biology of aggressive fibromatosis, only patients suffering from intolerable pain or functional impairment are considered for surgical treatment in these institutions. The safety and efficacy of conservative treatment is currently being evaluated in prospective trials in all three centres. For patients in the present study, a variety of treatments had been performed previously, and radical resection was deemed possible only with mutilating surgery or amputation. In these patients, TM-ILP treatment was considered. Before TM-ILP, all patients were discussed at a multidisciplinary tumour board meeting. Perfusion The ILP technique has been described extensively8,11. The procedure is performed under general anaesthesia. After heparinization, the targeted blood circuit is isolated by clamping and cannulation of the major artery and vein, and connected to an oxygenated extracorporeal circuit. A tourniquet compresses collateral vessels to prevent leakage. Using a precordial scintillation probe to detect technetium-labelled albumin, leakage is monitored throughout the procedure. TNF-α and melphalan are used; after 60–90 min of perfusion, the active compounds are rinsed from the vascular system of the limb and the circulation is restored. There were minimal differences in the ILP protocols of the contributing institutions; details of the procedures have been described previously9,12–13. Response and toxicity Response was evaluated by clinical examination and MRI 4–8 weeks after ILP treatment, and reported according to World Health Organization criteria14. Complete response (CR) was defined as complete disappearance of the tumour; partial response (PR) was a decrease in tumour size of more than 50 per cent; and stable disease (SD) was recorded when the criteria for neither PR nor progressive disease were met. Disease progression was defined as a 25 per cent increase in tumour size with no initial documentation of CR, PR or SD. In patients with incomplete radiographic measurements, the response was based on available radiological information and clinical judgement. Acute regional toxicity after perfusion was classified according to Wieberdink et al.15: grade I, no reaction; grade II, slight erythema or oedema; grade III, considerable erythema or oedema with some blistering, slightly disturbed motility permissible; grade IV, extensive epidermolysis, and threatening or manifest compartment syndrome; and grade V, reaction that may necessitate amputation. Functional outcome was based on clinical assessment and categorized as perfect, impairment without the necessity for medical aids, and impairment with the need for medical aids or amputation. Continuous data are presented as median (i.q.r.). Results A total of 25 patients received 28 TM-ILP treatments. The median age of the patients at the time of TM-ILP was 28 (19–34) years and median hospital stay after treatment was 8 (7–12) days. The majority of patients were female. Median length of follow-up was 84 (34–114) months. Baseline characteristics of the patients are summarized in Table 1. Five patients were treated for a primary tumour; two of these had previously received systemic anti-inflammatory treatment for inoperable disease, but this did not lead to sufficient reduction or relief of complaints. The indication for choosing ILP as a primary treatment in the other three patients was bone involvement, sciatic nerve involvement, and refusal of systemic treatment by the patient. Table 1 Baseline characteristics of patients with aggressive fibromatosis . No. of patients* (n = 25) . Sex ratio (M : F) 8 : 17 Age (years)† 28 (19–34) Localization Arm 4 Leg 21 Size of lesion (cm)† 12 (8.5–15) No. of tumours Single 14 Multiple 11 Tumour treated Primary tumour 5 Primary recurrence 10 Secondary recurrence 10 Previous treatment None 3 Surgery 7 Radiotherapy 2 Systemic treatment 2 Surgery + radiotherapy 3 Surgery + radiotherapy + systemic treatment 4 Surgery + systemic treatment 4 . No. of patients* (n = 25) . Sex ratio (M : F) 8 : 17 Age (years)† 28 (19–34) Localization Arm 4 Leg 21 Size of lesion (cm)† 12 (8.5–15) No. of tumours Single 14 Multiple 11 Tumour treated Primary tumour 5 Primary recurrence 10 Secondary recurrence 10 Previous treatment None 3 Surgery 7 Radiotherapy 2 Systemic treatment 2 Surgery + radiotherapy 3 Surgery + radiotherapy + systemic treatment 4 Surgery + systemic treatment 4 * Unless indicated otherwise; † values are median (i.q.r.). Open in new tab Table 1 Baseline characteristics of patients with aggressive fibromatosis . No. of patients* (n = 25) . Sex ratio (M : F) 8 : 17 Age (years)† 28 (19–34) Localization Arm 4 Leg 21 Size of lesion (cm)† 12 (8.5–15) No. of tumours Single 14 Multiple 11 Tumour treated Primary tumour 5 Primary recurrence 10 Secondary recurrence 10 Previous treatment None 3 Surgery 7 Radiotherapy 2 Systemic treatment 2 Surgery + radiotherapy 3 Surgery + radiotherapy + systemic treatment 4 Surgery + systemic treatment 4 . No. of patients* (n = 25) . Sex ratio (M : F) 8 : 17 Age (years)† 28 (19–34) Localization Arm 4 Leg 21 Size of lesion (cm)† 12 (8.5–15) No. of tumours Single 14 Multiple 11 Tumour treated Primary tumour 5 Primary recurrence 10 Secondary recurrence 10 Previous treatment None 3 Surgery 7 Radiotherapy 2 Systemic treatment 2 Surgery + radiotherapy 3 Surgery + radiotherapy + systemic treatment 4 Surgery + systemic treatment 4 * Unless indicated otherwise; † values are median (i.q.r.). Open in new tab Treatment outcomes An overall response was seen after 19 of 28 TM-ILP treatments (Table 2). A CR was achieved in two patients; both patients had TM-ILP as a single treatment modality for recurrent disease after previous surgical resections, and one patient had also undergone radiotherapy. These two patients had a sustained CR to the end of follow-up at 94 and 36 months. Table 2 Characteristics of the 28 treatments with isolated limb perfusion . No. of treatments . Type of ILP Axillar 3 Brachial 1 Iliacal 10 Femoral 10 Popliteal 4 Response Complete 2 Partial 17 Stable disease 8 Progression 1 Wieberdink grade I 4 II 16 III 6 IV 2 Duration of hospital stay (days)† 8 (7–12) Treatment after ILP‡ None 21 Resection 3 Radiotherapy 2 Amputation 3 Local recurrence after ILP No 17 Yes 11 Limb function§ Perfect 16 Limited 6 Amputated 3 . No. of treatments . Type of ILP Axillar 3 Brachial 1 Iliacal 10 Femoral 10 Popliteal 4 Response Complete 2 Partial 17 Stable disease 8 Progression 1 Wieberdink grade I 4 II 16 III 6 IV 2 Duration of hospital stay (days)† 8 (7–12) Treatment after ILP‡ None 21 Resection 3 Radiotherapy 2 Amputation 3 Local recurrence after ILP No 17 Yes 11 Limb function§ Perfect 16 Limited 6 Amputated 3 * Unless indicated otherwise; † values are median (i.q.r.). ‡ One patient had both amputation and radiotherapy. § For the 25 patients at end of follow-up. ILP, isolated limb perfusion. Open in new tab Table 2 Characteristics of the 28 treatments with isolated limb perfusion . No. of treatments . Type of ILP Axillar 3 Brachial 1 Iliacal 10 Femoral 10 Popliteal 4 Response Complete 2 Partial 17 Stable disease 8 Progression 1 Wieberdink grade I 4 II 16 III 6 IV 2 Duration of hospital stay (days)† 8 (7–12) Treatment after ILP‡ None 21 Resection 3 Radiotherapy 2 Amputation 3 Local recurrence after ILP No 17 Yes 11 Limb function§ Perfect 16 Limited 6 Amputated 3 . No. of treatments . Type of ILP Axillar 3 Brachial 1 Iliacal 10 Femoral 10 Popliteal 4 Response Complete 2 Partial 17 Stable disease 8 Progression 1 Wieberdink grade I 4 II 16 III 6 IV 2 Duration of hospital stay (days)† 8 (7–12) Treatment after ILP‡ None 21 Resection 3 Radiotherapy 2 Amputation 3 Local recurrence after ILP No 17 Yes 11 Limb function§ Perfect 16 Limited 6 Amputated 3 * Unless indicated otherwise; † values are median (i.q.r.). ‡ One patient had both amputation and radiotherapy. § For the 25 patients at end of follow-up. ILP, isolated limb perfusion. Open in new tab A PR was recorded after 17 TM-ILP treatments in 16 patients. For 11 patients, TM-ILP led to control of disease and symptoms that was sustained to the end of follow-up with no need for additional treatment; an example is shown in Fig. 1. Of the other five patients, one received adjuvant radiotherapy to achieve disease control. Surgical resection of residual disease was performed after the first TM-ILP in one patient presenting with recurrent disease. A second recurrence, treated by further TM-ILP 32 months after the first treatment, also resulted in a PR and disease control to the end of follow-up. Three patients with an initial PR after TM-ILP developed progression of remaining disease during follow-up. One of these patients had tumour deposits throughout one leg and had previously undergone extensive radiotherapy. The patient was included in a phase I study (GW786034) of treatment with paclitaxel and pazopanib. During this treatment, the patient developed necrosis of the foot and a below-knee amputation was performed 3 years after TM-ILP. Several tumour deposits were still in situ and stable at the end of follow-up. The other two patients with progression received systemic treatment for the progressive disease, which led to disease control. Fig. 1 Open in new tabDownload slide Patient with recurring aggressive fibromatosis of the right foot. a Macroscopic appearance and b axial view of contrast-enhanced, T1-weighted MRI before isolated limb perfusion (ILP); c macroscopic appearance and d axial view of contrast-enhanced, T1-weighted MRI 24 months after ILP SD was recorded after eight TM-ILP treatments in seven patients. The response to TM-ILP was sufficient to achieve disease control in four of these patients, which was sustained in two to the end of follow-up. One patient with initial disease control developed tumour progression after 24 months and received systemic treatment; the other developed a new lesion after 57 months, with SD of the primary tumour. After systemic treatment, both lesions remained stable to the end of follow-up. In the other three patients with SD, the response was sufficient to perform limb-sparing surgery. One of these patients, with a 30 per cent tumour response after TM-ILP, had macroscopically negative margins after surgical (R1) resection. Four years after resection, disease progression led to a second treatment with TM-ILP. After treatment, the leg was amputated above the knee owing to healing problems. However, a new recurrence occurred in the stump, for which ex-articulation of the hip was performed with postoperative adjuvant radiotherapy. At the end of follow-up, the patient was free from disease. Another patient, who had an R1 resection after TM-ILP, developed recurrence that was treated successfully with chemotherapy. In the last patient with SD, disease control was achieved after surgery (R1 resection). After 19 months a second recurrence was diagnosed. Systemic treatment was ineffective, and a second TM-ILP was performed 56 months after the first. Owing to continued disease progression, the patient underwent above-knee amputation and radiotherapy of the stump, which resulted in disease stabilization. The patient developed yet another recurrence, which was treated with systemic therapy. This was the only patient with progressive disease at the end of follow-up. Toxicity and function Local and systemic toxicity were modest. Wieberdink grade I was recorded after four procedures, grade II after 16 procedures and grade III after six procedures. More severe toxicity (Wieberdink grade IV) was seen after two TM-ILP treatments: one patient required fasciotomy and necrosectomy, and the other needed amputation as a result of rapid disease progression. Leakage of perfusate to the systemic circulation was seen after one treatment. The leakage was managed conservatively and did not result in systemic toxicity. Functional outcome was good. Sixteen patients had no limitations of physical function; six had some limitations of the limb, but without the need for medical aids. Amputation could not be prevented in three patients (Table 2). Recurrence and survival Local recurrence or disease progression after initial disease control was documented following 11 of the 28 TM-ILP treatments. The median time for tumour recurrence or progression was 27 (17–44) months. No patient died during follow-up. Discussion This multicentre study indicates that treatment with TM-ILP may be a limb-saving strategy for aggressive fibromatosis in patients for whom previous therapy has failed or where surgical treatment might result in severe limb impairment. A tumour response was seen after 19 of 28 TM-ILP treatments, and amputation was avoided in all but three of the 25 patients. A limitation of this study is the absence of pain scores before treatment, as a result of the retrospective study design. Although details are lacking, the indication for TM-ILP for each patient was assessed carefully during a multidisciplinary board meeting, which included evaluation of pain due to tumour growth. Tumour behaviour of aggressive fibromatosis varies greatly. At the time of diagnosis, a conservative approach with careful follow-up only is advocated, in agreement with a recent study6. Even minor tumour progression in the absence of patient complaints may justify a conservative approach. Aggressive local tumour growth or intolerable pain are reasons for active therapy. In these patients, surgery is the mainstay of treatment, but adequate surgical resection may lead to severe impairment of the limb or even amputation. In these patients, treatment with TM-ILP might be successful and achieve an excellent limb salvage rate, as shown in the present series. Chemotherapy, non-cytotoxic systemic treatment and radiotherapy are other options when the extent of disease excludes surgery as a primary treatment modality. Several different chemotherapy regimens have been proposed, with varied response rates16. In an overview by the French Sarcoma Group17, 62 patients were analysed after treatment with various regimens; the overall response rate was 21 per cent. The systemic side-effects of chemotherapy for a localized tumour disease warrant careful consideration before treatment. Less aggressive systemic treatment options are hormone therapy, non-steroid anti-inflammatory drugs or the tyrosine kinase inhibitor imatinib. Response rates of approximately 50 per cent have been reported18,19, but only small groups of patients were studied and with limited follow-up time. Side-effects vary depending on the type of treatment, but often include gastrointestinal complaints and fatigue. Long-term side-effects are unknown and follow-up protocols have yet to be established. Most of the experience with radiotherapy has been acquired in the adjuvant setting. Few reports discuss radiotherapy as a single-modality treatment for aggressive fibromatosis. Ballo and colleagues20 showed a 5-year progression-free survival rate of 69 per cent. A review21 reported local tumour control rates of 78 per cent. The most compelling data are from a recent EORTC study22, which included 44 patients with inoperable or incompletely resected disease with a median follow-up of 4·8 years; all patients received a dose of 56 Gy in 28 fractions. The 3-year local control rate was 81·5 per cent. A further response with tumour regression was seen after 3 years in three patients; two patients had a CR and one had a PR. Acute grade 3 side-effects were limited to the skin, mucosa and pain, whereas mild oedema was the late toxic effect in ten patients22. In the present study, nine patients had received radiotherapy before TM-ILP, excluding radiotherapy as a treatment option. For patients without previous radiotherapy, the present authors advocate treatment with TM-ILP because of the potential for surgical resection, and radiotherapy, or even repeat TM-ILP, after the initial treatment. In areas that are not amenable to TM-ILP, such as the groin or buttocks, radiotherapy might be an attractive alternative for unresectable and limb-threatening tumours. Local and systemic toxicity was limited in the present series. Toxicity in these patients was comparable to that reported in previous series9–10,12 of nearly 400 TM-ILP treatments of extremity sarcomas. Patient selection is essential in the current era of tailored treatment strategies. TNF-α-based ILP is an aggressive treatment for this relatively benign disease. Patients who are considered for TM-ILP have tumours with aggressive tumour biology and severe symptoms. Aetiological studies23,24 have shown the importance of the Wnt pathway and β-catenin in the development of aggressive fibromatosis. Specific mutations of the CTNNB1 gene (which encodes β-catenin) have been associated with aggressive fibromatosis25–27. The precise mechanism and the different effects exerted by the specific mutations are not well understood. Several studies28–31 have demonstrated a predictive value for the mutations on the risk of tumour recurrence after surgery. Whether the mutational status of the tumour has an effect on the outcome after TM-ILP is unknown. TNF-α-based ILP is an effective limb-sparing technique for the treatment of aggressive fibromatosis in selected patients. It should be considered after failure of initial therapy and where surgery for recurrent or progressive disease would lead to functional loss or amputation. Acknowledgements D.L.M.v.B. and J.P.D. are joint first authors of this article. Disclosure: The authors declare no conflict of interest. Snapshot quiz 14/15 References 1 Salas S , Dufresne A, Bui B, Blay JY, Terrier P, Ranchere-Vince D et al. Prognostic factors influencing progression-free survival determined from a series of sporadic desmoid tumors: a wait-and-see policy according to tumor presentation . J Clin Oncol 2011 ; 29 : 3553 – 3558 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Crago AM , Denton B, Salas S, Dufresne A, Mezhir JJ, Hameed M et al. A prognostic nomogram for prediction of recurrence in desmoid fibromatosis . Ann Surg 2013 ; 258 : 347 – 353 . Google Scholar Crossref Search ADS PubMed WorldCat 3 van Broekhoven DL , Verhoef C, Elias SG, Witkamp AJ, van Gorp JM, van Geel BA et al. Local recurrence after surgery for primary extra-abdominal desmoid-type fibromatosis . Br J Surg 2013 ; 100 : 1214 – 1219 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Quintini C , Ward G, Shatnawei A, Xhaja X, Hashimoto K, Steiger E et al. Mortality of intra-abdominal desmoid tumors in patients with familial adenomatous polyposis: a single center review of 154 patients . Ann Surg 2012 ; 255 : 511 – 516 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Mitchell G , Thomas JM, Harmer CL. Aggressive fibromatosis: evidence for a stable phase . Sarcoma 1998 ; 2 : 149 – 154 . Google Scholar Crossref Search ADS PubMed WorldCat 6 Gronchi A , Colombo C, Le Péchoux C, Dei Tos AP, Le Cesne A, Marrari A et al. Sporadic desmoids-type fibromatosis: a stepwise approach to a non-metastasising neoplasm – a position paper from the Italian and the French Sarcoma Group . Ann Oncol 2014 ; 25 : 578 – 583 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Creech O , Krementz ET, Ryan RF, Winbland JN. Chemotherapy of cancer: regional perfusion utilizing an extracorporeal circuit . Ann Surg 1958 ; 148 : 616 – 632 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Eggermont AM , Schraffordt Koops H, Klausner JM, Kroon BB, Schlag PM et al. Isolated limb perfusion with tumor necrosis factor and melphalan for limb salvage in 186 patients with locally advanced soft tissue extremity sarcomas. The cumulative multicenter European experience . Ann Surg 1996 ; 224 : 756 – 764 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Pennacchioli E , Deraco M, Mariani L, Fiore M, Mussi C, Collini P et al. Advanced extremity soft tissue sarcoma: prognosic effect of isolated limb perfusion in a series of 88 patients treated at a single institution . Ann Surg Oncol 2007 ; 14 : 553 – 559 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Deroose JP , Eggermont AM, van Geel AN, Burger JW, den Bakker MA, de Wilt JH et al. Long-term results of tumor necrosis factor alpha- and melphalan-based isolated limb perfusion in locally advanced extremity soft tissue sarcomas . J Clin Oncol 2011 ; 29 : 4036 – 4044 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Eggermont AM , Schraffordt Koops H, Klausner JM, Schlag PM, Kroon BB, Ben-Ari G et al. Isolated limb perfusion with high-dose tumor necrosis factor-alpha for locally advanced extremity soft tissue sarcomas . Cancer Treat Res 1997 ; 91 : 189 – 203 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Bonvalot S , Rimareix F, Causeret S, Le Péchoux C, Boulet B, Terrier P et al. Hyperthermic isolated limb perfusion in locally advanced soft tissue sarcoma and progressive desmoid-type fibromatosis with TNF 1 mg and melphalan (T1-M HILP) is safe and efficient . Ann Surg Oncol 2009 ; 16 : 3350 – 3357 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Grunhagen DJ , de Wilt JH, Verhoef C, van Geel AN, Eggermont AM. TNF-based isolated limb perfusion in unresectable extremity desmoid tumours . Eur J Surg Oncol 2005 ; 31 : 912 – 916 . Google Scholar Crossref Search ADS PubMed WorldCat 14 World Health Organization (WHO) . Handbook for Reporting Results of Cancer Treatment . WHO : Geneva , 1979 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 15 Wieberdink J , Benckhuysen C, Braat RP, van Slooten EA, Olthuis GA. Dosimetry in isolation perfusion of the limbs by assessment of perfused tissue volume and grading of toxic tissue reactions . Eur J Cancer Clin Oncol 1982 ; 18 : 905 – 910 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Constatinidou A , Jones RL, Scurr M, Al-Muderis O, Judson I. Advanced aggressive fibromatoses: effective palliation with chemotherapy . Acta Oncol 2011 ; 50 : 455 – 461 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Garbay D , Le Cesne A, Penel N, Chevreau C, Marec-Berard P, Blay JY et al. Chemotherapy in patients with desmoid tumors: a study from the French Sarcoma Group (FSG) . Ann Oncol 2012 ; 23 : 182 – 186 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Bocale D , Rotelli MT, Cavallini A, Altomare DF. Anti-oestrogen therapy in the treatment of desmoid tumours: a systematic review . Colorectal Dis 2011 ; 13 : e388 – e395 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Penel N , Le Cesne A, Bui BN, Perol D, Brain EG, Ray-Coquard I et al. Imatinib for progressive and recurrent aggressive fibromatosis (desmoid tumors): an FNCLCC/French Sarcoma Group phase II trial with a long-term follow-up . Ann Oncol 2011 ; 22 : 452 – 457 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Ballo MT , Zagars GK, Pollack A. Radiation therapy in the management of desmoid tumors . Int J Radiat Oncol Biol Phys 1998 ; 42 : 1007 – 1014 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Nuyttens JJ , Rust PF, Thomas CR, Turrisi AT. Surgery versus radiation therapy for patients with aggressive fibromatosis or desmoid tumors: a comparative review of 22 articles . Cancer 2000 ; 88 : 1517 – 1523 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Keus RB , Nout RA, Blay JY, de Jong JM, Hennig I, Saran F et al. Results of a phase II pilot study of moderate dose radiotherapy for inoperable desmoid-type fibromatosis – an EORTC STBSG and ROG study (EORTC 62991–22998) . Ann Oncol 2013 ; 24 : 2672 – 2676 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Tejpar S , Nollet F, Li C, Wunder JS, Michils G, dal Cin P et al. Predominance of beta-catenin mutations and beta-catenin dysregulation in sporadic aggressive fibromatosis (desmoid tumor) . Oncogene 1999 ; 18 : 6615 – 6620 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Alman BA , Li C, Pajerski ME, Diaz-Cano S, Wolfe HJ. Increased beta-catenin protein and somatic APC mutations in sporadic aggressive fibromatoses (desmoid tumors) . Am J Pathol 1997 ; 151 : 329 – 334 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 25 Lips DJ , Barker N, Clevers H, Hennipman A. The role of APC and beta-catenin in the aetiology of aggressive fibromatosis (desmoid tumors) . Eur J Surg Oncol 2009 ; 35 : 3 – 10 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Tejpar S , Michils G, Denys H, Van Dam K, Nik SA, Jadidizadeh A et al. Analysis of Wnt/Beta catenin signalling in desmoid tumors . Acta Gastroenterol Belg 2005 ; 68 : 5 – 9 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 27 Miyoshi Y , Iwao K, Nawa G, Yoshikawa H, Ochi T, Nakamura Y. Frequent mutations in the beta-catenin gene in desmoid tumors from patients without familial adenomatous polyposis . Oncol Res 1998 ; 10 : 591 – 594 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 28 Lazar AJ , Tuvin D, Hajibashi S, Habeeb S, Bolshakov S, Mayordomo-Aranda E et al. Specific mutations in the beta-catenin gene (CTNNB1) correlate with local recurrence in sporadic desmoid tumors . Am J Pathol 2008 ; 173 : 1518 – 1527 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Dômont J , Salas S, Lacroix L, Brouste V, Saulnier P, Terrier P et al. High frequency of β-catenin heterozygous mutations in extra-abdominal fibromatosis: a potential molecular tool for disease management . Br J Cancer 2010 : 102 ; 1032 – 1036 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Colombo C , Miceli R, Lazar AJ, Perrone F, Pollock RE, Le Cesne A et al. CTNNB1 45 F mutation is a molecular prognosticator of increased postoperative primary desmoid tumor recurrence: an independent, multicenter validation study . Cancer 2013 ; 119 : 3696 – 3702 . Google Scholar Crossref Search ADS PubMed WorldCat 31 van Broekhoven DLM , Verhoef C, Grünhagen DJ, Van Gorp J, Van Dalen T. Prognostic value of CTNNB1-gene mutation in primary extra-abdominal aggressive fibromatosis. Society of Surgical Oncology 67th Annual Cancer Symposium , Phoenix, Arizona , 12–15 March 2014 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd

Brunner, S M; Kesselring, R; Rubner, C; Martin, M; Jeiter, T; Boerner, T; Ruemmele, P; Schlitt, H J; Fichtner-Feigl, S

doi: 10.1002/bjs.9627pmid: 25331841

Abstract Background Liver metastases occur in 40–50 per cent of patients with colorectal cancer and determine long-term survival. The aim of this study was to examine the immunological architecture of colorectal liver metastases and its impact on patient survival. Methods Specimens from patients with colorectal liver metastases were stained with haematoxylin and eosin and Masson trichrome, immunostained for α-smooth muscle actin, CD4, CD45RO and CD8, and analysed by flow cytometry. In addition to histomorphological evaluation, immunohistochemically stained sections were analysed for cell numbers in the tumour area, infiltrative margin and distant liver stroma separately. These findings were correlated with clinical data and patient outcome. Results Tumour containment by a fibrotic capsule around liver metastases was observed in 37·8 per cent of 201 patients and was prognostic for improved survival (median (s.e.) survival 64 (6) and 31 (4) months for patients with capsule and no capsule respectively; P < 0·001) and independently led to higher R0 resection rates (P = 0·040). In multivariable analysis, CD45RO+ cell infiltration at the peritumoral margin with low CD45RO+ cell infiltration in the distant liver stroma (P = 0·001) and fibrotic capsule formation (P = 0·008) both independently prolonged patient survival. Using these two factors, a cellular immune score was designed and shown to stratify patient survival in test and validation samples (both P < 0·001). Conclusion Fibrotic capsule formation and localized cell infiltration of colorectal liver metastases by CD45RO+ cells were related to prolonged patient survival. Based on these immunological criteria a cellular immune score was developed to stratify patients according to prognosis. Introduction Primary colorectal carcinomas are currently classified according to the International Union Against Cancer (UICC)/American Joint Committee on Cancer (AJCC) TNM system1, which stages these tumours by histological criteria based on the extent of wall infiltration of resected tissue and the presence of metastases. This system is designed to code the extent of cancer in order to assess prognosis for individual patients1. However, it is recognized that outcome can vary considerably within the same histological cancer stages2. To explain variations in cancer outcome, many studies3–9 have investigated the interplay between tumours and their immunological microenvironment. These studies have shown the importance of tumour-infiltrating immune cells to patient survival. In particular, infiltration of primary tumours by CD4+ Th1 helper, CD8+ cytotoxic and CD45RO+ memory T cells was demonstrated to have beneficial effects with regard to overall survival and also metastasis formation10–12. In contrast, strong infiltration by Th17 cells was related to poor prognosis, whereas Th2 clustering in colorectal tumours did not influence patient outcome13. Moreover, various scores have been developed to predict patient survival using infiltration patterns of different immune cells3,14. As these immune scores for primary colorectal tumours have shown better prognostic ability than the UICC/AJCC TNM staging system, it has been proposed to use immune scoring as a new component for the classification of cancer2,7,15. For colorectal liver metastases, however, there is no ideal scoring system with which to stratify the risk and survival chances of patients16. Evaluation of outcome following liver resection for colorectal metastases using the UICC/AJCC TNM system has been demonstrated to be inappropriate, whereas the use of clinical risk scores might be helpful for risk stratification of patients17–19. However, in a review20 that systematically analysed the impact of 15 clinical scoring systems, none of the existing predictive models was able to provide an accurate risk stratification for patients with colorectal liver metastases. Prognostic models for patients with liver metastases that are based on immune cell infiltration patterns analogous to those evaluated in examination of primary colorectal carcinomas are currently lacking. To address this problem, the infiltration of colorectal liver metastases by CD4+, CD8+ and CD45RO+ cells was examined in the present study at different locations: tumour area, infiltrative margin and distant, non-tumorous liver stroma. Cellular infiltration patterns that correlate with overall survival depending on the location of cell infiltration were identified. Based on these findings, a scoring system was developed and the prognostic relevance of this cellular immune score (CIS) for patient survival was evaluated. Methods This study was performed at the University Medical Centre of Regensburg, Germany and was approved by the local ethics committee (no. 12-101-0009). Patient selection and data collection Patients with colorectal liver metastases (confirmed by histology) who underwent liver surgery at the University Medical Centre of Regensburg between 2004 and 2010 were identified from the hospital computer database. Only patients who had surgery with curative intent were included. If there was not enough tissue for histological evaluation, patients were excluded from this study. General patient information, TNM classification of primary tumours and histological data on colorectal liver metastases were obtained from the routine pathology report. Chemotherapy within 3 months before liver resection was considered as preoperative chemotherapy. Response to chemotherapy was quantified using Response Evaluation Criteria In Solid Tumours (RECIST) criteria and tumour regression grade21,22. Survival data were collected using the database of the Tumorzentrum Regensburg (regional tumour centre of East Bavaria), Germany. Only patients receiving primary liver resection were included in survival analyses. Specimens from patients undergoing liver resection for recurrence were compared with samples from the primary liver resection to test the consistency of the tumour microenvironment within an individual over time. Samples of multiple metastases in individual patients were compared with each other to ensure consistency of immune cell infiltration. To provide an internal control for survival analyses and evaluation of the CIS, patients were stratified by operation date and divided randomly into a test and a validation sample. To confirm the results from this retrospective histological analysis, fresh liver tissue from patients undergoing liver resection was collected prospectively after informed consent had been given. Histological analysis Tissue cross-sections were stained using haematoxylin and eosin and Masson trichrome, and immunostained for CD4, CD8, CD45RO and α-alpha smooth muscle actin (α-SMA), as described in Appendix S1 (supporting information). Digital images of liver metastases were obtained at 10 × and 20 × magnification on light microscopy (Carl Zeiss MicroImaging, Jena, Germany). Specimens were evaluated for a fibrotic capsule around tumours using haematoxylin and eosin, Masson trichrome and α-SMA staining. Numbers of CD4+, CD45RO+ and CD8+ cells were counted manually in three randomly selected areas (20 × magnification) and the mean taken separately for three areas: tumour (Tu), near stroma (Sn) and distant stroma (Sd) using ImageJ 1·45s software (Wayne Rasband, National Institutes of Health, Bethesda, Maryland, USA); an example is shown in Appendix 1 (supporting information). Sn was defined as the area within one field of view (20 ×) next to the tumour, and Sd as the outmost area of the specimen in a tumour-free region. For survival analysis of both the test and validation samples, patients were divided into two groups depending on numbers of CD4+, CD45RO+ and CD8+ cells in the different areas. Cell numbers larger than the median value for each cell type were defined as high, and numbers smaller than or equal to the median as low. To express tumour-specific immune cell infiltration, ratios of CD4+, CD45RO+ and CD8+ cell numbers in the Sn region relative to cell numbers in the Sd area were calculated. For further analysis, a ratio of 2 or greater was defined as ‘high’, and less than 2 as ‘low’. Cellular immune score A score that took into account factors that influence patient survival significantly in univariable and multivariable analysis was developed to enable patient risk stratif ication. A high CD45RO+ cell ratio and capsule formation were each assigned 1 point. Patients with 2 points were considered to be at low risk, 1 point signified medium risk and 0 points defined a high-risk group. This score was calculated separately for the test and validation samples. Cell isolation and flow cytometry Fresh tissue from tumour, infiltrative margin and distant liver was obtained from surgically resected colorectal liver metastases in a subgroup of patients and analysed separately (Appendix 1, supporting information). Statistical analysis Kaplan–Meier graphs were plotted for survival analyses. Group comparisons were made using the log rank test. Factors that were statistically significant in univariable analysis were included in a multivariable Cox proportional hazards model. Pearson's correlation was used to evaluate correlations with cellular expression profiles of liver metastases, and the Mann–Whitney U test for evaluation of continuous variables. The level of significance was set at P < 0·050. All histological evaluations were performed independently in a blinded fashion by two examiners. In the event of disagreement, consensus was achieved by re-evaluating the cross-sections together. To explore reproducibility, after 3 months random samples were drawn and re-evaluated. In all histological evaluations the concordance rate between the observers and the reproducibility of the results exceeded 95 per cent. Results A total of 201 patients (83 women, 41·3 per cent; 118 men, 58·7 per cent) with a median age of 62 (i.q.r. 53–68) years were enrolled. The primary tumour was colonic in 119 patients (59·2 per cent) and rectal in 82 (40·8 per cent). Liver metastases occurred synchronously in 93 patients (46·3 per cent) and metachronously in 108 (53·7 per cent). No chemotherapy had been administered within 3 months of operation in 111 patients (55·2 per cent), whereas 90 (44·8 per cent) had received preoperative chemotherapy. The chemotherapy regimens used were 5-fluorouracil (5-FU; 13 patients), folinic acid–5-FU–oxaliplatin (FOLFOX; 41), folinic acid–5-FU–irinotecan (FOLFIRI; 9), FOLFOX + antibody (13) and FOLFIRI + antibody (14). Overall median survival was 37 months after liver surgery with a 3-year survival rate of 51·1 per cent (Fig. S1, supporting information). Clinical descriptors or TNM classification of primary tumours and risk stratification of patients with liver metastases To evaluate the use of the TNM classification of primary tumours or descriptive clinical parameters to assess patient survival after resection of colorectal liver metastases, a Kaplan–Meier survival analysis was performed for: colonic versus rectal carcinoma; tumour category, node status and grade of primary tumours; primary versus recurrent, synchronous versus metachronous and single versus multiple liver metastases; and R0 versus R1 liver resections. Only node status of the primary tumour (P = 0·003) and growth pattern with single/multiple metastases (P = 0·004) correlated with survival (Fig. S1, supporting information). Fibrotic capsule formation around liver metastasis Histological analysis by haematoxylin and eosin staining showed capsule formation around liver metastases in 76 patients (37·8 per cent), whereas in 125 (62·2 per cent) a capsule was absent (Fig. 1a). Analysis of recurrent liver metastases and also multiple metastases demonstrated that this encapsulation remained consistent within a single patient over time (Figs S2 and S3, supporting information). Masson trichrome staining demonstrated that capsules mainly consisted of collagen fibres (Fig. 1b). Immunohistochemical staining for α-SMA was strongly positive in the area of the capsule showing that activated myofibroblasts as potential source of collagen are located in this area (Fig. 1c). Fig. 1 Open in new tabDownload slide Histomorphological analysis showing fibrotic capsule formation around some liver metastases. a Representative haematoxylin and eosin-stained sections of liver metastases with and without capsule formation. b Masson trichrome staining demonstrates that this capsule consists mainly of collagen fibres (blue) and c is strongly positive for α-smooth muscle actin (brown) which depicts activated myofibroblasts located in this area (haematoxylin counterstain in c). Fig. 1 continues on next page Capsule formation, patient survival and R0 resection rate Kaplan–Meier analysis demonstrated that patients with fibrotic capsules around liver metastases survived longer than those without capsule formation (median (s.e.) 64 (6) versus 31 (4) months; hazard ratio 2·53, 95 per cent c.i. 1·52 to 4·22; P < 0·001) (Fig. 2a). The R0 resection rate was higher among patients with capsules than in those without (92 versus 82 per cent; P = 0·040). To determine whether longer survival of patients with capsule formation was due to the higher rate of R0 resection, an additional analysis was performed separately for patients who had R0 and R1 resection. Among R0 resections, patients with capsule formation survived longer than those without (undefined versus median (s.e.) 34 (6) months; hazard ratio 2·78, 1·58 to 4·91; P < 0·001) (Fig. 2b). If only R1 resection was achieved, survival was not significantly associated with capsule formation (P = 0·236) (Fig. 2c). Fig. 2 Open in new tabDownload slide Kaplan–Meier analysis of patient survival in relation to capsule formation in a all patients, b patients who had R0 resection of liver metastases and c those who had R1 resection. a,bP < 0·001, cP = 0·236 (log rank test) Peritumoral infiltration of CD4+, CD45RO+ and CD8+ cells and patient survival Immunohistochemical analysis for CD4+ (Fig. S4a,b, supporting information), CD45RO+ (Fig. 3a) and CD8+ (Fig. S5a,b, supporting information) cells revealed that significantly more of these cells were localized in the Sn region around the tumour (median counts: CD4+, 210 cells per high-power field (HPF); CD45RO+, 220 cells/HPF; CD8+, 109 cells/HPF) than in the Tu area (24, 36 and 10 cells/HPF respectively; P < 0·001 for all 3 T cell phenotypes) and in the Sd region (49, 56 and 45 cells/HPF respectively; P < 0·001 for all 3 T cell phenotypes). The distribution of these cells in analyses of resection samples from recurrent liver metastases remained consistent within individual patients over time (Fig. S2b–d, supporting information). Furthermore, the cellular infiltration pattern of patients with multiple metastases was shown to be homogeneous between the different metastatic lesions (Fig. S3b–d, supporting information). The immune cell infiltration of the colorectal liver metastases was not influenced significantly by any kind of chemotherapy administered within 3 months before liver resection, or by response to chemotherapy quantified by RECIST criteria or tumour regression grade (Tables S1–S4, supporting information). Fig. 3 Open in new tabDownload slide Association between localized infiltration of CD45RO+ cells at the infiltrative margin and patient survival. a Representative immunohistochemical staining for CD45RO+ cells (brown) in tumour (Tu), near stroma (Sn) and distant stroma (Sd) (normal liver) (haematoxylin counterstain). b–d Survival in relation to cell numbers in the b Tu c Sn and d Sd regions. e Survival in relation to CD45RO+ ratio (cell numbers in Sn relative to Sd region). bP = 0·349, cP < 0·001, dP = 0·294, eP = 0·003 (log rank test) Flow cytometry demonstrated that, of all CD45+ leucocytes, the highest numbers of CD4+ cells were found within the Tu area (36 per cent) (P = 0·011) (Fig. S6a,c, supporting information). In contrast, more CD45+ leucocytes were CD45RO+ within the Sn (54 per cent) and Sd (52 per cent) regions (P < 0·001) (Fig. S6b,d, supporting information). Comparable to the results for CD45RO+ cells, the highest relative numbers of CD8+ cells were found within the Sn (29 per cent) and Sd (30 per cent) regions (P < 0·001) (Fig. S6a,e, supporting information). For survival analysis the patients were divided into groups with high and low levels of infiltration by each cell type, based on the median value for that cell type. Survival of patients in the test sample with high CD4+ (P = 0·317) (Fig. S4c, supporting information), CD45RO+ (P = 0·349) (Fig. 3b) and CD8+ (P = 0·713) (Fig. S5c, supporting information) cell numbers in the Tu region was no different from that of patients with low cell infiltration. However, significantly increased survival was noted for patients with high CD4+ (median (s.e.) 62 (4) versus 34 (5) months; hazard ratio 2·16, 95 per cent c.i. 1·29 to 3·61; P = 0·003) (Fig. S4d, supporting information), CD45RO+ (53 (9) versus 37 (8) months; hazard ratio 1·29, 0·76 to 2·19; P < 0·001) (Fig. 3c) and CD8+ (64 (4) versus 36 (6) months; hazard ratio 2·32, 1·38 to 3·91; P = 0·002) (Fig. S5d, supporting information) cell numbers in the Sn region compared with patients with low infiltration by these respective cell types. In contrast, no statistically significant survival differences based on CD4+ (P = 0·372) (Fig. S4e, supporting information), CD45RO+ (P = 0·294) (Fig. 3d) and CD8+ (P = 0·773) (Fig. S4d, supporting information) cell infiltration in the Sd area was detected. Localized peritumoral dense CD4+, CD45RO+ and CD8+ cell infiltration can be expressed by high CD4+, CD45RO+ and CD8+ ratios (cell numbers in Sn relative to Sd region). If this localized dense cell infiltration (high cell ratios) was present, patients in the test sample survived significantly longer than those with low ratios of these cells (P < 0·001, P = 0·003 and P < 0·001 respectively) (Fig. 3e; Figs S4f and S5f, supporting information). To confirm these results, patients in the validation sample were analysed in the same manner. This cohort also showed better survival if capsule formation (P = 0·008), and a high ratio of CD4+, CD45RO+ and CD8+ cells (P = 0·002, P = 0·020 and P < 0·001 respectively) (Fig. S7, supporting information) was detected. Peritumoral infiltration by CD4+, CD45RO+ and CD8+ cells and capsule formation A correlation analysis was performed to investigate further whether the cellular infiltration pattern influences the development of capsule formation around liver metastases. This revealed that a capsule typically occurred when there was a high level of infiltration by CD4+, CD45RO+ and CD8+ cells in the Sn region, combined with low numbers of the respective cells in the Sd area (Fig. S8a–c, supporting information). More specifically, this was demonstrated by the fact that capsule formation enclosing the liver metastases almost only occurred in patients with high CD4+, CD45RO+ and CD8+ cell ratios; those with low CD4+, CD45RO+ and CD8+ cell ratios hardly ever showed capsule formation (P < 0·001) (Fig. S8d–i, supporting information). Further analysis showed no additional associations between capsule and origin, growth pattern and size of colorectal liver metastases, previous chemotherapy, tumour category or grade, apart from a lower rate of lymph node metastases from the primary tumour (Table S5, supporting information). Cellular immune score and survival of patients with liver metastases All factors that significantly influenced patient survival in the test sample in univariable analysis were tested by multivariable analysis (Table 1). A high CD45RO+ cell ratio (P = 0·001) and the presence of capsule formation (P = 0·008) were found to correlate significantly with prolonged patient survival, and 1 point was assigned to each. For risk stratification, patients in the sample were divided into three groups according to the CIS. The low-risk group (CIS 2) showed significantly better survival than the medium-risk (CIS 1) and high-risk (CIS 0) groups (median (s.e.) 62 (5), 48 (8) and 18 (4) months respectively; P < 0·001) (Fig. 4a). To confirm these results, patients in the randomly selected validation sample were analysed in the same manner. Analogous to the test sample, multivariable analysis identified a high CD45RO+ cell ratio (P = 0·004) and capsule formation (P = 0·031) as independent factors influencing patient survival (Table 2). The CIS also enabled risk stratification of patients in this cohort (median survival 68 (23), 37 (8) and 29 (6) months for patients at low, medium and high risk respectively; P = 0·001) (Fig. 4b). Table 1 Multivariable Cox proportional hazards analysis of factors that were significant predictors of survival in univariable analysis in the test sample . Hazard ratio . P . Capsule (no versus yes) 2·11 (1·22, 3·65) 0·008 CD45RO ratio (low versus high) 2·19 (1·37, 3·52) 0·001 CD4 ratio (low versus high) 1·31 (0·80, 2·13) 0·281 CD8 ratio (low versus high) 0·83 (0·50, 1·36) 0·449 . Hazard ratio . P . Capsule (no versus yes) 2·11 (1·22, 3·65) 0·008 CD45RO ratio (low versus high) 2·19 (1·37, 3·52) 0·001 CD4 ratio (low versus high) 1·31 (0·80, 2·13) 0·281 CD8 ratio (low versus high) 0·83 (0·50, 1·36) 0·449 Values in parentheses are 95 per cent c.i. Open in new tab Table 1 Multivariable Cox proportional hazards analysis of factors that were significant predictors of survival in univariable analysis in the test sample . Hazard ratio . P . Capsule (no versus yes) 2·11 (1·22, 3·65) 0·008 CD45RO ratio (low versus high) 2·19 (1·37, 3·52) 0·001 CD4 ratio (low versus high) 1·31 (0·80, 2·13) 0·281 CD8 ratio (low versus high) 0·83 (0·50, 1·36) 0·449 . Hazard ratio . P . Capsule (no versus yes) 2·11 (1·22, 3·65) 0·008 CD45RO ratio (low versus high) 2·19 (1·37, 3·52) 0·001 CD4 ratio (low versus high) 1·31 (0·80, 2·13) 0·281 CD8 ratio (low versus high) 0·83 (0·50, 1·36) 0·449 Values in parentheses are 95 per cent c.i. Open in new tab Fig. 4 Open in new tabDownload slide Correlation between cellular immune score and survival of patients with colorectal liver metastases: a test sample and b validation sample. aP < 0·001, bP = 0·001 (log rank test) Table 2 Multivariable Cox proportional hazards analysis of factors that were significant predictors of survival in univariable analysis in the validation sample . Hazard ratio . P . Capsule (no versus yes) 3·09 (1·11, 8·61) 0·031 CD45RO ratio (low versus high)  4·00 (1·56, 10·28) 0·004 CD4 ratio (low versus high) 0·91 (0·33, 2·69) 0·907 CD8 ratio (low versus high) 0·47 (0·17, 1·28) 0·139 . Hazard ratio . P . Capsule (no versus yes) 3·09 (1·11, 8·61) 0·031 CD45RO ratio (low versus high)  4·00 (1·56, 10·28) 0·004 CD4 ratio (low versus high) 0·91 (0·33, 2·69) 0·907 CD8 ratio (low versus high) 0·47 (0·17, 1·28) 0·139 Values in parentheses are 95 per cent c.i. Open in new tab Table 2 Multivariable Cox proportional hazards analysis of factors that were significant predictors of survival in univariable analysis in the validation sample . Hazard ratio . P . Capsule (no versus yes) 3·09 (1·11, 8·61) 0·031 CD45RO ratio (low versus high)  4·00 (1·56, 10·28) 0·004 CD4 ratio (low versus high) 0·91 (0·33, 2·69) 0·907 CD8 ratio (low versus high) 0·47 (0·17, 1·28) 0·139 . Hazard ratio . P . Capsule (no versus yes) 3·09 (1·11, 8·61) 0·031 CD45RO ratio (low versus high)  4·00 (1·56, 10·28) 0·004 CD4 ratio (low versus high) 0·91 (0·33, 2·69) 0·907 CD8 ratio (low versus high) 0·47 (0·17, 1·28) 0·139 Values in parentheses are 95 per cent c.i. Open in new tab Discussion In this study a fibrotic, α-SMA-expressing, activated fibroblast-containing capsule was detected by histomorphological evaluation in about one-third of patients with colorectal liver metastases. The presence of a fibrotic capsule occurred independently of preoperative chemotherapy and correlated with patient survival, which was more than twice as long for patients with capsule formation than for those without. This finding confirms the results of previous studies23–25 which also described capsule formation in 20–60 per cent of patients that was associated with a reduced local recurrence rate and better survival. An additional finding of the present study was a higher R0 resection rate if a capsule was present around liver metastases. Furthermore, the longer survival of patients with capsules in the R0 resection group demonstrated that better survival in the event of capsule formation and the higher R0 resection rate are two independent factors. The higher R0 resection rate could be due to better visualization of the metastasis margins by the surgeon, or to a biochemical and mechanical barrier function of the capsule24. It was also shown that localized infiltration by CD45RO+ cells at the infiltrative margin of colorectal liver metastases combined with normal cell numbers in the periphery of the liver was an independent positive predictor of patient survival. Patients with high cell ratios (at least 2) survived longer than those with low ratios. This parallels studies8,10–12,26 of primary colorectal carcinomas that have shown survival benefits for patients with high levels of tumour-infiltrating lymphocytes, especially CD45RO+ memory cells, adjacent to CD8+ cytotoxic T lymphocytes and CD4+ Th1 lymphocytes around the tumour. Further analysis revealed a correlation between the presence of a fibrotic capsule and a high ratio of CD4+, CD45RO+ and CD8+ cells. More specifically, 48–84 per cent of patients with a ratio of CD4+, CD45RO+ and CD8+ cells of at least 2 developed a fibrotic capsule whereas a fibrotic capsule was rarely detected in those with lower ratios. These factors, in combination with data from previous studies demonstrating higher levels of mononuclear cells in patients with capsule formation, suggest that CD4+, CD45RO+ and CD8+ cells might be involved directly in the generation of the fibrotic capsule24. In the setting of liver fibrosis, hepatic stellate cells (HSC) have been identified as key players in the production of extracellular matrix proteins27. These HSC differentiate into α-SMA-positive myofibroblasts, which were present in the area of the fibrotic capsules in this and previous studies23,24. CD4+, CD45RO+ and CD8+ cells possibly influence HSC to produce extracellular matrix proteins in an interleukin 13- and transforming growth factor β1-dependent manner28,29. This study showed that peritumoral infiltration of CD45RO+ cells and the development of a fibrotic capsule are independent prognostic factors for increased survival of patients with colorectal liver metastases. Most other parameters describing or categorizing liver metastases were found not to be relevant to patient survival once liver metastases have occurred. Clinical scoring systems potentially predict the feasibility of resection and indicate operative risks, but have been demonstrated not to be suitable for prognostic assessment in patients with colorectal liver metastases1,16–17,19,30–32. Therefore, a CIS was developed that incorporates the relevant immunological predictors of outcome and provided a survival stratification of patients with colorectal liver metastases in test and validation samples. The criteria incorporated in this new CIS for colorectal liver metastases are comparable to factors that have been shown to influence response to chemotherapy of colorectal liver metastases and survival of patients with primary colorectal carcinomas3,6,10,12,33. Based on these studies, a worldwide task force is currently implementing a cancer classification system for primary colorectal tumours using an immune score in addition to the AJCC/UICC TNM classification, the so-called TNM-Immune. A similar approach has been undertaken in the present study for colorectal liver metastases. Possible limitations of this study are its retrospective nature and the fact that the prognostic ability of the CIS was not confirmed in an independent patient cohort, although it was tested in test and validation samples. However, this study has identified immunological factors that influence survival in patients with colorectal liver metastases and immune scoring can now be extended to this patient group. An interesting point to address would be a comparison of these results with the immune infiltrate of the primary tumour. However, as the authors' unit serves as a tertiary hepatobiliary referral centre, the majority of primary colonic carcinomas were resected in the referring hospitals. Acknowledgements This study was supported by the University of Regensburg. Disclosure: The authors declare no conflict of interest. References 1 Edge SB , Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM . Ann Surg Oncol 2010 ; 17 : 1471 – 1474 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Galon J , Pages F, Marincola FM, Thurin M, Trinchieri G, Fox BA et al. The immune score as a new possible approach for the classification of cancer . J Transl Med 2012 ; 10 : 1 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Bindea G , Mlecnik B, Fridman WH, Galon J. The prognostic impact of anti-cancer immune response: a novel classification of cancer patients . Semin Immunopathol 2011 ; 33 : 335 – 340 . Google Scholar Crossref Search ADS PubMed WorldCat 4 Camus M , Tosolini M, Mlecnik B, Pagès F, Kirilovsky A, Berger A et al. Coordination of intratumoral immune reaction and human colorectal cancer recurrence . Cancer Res 2009 ; 69 : 2685 – 2693 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Fridman WH , Galon J, Dieu-Nosjean MC, Cremer I, Fisson S, Damotte D et al. Immune infiltration in human cancer: prognostic significance and disease control . Curr Top Microbiol Immunol 2011 ; 344 : 1 – 24 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 6 Galon J , Costes A, Sanchez-Cabo F, Kirilovsky A, Mlecnik B, Lagorce-Pagès C et al. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome . Science 2006 ; 313 : 1960 – 1964 . Google Scholar Crossref Search ADS PubMed WorldCat 7 Galon J , Pagès F, Marincola FM, Angell HK, Thurin M, Lugli A et al. Cancer classification using the Immunoscore: a worldwide task force . J Transl Med 2012 ; 10 : 205 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Mlecnik B , Tosolini M, Kirilovsky A, Berger A, Bindea G, Meatchi T et al. Histopathologic-based prognostic factors of colorectal cancers are associated with the state of the local immune reaction . J Clin Oncol 2011 ; 29 : 610 – 618 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Pagès F , Galon J, Fridman WH. The essential role of the in situ immune reaction in human colorectal cancer . J Leukoc Biol 2008 ; 84 : 981 – 987 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Pagès F , Berger A, Camus M, Sanchez-Cabo F, Costes A, Molidor R et al. Effector memory T cells, early metastasis, and survival in colorectal cancer . N Engl J Med 2005 ; 353 : 2654 – 2666 . Google Scholar Crossref Search ADS PubMed WorldCat 11 Pagès F , Galon J, Dieu-Nosjean MC, Tartour E, Sautès-Fridman C, Fridman WH. Immune infiltration in human tumors: a prognostic factor that should not be ignored . Oncogene 2010 ; 29 : 1093 – 1102 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Pagès F , Kirilovsky A, Mlecnik B, Asslaber M, Tosolini M, Bindea G et al. In situ cytotoxic and memory T cells predict outcome in patients with early-stage colorectal cancer . J Clin Oncol 2009 ; 27 : 5944 – 5951 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Tosolini M , Kirilovsky A, Mlecnik B, Fredriksen T, Mauger S, Bindea G et al. Clinical impact of different classes of infiltrating T cytotoxic and helper cells (Th1, th2, treg, th17) in patients with colorectal cancer . Cancer Res 2011 ; 71 : 1263 – 1271 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Khanh do T , Mekata E, Mukaisho K, Sugihara H, Shimizu T, Shiomi H et al. Prognostic role of CD10+ myeloid cells in association with tumor budding at the invasion front of colorectal cancer . Cancer Sci 2011 ; 102 : 1724 – 1733 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Galon J , Fridman WH, Pagès F. The adaptive immunologic microenvironment in colorectal cancer: a novel perspective . Cancer Res 2007 ; 67 : 1883 – 1886 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Gomez D , Cameron IC. Prognostic scores for colorectal liver metastasis: clinically important or an academic exercise? HPB (Oxford) 2010 ; 12 : 227 – 238 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Roxburgh CS , Richards CH, Moug SJ, Foulis AK, McMillan DC, Horgan PG. Determinants of short- and long-term outcome in patients undergoing simultaneous resection of colorectal cancer and synchronous colorectal liver metastases . Int J Colorectal Dis 2012 ; 27 : 363 – 369 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Shin SJ , Ahn JB, Choi JS, Choi GH, Lee KY, Baik SH et al. Implications of clinical risk score to predict outcomes of liver-confined metastasis of colorectal cancer . Surg Oncol 2012 ; 21 : e125 – e130 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Hill CR , Chagpar RB, Callender GG, Brown RE, Gilbert JE, Martin RC et al. Recurrence following hepatectomy for metastatic colorectal cancer: development of a model that predicts patterns of recurrence and survival . Ann Surg Oncol 2012 ; 19 : 139 – 144 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Spelt L , Andersson B, Nilsson J, Andersson R. Prognostic models for outcome following liver resection for colorectal cancer metastases: a systematic review . Eur J Surg Oncol 2012 ; 38 : 16 – 24 . Google Scholar Crossref Search ADS PubMed WorldCat 21 Eisenhauer EA , Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1·1) . Eur J Cancer 2009 ; 45 : 228 – 247 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Rubbia-Brandt L , Giostra E, Brezault C, Roth AD, Andres A, Audard V et al. Importance of histological tumor response assessment in predicting the outcome in patients with colorectal liver metastases treated with neo-adjuvant chemotherapy followed by liver surgery . Ann Oncol 2007 ; 18 : 299 – 304 . Google Scholar Crossref Search ADS PubMed WorldCat 23 Gulubova MV , Vlaykova TI. Significance of tenascin-C, fibronectin, laminin, collagen IV, alpha5beta1 and alpha9beta1 integrins and fibrotic capsule formation around liver metastases originating from cancers of the digestive tract . Neoplasma 2006 ; 53 : 372 – 383 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 24 Lunevicius R , Nakanishi H, Ito S, Kozaki K, Kato T, Tatematsu M et al. Clinicopathological significance of fibrotic capsule formation around liver metastasis from colorectal cancer . J Cancer Res Clin Oncol 2001 ; 127 : 193 – 199 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Ueno H , Mochizuki H, Hashiguchi Y, Hatsuse K, Fujimoto H, Hase K. Predictors of extrahepatic recurrence after resection of colorectal liver metastases . Br J Surg 2004 ; 91 : 327 – 333 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Okano K , Maeba T, Moroguchi A, Ishimura K, Karasawa Y, Izuishi K et al. Lymphocytic infiltration surrounding liver metastases from colorectal cancer . J Surg Oncol 2003 ; 82 : 28 – 33 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Bian EB , Huang C, Wang H, Wu BM, Zhang L, Lv XW et al. DNA methylation: new therapeutic implications for hepatic fibrosis . Cell Signal 2013 ; 25 : 355 – 358 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Fichtner-Feigl S , Strober W, Kawakami K, Puri RK, Kitani A. IL-13 signaling through the IL-13alpha2 receptor is involved in induction of TGF-beta1 production and fibrosis . Nat Med 2006 ; 12 : 99 – 106 . Google Scholar Crossref Search ADS PubMed WorldCat 29 Liu Y , Munker S, Müllenbach R, Weng HL. IL-13 signaling in liver fibrogenesis . Front Immunol 2012 ; 3 : 116 . Google Scholar Crossref Search ADS PubMed WorldCat 30 Ivanecz A , Potrc S, Horvat M, Jagric T, Gadzijev E. The validity of clinical risk score for patients undergoing liver resection for colorectal metastases . Hepatogastroenterology 2009 ; 56 : 1452 – 1458 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 31 Kobayashi H , Kotake K, Sugihara K. Prognostic scoring system for stage IV colorectal cancer: is the AJCC sub-classification of stage IV colorectal cancer appropriate? Int J Clin Oncol 2013 ; 18 : 696 – 703 . Google Scholar Crossref Search ADS PubMed WorldCat 32 Lee WS , Kim MJ, Yun SH, Chun HK, Lee WY, Kim SJ et al. Risk factor stratification after simultaneous liver and colorectal resection for synchronous colorectal metastasis . Langenbecks Arch Surg 2008 ; 393 : 13 – 19 . Google Scholar Crossref Search ADS PubMed WorldCat 33 Halama N , Michel S, Kloor M, Zoernig I, Benner A, Spille A et al. Localization and density of immune cells in the invasive margin of human colorectal cancer liver metastases are prognostic for response to chemotherapy . Cancer Res 2011 ; 71 : 5670 – 5677 . Google Scholar Crossref Search ADS PubMed WorldCat © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd

Jacobs, M; Henselmans, I; Macefield, R C; Blencowe, N S; Smets, E M A; de Haes, J C J M; Sprangers, M A G; Blazeby, J M; van Berge Henegouwen, M I

doi: 10.1002/bjs.9647pmid: 25319127

Abstract Background There is no consensus among patients and healthcare professionals (HCPs) on the topics that need to be addressed after oesophageal cancer surgery. The aim of this study was to identify these topics, using a two-round Delphi survey. Methods In round 1, patients and HCPs (surgeons, dieticians, nurses) were invited to rate the importance of 49 topics. The proportion of panellists that considered a topic to be of low, moderate or high importance was then calculated for each of these two groups. Based on these proportions and the i.q.r., topics were categorized as: ‘consensus to be included’, ‘consensus to be excluded’ and ‘no consensus’. Only topics in the first category were included in the second round. In round 2, panellists were provided with individual and group feedback. To be included in the final list, topics had to meet criteria for consensus and stability. Results There were 108 patients and 77 HCPs in the round 2 analyses. In general, patients and HCPs considered the same topics important. The final list included 23 topics and revealed that it was most important to address: cancer removed/lymph nodes, the new oesophagus, eating and drinking, surgery, alarming new complaints and the recovery period. Conclusion The study provides surgeons with a list of topics selected by patients and HCPs that may be addressed systematically at the initial follow-up consultation after oesophageal cancer surgery. Introduction Surgery, preceded by neoadjuvant therapy, is the main curative treatment option for oesophageal cancer1. Overall prognosis is poor, however, with 5-year survival rates of between 15 and 50 per cent2,3. Surgery has major short- and long-term negative effects on patients' health and well-being4. Most experience a clinically relevant and long-lasting deterioration in health-related quality of life (HRQL) after oesophageal cancer surgery5,6. Informing patients about these prospects may help them create realistic expectations about their recovery period and enable them to cope with difficulties7,8. Clinicians need to inform patients systematically about important topics after this type of surgery. Three studies7,9–10 have described the information needs of patients after oesophageal cancer surgery. They indicated that patients wanted to be informed about HRQL and nutrition-related topics, medical care and prognosis. Although the patient perspective is important11, the studies did not take the view of healthcare professionals (HCPs; surgeons, nurses, dieticians) into account. Their perspective is also needed to define optimal information for patients after oesophageal cancer surgery12. No formal consensus exists among patients and HCPs on the topics that need to be addressed after oesophageal cancer surgery. Delphi methodology applies a structured and iterative process to collect information from a large number of panellists across locations and expertise13–15. By keeping the identity of the participants anonymous and preventing them from interacting directly with one another, the influence of dominant panellists on the opinion of others is limited16–19. Consensus can be encouraged by providing panellists with individual and group feedback between rounds12,20–22. The aim of this study was, therefore, to identify important topics that need to be addressed in the initial follow-up consultation after potentially curative oesophagectomy for cancer, using a two-round Delphi survey of patients and HCPs. Methods The preliminary work and methodology used in this Delphi survey was obtained from previous literature work, interviews and surveys conducted in the UK and the Netherlands with the aim of creating a core information set and a core outcome set for patients before oesophageal cancer surgery23–25. The Medical Ethics Committee of the Academic Medical Centre (AMC), Amsterdam, The Netherlands, provided an exemption for the study to seek formal approval. The survey included the list of topics used to inform the round 1 questionnaire of the previous Delphi survey23 after reducing it to make it relevant to the postoperative setting. This list provided topics that were possibly relevant to patients with oesophageal cancer before surgery, and was based on systematic reviews of clinical24 and HRQL25 topics, the 2010 Association of Upper Gastrointestinal Surgeons of Great Britain and Ireland National Oesophago-Gastric Cancer Audit26, information included in information leaflets in the UK27, and observations from consultations and interviews with patients following preoperative consultations. These data sources were extended by including topics that were possibly relevant after operation by conducting interviews with Dutch patients following surgery7, analysing the content of 18 consecutively sampled audiotapes that recorded the initial follow-up consultation in the Netherlands, and using the expertise of upper gastrointestinal surgeons, and experts in medical communication and HRQL. Overlapping topics, and topics deemed to be too rare and specific (such as fever) or too personal (for example a specific hobby), were removed. Some 49 topics (Fig. 1) were identified and grouped in three domains: oesophageal cancer surgery and prognosis (7 clinical topics), new oesophagus and eating/drinking (13) and HRQL (29). Each topic was formulated as an item for the Delphi survey questionnaire with examples in parentheses for some topics. Topics obtained from English data sources were translated into Dutch by a professional translator. To prevent the suggestion of a ranking of importance, topics in each domain were listed in either temporal (oesophageal cancer and prognosis) or alphabetical (new oesophagus and eating/drinking, HRQL) order. Fig. 1 Open in new tabDownload slide Results of the two-round Delphi survey: a round 1 and b round 2. The dashed line represents the level of agreement for consensus to be included (at least 70 per cent scored 7–9 and no more than 15 per cent scored 1–3). Lightly shaded bars indicate topics with an i.q.r. larger than 2. Topics with no bar in round 1 were additional topics identified and carried forward automatically to round 2. Topics with no bar in round 2 were those that were not carried forward from round 1. HCP, healthcare professional A modified two-round Delphi survey was chosen28, based on the preliminary results of the previous Delphi study23,29 and the expected high attrition rate. As the aim was to recruit a heterogeneous group of individuals based on experience, interest and expertise30–32, and, because no sample size guidelines exist for Delphi surveys12,17, it was decided to include at least 100 patients and 50 HCPs in the first round. Within the latter group, the aim was to include at least ten surgeons, ten dieticians and ten nurses. At the time of invitation, patients had to be 18 years or older, had treatment that involved oesophagectomy, and be considered free from distant metastases as determined on clinical grounds with supportive imaging or histopathology. The patient population was drawn from those attending the AMC and patient support groups in the Netherlands. Patients were identified from a prospectively maintained database of individuals who underwent oesophageal cancer surgery at the AMC between January 2011 and March 2013. Patients in support groups received a postal invitation on the forum of an AMC-supported website for patients with oesophageal cancer, along with e-mails to members of the national upper gastrointestinal patient support group. HCPs had to be currently active in the postoperative treatment of patients with oesophageal cancer. Surgeons had to be actively performing oesophageal cancer resections. A list of potentially eligible HCPs was obtained from the gastrointestinal staff of the AMC. These HCPs were invited to participate and asked to identify other potentially eligible HCPs. Members of the Dutch Society for Gastroenterology, nurses and dieticians who were members of a special interest group for gastroenterology were also invited to participate. The first round of the Delphi study was used to identify topics considered important by patients and/or HCPs, create feedback material for the second round and identify possible omissions. The second round was used to obtain final ratings on topics considered important in round 1, assess the stability of ratings and create a final list of important topics. Patients and HCPs were informed about the purpose and procedure of the study. Both paper and electronic questionnaires were created. Those willing to participate were given an information sheet and a paper or electronic questionnaire. Panellists were instructed to return the questionnaire within 2 weeks and non-responders received up to three reminders. The questionnaire in the first round contained five parts with the aim of obtaining sociodemographic and/or clinical background information, a list of five topics deemed important by panellists before reading the long list of topics, importance ratings on a Likert scale from 1 to 9 (1, not important; 9, very important) of the 49 topics from the long list (‘How important do you think it is that these topics are addressed in the first consultation with the surgeon after surgery for oesophageal cancer?’), additional important topics not included in the long list, and a list of five topics deemed important after rating topics from the long list. Patients were resurveyed if they provided at least some information on one of the five parts in the first-round questionnaire, and were still alive and free from disease recurrence. HCPs were resurveyed if they still treated patients with oesophageal cancer. After assessment of the first-round questionnaires, a second-round questionnaire was created based on the importance ratings of the 49 original topics, supplemented by topics identified elsewhere in the first round that could not be combined with any of the 49 original topics. These potentially unique topics were classified as ‘suitable for inclusion’, ‘uncertain suitability for inclusion’ or ‘unsuitable for inclusion’ (for example because of overlap) in the proposed list of topics. These classifications were proposed by one author and cross-checked by a second. Discrepancies were discussed with two upper gastrointestinal surgeons or a HRQL expert. Topics classified as suitable for inclusion were included in the second-round questionnaire. Each panellist was provided with their own rating and summary rating (median) of the patient and HCP groups from round 133,34. Each panellist was asked to consider this feedback and rerate the importance of each topic. Statistical analysis To observe group differences and give both groups equal weight in the analysis, data for patients and HCPs were analysed separately. Panellists who did not provide any importance ratings, but provided background information, were included in the first-round analyses. Panellists who did not provide any importance ratings in round 1 or 2 were excluded from the second-round analyses. All statistical analyses were conducted using the SPSS® version 19.0 (IBM, Armonk, New York, USA). Patterns of missing information were assumed to be random in both rounds35–37. Missing data were imputed separately for rounds 1 and 2, and for patients and HCPs, by using different imputation models38,39. Each model created a multiply imputed data set40–42 by adding 20 imputed sets to the original42,43. For some statistics (such as frequencies), imputed data sets could be combined into a pooled analysis, whereas for others (such as median with i.q.r.), a result was used if found in at least 15 of 20 imputed data sets. Sensitivity analyses were conducted to estimate the change in results when multiple imputation was compared with complete-case analysis41. In the end, all available variables were used to build a fully conditional specification model44, and to present the results obtained by multiple imputation43. Consensus In the absence of a formal guideline to define and determine consensus22,45, this was defined at the outset as the extent to which panellists collectively agree with one another on the importance of a topic (level of agreement). Consensus was determined using a combination of the level of agreement and the i.q.r.46. To determine the level of agreement, the proportion of panellists that rated a topic 1–3 (low importance), 4–6 (moderate importance) and 7–9 (high importance) was calculated11,13. Each topic was then assigned to one of three predefined categories: consensus to be included (at least 70 per cent rated a topic as 7–9 and no more than 15 per cent rated it as 1–3), consensus to be excluded (at least 70 per cent scored a topic as 1–3 and no more than 15 per cent scored it as 7–9) and no consensus (any other combination). An i.q.r. of 2 or less was considered to reflect consensus22. Topics were included in the second round and final list only if both criteria had been met in at least one group. Stability of results To estimate the influence of panellist composition34 and sample size47 on the stability of ratings per round, a bias-corrected accelerated bootstrapping sampling technique was applied to complete cases to create 95 per cent c.i. based on 2000 sampling iterations48. Results were considered to be unstable if the 95 per cent c.i. included values that would alter the conclusion when applying criteria to determine consensus. The stability of importance ratings across rounds was also determined to establish whether consensus had been achieved for a topic30,49–50, and whether a hypothetical third round might have changed the results. Stability was defined as the consistency of responses between successive rounds50. To determine the stability of importance ratings across rounds, both descriptive data47,51 (change in median, i.q.r., level of agreement) and significance tests52,53 (Wilcoxon signed-rank test, Spearman rank correlation, P ≤ 0·050) were used. At the outset ratings were considered stable across rounds if at least four of the following five criteria were met in at least one group: a stable median, a stable or decreased i.q.r.51, stable category of consensus (included, excluded, no consensus), a non-significant Wilcoxon signed-rank test result, and a significant Spearman's rank correlation result52. The stable category of consensus was a mandatory criterion. Estimates of the stability of importance ratings across rounds were confined to topics that were included in both rounds. For a topic to be included in the final list, ratings had to meet the above stated criteria for consensus, be stable in round 2, and be considered stable across rounds in at least one group. As the stability of ratings across rounds for the four additional topics included in round 2 could not be estimated, these topics were included in the final list if, in both groups, at least 80 per cent assigned a rating of between 7 and 9, and this result was stable in round 2. The final identification and subsequent ranking of each important topic was based on the group (patients or HCPs) that provided the highest proportion of panellists rating a topic as ‘high importance’. Topics were assigned to one of three defined categories: most important (at least 90 per cent 7–9 and no more than 10 per cent 1–3), very important (at least 80 per cent 7–9 and no more than 15 per cent 1–3) and important (at least 70 per cent 7–9 and no more than 15 per cent 1–3). To explore the possible influence of one subgroup on the total ratings, post hoc subgroup and sensitivity analyses were conducted. Results Rounds 1 and 2 were conducted in March and July 2013. The amount of missing data was small (83 and 97 per cent complete for patients and HCPs respectively) and no variable had more than 3 per cent missing. Although more data were missing in round 2 (78 and 88 per cent complete for patients and HCPs; no variable more than 12 per cent missing), this was mainly due to death/recurrence, and not because of any dissonance between individual scores and the group score37. Panellists In round 1, of 94 patients invited from the AMC, 71 (76 per cent) accepted, along with positive responses from 37 patients affiliated to a patient support group and 78 HCPs. In round 2, questionnaires were returned by 98 patients (91 per cent) and 69 HCPs (88 per cent). Information provided by 108 patients and 77 HCPs was used in the final analysis. Characteristics of patients and HCPs are summarized in Table 1 and further patient details are provided in Table S1 (supporting information). Of the HCPs, 29 were surgeons, 27 dieticians and 21 nurses. Most patients were men and most HCPs were women. The patients had a median age of 66 years and most HCPs were aged over 40 years. Most patients (67·6 per cent) used the paper questionnaire, whereas all HCPs used the electronic version. Table 1 Characteristics of panellists included in the Delphi process . Patients (n = 108) . Healthcare professionals (n = 77) . Questionnaire Paper 73 (67·6) 0 (0) Website 35 (32·4)  77 (100) Age (years) Median (i.q.r.)   66 (59–71) – 21–30 1 (0·9) 10 (13) 31–40 0 (0)  20 (26) 41–50 4 (3·7) 29 (38) 51–60 25 (23·1) 15 (19) > 60 78 (72·2) 3 (4) Sex ratio (M : F) 74 : 34 30 : 47 . Patients (n = 108) . Healthcare professionals (n = 77) . Questionnaire Paper 73 (67·6) 0 (0) Website 35 (32·4)  77 (100) Age (years) Median (i.q.r.)   66 (59–71) – 21–30 1 (0·9) 10 (13) 31–40 0 (0)  20 (26) 41–50 4 (3·7) 29 (38) 51–60 25 (23·1) 15 (19) > 60 78 (72·2) 3 (4) Sex ratio (M : F) 74 : 34 30 : 47 Values in parentheses are percentages unless indicated otherwise. Open in new tab Table 1 Characteristics of panellists included in the Delphi process . Patients (n = 108) . Healthcare professionals (n = 77) . Questionnaire Paper 73 (67·6) 0 (0) Website 35 (32·4)  77 (100) Age (years) Median (i.q.r.)   66 (59–71) – 21–30 1 (0·9) 10 (13) 31–40 0 (0)  20 (26) 41–50 4 (3·7) 29 (38) 51–60 25 (23·1) 15 (19) > 60 78 (72·2) 3 (4) Sex ratio (M : F) 74 : 34 30 : 47 . Patients (n = 108) . Healthcare professionals (n = 77) . Questionnaire Paper 73 (67·6) 0 (0) Website 35 (32·4)  77 (100) Age (years) Median (i.q.r.)   66 (59–71) – 21–30 1 (0·9) 10 (13) 31–40 0 (0)  20 (26) 41–50 4 (3·7) 29 (38) 51–60 25 (23·1) 15 (19) > 60 78 (72·2) 3 (4) Sex ratio (M : F) 74 : 34 30 : 47 Values in parentheses are percentages unless indicated otherwise. Open in new tab Delphi results Round 1 Patients and HCPs considered 14 and 20 of the 49 topics on the long list to be important (Fig. 1; Table S2, supporting information). Of these, ten were considered important by both patients and HCPs, and 14 were considered important by either patients or HCPs. From the 1672 topics included elsewhere in the first-round questionnaire, four additional items were identified: hospital admission, pneumonia, contact patient support group, alarming new complaints. Hence, panellists had to rate 28 topics in round 2. Round 2 Seventeen topics were considered important by both patients and HCPs, and seven were considered important by either patients or HCPs (Fig. 1; Table S3, supporting information). Both patients and HCPs considered the following topics important: cancer removed/lymph nodes, surgery, tightening scar tissue at surgical joint, recovery period, follow-up, cancer recurrence, life expectancy, eating/drinking, information about new oesophagus, choking when swallowing, dysphagia, general health, quality of life, physical functioning, dumping syndrome, sleeping problems and alarming new complaints. Items deemed important by patients only were vitamin deficiency and feeding supplements, whereas those considered important by HCPs only were feeding tube, enteral nutrition, weight loss, physical pain or discomfort, and other care. Stability of results The stability of ratings increased in both the patient (32 of 49 versus 20 of 28) and HCP (29 of 49 versus 22 of 28) groups between the two rounds (Tables S2 and S3, supporting information). Owing to unstable results in round 2 (Table S3, supporting information), feeding tube, enteral nutrition, weight loss, physical functioning, physical pain or discomfort, other care, hospital admission and pneumonia were excluded in the patient group. Cancer recurrence, life expectancy, dysphagia, general health, wounds and contact patient support group were excluded in the HCP group. Many of these topics also did not reach the criteria for consensus to be included. In the patient group, unstable ratings across rounds were identified for vitamin deficiency (Tables S4–S9, supporting information). In the HCP group, unstable ratings across rounds were identified for cancer recurrence, life expectancy and wounds. For these topics, responses were still shifting in round 2, indicating that a third round might have provided a different result. In the end, only vitamin deficiency was excluded from the final list, as both the HCP and patient groups did not provide the stable ratings needed to meet the criteria for inclusion in the final list (Table S10, supporting information). Post hoc subgroup and sensitivity analyses included the same topics in the final list, although cancer recurrence would have received a higher rating if only surgeons had been included (Tables S11–S14, supporting information). Final list of topics Based on the level of agreement and stability of each topic across groups, 23 topics were included in the final list (Table S14, supporting information). Six were considered most important, 13 very important and four important (Table 2). Topics deemed most important were: cancer removed/lymph nodes, information about the new oesophagus, eating/drinking, surgery, alarming new complaints and the recovery period. Table 2 Final list of important topics Most important* Cancer removed/lymph nodes New oesophagus Eating/drinking Surgery Alarming new complaints† Recovery period Very important‡ Quality of life Tightening scar at tissue surgical joint Cancer recurrence Feeding tube General health Physical functioning Choking when swallowing Weight loss Life expectancy Sleeping problems Enteral nutrition Dumping syndrome Dysphagia Important§ Follow-up Other care Physical pain or discomfort Feeding supplements Most important* Cancer removed/lymph nodes New oesophagus Eating/drinking Surgery Alarming new complaints† Recovery period Very important‡ Quality of life Tightening scar at tissue surgical joint Cancer recurrence Feeding tube General health Physical functioning Choking when swallowing Weight loss Life expectancy Sleeping problems Enteral nutrition Dumping syndrome Dysphagia Important§ Follow-up Other care Physical pain or discomfort Feeding supplements * At least 90 per cent rated topic as 7–9 and no more than 10 per cent 1–3 in at least one group; † at least 80 per cent 7–9 in both groups; ‡ at least 80 per cent 7–9 and no more than 15 per cent 1–3 in at least one group; § at least 70 per cent 7–9 and no more than 15 per cent 1–3 in at least one group. Open in new tab Table 2 Final list of important topics Most important* Cancer removed/lymph nodes New oesophagus Eating/drinking Surgery Alarming new complaints† Recovery period Very important‡ Quality of life Tightening scar at tissue surgical joint Cancer recurrence Feeding tube General health Physical functioning Choking when swallowing Weight loss Life expectancy Sleeping problems Enteral nutrition Dumping syndrome Dysphagia Important§ Follow-up Other care Physical pain or discomfort Feeding supplements Most important* Cancer removed/lymph nodes New oesophagus Eating/drinking Surgery Alarming new complaints† Recovery period Very important‡ Quality of life Tightening scar at tissue surgical joint Cancer recurrence Feeding tube General health Physical functioning Choking when swallowing Weight loss Life expectancy Sleeping problems Enteral nutrition Dumping syndrome Dysphagia Important§ Follow-up Other care Physical pain or discomfort Feeding supplements * At least 90 per cent rated topic as 7–9 and no more than 10 per cent 1–3 in at least one group; † at least 80 per cent 7–9 in both groups; ‡ at least 80 per cent 7–9 and no more than 15 per cent 1–3 in at least one group; § at least 70 per cent 7–9 and no more than 15 per cent 1–3 in at least one group. Open in new tab Discussion This study included the views of both patients and HCPs to identify topics that need to be addressed after potentially curative oesophageal cancer surgery. Topics considered to be most important to address were: the cancer removed/ lymph nodes, information about the new oesophagus, eating/drinking, surgery, alarming new complaints and the recovery period. These findings confirm the results of earlier studies7,9–10 on patient information needs. Addressing patients' information needs in postoperative consultations is essential, particularly when the surgery has limited clinical effectiveness3 and has a detrimental impact on HRQL4,5, as is the case with oesophagectomy. The list of important topics resulting from this study may guide the provision of information in consultations after surgery for oesophageal cancer. If a systematic discussion of all topics is not feasible, the list could also serve as a checklist at the beginning of the consultation. These findings do not mean that surgeons can adopt a one-size-fits-all approach to this consultation. Surgeons need to tailor the content as well as the way in which information is given to the individual patient's needs and abilities. For instance, answering patient questions about pathology results and prognosis can be complex. One possible approach to tailoring information-giving is the ask–tell–ask approach54, whereby the surgeon finds out what information the patient is actually looking for (ask), adjusts their information to the patient's needs (tell) and then checks whether the patient has understood (ask). HCPs do not always attribute the same level of importance as patients to some topics. The most notable difference was found in those related to prognosis, as identified by others55. One way of supporting patients in discussing difficult topics such as prognosis, and also of helping raise HCPs' awareness of this area, is to provide patients with a question prompt sheet7,56–57. By using a structured list of example questions, patients may be encouraged to address prognosis, and also show HCPs that they do want to receive honest and realistic information10,58. Both patients and HCPs considered it very important to address issues related to HRQL. Ideally, HCPs should be able to inform patients about the incidence and course of HRQL7, but for many topics in this area there is limited information on the effect of treatment5–6,59, and HRQL data are not always presented in a way that is understandable to clinicians and patients60,61. One method of increasing the use of HRQL data is to provide HCPs with an easily accessible, electronic support tool giving an evidence-based summary of important topics in a patient-friendly and clinically applicable summary format60,61, such as simple line graphs62–64. This study has some limitations. It is possible that panellists differed in their view of what constitutes an initial follow-up consultation, as this was specified as the first consultation with the surgeon following oesophageal cancer surgery. Reasons for missing data from HCPs, and from patients affiliated to a patient support group, were not obtained. There is possible selection bias and quasi-anonymity12, as panellists not only selected themselves but also identified other potentially eligible panellists. Although self-selection may increase the inclusion of motivated panellists this does decrease attrition rates31. Both the patient sample (different accrual methods) and the HCP sample (different disciplines) consisted of subgroups that may have influenced the results. An additional third round might have changed the results and the ranking for some of the topics included in the final list. Because patients did not complete the questionnaires in the early postoperative period some recall bias is possible, leading to issues that were over- or under-emphasized depending on their individual recovery period. The fact that the results were stable across panellists, however, would suggest that any such influence was small. The strengths of the study are the various and extensive data sources used to create possibly relevant topics, the low rates of attrition and missing data, combining several methods to create a final list of important topics, and the stability of the results across a large and heterogeneous panel. On this basis it seems likely that these results are robust and suitable for use in guiding surgeons and other HCPs in communicating with patients. Acknowledgements The authors thank H. Pieterse for constructing the electronic questionnaire, and S. S. Gisbertz for help in creating the long list of topics. Disclosure: The authors declare no conflict of interest. References 1 Gebski V , Burmeister B, Smithers BM, Foo K, Zalcberg J, Simes J; Australasian Gastro-Intestinal Trials Group . Survival benefits from neoadjuvant chemoradiotherapy or chemotherapy in oesophageal carcinoma: a meta-analysis . Lancet Oncol 2007 ; 8 : 226 – 234 . Google Scholar OpenURL Placeholder Text WorldCat 2 Rutegård M , Charonis K, Lu Y, Lagergren P, Lagergren J, Rouvelas I. Population-based esophageal cancer survival after resection without neoadjuvant therapy: an update . Surgery 2012 ; 152 : 903 – 910 . Google Scholar OpenURL Placeholder Text WorldCat 3 van Hagen P , Hulshof MC, van Lanschot JJ, Steyerberg EW, van Berge Henegouwen MI, Wijnhoven BP et al. Preoperative chemoradiotherapy for esophageal or junctional cancer . N Engl J Med 2012 ; 366 : 2074 – 2084 . Google Scholar OpenURL Placeholder Text WorldCat 4 Djärv T , Lagergren P. Quality of life after esophagectomy for cancer . Expert Rev Gastroenterol Hepatol 2012 ; 6 : 115 – 122 . Google Scholar OpenURL Placeholder Text WorldCat 5 Jacobs M , Macefield RC, Elbers RG, Sitnikova K, Korfage IJ, Smets EM et al. Meta-analysis shows clinically relevant and long-lasting deterioration in health-related quality of life after esophageal cancer surgery . Qual Life Res 2014 ; 23 : 1155 – 1176 . Google Scholar OpenURL Placeholder Text WorldCat 6 Jacobs M , Macefield RC, Blazeby JM, Korfage IJ, van Berge Henegouwen MI, de Haes HC et al. Systematic review reveals limitations of studies evaluating health-related quality of life after potentially curative treatment for esophageal cancer . Qual Life Res 2013 ; 22 : 1787 – 1803 . Google Scholar OpenURL Placeholder Text WorldCat 7 Henselmans I , Jacobs M, van Berge Henegouwen MI, de Haes HC, Sprangers MA, Smets EM. Postoperative information needs and communication barriers of esophageal cancer patients . Patient Educ Couns 2012 ; 88 : 138 – 146 . Google Scholar OpenURL Placeholder Text WorldCat 8 Trask PC , Hsu MA, McQuellon R. Other paradigms: health-related quality of life as a measure in cancer treatment: its importance and relevance . Cancer J 2009 ; 15 : 435 – 440 . Google Scholar OpenURL Placeholder Text WorldCat 9 McNair AG , Brookes ST, Kinnersley P, Blazeby JM. What surgeons should tell patients with oesophago-gastric cancer: a cross sectional study of information needs . Eur J Surg Oncol 2013 ; 39 : 1278 – 1286 . Google Scholar OpenURL Placeholder Text WorldCat 10 Lagarde SM , Franssen SJ, van Werven JR, Smets EM, Tran TC, Tilanus HW et al. Patient preferences for the disclosure of prognosis after esophagectomy for cancer with curative intent . Ann Surg Oncol 2008 ; 15 : 3289 – 3298 . Google Scholar OpenURL Placeholder Text WorldCat 11 Guyatt GH , Oxman AD, Kunz R, Atkins D, Brozek J, Vist G et al. GRADE guidelines: 2. Framing the question and deciding on important outcomes . J Clin Epidemiol 2011 ; 64 : 395 – 400 . Google Scholar OpenURL Placeholder Text WorldCat 12 Sinha IP , Smyth RL, Williamson PR. Using the Delphi technique to determine which outcomes to measure in clinical trials: recommendations for the future based on a systematic review of existing studies . PLoS Med 2011 ; 8 : e1000393 . Google Scholar OpenURL Placeholder Text WorldCat 13 Jones J , Hunter D. Consensus methods for medical and health services research . BMJ 1995 ; 311 : 376 – 380 . Google Scholar OpenURL Placeholder Text WorldCat 14 Rowe G , Wright G. The Delphi technique as a forecasting tool: issues and analysis . Int J Forecast 1999 ; 15 : 353 – 375 . Google Scholar OpenURL Placeholder Text WorldCat 15 De Vet E , Brug J, De Nooijer J, Dijkstra A, De Vries NK. Determinants of forward stage transitions: a Delphi study . Health Educ Res 2005 ; 20 : 195 – 205 . Google Scholar OpenURL Placeholder Text WorldCat 16 Williams PL , Webb C. The Delphi technique: a methodological discussion . J Adv Nurs 1994 ; 19 : 180 – 186 . Google Scholar OpenURL Placeholder Text WorldCat 17 Keeney S , Hasson F, McKenna H. Consulting the oracle: ten lessons from using the Delphi technique in nursing research . J Adv Nurs 2006 ; 53 : 205 – 212 . Google Scholar OpenURL Placeholder Text WorldCat 18 Weir CR , Hicken BL, Rappaport HS, Nebeker JR. Crossing the quality chasm: the role of information technology departments . Am J Med Qual 2006 ; 21 : 382 – 393 . Google Scholar OpenURL Placeholder Text WorldCat 19 Strauss HJ , Zeigler LH. Delphi, political philosophy and the future . Futures 1975 ; 7 : 184 – 196 . Google Scholar OpenURL Placeholder Text WorldCat 20 Okoli C , Pawlowski SD. The Delphi method as a research tool: an example, design considerations and applications . Information and Management 2004 ; 42 : 15 – 29 . Google Scholar OpenURL Placeholder Text WorldCat 21 Sandford BA , Hsu C. The Delphi technique: making sense of consensus . Practical Assessment, Research & Evaluation 2007 ; 12 : 1 – 8 . Google Scholar OpenURL Placeholder Text WorldCat 22 van der Gracht HA . Consensus measurement in Delphi studies: review and implications for future quality assurance . Technol Forecast Soc Change 2012 ; 79 : 1525 – 1536 . Google Scholar OpenURL Placeholder Text WorldCat 23 Macefield RC , Jacobs M, Blencowe NS, Korfage IJ, Nicklin J, Brookes ST et al. The case for a HRQL core outcome set: outcome reporting bias in oesophageal cancer studies . Trials 2011 ; 12 ( Suppl ): A77 . Google Scholar OpenURL Placeholder Text WorldCat 24 Blencowe NS , Strong S, McNair AG, Brookes ST, Crosby T, Griffin SM et al. Reporting of short-term clinical outcomes after esophagectomy: a systematic review . Ann Surg 2012 ; 255 : 658 – 666 . Google Scholar OpenURL Placeholder Text WorldCat 25 Macefield RC , Jacobs M, Korfage IJ, Nicklin J, Whistance RN, Brookes ST et al. Developing core outcomes sets: methods for identifying and including patient-reported outcomes (PROs) . Trials 2014 ; 15 : 49 . Google Scholar OpenURL Placeholder Text WorldCat 26 Association of Upper Gastrointestinal Surgeons of Great Britain and Ireland . National Oesophago-Gastric Cancer Audit 2010 . http://www.hscic.gov.uk/catalogue/PUB02758/clin-audi-supp-prog-oeso-gast-2010-rep1.pdf [accessed 1 January 2013]. 27 Blencowe NS , Strong S, Howes N, Elliot J, Avery K, Blazeby J. The need for better and accurate written patient information for oesophageal cancer surgery . Br J Surg 2013 ; 41 : 100 . Google Scholar OpenURL Placeholder Text WorldCat 28 Hasson F , Keeney S. Enhancing rigour in the Delphi technique research . Technol Forecast Soc Change 2011 ; 78 : 1695 – 1704 . Google Scholar OpenURL Placeholder Text WorldCat 29 Blencowe NS , Macefield RC, Jacobs M, Sprangers MAG, Brookes ST, Blazeby JM. Which outcomes are important to patients and surgeons? Core information prior to oesophageal cancer surgery . AUGIS 17th Annual Scientific Meeting , Gateshead, 2013. OpenURL Placeholder Text WorldCat 30 Goodman CM . The Delphi technique: a critique . J Adv Nurs 1987 ; 12 : 729 – 734 . Google Scholar OpenURL Placeholder Text WorldCat 31 Keeney S , Hasson F, McKenna HP. A critical review of the Delphi technique as a research methodology for nursing . Int J Nurs Stud 2001 ; 38 : 195 – 200 . Google Scholar OpenURL Placeholder Text WorldCat 32 Delbecq A , Van de Ven A, Gustafson D. Group Techniques for Program Planning: a Guide to Nominal Group and Delphi Processes . Scott, Foresman : Glenview , 1975 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 33 Nelson BW . Statistical manipulation of Delphi statements: its success and effects on convergence and stability . Technol Forecast Soc Change 1978 ; 12 : 41 – 60 . Google Scholar OpenURL Placeholder Text WorldCat 34 Macefield RC , Blencowe NS, Brookes ST, Jacobs M, Sprangers M, Williamson PR et al. Core outcome set development: the effect of Delphi panel composition and feedback on prioritisation of outcomes . Trials 2013 ; 14 ( Suppl ): P77. Google Scholar OpenURL Placeholder Text WorldCat 35 Steyerberg EW . Clinical Prediction Models. A Practical Approach to Development, Validation, and Updating . Springer Science + Business Media, LLC : New York , 2009 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 36 Little RJ , Rubin DB. Statistical Analysis with Missing Data (2nd edn). Wiley : New York , 2002 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 37 Bardecki MJ . Participants' response to the Delphi method: an attitudinal perspective . Technol Forecast Soc Change 1984 ; 25 : 281 – 292 . Google Scholar OpenURL Placeholder Text WorldCat 38 Royston P Multiple imputation of missing values: update of ice . Stata J 2005 ; 5 : 527 – 536 . Google Scholar OpenURL Placeholder Text WorldCat 39 Royston P Multiple imputation of missing values . Stata J 2004 ; 4 : 227 – 241 . Google Scholar OpenURL Placeholder Text WorldCat 40 de Goeij MC , van Diepen M, Jager KJ, Tripepi G, Zoccali C, Dekker FW. Multiple imputation: dealing with missing data . Nephrol Dial Transplant 2013 ; 28 : 2415 – 2420 . Google Scholar OpenURL Placeholder Text WorldCat 41 He Y Missing data analysis using multiple imputation: getting to the heart of the matter . Circ Cardiovasc Qual Outcomes 2010 ; 3 : 98 – 105 . Google Scholar OpenURL Placeholder Text WorldCat 42 Graham JW . Missing data analysis: making it work in the real world . Annu Rev Psychol 2009 ; 60 : 549 – 576 . Google Scholar OpenURL Placeholder Text WorldCat 43 Sterne JA , White IR, Carlin JB, Spratt M, Royston P, Kenward MG et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls . BMJ 2009 ; 338 : b2393 . Google Scholar OpenURL Placeholder Text WorldCat 44 Lee KJ , Carlin JB. Multiple imputation for missing data: fully conditional specification versus multivariate normal imputation . Am J Epidemiol 2010 ; 171 : 624 – 632 . Google Scholar OpenURL Placeholder Text WorldCat 45 Meijering JV , Kampen JK, Tobi H. Quantifying the development of agreement among experts in Delphi studies . Technol Forecast Soc Change 2013 ; 80 : 1607 – 1614 . Google Scholar OpenURL Placeholder Text WorldCat 46 Diamond IR , Grant RC, Feldman BM, Pencharz PB, Ling SC, Moore AM et al. Defining consensus: a systematic review recommends methodologic criteria for reporting of Delphi studies . J Clin Epidemiol 2014 ; 67 : 401 – 409 . Google Scholar OpenURL Placeholder Text WorldCat 47 Holey EA , Feeley JL, Dixon J, Whittaker VJ. An exploration of the use of simple statistics to measure consensus and stability in Delphi studies . BMC Med Res Methodol 2007 ; 7 : 52 . Google Scholar OpenURL Placeholder Text WorldCat 48 Akins RB , Tolson H, Cole BR. Stability of response characteristics of a Delphi panel: application of bootstrap data expansion . BMC Med Res Methodol 2005 ; 5 : 37 . Google Scholar OpenURL Placeholder Text WorldCat 49 Scheibe M , Skutsch J, Schofer J. Experiments in Delphi methodology. In The Delphi Method – Techniques and Applications , Linstone HA, Turoff M (eds). Addison-Wesley : Reading , 1975 ; 262 – 287 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 50 Dajani JS , Sincoff MZ, Talley WK. Stability and agreement criteria for the termination of Delphi studies . Technol Forecast Soc Change 1979 ; 13 : 83 – 90 . Google Scholar OpenURL Placeholder Text WorldCat 51 Greatorex J , Dexter T. An accessible analytical approach for investigating what happens between the rounds of a Delphi study . J Adv Nurs 2000 ; 32 : 1016 – 1024 . Google Scholar OpenURL Placeholder Text WorldCat 52 Kalaian SA , Kasim RM. Terminating sequential Delphi survey data collection . Practical Assessment, Research & Evaluation 2012 ; 17 : 1 – 10 . Google Scholar OpenURL Placeholder Text WorldCat 53 De Vet E , Brug J, De Nooijer J, Dijkstra A, De Vries NK. Determinants of forward stage transitions: a Delphi study . Health Educ Res 2005 ; 20 : 195 – 205 . Google Scholar OpenURL Placeholder Text WorldCat 54 Back AL , Arnold RM, Baile WF, Tulsky JA, Fryer-Edwards K. Approaching difficult communication tasks in oncology . CA Cancer J Clin 2005 ; 55 : 164 – 177 . Google Scholar OpenURL Placeholder Text WorldCat 55 Hagerty RG , Butow PN, Ellis PM, Dimitry S, Tattersall MH. Communicating prognosis in cancer care: a systematic review of the literature . Ann Oncol 2005 ; 16 : 1005 – 1053 . Google Scholar OpenURL Placeholder Text WorldCat 56 Smets EM , van Heijl M, van Wijngaarden AK, Henselmans I, van Berge Henegouwen MI. Addressing patients' information needs: a first evaluation of a question prompt sheet in the pretreatment consultation for patients with esophageal cancer . Dis Esophagus 2012 ; 25 : 512 – 519 . Google Scholar OpenURL Placeholder Text WorldCat 57 Clayton JM , Butow PN, Tattersall MH, Devine RJ, Simpson JM, Aggarwal G et al. Randomized controlled trial of a prompt list to help advanced cancer patients and their caregivers to ask questions about prognosis and end-of-life care . J Clin Oncol 2007 ; 25 : 715 – 723 . Google Scholar OpenURL Placeholder Text WorldCat 58 Mack JW , Smith TJ. Reasons why physicians do not have discussions about poor prognosis, why it matters, and what can be improved . J Clin Oncol 2012 ; 30 : 2715 – 2717 . Google Scholar OpenURL Placeholder Text WorldCat 59 Blencowe NS , McNair AG, Davis CR, Brookes ST, Blazeby JM. Standards of outcome reporting in surgical oncology: a case study in esophageal cancer . Ann Surg Oncol 2012 ; 19 : 4012 – 4018 . Google Scholar OpenURL Placeholder Text WorldCat 60 Brundage M , Bass B, Jolie R, Foley K. A knowledge translation challenge: clinical use of quality of life data from cancer clinical trials . Qual Life Res 2011 ; 20 : 979 – 985 . Google Scholar OpenURL Placeholder Text WorldCat 61 Calvert M , Blazeby J, Altman DG, Revicki DA, Moher D, Brundage MD; CONSORT PRO Group . Reporting of patient-reported outcomes in randomized trials: the CONSORT PRO extension . JAMA 2013 ; 309 : 814 – 822 . Google Scholar OpenURL Placeholder Text WorldCat 62 McNair AG , Brookes ST, Davis CR, Argyropoulos M, Blazeby JM. Communicating the results of randomized clinical trials: do patients understand multidimensional patient-reported outcomes? J Clin Oncol 2010 ; 28 : 738 – 743 . Google Scholar OpenURL Placeholder Text WorldCat 63 Brundage M , Feldman-Stewart D, Leis A, Bezjak A, Degner L, Velji K et al. Communicating quality of life information to cancer patients: a study of six presentation formats . J Clin Oncol 2005 ; 23 : 6949 – 6956 . Google Scholar OpenURL Placeholder Text WorldCat 64 Brundage M , Leis A, Bezjak A, Feldman-Stewart D, Degner L, Velji K et al. Cancer patients' preferences for communicating clinical trial quality of life information: a qualitative study . Qual Life Res 2003 ; 12 : 395 – 404 . Google Scholar OpenURL Placeholder Text WorldCat Author notes Presented in part to the Dutch Conference for Gastroenterology, Veldhoven, The Netherlands, October 2013, the 20th Annual Conference of the International Society for Quality of Life Research, Miami, Florida, USA, October 2013, and the 15th World Congress of Psycho-Social Oncology and Psychosocial Academy, Rotterdam, The Netherlands, November 2013 © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © 2014 BJS Society Ltd. Published by John Wiley & Sons Ltd