Letters to the Editor Annals of Surgery Volume 267, Number 3, March 2018 would facilitate the use of miRNA-based easy to score, this is a powerful design. The Outcomes Research (DM) for ACELITY, screening strategy in clinical settings. risk of biased selection is almost completely we acknowledge and value addressing the avoided by consecutively including all ‘‘clinical data gap’’ in ventral hernia repair. Disclosure: The authors declare no patients, incomplete blinding is never an issue We believe it is equally important that studies conflicts of interest. and optimal external validity ensured. The are conducted and reported with appropriate magnitude of effect ‘across the nation’ is also methodological rigor. The aim of our letter is much better preserved in the design used by to provide surgeons and other study Kecheng Zhang, MD McMillan’s trial and this is an evidence reviewers a deeper and clearer understanding Lin Chen, MD modality that has received too little attention of the recently published complex open bio- Department of General Surgery, People’s compared with evidence of ‘cause-and-effect’ absorbable reconstruction of the abdominal Liberation Army General Hospital that is the raison d’etre of the explanatory wall (COBRA) study. In particular, we high- Beijing, China RCT. light the limitations of the study design email@example.com The devastating number of trial viola- employed, the limitations in the statistical tions in the German trial illustrates lack of analysis and reporting, and the inappropriate- surgeon equipoise immediately after resection ness of the conclusions drawn by the authors. REFERENCES and shows that the timing of randomization To be clear, W.L. Gore & Associates, the 1. Carter JV, Roberts HL, Pan J, et al. A highly was wrong. We have previously conducted a sponsor of The COBRA study, and LifeCell, predictive model for diagnosis of colorectal neoplasms using plasma microrna: improving large RCT on intraoperative insertion of an ACELITY Company, are competitors, and specificity and sensitivity. Ann Surg. 2016;264: feeding catheters and anticipated this phenom- it would be disingenuous not to explicitly 575–584. enon. Postresection randomization per- declare this before reviewing the aforemen- 2. Dong H, Lei J, Ding L, et al. MicroRNA: function, formed by computer from the operating tioned study. Our hope is that this critical detection, and bioanalysis. Chem Rev. 2013;113: theatre immediately before closure of the assessment might assist surgeons in their 6207–6233. abdomen is effortless today and would have evaluation of this and future studies. 3. Moldovan L, Batte KE, Trgovcich J, et al. Meth- odological challenges in utilizing miRNAs as cir- avoided this problem in the PANDRA trial. The stated objective was to evaluate the culating biomarkers. J Cell Mol Med. 2014; use and performance of a new tissue reinforce- 18:371–390. Disclosure: The author declares no ment material. The absence of a control group 4. Schultz NA, Dehlendorff C, Jensen BV, et al. conflicts of interest. precludes statements attributing treatment MicroRNA biomarkers in whole blood for detec- tion of pancreatic cancer. JAMA. 2014;311:392– effect to the mesh. In a single-arm trial, eval- uating new mesh reinforcement cannot be Kristoffer Lassen, MD, PhD 5. Zhu C, Ren C, Han J, et al. A five-microRNA panel separated from the surgical treatment effect. Oslo University Hospital, Oslo, Norway. in plasma was identified as potential biomarker for Other variables, both measured (surgical tech- firstname.lastname@example.org; email@example.com. early detection of gastric cancer. Br J Cancer. nique, mesh placement, presence of infection, 2014;110:2291–2299. enrollment criteria, size of wound) and unmeasured (surgeon experience, care path- REFERENCES ways, time) all have a substantial effect on 1. Witzigmann H, Diener MK, Kienkotter S, et al. No need for routine drainage after pancreatic head response to treatment and importantly are all Drain After resection: the dual-center, randomized, controlled uncontrolled. As a result, from the outset, the PANDRA trial (ISRCTN04937707). Ann Surg. Pancreatoduodenectomy: objective of evaluating a new mesh was incon- 2016;264:528–537. gruent with a single-arm design. Methodological Issues 2. McMillan MT, Malleo G, Bassi C, et al. Multicenter, The sample size was calculated from a prospective trial of selective drain management for background event rate of 50% recurrence at 2 pancreatoduodenectomy using risk stratification. Ann Surg. 2016 [Epub ahead of print]. years, based on the study by de Vries Rei- 3. Lassen K, Høye A, Myrmel T. Randomised trials in To the Editor: lingh et al, which recruited patients between surgery: the burden of evidence. Rev Recent Clin he Annals of Surgery have recently pub- 1999 and 2001, a full 12 to 13 years before Trials. 2012;7:244–248. T lished 2 interesting articles addressing the COBRA study. Moreover, the defect size 4. Lassen K, Kjaeve J, Fetveit T, et al. Allowing normal routine use of intraabdominal drains after was 2 times larger in the study by de Vries food at will after major upper gastrointestinal 1,2 pancreatoduodenectomies (PD). Both Reilingh et al. Differences in patient charac- surgery does not increase morbidity: a randomized multicenter trial. Ann Surg. 2008; 247:721–729. articles provide significant insight into the teristics from the original study, and advances clinical dilemma under scrutiny but they also in surgical technique and patient care strat- demonstrate some equally interesting meth- egies in the intervening time would suggest odological issues. The PANDRA trial, a that 50% recurrence is inappropriate for the randomized controlled trial (RCT), included current timeframe and study. This likely only 13% of eligible patients and 40 patients underestimates the necessary sample size Questions Regarding in the no-drain group had drains placed for an informative study and of its relevance Statistical Inferences, because of surgeons deliberately violating to current practice. trial protocol. Both issues are known The purpose of statistical inference is to Quality of Life, and Achilles heels of randomized trials in use a sample of patients to draw conclusions surgery. The low fraction of eligible patients Conclusions From the that were included severely threatens exter- This is an open-access article distributed under the COBRA Study nal validity (generalizability) of trial results. terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY- The 2-center trial by McMillan et al used a NC-ND), where it is permissible to download and To the Editor: prospective (nonrandomized) cohort design. share the work provided it is properly cited. The s the Chief Medical Officer (RS) and As long as complete series are ensured and work cannot be changed in any way or used com- A Head of Health Economics and the trial outcome is predefined, robust, and mercially without permission from the journal. e58 | www.annalsofsurgery.com 2017 Wolters Kluwer Health, Inc. All rights reserved. Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.
Annals of Surgery – Wolters Kluwer Health
Published: Mar 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.
All for just $49/month
It’s easy to organize your research with our built-in tools.
All the latest content is available, no embargo periods.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud