Comparative effectiveness of elemental formula in the early enteral nutrition management of acute pancreatitis: a retrospective cohort study

Comparative effectiveness of elemental formula in the early enteral nutrition management of acute... Background: Although enteral nutrition has become one of the standard therapies for patients with acute pancrea- titis, the optimal formulae for enteral nutrition have been under debate. Elemental formula is assumed to be suitable in the treatment of patients with acute pancreatitis because it has less stimulating effects for exocrine secretions of the pancreas, simultaneously maintaining gut immunity; however, clinical studies corroborating this assumption have been scarce. Methods: We conducted a retrospective cohort study using a Japanese national administrative database between 2010 and 2015. Patients with acute pancreatitis who received enteral feeding within 3 days of admission were identi- fied and divided into two groups according to whether elemental formula was administered. We assessed the impact of elemental formula for the outcomes (primary, in-hospital mortality; secondary, development of sepsis, hospital-free days at 90 days, and total health-care costs) using a multivariate mixed-effect regression analysis and propensity score matching analysis adjusted by a well-validated case-mix adjustment model. Analysis for the subpopulation of patients with severe acute pancreatitis was also performed. Results: Of 243,312 patients with acute pancreatitis, 948 patients were identified and classified into the elemental formula group (N = 382) and the control group (N = 566). No significant differences were observed for in-hospital mortality [10.2% in the elemental formula group vs. 11.0% in the control group; adjusted adds ratio (95% confidence interval; CI) = 0.94 (0.53–1.67)], sepsis development [5.0 vs. 7.1%; adjusted adds ratio (95% CI) = 0.66 (0.34–1.28)], mean hospital-free days [54 days vs. 51 days; adjusted difference (95% CI) = 2 days (− 2 to 5)], and mean total health-care costs [$29,360 vs. $34,214; adjusted difference (95% CI) = − $4250 (− 8643 to 141)]. Similar results were also observed in patients with severe acute pancreatitis. Conclusions: The results of our retrospective cohort study using a large-scale national database did not demonstrate the benefit of elemental formula compared to semi-elemental and polymeric formulae in patients with acute pan- creatitis. Further assessment of alternative nutritional strategy is expected. Keywords: Acute pancreatitis, Enteral nutrition, Elemental diet, Polymeric formula, Mortality *Correspondence: eraeaccm@tmd.ac.jp Trauma and Acute Critical Care Medical Center, Tokyo Medical and Dental University Hospital of Medicine, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan Full list of author information is available at the end of the article © The Author(s) 2018. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Endo et al. Ann. Intensive Care (2018) 8:69 Page 2 of 8 comorbidities at admission, and post-admission compli- Background cations are independently recorded using the relevant Acute pancreatitis is a life-threatening inflammatory dis - codes from the International Classification of Diseases, ease characterized by autodigestion and destruction of 10th revision (ICD-10). In addition, the DPC database the pancreas due to self-producing proteases. The stand - includes baseline characteristics of patients and informa- ard treatment for acute pancreatitis mainly comprises tion regarding the treating hospital. Furthermore, infor- supportive therapy, such as adequate fluid resuscitation mation regarding the severity of acute pancreatitis was and respiratory care, because evidence of disease-specific recorded using the Japanese severity scoring system for therapy has been limited [1–3]. acute pancreatitis of the Ministry of Health, Labour, and Adequate nutritional strategy has been one of the key Welfare of Japan [14], as this comprises the prognostic factors during systematic support in patients with acute factor scores and computed tomography (CT) sever- pancreatitis [4, 5]. The concept of “pancreatic rest” had ity scores based on contrast-enhanced CT (Additional been widely believed to be the standard nutritional strat- file  1). Further details regarding the DPC database have egy in the management of acute pancreatitis; therefore, been described elsewhere [15]. total parenteral nutrition had been widely used up to the This study was conducted in accordance with the prin - 1990s [6, 7]. However, several randomized controlled tri- ciples of the 1964 Declaration of Helsinki and its later als (RCTs) conducted in the late 1990s demonstrated the amendments. The institutional review board of the Tokyo consistent superiority of enteral nutrition over parenteral Medical and Dental University approved this study nutrition [8–11]. Among the enteral nutrition formulae, (#788). Informed consent from each patient was waived elemental formula has been believed to elicit theoreti- because of the retrospective design of the study and the cal advantages owing to a lower degree of exocrine pan- use of anonymized patient and hospital data. creatic stimulation and because it is fat-free. A previous RCT [12] comparing the efficacy of a semi-elemental formula with a polymeric formula was underpowered Study population (N = 30) and failed to demonstrate the superiority of a We included patients who were admitted to the hospital semi-elemental formula. Because few RCTs or large-scale because of acute pancreatitis between April 1, 2010, and cohort studies have been reported in this theme, Petrov March 31, 2015, and patients who received nasogastric et  al. [13] performed an indirect adjusted meta-analysis, feeding or nasojejunal feeding within 3 days of admission. in which parenteral nutrition groups were used as the We excluded patients younger than 16  years and those reference and reported that significant difference regard - who were pregnant. Patients who were discharged within ing survival benefit and adverse events were not observed 3  days of admission were also excluded considering the between the groups of (semi)elemental formulae and pol- issue of immortal-time bias. In addition, we excluded ymeric formula. However, to our knowledge, a large RCT patients who had missing values in any variables used in or cohort study that directly compared the efficacy of ele - the analyses (i.e., complete case analyses). mental formula has not been reported in the treatment of acute pancreatitis. In this study, we aimed to assess Data collection the clinical benefit of elemental formula compared to the We collected the following information from the DPC other formulae (semi-elemental and polymeric formulae) database: age; sex; ICD-10 codes for four primary diag- in the initiation of enteral nutrition management in acute noses, four concurrent diagnoses at admission, and four pancreatitis, using a large-scale national administrative post-admission complications; the aforementioned prog- database. nostic factor score and CT severity score; unique hos- pital identifier; annual number of patients with acute pancreatitis per hospital; presence or absence of specific Methods reimbursement claims for enteral feeding; presence or Study design and data sources absence of specific reimbursement claims for use of ele - We conducted a retrospective cohort study to evaluate mental formula; status at hospital discharge (i.e., survived the efficacy of elemental formula in patients with acute or deceased); total health-care costs per admission; and pancreatitis, using the Japanese Diagnosis Procedure length of hospital stay. Furthermore, we collected infor- Combination (DPC) database. The database is a case- mation on whether the following interventions were mix classification system that is linked to the reimburse - performed within 3 days of admission: administration of ment system for inpatient cases in Japanese hospitals and vasopressors (dopamine, norepinephrine, epinephrine, contains administrative claims for every drug adminis- or vasopressin), mechanical ventilation, renal replace- tered and every procedure and care performed at more ment therapy, and transfusion. Patient comorbidities than 1500 hospitals. Each patient’s primary diagnosis, were assessed using the Charlson comorbidity index [16] Endo et al. Ann. Intensive Care (2018) 8:69 Page 3 of 8 based on a previously reported method for extracting the classification model established, with the random effects ICD-10 codes [17]. of hospital-level clustering. In the linear mixed-effects model, the case-mix classification model was inverse- logit-transformed to satisfy the homoscedasticity Definitions and outcomes requirement for linear regression. Patients who were administered elemental formula We further compared the outcomes in the elemental were identified by the presence of specific reimburse - formula group and the control group using a propensity ment claims of elemental formula (Elental , Ajinomoto score matching analysis [22] as the secondary analy- Pharmaceutical Ltd., Tokyo, Japan; and Hepan ED , EA sis. The propensity score for predicting administration pharma Ltd., Tokyo, Japan; those are all the elemental of elemental formula was calculated through a logistic formula products approved in Japan). The control group regression analysis using the variables used for the estab- was defined as patients who were administered other lishment of the aforementioned prognosis model and types of enteral nutritional formulae within 3  days of the annual number of acute pancreatitis cases per hos- admission. The primary outcome was in-hospital mortal - pital as a variable to account for the differences in treat - ity. Secondary outcomes were development of sepsis after ment quality at each hospital. Propensity score matching admission, hospital-free days, and total health-care costs extracted 1:1 matched pairs from the elemental formula per admission. The ICD-10 codes used to identify sepsis group and the control group. A match balance between are presented in Additional file  2. Hospital-free days were the two groups was assessed using the absolute standard- defined as days survived and days free from hospitaliza - ized mean difference (ASMD) of all variables; values < 0.1 tion within 90  days from initial hospital admission. A were regarded as acceptable. To achieve balanced match- recent clinical trial group consensus recommended that ing, the caliper width for matching was set as the stand- hospital-free days should represent composite meas- ard deviation (SD) of the logit-transformed propensity ures compared to length of hospital stay, which could score multiplied by 0.3. Intergroup comparison of the be highly influenced by mortality [18]. Total health-care outcomes with propensity score-matched subjects was cost was defined as all aggregated payments (except for performed using a Chi-square test. boarding costs) to the hospital per discharge, and these In addition, we performed analyses using the afore- payments were estimated using the reference prices in mentioned two models only in patients with severe acute the Japanese fee schedule, which lists reimbursement pancreatitis (SAP) diagnosed by the severity diagnosis rates for surgical, pharmacological, laboratory, and other criteria in Japan, to assess the efficacy of elemental for - inpatient services. The cost data were provided after con - mula in patients with SAP. verting the cost in yen to US dollars (100 yen = $1 USD). All statistical analyses were performed using R software (version 3.4.1; R Foundation for Statistical Computing, Statistical analysis Vienna, Austria). The level of significance was defined as We developed a risk adjustment model for in-hospital p < 0.05. mortality using the variables of age, sex, Charlson comor- bidity index, prognostic factor score, CT severity score, mechanical ventilation within 3  days of admission, renal Results replacement therapy within 3  days of admission, trans- Study population fusion within 3  days of admission, and vasopressor use The flow diagram of the patient selection process is within 3  days of admission by applying a logistic regres- presented in Fig.  1. During the study period, a total of sion model that included a random sample of 80% of the 243,312 patients with acute pancreatitis were hospital- entire study cohort. The covariables used were selected ized at the DPC participating hospitals. Of these, 948 based on clinical experience and previous studies [19, patients were identified according to the inclusion and 20]. Issues with variable multicollinearity were assessed exclusion criteria. Among these patients, 382 (approxi- using a variance inflation factor (VIF), and the toler - mately 40.3%) were administered elemental formula. ance value was set at < 2. We validated the model for the Patients’ characteristics according to whether elemen- remaining 20% of the cohort using the area under the tal formula was administered are presented in Table  1. receiver operating curve (AUROC) and a Hosmer–Leme- The in-hospital mortality rate was 10.2% (39/382) in the show goodness-of-fit test. elemental formula group and 11.0% (62/566) in the con- We then compared the outcomes between the elemen- trol group. Prescription doses of elemental formula from tal formula group and the control group using a mixed- the day of admission to day 14  days are shown in Addi- effects logistic regression model [21] for binary outcomes tional file  3. Median duration of enteral feeding (25th– and a linear-mixed regression model for continuous out- 75th percentiles) was 9  days (5, 17) in the elemental comes as the primary analysis, adjusted by the case-mix Endo et al. Ann. Intensive Care (2018) 8:69 Page 4 of 8 Fig. 1 Flow diagram of patient selection. DPC Diagnosis Procedure Combination Table 1 Patient characteristics in our model. The case-mix classification model that we established had high accuracy regarding in-hospital Characteristics Elemental formula Control group mortality, with an AUROC of 0.90 for the establishment group cohort (Fig.  2). Furthermore, the established model was Number of subjects, n 382 566 well calibrated for the validation cohort (AUROC, 0.93; Age (years) 62 [45, 74] 63 [45, 75] Hosmer–Lemeshow goodness-of-fit test, p = 0.704) Sex, female, n (%) 134 (35.1) 204 (36.0) (Additional file 4). Charlson comorbidity 0 [0, 1] 0 [0, 1] index Multivariate mixed‑effect regression model Prognostic factor score 3 [1, 4] 3 [1, 4] Results of the multivariate mixed-effects model are CT severity score 2 [2, 2] 2 [1, 3] summarized in Table  2. No significant difference was Mechanical ventilation use, 101 (26.4) 201 (35.5) n (%) observed for in-hospital mortality [adjusted odds ratio Renal replacement 78 (20.4) 125 (22.1) (95% confidence interval; CI) = 0.94 (0.53–1.67)] and all therapy, n (%) secondary outcomes. Vasopressors use, n (%) 61 (16.0) 114 (20.1) Transfusion, n (%) 101 (26.4) 158 (27.9) Propensity score matching Annual number of acute 61.8 [47.7, 85.5] 73.3 [48.2, 92.5] Among all 948 eligible patients, 380 propensity score- pancreatitis per hospital matched pairs were generated via the matching process. Numeric variables are expressed as median [25th–75th percentiles] The ASMD in the variables indicated a well-matched bal - CT computed tomography ance (Additional file  5). The in-hospital mortality rate was 10.3% (39/380) for the elemental formula group and 8.7% (33/380) for the control group in the propensity score- formula group and 10  days (5, 18) in the control group, matched cohort. Results of the propensity score match- respectively. ing analysis are summarized in Table  3. Similar to the results of the multivariate mixed-effects model, no sig - Case‑mix adjustment nificant difference was observed for in-hospital mortality All VIFs of the variables used in the regression analysis [adjusted odds ratio (95% confidence interval: CI) = 1.20 were < 2, which eliminated the issue of multicollinearity (0.74–1.96)] and in all the secondary outcomes. Endo et al. Ann. Intensive Care (2018) 8:69 Page 5 of 8 was observed for all outcomes in both models as well as in patients with SAP. Discussion In this retrospective cohort study using a Japanese national database, we assessed the efficacy of elemental formula compared to other formulae of enteral nutri- tion for the outcomes of in-hospital mortality, develop- ment of sepsis, hospital-free days, and total health-care costs per admission. The results demonstrated that ele - mental formula use had few associations with all out- comes regarding early enteral nutrition in the treatment of acute pancreatitis. To the best of our knowledge, this study is the first to use large-scale data to directly com - pare the impact of elemental formula and other formulae in patients with acute pancreatitis. Several RCTs demonstrated the superiority of enteral nutrition compared to parenteral nutrition in late 1990s [8–11], resulting in a paradigm shift in the nutri- Fig. 2 Receiver operating curves of the risk adjustment model in the tional management of patients with acute pancreatitis establishment and validation cohort. AUROC area under the receiver toward enteral nutrition. The main mechanism of acute operating curve pancreatitis is autodigestion of the pancreas due to self-producing proteases. Enteral feeding increases pan- creatic secretion by stimulating the cephalic and gastric Analyses in patients with SAP phases, and early oral feeding may lead to recurrence of The results of analyses using the mixed-effect and pro - symptoms, elevation of serum amylase and lipase, and pensity score matching models in patients with SAP are delayed complications [23, 24]. On the other hand, it is presented in Additional file  6. No significant difference widely recognized that enteral nutrition decreases gut Table 2 Results of multivariate mixed-effects regression analysis Outcomes Elemental Control (N = 566) Adjusted odds Adjusted difference [95% CI] p value formula ratio [95% CI] (N = 382) Primary outcome In-hospital mortality, % 10.2 11.0 0.94 (0.53–1.67) – 0.823 Secondary outcomes Sepsis development, % 5.0 7.1 0.66 (0.34–1.28) – 0.218 Mean hospital-free days at 90 days, days 54 51 – 2 days (−2 to 5) 0.331 Mean total health-care costs, $ $29,360 $34,214 – −$4250 (−8643 to 141) 0.872 CI confidence interval Table 3 Results of propensity score matching analysis Outcomes Elemental Control (N = 380) Adjusted odds Adjusted difference [95% CI] p value formula ratio [95% CI] (N = 380) Primary outcome In-hospital mortality, % 10.3 8.7 1.20 (0.74–1.96) – 0.458 Secondary outcomes Sepsis development, % 5.0 6.8 0.72 (0.39–1.32) – 0.284 Mean hospital-free days at 90 days, days 54 53 – 1.1 days (−3.0 to 5.2) 0.596 Mean total health-care costs, $ $29,450 $32,366 – −$2916 (−8267 to 2435) 0.286 CI confidence interval Endo et al. Ann. Intensive Care (2018) 8:69 Page 6 of 8 permeability, reduces bacterial translocation, and acti- the advantage of a semi-elemental formula compared to vates mucosal immunity [25–27]. Considering the results elemental formula from the perspective of absorption, of clinical studies that suggested the superiority of enteral tolerability, and maintenance of gut immunity. Thus, ele - feeding compared to parenteral feeding for patients with mental formula may not always be beneficial compared acute pancreatitis, it is expected that the benefit of main - to a polymeric formula in experimental settings. taining gut immunity would overcome the drawbacks of Because this study was a retrospective study conducted an increase in the exocrine secretions of the pancreas. for a limited duration, the possibility of under power was Several clinical studies concerning the type of enteral a concern. Therefore, we performed a sample size calcu - nutrition have been conducted since the 2000s [28–30]; lation for in-hospital mortality based on the effect size however, there has not been sufficient evidence to justify estimated in propensity score matching analysis. The the use of specific nutrition, such as probiotics, fiber- results of power analysis, assuming α = 0.05 (two-sided) enriched formulae, and n − 3 fatty acids. In response to and β = 0.2 (power = 80%), showed that 5270 patients in those results, recent guidelines [31–35] generally recom- each group were required to demonstrate statistical sig- mended the use of standard polymeric formulae; how- nificance. This result implied that the effect of elemental ever, the issue of a shortage of sufficient evidence was formula was clinically limited for the outcome of mortal- also mentioned in their comments (Additional file  7). ity in patients with acute pancreatitis. Despite the recommendation of guidelines, in this study, The strength of our study was that we directly evalu - the proportion of patients who received elemental for- ated the effect of elemental formula in a large number mula reached approximately 40% among patients who of patients with acute pancreatitis compared to those received enteral feeding via nasogastric tube. This sug - in previous studies. Furthermore, we used a statistical gested that clinicians still expect the comparative effec - model with well-calibrated case-mix adjustment, which tiveness of elemental formula and that further evidence simultaneously accounted for hospital-level clustering. In regarding this is required. addition, health-care costs are uniform across all hospi- A study on volunteers indicated that elemental formula tals and individuals in Japan, which is often not the case has a less stimulatory effect on the secretion of pancre - in the other countries. However, our study had several atic lipase and chymotrypsin compared to food homoge- limitations. First, the possibility of residual confound- nate [36]. Elemental formula is expected to achieve both ing existed because of the retrospective nature of our maintenance of gut immunity and suppression of exo- study; however, the results were sufficiently informative crine secretions in the pancreas, suggesting that elemen- because the case-mix classification model used demon - tal formula is theoretically beneficial for patients with strated high accuracy for predicting in-hospital mor- acute pancreatitis. However, contrary to this assumption, tality. Second, details of the administration route were the results of this study failed to show the superiority not recorded in the DPC database; therefore, we could of elemental formula compared to semi-elemental and not distinguish between nasogastric feeding and naso- polymeric formulae, and several reasons for this can be jejunal feeding in this study. However, meta-analysis by considered. It has been reported that trans-jejunal feed- Chang et al. [40] showed that this difference had no effect ing of polymeric formulae was well tolerated by patients on the outcomes. Third, because this was not a parallel with acute pancreatitis and can potentially be used to RCT, standardization of the nutrition regimen between facilitate pancreatic rest [37]. This means that standard the study groups was not possible. Therefore, a nutri - formulae also could achieve pancreatic rest by control- tion management, such as administration dose, rate, ling the route of administration; however, information on and methods (i.e., continuous or intermitted), for each the dosage regimen and administration route could not patient varied according to respective patient status and be evaluated in this study. In addition, a study comparing hospital standard operation procedure. Fourth, a num- the absorption of nutrients in patients with cystic fibro - ber of patients were excluded through a patient selec- sis and pancreatic insufficiency who were administered tion process, and this may have caused a selection bias. elemental or polymeric formulae with and without pan- In Japan, proportion of severe acute pancreatitis patients creatic enzymes showed that no benefit was experienced who received enteral feeding within 48  h of admission by patients who were administered elemental formula, was reported to be approximately 10% [41]. In addition, compared to those administered the polymeric formula because the present study analyzed only patients who with pancreatic enzymes [38]. This result suggested that received enteral nutrition via nasogastric tube, less severe patients with pancreatic insufficiency could utilize poly - patients who were allowed oral intake were not analyzed. meric formulae as opposed to elemental formulae, with Furthermore, registering information on the severity of pancreatic enzyme supplementation as needed. Further- pancreatitis was not mandatory in the DPC database. more, a recent experimental study in mice [39] showed Finally, although every use of elemental formula product Endo et al. Ann. Intensive Care (2018) 8:69 Page 7 of 8 Authors’ contributions could be specified because they required prescriptions, AE helped in drafting/revising the manuscript, study concept and design, sta- some products of semi-elemental and polymeric formu- tistical analysis and interpretation of data, and obtaining funding and accepts lae could not be specified in the DPC database because responsibility for conduct of research and final approval, study supervision; AS contributed to revising the manuscript, study concept and design, statistical they were treated as meals and did not require prescrip- analysis and interpretation of data and accepts responsibility for conduct of tions in Japan. This prevented evaluation of the detailed research and final approval; KF was involved in acquisition of data, revising the differences among the type of semi-elemental and poly - manuscript, and obtaining funding and accepts responsibility for conduct of research and final approval; KM and YO contributed to revising the manuscript meric formulae. However, apparent overfeeding was not and interpretation of data and accepts responsibility for conduct of research observed, at least in the elemental formula group, and the and final approval. All authors read and approved the final manuscript. duration of enteral feeding was not markedly different Author details among groups. Trauma and Acute Critical Care Medical Center, Tokyo Medical and Dental Despite these limitations, the design of our study that University Hospital of Medicine, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, compared elemental formula and other formulae was Japan. Emergency and Trauma Center, Kameda Medical Center, 929 Higashi- cho, Kamogawa, Chiba, Japan. Department of Health Policy and Informatics, reasonable to show the difference of the effect of enteral Tokyo Medical and Dental University Graduate School of Medicine, 1-5-45 nutrition formulae from the perspective of pancre- Yushima, Bunkyo-ku, Tokyo, Japan. The Shock Trauma and Emergency Medi- atic rest, since only elemental formula does not require cal Center, Matsudo City Hospital, 4005 Kamihongo, Matsudo, Chiba, Japan. digestion process. Elemental formula is generally more Acknowledgements expensive compared to polymeric formulae; however, the The authors thank Editage (https ://www.edita ge.jp) for English language results of this study did not demonstrate any significant editing. difference in outcomes between the two. From the results Competing interests of this study and previous studies, there is not sufficient The authors declare that they have no competing interests. evidence to justify the routine use of elemental formula Availability of data and materials in the initiation of enteral nutrition for patients with The datasets generated and analyzed during the current study are not acute pancreatitis. publicly available due to the potential risk of leakage of personally identifiable information. Conclusions Consent for publication Not applicable. The results of a retrospective cohort study using a large- scale national database did not show the benefits of Ethics approval and consent to participate elemental formula compared to semi-elemental and The institutional review board of the Tokyo Medical and Dental University approved this study (#788). Informed consent from each patient was waived polymeric formulae in patients with acute pancreatitis. because of the retrospective design of the study and the use of anonymized Further assessment of alternative nutritional strategy is patient and hospital data. expected. Funding AE received the Grants-in-Aid for Scientific Research (#17K17045) from Japan Additional files Society for the Promotion of Science. KF received a Grant-in-Aid for Research on Policy Planning and Evaluation (#H28-Seisaku-Shitei-009) from the Ministry Additional file 1. Severity Scoring System for Acute Pancreatitis of the of Health, Labour and Welfare, Japan. Remaining authors received no funding. Japanese Ministry of Health, Labour and Welfare (2008). Additional file 2. International Classification of Diseases, 10th Revision Publisher’s Note codes used in the study. Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Additional file 3. Mean prescription dose of elemental formula from the day of admission to day 14. Received: 16 November 2017 Accepted: 28 May 2018 Additional file 4. Calibration plot. Additional file 5. Patient characteristics before and after propensity score matching. Additional file 6. Results of analyses using mixed-effect model and pro - References pensity score matching model in patients with severe acute pancreatitis. 1. Wall I, Badalov N, Baradarian R, Iswara K, Li JJ, Tenner S. Decreased mortal- Additional file 7. Recommendations in clinical practice guidelines for the ity in acute pancreatitis related to early aggressive hydration. Pancreas. type of formulae in acute pancreatitis. 2011;40:547–50. 2. Wu BU, Banks PA. Clinical management of patients with acute pancreati- tis. Gastroenterology. 2013;144:1272–81. 3. Greenberg JA, Hsu J, Bawazeer M, Marshall J, Friedrich JO, Nathens A, et al. Abbreviations Clinical practice guideline: management of acute pancreatitis. Can J Surg. RCT : randomized controlled trial; DPC: Diagnosis Procedure Combination; 2016;59:128–40. ICD-10: International Classification of Diseases, 10th revision; CT: computed 4. Abou-Assi S, O’Keefe SJ. Nutrition support during acute pancreatitis. tomography; VIF: variance inflation factor; AUROC: area under the receiver Nutrition. 2002;18:938–43. operating curve; ASMD: absolute standardized mean difference; SD: standard 5. Poropat G, Giljaca V, Hauser G, Štimac D. Enteral nutrition formulations for deviation; SAP: severe acute pancreatitis; CI: confidence interval. acute pancreatitis. Cochrane Database Syst Rev. 2015;3:CD010605. Endo et al. Ann. Intensive Care (2018) 8:69 Page 8 of 8 6. Grant JP, James S, Grabowski V, Trexler KM. Total parenteral nutrition in 26. Wang G, Wen J, Xu L, Zhou S, Gong M, Wen P, et al. Eec ff t of enteral pancreatic disease. Ann Surg. 1984;200:627–31. nutrition and ecoimmunonutrition on bacterial translocation and 7. Pisters PW, Ranson JH. Nutritional support for acute pancreatitis. Surg cytokine production in patients with severe acute pancreatitis. J Surg Res. Gynecol Obstet. 1992;175:275–84. 2013;183:592–7. 8. McClave SA, Greene LM, Snider HL, Makk LJ, Cheadle WG, Owens NA, 27. Sigalet DL, Mackenzie SL, Hameed SM. Enteral nutrition and mucosal et al. Comparison of the safety of early enteral vs parenteral nutrition in immunity: implications for feeding strategies in surgery and trauma. Can mild acute pancreatitis. JPEN J Parenter Enteral Nutr. 1997;21:14–20. J Surg. 2004;47:109–16. 9. Kalfarentzos F, Kehagias J, Mead N, Kokkinis K, Gogos CA. Enteral nutrition 28. Besselink MG, van Santvoort HC, Buskens E, Boermeester MA, van Goor is superior to parenteral nutrition in severe acute pancreatitis: results of a H, Timmerman HM, et al. Probiotic prophylaxis in predicted severe acute randomized prospective trial. Br J Surg. 1997;84:1665–9. pancreatitis: a randomised, double-blind, placebo-controlled trial. Lancet. 10. Windsor AC, Kanwar S, Li AG, Barnes E, Guthrie JA, Spark JI, et al. Com- 2008;371:651–9. pared with parenteral nutrition, enteral feeding attenuates the acute 29. Oláh A, Belágyi T, Issekutz A, Gamal ME, Bengmark S. Randomized clinical phase response and improves disease severity in acute pancreatitis. Gut. trial of specific lactobacillus and fibre supplement to early enteral nutri- 1998;42:431–5. tion in patients with acute pancreatitis. Br J Surg. 2002;89:1103–7. 11. Marik PE, Zaloga GP. Meta-analysis of parenteral nutrition versus enteral 30. Pearce CB, Sadek SA, Walters AM, Goggin PM, Somers SS, Toh SK, et al. nutrition in patients with acute pancreatitis. BMJ. 2004;328:1407. A double-blind, randomised, controlled trial to study the effects of an 12. Tiengou LE, Gloro R, Pouzoulet J, Bouhier K, Read MH, Arnaud-Battandier enteral feed supplemented with glutamine, arginine, and omega-3 fatty F, et al. Semi-elemental formula or polymeric formula: is there a better acid in predicted acute severe pancreatitis. JOP. 2006;7:361–71. choice for enteral nutrition in acute pancreatitis? Randomized compara- 31. Meier R, Ockenga J, Pertkiewicz M, Pap A, Milinic N, Macfie J, et al. ESPEN tive study. JPEN J Parenter Enteral Nutr. 2006;30:1–5. guidelines on enteral nutrition: pancreas. Clin Nutr. 2006;25:275–84. 13. Petrov MS, Loveday BP, Pylypchuk RD, McIlroy K, Phillips AR, Windsor JA. 32. Tenner S, Baillie J, DeWitt J, Vege SS. American College of Gastroenterol- Systematic review and meta-analysis of enteral nutrition formulations in ogy guideline: management of acute pancreatitis. Am J Gastroenterol. acute pancreatitis. Br J Surg. 2009;96:1243–52. 2013;108:1400–15. 14. Takeda K, Yokoe M, Takada T, Kataoka K, Yoshida M, Gabata T, et al. Assess- 33. Yokoe M, Takada T, Mayumi T, Yoshida M, Isaji S, Wada K, et al. Japanese ment of severity of acute pancreatitis according to new prognostic guidelines for the management of acute pancreatitis: Japanese Guide- factors and CT grading. J Hepatobiliary Pancreat Sci. 2010;17:37–44. lines 2015. J Hepatobiliary Pancreat Sci. 2015;22:405–32. 15. Yasunaga H, Matsui H, Horiguchi H, Fushimi K, Matsuda S. Clinical epi- 34. Taylor BE, McClave SA, Martindale RG, Warren MM, Johnson DR, Braun- demiology and health services research using the diagnosis procedure schweig C, et al. Guidelines for the Provision and Assessment of Nutrition combination database in Japan. Asian Pac J Dis Manage. 2013;7:19–24. Support Therapy in the Adult Critically Ill Patient: Society of Critical Care 16. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of clas- Medicine (SCCM) and American Society for Parenteral and Enteral Nutri- sifying prognostic comorbidity in longitudinal studies: development and tion (A.S.P.E.N.). Crit Care Med. 2016;44:390–438. validation. J Chronic Dis. 1987;40:373–83. 35. Crockett SD, Wani S, Gardner TB, Falck-Ytter Y, Barkun AN. American Gas- 17. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, et al. troenterological Association Institute Guideline on initial management of Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 acute pancreatitis. Gastroenterology. 2018;154:1096–101. administrative data. Med Care. 2005;43:1130–9. 36. Vison N, Hecketsweiler P, Butel J, Bernier JJ. Eec ff t of continuous jejunal 18. Young P, Hodgson C, Dulhunty J, Saxena M, Bailey M, Bellomo R, et al. perfusion of elemental and complex nutritional solutions on pancreatic End points for phase II trials in intensive care: recommendations from the enzyme secretion in human subjects. Gut. 1978;19:194–8. Australian and New Zealand Clinical Trials Group consensus panel meet- 37. Powell JJ, Murchison JT, Fearon KC, Ross JA, Siriwardena AK. Randomized ing. Crit Care Resusc. 2012;14:211–5. controlled trial of the effect of early enteral nutrition on markers of the 19. Hamada T, Yasunaga H, Nakai Y, Isayama H, Horiguchi H, Matsuda S, et al. inflammatory response in predicted severe acute pancreatitis. Br J Surg. Continuous regional arterial infusion for acute pancreatitis: a propensity 2000;87:1375–81. score analysis using a nationwide administrative database. Crit Care. 38. Erskine JM, Lingard CD, Sontag MK, Accurso FJ. Enteral nutrition for 2013;17:R214. patients with cystic fibrosis: comparison of a semi-elemental and nonel- 20. Horibe M, Sasaki M, Sanui M, Sugiyama D, Iwasaki E, Yamagishi Y, et al. emental formula. J Pediatr. 1998;132:265–9. Continuous regional arterial infusion of protease inhibitors has no 39. Moriya T, Fukatsu K, Noguchi M, Nishikawa M, Miyazaki H, Saitoh D, et al. efficacy in the treatment of severe acute pancreatitis: a retrospective Eec ff ts of semielemental diet containing whey peptides on Peyer’s patch multicenter cohort study. Pancreas. 2017;46:510–7. lymphocyte number, immunoglobulin A levels, and intestinal morphol- 21. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear ogy in mice. J Surg Res. 2018;222:153–9. models. Biometrika. 1986;73:13–22. 40. Chang YS, Fu HQ, Xiao YM, Liu JC. Nasogastric or nasojejunal feed- 22. D’Agostino RB Jr. Propensity score methods for bias reduction in the ing in predicted severe acute pancreatitis: a meta-analysis. Crit Care. comparison of a treatment to a non-randomized control group. Stat Med. 2013;17:R118. 1998;17:2265–81. 41. Arata S, Ikeda T, Soma K, Miyazawa S, Ikeda K, Shiga H, et al. The safety and 23. Ranson JHC. Acute pancreatitis: pathogenesis, outcome and treatment. tolerance of enteral nutrition in severe acute pancreatitis: a multicenter Clin Gastroenterol. 1984;13:843–63. prospective case series study. J Jpn Assoc Acute Med. 2012;23:233–41 (in 24. Wu XN. The mechanism of actions of Octreotide, Bupleurum-Peony Japanese). Cheng Qi decoction and Dan Shan in severe acute pancretitis. World J Gastroenterol. 1999;5:249–51. 25. 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:S46–51. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Annals of Intensive Care Springer Journals

Comparative effectiveness of elemental formula in the early enteral nutrition management of acute pancreatitis: a retrospective cohort study

Free
8 pages
Loading next page...
 
/lp/springer_journal/comparative-effectiveness-of-elemental-formula-in-the-early-enteral-ENll51V3vi
Publisher
Springer International Publishing
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Intensive / Critical Care Medicine; Emergency Medicine; Anesthesiology
eISSN
2110-5820
D.O.I.
10.1186/s13613-018-0414-6
Publisher site
See Article on Publisher Site

Abstract

Background: Although enteral nutrition has become one of the standard therapies for patients with acute pancrea- titis, the optimal formulae for enteral nutrition have been under debate. Elemental formula is assumed to be suitable in the treatment of patients with acute pancreatitis because it has less stimulating effects for exocrine secretions of the pancreas, simultaneously maintaining gut immunity; however, clinical studies corroborating this assumption have been scarce. Methods: We conducted a retrospective cohort study using a Japanese national administrative database between 2010 and 2015. Patients with acute pancreatitis who received enteral feeding within 3 days of admission were identi- fied and divided into two groups according to whether elemental formula was administered. We assessed the impact of elemental formula for the outcomes (primary, in-hospital mortality; secondary, development of sepsis, hospital-free days at 90 days, and total health-care costs) using a multivariate mixed-effect regression analysis and propensity score matching analysis adjusted by a well-validated case-mix adjustment model. Analysis for the subpopulation of patients with severe acute pancreatitis was also performed. Results: Of 243,312 patients with acute pancreatitis, 948 patients were identified and classified into the elemental formula group (N = 382) and the control group (N = 566). No significant differences were observed for in-hospital mortality [10.2% in the elemental formula group vs. 11.0% in the control group; adjusted adds ratio (95% confidence interval; CI) = 0.94 (0.53–1.67)], sepsis development [5.0 vs. 7.1%; adjusted adds ratio (95% CI) = 0.66 (0.34–1.28)], mean hospital-free days [54 days vs. 51 days; adjusted difference (95% CI) = 2 days (− 2 to 5)], and mean total health-care costs [$29,360 vs. $34,214; adjusted difference (95% CI) = − $4250 (− 8643 to 141)]. Similar results were also observed in patients with severe acute pancreatitis. Conclusions: The results of our retrospective cohort study using a large-scale national database did not demonstrate the benefit of elemental formula compared to semi-elemental and polymeric formulae in patients with acute pan- creatitis. Further assessment of alternative nutritional strategy is expected. Keywords: Acute pancreatitis, Enteral nutrition, Elemental diet, Polymeric formula, Mortality *Correspondence: eraeaccm@tmd.ac.jp Trauma and Acute Critical Care Medical Center, Tokyo Medical and Dental University Hospital of Medicine, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan Full list of author information is available at the end of the article © The Author(s) 2018. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Endo et al. Ann. Intensive Care (2018) 8:69 Page 2 of 8 comorbidities at admission, and post-admission compli- Background cations are independently recorded using the relevant Acute pancreatitis is a life-threatening inflammatory dis - codes from the International Classification of Diseases, ease characterized by autodigestion and destruction of 10th revision (ICD-10). In addition, the DPC database the pancreas due to self-producing proteases. The stand - includes baseline characteristics of patients and informa- ard treatment for acute pancreatitis mainly comprises tion regarding the treating hospital. Furthermore, infor- supportive therapy, such as adequate fluid resuscitation mation regarding the severity of acute pancreatitis was and respiratory care, because evidence of disease-specific recorded using the Japanese severity scoring system for therapy has been limited [1–3]. acute pancreatitis of the Ministry of Health, Labour, and Adequate nutritional strategy has been one of the key Welfare of Japan [14], as this comprises the prognostic factors during systematic support in patients with acute factor scores and computed tomography (CT) sever- pancreatitis [4, 5]. The concept of “pancreatic rest” had ity scores based on contrast-enhanced CT (Additional been widely believed to be the standard nutritional strat- file  1). Further details regarding the DPC database have egy in the management of acute pancreatitis; therefore, been described elsewhere [15]. total parenteral nutrition had been widely used up to the This study was conducted in accordance with the prin - 1990s [6, 7]. However, several randomized controlled tri- ciples of the 1964 Declaration of Helsinki and its later als (RCTs) conducted in the late 1990s demonstrated the amendments. The institutional review board of the Tokyo consistent superiority of enteral nutrition over parenteral Medical and Dental University approved this study nutrition [8–11]. Among the enteral nutrition formulae, (#788). Informed consent from each patient was waived elemental formula has been believed to elicit theoreti- because of the retrospective design of the study and the cal advantages owing to a lower degree of exocrine pan- use of anonymized patient and hospital data. creatic stimulation and because it is fat-free. A previous RCT [12] comparing the efficacy of a semi-elemental formula with a polymeric formula was underpowered Study population (N = 30) and failed to demonstrate the superiority of a We included patients who were admitted to the hospital semi-elemental formula. Because few RCTs or large-scale because of acute pancreatitis between April 1, 2010, and cohort studies have been reported in this theme, Petrov March 31, 2015, and patients who received nasogastric et  al. [13] performed an indirect adjusted meta-analysis, feeding or nasojejunal feeding within 3 days of admission. in which parenteral nutrition groups were used as the We excluded patients younger than 16  years and those reference and reported that significant difference regard - who were pregnant. Patients who were discharged within ing survival benefit and adverse events were not observed 3  days of admission were also excluded considering the between the groups of (semi)elemental formulae and pol- issue of immortal-time bias. In addition, we excluded ymeric formula. However, to our knowledge, a large RCT patients who had missing values in any variables used in or cohort study that directly compared the efficacy of ele - the analyses (i.e., complete case analyses). mental formula has not been reported in the treatment of acute pancreatitis. In this study, we aimed to assess Data collection the clinical benefit of elemental formula compared to the We collected the following information from the DPC other formulae (semi-elemental and polymeric formulae) database: age; sex; ICD-10 codes for four primary diag- in the initiation of enteral nutrition management in acute noses, four concurrent diagnoses at admission, and four pancreatitis, using a large-scale national administrative post-admission complications; the aforementioned prog- database. nostic factor score and CT severity score; unique hos- pital identifier; annual number of patients with acute pancreatitis per hospital; presence or absence of specific Methods reimbursement claims for enteral feeding; presence or Study design and data sources absence of specific reimbursement claims for use of ele - We conducted a retrospective cohort study to evaluate mental formula; status at hospital discharge (i.e., survived the efficacy of elemental formula in patients with acute or deceased); total health-care costs per admission; and pancreatitis, using the Japanese Diagnosis Procedure length of hospital stay. Furthermore, we collected infor- Combination (DPC) database. The database is a case- mation on whether the following interventions were mix classification system that is linked to the reimburse - performed within 3 days of admission: administration of ment system for inpatient cases in Japanese hospitals and vasopressors (dopamine, norepinephrine, epinephrine, contains administrative claims for every drug adminis- or vasopressin), mechanical ventilation, renal replace- tered and every procedure and care performed at more ment therapy, and transfusion. Patient comorbidities than 1500 hospitals. Each patient’s primary diagnosis, were assessed using the Charlson comorbidity index [16] Endo et al. Ann. Intensive Care (2018) 8:69 Page 3 of 8 based on a previously reported method for extracting the classification model established, with the random effects ICD-10 codes [17]. of hospital-level clustering. In the linear mixed-effects model, the case-mix classification model was inverse- logit-transformed to satisfy the homoscedasticity Definitions and outcomes requirement for linear regression. Patients who were administered elemental formula We further compared the outcomes in the elemental were identified by the presence of specific reimburse - formula group and the control group using a propensity ment claims of elemental formula (Elental , Ajinomoto score matching analysis [22] as the secondary analy- Pharmaceutical Ltd., Tokyo, Japan; and Hepan ED , EA sis. The propensity score for predicting administration pharma Ltd., Tokyo, Japan; those are all the elemental of elemental formula was calculated through a logistic formula products approved in Japan). The control group regression analysis using the variables used for the estab- was defined as patients who were administered other lishment of the aforementioned prognosis model and types of enteral nutritional formulae within 3  days of the annual number of acute pancreatitis cases per hos- admission. The primary outcome was in-hospital mortal - pital as a variable to account for the differences in treat - ity. Secondary outcomes were development of sepsis after ment quality at each hospital. Propensity score matching admission, hospital-free days, and total health-care costs extracted 1:1 matched pairs from the elemental formula per admission. The ICD-10 codes used to identify sepsis group and the control group. A match balance between are presented in Additional file  2. Hospital-free days were the two groups was assessed using the absolute standard- defined as days survived and days free from hospitaliza - ized mean difference (ASMD) of all variables; values < 0.1 tion within 90  days from initial hospital admission. A were regarded as acceptable. To achieve balanced match- recent clinical trial group consensus recommended that ing, the caliper width for matching was set as the stand- hospital-free days should represent composite meas- ard deviation (SD) of the logit-transformed propensity ures compared to length of hospital stay, which could score multiplied by 0.3. Intergroup comparison of the be highly influenced by mortality [18]. Total health-care outcomes with propensity score-matched subjects was cost was defined as all aggregated payments (except for performed using a Chi-square test. boarding costs) to the hospital per discharge, and these In addition, we performed analyses using the afore- payments were estimated using the reference prices in mentioned two models only in patients with severe acute the Japanese fee schedule, which lists reimbursement pancreatitis (SAP) diagnosed by the severity diagnosis rates for surgical, pharmacological, laboratory, and other criteria in Japan, to assess the efficacy of elemental for - inpatient services. The cost data were provided after con - mula in patients with SAP. verting the cost in yen to US dollars (100 yen = $1 USD). All statistical analyses were performed using R software (version 3.4.1; R Foundation for Statistical Computing, Statistical analysis Vienna, Austria). The level of significance was defined as We developed a risk adjustment model for in-hospital p < 0.05. mortality using the variables of age, sex, Charlson comor- bidity index, prognostic factor score, CT severity score, mechanical ventilation within 3  days of admission, renal Results replacement therapy within 3  days of admission, trans- Study population fusion within 3  days of admission, and vasopressor use The flow diagram of the patient selection process is within 3  days of admission by applying a logistic regres- presented in Fig.  1. During the study period, a total of sion model that included a random sample of 80% of the 243,312 patients with acute pancreatitis were hospital- entire study cohort. The covariables used were selected ized at the DPC participating hospitals. Of these, 948 based on clinical experience and previous studies [19, patients were identified according to the inclusion and 20]. Issues with variable multicollinearity were assessed exclusion criteria. Among these patients, 382 (approxi- using a variance inflation factor (VIF), and the toler - mately 40.3%) were administered elemental formula. ance value was set at < 2. We validated the model for the Patients’ characteristics according to whether elemen- remaining 20% of the cohort using the area under the tal formula was administered are presented in Table  1. receiver operating curve (AUROC) and a Hosmer–Leme- The in-hospital mortality rate was 10.2% (39/382) in the show goodness-of-fit test. elemental formula group and 11.0% (62/566) in the con- We then compared the outcomes between the elemen- trol group. Prescription doses of elemental formula from tal formula group and the control group using a mixed- the day of admission to day 14  days are shown in Addi- effects logistic regression model [21] for binary outcomes tional file  3. Median duration of enteral feeding (25th– and a linear-mixed regression model for continuous out- 75th percentiles) was 9  days (5, 17) in the elemental comes as the primary analysis, adjusted by the case-mix Endo et al. Ann. Intensive Care (2018) 8:69 Page 4 of 8 Fig. 1 Flow diagram of patient selection. DPC Diagnosis Procedure Combination Table 1 Patient characteristics in our model. The case-mix classification model that we established had high accuracy regarding in-hospital Characteristics Elemental formula Control group mortality, with an AUROC of 0.90 for the establishment group cohort (Fig.  2). Furthermore, the established model was Number of subjects, n 382 566 well calibrated for the validation cohort (AUROC, 0.93; Age (years) 62 [45, 74] 63 [45, 75] Hosmer–Lemeshow goodness-of-fit test, p = 0.704) Sex, female, n (%) 134 (35.1) 204 (36.0) (Additional file 4). Charlson comorbidity 0 [0, 1] 0 [0, 1] index Multivariate mixed‑effect regression model Prognostic factor score 3 [1, 4] 3 [1, 4] Results of the multivariate mixed-effects model are CT severity score 2 [2, 2] 2 [1, 3] summarized in Table  2. No significant difference was Mechanical ventilation use, 101 (26.4) 201 (35.5) n (%) observed for in-hospital mortality [adjusted odds ratio Renal replacement 78 (20.4) 125 (22.1) (95% confidence interval; CI) = 0.94 (0.53–1.67)] and all therapy, n (%) secondary outcomes. Vasopressors use, n (%) 61 (16.0) 114 (20.1) Transfusion, n (%) 101 (26.4) 158 (27.9) Propensity score matching Annual number of acute 61.8 [47.7, 85.5] 73.3 [48.2, 92.5] Among all 948 eligible patients, 380 propensity score- pancreatitis per hospital matched pairs were generated via the matching process. Numeric variables are expressed as median [25th–75th percentiles] The ASMD in the variables indicated a well-matched bal - CT computed tomography ance (Additional file  5). The in-hospital mortality rate was 10.3% (39/380) for the elemental formula group and 8.7% (33/380) for the control group in the propensity score- formula group and 10  days (5, 18) in the control group, matched cohort. Results of the propensity score match- respectively. ing analysis are summarized in Table  3. Similar to the results of the multivariate mixed-effects model, no sig - Case‑mix adjustment nificant difference was observed for in-hospital mortality All VIFs of the variables used in the regression analysis [adjusted odds ratio (95% confidence interval: CI) = 1.20 were < 2, which eliminated the issue of multicollinearity (0.74–1.96)] and in all the secondary outcomes. Endo et al. Ann. Intensive Care (2018) 8:69 Page 5 of 8 was observed for all outcomes in both models as well as in patients with SAP. Discussion In this retrospective cohort study using a Japanese national database, we assessed the efficacy of elemental formula compared to other formulae of enteral nutri- tion for the outcomes of in-hospital mortality, develop- ment of sepsis, hospital-free days, and total health-care costs per admission. The results demonstrated that ele - mental formula use had few associations with all out- comes regarding early enteral nutrition in the treatment of acute pancreatitis. To the best of our knowledge, this study is the first to use large-scale data to directly com - pare the impact of elemental formula and other formulae in patients with acute pancreatitis. Several RCTs demonstrated the superiority of enteral nutrition compared to parenteral nutrition in late 1990s [8–11], resulting in a paradigm shift in the nutri- Fig. 2 Receiver operating curves of the risk adjustment model in the tional management of patients with acute pancreatitis establishment and validation cohort. AUROC area under the receiver toward enteral nutrition. The main mechanism of acute operating curve pancreatitis is autodigestion of the pancreas due to self-producing proteases. Enteral feeding increases pan- creatic secretion by stimulating the cephalic and gastric Analyses in patients with SAP phases, and early oral feeding may lead to recurrence of The results of analyses using the mixed-effect and pro - symptoms, elevation of serum amylase and lipase, and pensity score matching models in patients with SAP are delayed complications [23, 24]. On the other hand, it is presented in Additional file  6. No significant difference widely recognized that enteral nutrition decreases gut Table 2 Results of multivariate mixed-effects regression analysis Outcomes Elemental Control (N = 566) Adjusted odds Adjusted difference [95% CI] p value formula ratio [95% CI] (N = 382) Primary outcome In-hospital mortality, % 10.2 11.0 0.94 (0.53–1.67) – 0.823 Secondary outcomes Sepsis development, % 5.0 7.1 0.66 (0.34–1.28) – 0.218 Mean hospital-free days at 90 days, days 54 51 – 2 days (−2 to 5) 0.331 Mean total health-care costs, $ $29,360 $34,214 – −$4250 (−8643 to 141) 0.872 CI confidence interval Table 3 Results of propensity score matching analysis Outcomes Elemental Control (N = 380) Adjusted odds Adjusted difference [95% CI] p value formula ratio [95% CI] (N = 380) Primary outcome In-hospital mortality, % 10.3 8.7 1.20 (0.74–1.96) – 0.458 Secondary outcomes Sepsis development, % 5.0 6.8 0.72 (0.39–1.32) – 0.284 Mean hospital-free days at 90 days, days 54 53 – 1.1 days (−3.0 to 5.2) 0.596 Mean total health-care costs, $ $29,450 $32,366 – −$2916 (−8267 to 2435) 0.286 CI confidence interval Endo et al. Ann. Intensive Care (2018) 8:69 Page 6 of 8 permeability, reduces bacterial translocation, and acti- the advantage of a semi-elemental formula compared to vates mucosal immunity [25–27]. Considering the results elemental formula from the perspective of absorption, of clinical studies that suggested the superiority of enteral tolerability, and maintenance of gut immunity. Thus, ele - feeding compared to parenteral feeding for patients with mental formula may not always be beneficial compared acute pancreatitis, it is expected that the benefit of main - to a polymeric formula in experimental settings. taining gut immunity would overcome the drawbacks of Because this study was a retrospective study conducted an increase in the exocrine secretions of the pancreas. for a limited duration, the possibility of under power was Several clinical studies concerning the type of enteral a concern. Therefore, we performed a sample size calcu - nutrition have been conducted since the 2000s [28–30]; lation for in-hospital mortality based on the effect size however, there has not been sufficient evidence to justify estimated in propensity score matching analysis. The the use of specific nutrition, such as probiotics, fiber- results of power analysis, assuming α = 0.05 (two-sided) enriched formulae, and n − 3 fatty acids. In response to and β = 0.2 (power = 80%), showed that 5270 patients in those results, recent guidelines [31–35] generally recom- each group were required to demonstrate statistical sig- mended the use of standard polymeric formulae; how- nificance. This result implied that the effect of elemental ever, the issue of a shortage of sufficient evidence was formula was clinically limited for the outcome of mortal- also mentioned in their comments (Additional file  7). ity in patients with acute pancreatitis. Despite the recommendation of guidelines, in this study, The strength of our study was that we directly evalu - the proportion of patients who received elemental for- ated the effect of elemental formula in a large number mula reached approximately 40% among patients who of patients with acute pancreatitis compared to those received enteral feeding via nasogastric tube. This sug - in previous studies. Furthermore, we used a statistical gested that clinicians still expect the comparative effec - model with well-calibrated case-mix adjustment, which tiveness of elemental formula and that further evidence simultaneously accounted for hospital-level clustering. In regarding this is required. addition, health-care costs are uniform across all hospi- A study on volunteers indicated that elemental formula tals and individuals in Japan, which is often not the case has a less stimulatory effect on the secretion of pancre - in the other countries. However, our study had several atic lipase and chymotrypsin compared to food homoge- limitations. First, the possibility of residual confound- nate [36]. Elemental formula is expected to achieve both ing existed because of the retrospective nature of our maintenance of gut immunity and suppression of exo- study; however, the results were sufficiently informative crine secretions in the pancreas, suggesting that elemen- because the case-mix classification model used demon - tal formula is theoretically beneficial for patients with strated high accuracy for predicting in-hospital mor- acute pancreatitis. However, contrary to this assumption, tality. Second, details of the administration route were the results of this study failed to show the superiority not recorded in the DPC database; therefore, we could of elemental formula compared to semi-elemental and not distinguish between nasogastric feeding and naso- polymeric formulae, and several reasons for this can be jejunal feeding in this study. However, meta-analysis by considered. It has been reported that trans-jejunal feed- Chang et al. [40] showed that this difference had no effect ing of polymeric formulae was well tolerated by patients on the outcomes. Third, because this was not a parallel with acute pancreatitis and can potentially be used to RCT, standardization of the nutrition regimen between facilitate pancreatic rest [37]. This means that standard the study groups was not possible. Therefore, a nutri - formulae also could achieve pancreatic rest by control- tion management, such as administration dose, rate, ling the route of administration; however, information on and methods (i.e., continuous or intermitted), for each the dosage regimen and administration route could not patient varied according to respective patient status and be evaluated in this study. In addition, a study comparing hospital standard operation procedure. Fourth, a num- the absorption of nutrients in patients with cystic fibro - ber of patients were excluded through a patient selec- sis and pancreatic insufficiency who were administered tion process, and this may have caused a selection bias. elemental or polymeric formulae with and without pan- In Japan, proportion of severe acute pancreatitis patients creatic enzymes showed that no benefit was experienced who received enteral feeding within 48  h of admission by patients who were administered elemental formula, was reported to be approximately 10% [41]. In addition, compared to those administered the polymeric formula because the present study analyzed only patients who with pancreatic enzymes [38]. This result suggested that received enteral nutrition via nasogastric tube, less severe patients with pancreatic insufficiency could utilize poly - patients who were allowed oral intake were not analyzed. meric formulae as opposed to elemental formulae, with Furthermore, registering information on the severity of pancreatic enzyme supplementation as needed. Further- pancreatitis was not mandatory in the DPC database. more, a recent experimental study in mice [39] showed Finally, although every use of elemental formula product Endo et al. Ann. Intensive Care (2018) 8:69 Page 7 of 8 Authors’ contributions could be specified because they required prescriptions, AE helped in drafting/revising the manuscript, study concept and design, sta- some products of semi-elemental and polymeric formu- tistical analysis and interpretation of data, and obtaining funding and accepts lae could not be specified in the DPC database because responsibility for conduct of research and final approval, study supervision; AS contributed to revising the manuscript, study concept and design, statistical they were treated as meals and did not require prescrip- analysis and interpretation of data and accepts responsibility for conduct of tions in Japan. This prevented evaluation of the detailed research and final approval; KF was involved in acquisition of data, revising the differences among the type of semi-elemental and poly - manuscript, and obtaining funding and accepts responsibility for conduct of research and final approval; KM and YO contributed to revising the manuscript meric formulae. However, apparent overfeeding was not and interpretation of data and accepts responsibility for conduct of research observed, at least in the elemental formula group, and the and final approval. All authors read and approved the final manuscript. duration of enteral feeding was not markedly different Author details among groups. Trauma and Acute Critical Care Medical Center, Tokyo Medical and Dental Despite these limitations, the design of our study that University Hospital of Medicine, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, compared elemental formula and other formulae was Japan. Emergency and Trauma Center, Kameda Medical Center, 929 Higashi- cho, Kamogawa, Chiba, Japan. Department of Health Policy and Informatics, reasonable to show the difference of the effect of enteral Tokyo Medical and Dental University Graduate School of Medicine, 1-5-45 nutrition formulae from the perspective of pancre- Yushima, Bunkyo-ku, Tokyo, Japan. The Shock Trauma and Emergency Medi- atic rest, since only elemental formula does not require cal Center, Matsudo City Hospital, 4005 Kamihongo, Matsudo, Chiba, Japan. digestion process. Elemental formula is generally more Acknowledgements expensive compared to polymeric formulae; however, the The authors thank Editage (https ://www.edita ge.jp) for English language results of this study did not demonstrate any significant editing. difference in outcomes between the two. From the results Competing interests of this study and previous studies, there is not sufficient The authors declare that they have no competing interests. evidence to justify the routine use of elemental formula Availability of data and materials in the initiation of enteral nutrition for patients with The datasets generated and analyzed during the current study are not acute pancreatitis. publicly available due to the potential risk of leakage of personally identifiable information. Conclusions Consent for publication Not applicable. The results of a retrospective cohort study using a large- scale national database did not show the benefits of Ethics approval and consent to participate elemental formula compared to semi-elemental and The institutional review board of the Tokyo Medical and Dental University approved this study (#788). Informed consent from each patient was waived polymeric formulae in patients with acute pancreatitis. because of the retrospective design of the study and the use of anonymized Further assessment of alternative nutritional strategy is patient and hospital data. expected. Funding AE received the Grants-in-Aid for Scientific Research (#17K17045) from Japan Additional files Society for the Promotion of Science. KF received a Grant-in-Aid for Research on Policy Planning and Evaluation (#H28-Seisaku-Shitei-009) from the Ministry Additional file 1. Severity Scoring System for Acute Pancreatitis of the of Health, Labour and Welfare, Japan. Remaining authors received no funding. Japanese Ministry of Health, Labour and Welfare (2008). Additional file 2. International Classification of Diseases, 10th Revision Publisher’s Note codes used in the study. Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Additional file 3. Mean prescription dose of elemental formula from the day of admission to day 14. Received: 16 November 2017 Accepted: 28 May 2018 Additional file 4. Calibration plot. Additional file 5. Patient characteristics before and after propensity score matching. Additional file 6. Results of analyses using mixed-effect model and pro - References pensity score matching model in patients with severe acute pancreatitis. 1. Wall I, Badalov N, Baradarian R, Iswara K, Li JJ, Tenner S. Decreased mortal- Additional file 7. Recommendations in clinical practice guidelines for the ity in acute pancreatitis related to early aggressive hydration. Pancreas. type of formulae in acute pancreatitis. 2011;40:547–50. 2. Wu BU, Banks PA. Clinical management of patients with acute pancreati- tis. Gastroenterology. 2013;144:1272–81. 3. Greenberg JA, Hsu J, Bawazeer M, Marshall J, Friedrich JO, Nathens A, et al. Abbreviations Clinical practice guideline: management of acute pancreatitis. Can J Surg. RCT : randomized controlled trial; DPC: Diagnosis Procedure Combination; 2016;59:128–40. ICD-10: International Classification of Diseases, 10th revision; CT: computed 4. Abou-Assi S, O’Keefe SJ. Nutrition support during acute pancreatitis. tomography; VIF: variance inflation factor; AUROC: area under the receiver Nutrition. 2002;18:938–43. operating curve; ASMD: absolute standardized mean difference; SD: standard 5. Poropat G, Giljaca V, Hauser G, Štimac D. Enteral nutrition formulations for deviation; SAP: severe acute pancreatitis; CI: confidence interval. acute pancreatitis. Cochrane Database Syst Rev. 2015;3:CD010605. Endo et al. Ann. Intensive Care (2018) 8:69 Page 8 of 8 6. Grant JP, James S, Grabowski V, Trexler KM. Total parenteral nutrition in 26. Wang G, Wen J, Xu L, Zhou S, Gong M, Wen P, et al. Eec ff t of enteral pancreatic disease. Ann Surg. 1984;200:627–31. nutrition and ecoimmunonutrition on bacterial translocation and 7. Pisters PW, Ranson JH. Nutritional support for acute pancreatitis. Surg cytokine production in patients with severe acute pancreatitis. J Surg Res. Gynecol Obstet. 1992;175:275–84. 2013;183:592–7. 8. McClave SA, Greene LM, Snider HL, Makk LJ, Cheadle WG, Owens NA, 27. Sigalet DL, Mackenzie SL, Hameed SM. Enteral nutrition and mucosal et al. Comparison of the safety of early enteral vs parenteral nutrition in immunity: implications for feeding strategies in surgery and trauma. Can mild acute pancreatitis. JPEN J Parenter Enteral Nutr. 1997;21:14–20. J Surg. 2004;47:109–16. 9. Kalfarentzos F, Kehagias J, Mead N, Kokkinis K, Gogos CA. Enteral nutrition 28. Besselink MG, van Santvoort HC, Buskens E, Boermeester MA, van Goor is superior to parenteral nutrition in severe acute pancreatitis: results of a H, Timmerman HM, et al. Probiotic prophylaxis in predicted severe acute randomized prospective trial. Br J Surg. 1997;84:1665–9. pancreatitis: a randomised, double-blind, placebo-controlled trial. Lancet. 10. Windsor AC, Kanwar S, Li AG, Barnes E, Guthrie JA, Spark JI, et al. Com- 2008;371:651–9. pared with parenteral nutrition, enteral feeding attenuates the acute 29. Oláh A, Belágyi T, Issekutz A, Gamal ME, Bengmark S. Randomized clinical phase response and improves disease severity in acute pancreatitis. Gut. trial of specific lactobacillus and fibre supplement to early enteral nutri- 1998;42:431–5. tion in patients with acute pancreatitis. Br J Surg. 2002;89:1103–7. 11. Marik PE, Zaloga GP. Meta-analysis of parenteral nutrition versus enteral 30. Pearce CB, Sadek SA, Walters AM, Goggin PM, Somers SS, Toh SK, et al. nutrition in patients with acute pancreatitis. BMJ. 2004;328:1407. A double-blind, randomised, controlled trial to study the effects of an 12. Tiengou LE, Gloro R, Pouzoulet J, Bouhier K, Read MH, Arnaud-Battandier enteral feed supplemented with glutamine, arginine, and omega-3 fatty F, et al. Semi-elemental formula or polymeric formula: is there a better acid in predicted acute severe pancreatitis. JOP. 2006;7:361–71. choice for enteral nutrition in acute pancreatitis? Randomized compara- 31. Meier R, Ockenga J, Pertkiewicz M, Pap A, Milinic N, Macfie J, et al. ESPEN tive study. JPEN J Parenter Enteral Nutr. 2006;30:1–5. guidelines on enteral nutrition: pancreas. Clin Nutr. 2006;25:275–84. 13. Petrov MS, Loveday BP, Pylypchuk RD, McIlroy K, Phillips AR, Windsor JA. 32. Tenner S, Baillie J, DeWitt J, Vege SS. American College of Gastroenterol- Systematic review and meta-analysis of enteral nutrition formulations in ogy guideline: management of acute pancreatitis. Am J Gastroenterol. acute pancreatitis. Br J Surg. 2009;96:1243–52. 2013;108:1400–15. 14. Takeda K, Yokoe M, Takada T, Kataoka K, Yoshida M, Gabata T, et al. Assess- 33. Yokoe M, Takada T, Mayumi T, Yoshida M, Isaji S, Wada K, et al. Japanese ment of severity of acute pancreatitis according to new prognostic guidelines for the management of acute pancreatitis: Japanese Guide- factors and CT grading. J Hepatobiliary Pancreat Sci. 2010;17:37–44. lines 2015. J Hepatobiliary Pancreat Sci. 2015;22:405–32. 15. Yasunaga H, Matsui H, Horiguchi H, Fushimi K, Matsuda S. Clinical epi- 34. Taylor BE, McClave SA, Martindale RG, Warren MM, Johnson DR, Braun- demiology and health services research using the diagnosis procedure schweig C, et al. Guidelines for the Provision and Assessment of Nutrition combination database in Japan. Asian Pac J Dis Manage. 2013;7:19–24. Support Therapy in the Adult Critically Ill Patient: Society of Critical Care 16. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of clas- Medicine (SCCM) and American Society for Parenteral and Enteral Nutri- sifying prognostic comorbidity in longitudinal studies: development and tion (A.S.P.E.N.). Crit Care Med. 2016;44:390–438. validation. J Chronic Dis. 1987;40:373–83. 35. Crockett SD, Wani S, Gardner TB, Falck-Ytter Y, Barkun AN. American Gas- 17. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, et al. troenterological Association Institute Guideline on initial management of Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 acute pancreatitis. Gastroenterology. 2018;154:1096–101. administrative data. Med Care. 2005;43:1130–9. 36. Vison N, Hecketsweiler P, Butel J, Bernier JJ. Eec ff t of continuous jejunal 18. Young P, Hodgson C, Dulhunty J, Saxena M, Bailey M, Bellomo R, et al. perfusion of elemental and complex nutritional solutions on pancreatic End points for phase II trials in intensive care: recommendations from the enzyme secretion in human subjects. Gut. 1978;19:194–8. Australian and New Zealand Clinical Trials Group consensus panel meet- 37. Powell JJ, Murchison JT, Fearon KC, Ross JA, Siriwardena AK. Randomized ing. Crit Care Resusc. 2012;14:211–5. controlled trial of the effect of early enteral nutrition on markers of the 19. Hamada T, Yasunaga H, Nakai Y, Isayama H, Horiguchi H, Matsuda S, et al. inflammatory response in predicted severe acute pancreatitis. Br J Surg. Continuous regional arterial infusion for acute pancreatitis: a propensity 2000;87:1375–81. score analysis using a nationwide administrative database. Crit Care. 38. Erskine JM, Lingard CD, Sontag MK, Accurso FJ. Enteral nutrition for 2013;17:R214. patients with cystic fibrosis: comparison of a semi-elemental and nonel- 20. Horibe M, Sasaki M, Sanui M, Sugiyama D, Iwasaki E, Yamagishi Y, et al. emental formula. J Pediatr. 1998;132:265–9. Continuous regional arterial infusion of protease inhibitors has no 39. Moriya T, Fukatsu K, Noguchi M, Nishikawa M, Miyazaki H, Saitoh D, et al. efficacy in the treatment of severe acute pancreatitis: a retrospective Eec ff ts of semielemental diet containing whey peptides on Peyer’s patch multicenter cohort study. Pancreas. 2017;46:510–7. lymphocyte number, immunoglobulin A levels, and intestinal morphol- 21. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear ogy in mice. J Surg Res. 2018;222:153–9. models. Biometrika. 1986;73:13–22. 40. Chang YS, Fu HQ, Xiao YM, Liu JC. Nasogastric or nasojejunal feed- 22. D’Agostino RB Jr. Propensity score methods for bias reduction in the ing in predicted severe acute pancreatitis: a meta-analysis. Crit Care. comparison of a treatment to a non-randomized control group. Stat Med. 2013;17:R118. 1998;17:2265–81. 41. Arata S, Ikeda T, Soma K, Miyazawa S, Ikeda K, Shiga H, et al. The safety and 23. Ranson JHC. Acute pancreatitis: pathogenesis, outcome and treatment. tolerance of enteral nutrition in severe acute pancreatitis: a multicenter Clin Gastroenterol. 1984;13:843–63. prospective case series study. J Jpn Assoc Acute Med. 2012;23:233–41 (in 24. Wu XN. The mechanism of actions of Octreotide, Bupleurum-Peony Japanese). Cheng Qi decoction and Dan Shan in severe acute pancretitis. World J Gastroenterol. 1999;5:249–51. 25. 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:S46–51.

Journal

Annals of Intensive CareSpringer Journals

Published: Jun 5, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off