Mission Statement: To promote the art and science of surgery by publishing relevant peer-reviewed research to assist the surgeon in optimizing patient care. JAMA Surgery will also serve as a forum for the discussion of issues pertinent to surgery, such as the education and training of the surgical workforce, quality improvement, and the ethics and economics of health care delivery.

In the Original Investigation titled “Association of Postoperative Infection With Risk of Long-term Infection and Mortality,”1 published online November 6, 2019, there was an error in a percentage in Figure 1. At the split, the percentage in the right box should have been 96.4% instead of 46.4% so the sentence reads, “635 671 Patients (96.4%) had no infection within 30 d postoperatively.” This article was corrected online.

To help reduce surgical site infections, a large multicenter health care system implemented a policy that required personnel who had scrubbed to use disposable perioperative jackets. Stapleton and colleagues retrospectively reviewed the cases of 60 009 patients who underwent these clean procedures and compared the rate of surgical site infections before and after policy implementation. Despite spending more than $1.7 million on disposable jackets, there was no significant reduction in surgical site infections, suggesting that perioperative attire has no association with such infections. Invited Commentary Continuing Medical Education

In the Surgical Innovations article, “Primary Fascial Closure During Minimally Invasive Ventral Hernia Repair,”1 published online December 26, 2019, an error appeared in the text. The 95% CI of a relative risk for a study by Ahonens et al was misreported as 2.8 to 13.3. The correct 95% CI is 0.08 to 0.36. The article has been corrected online.

In Reply Thank you for your interest in our article, “Effect of Low-Dose Supplementation of Arginine Vasopressin on Need for Blood Product Transfusions in Patients With Trauma and Hemorrhagic Shock.”1 We believe Gauss et al have misread our statistical approach and are incorrectly interpreting the absolute standardized differences (ASD) as P values. There is a growing momentum to use P value alternatives, such as ASDs, in clinical research.2,3 Absolute standardized differences assess the magnitude of differences between groups (specifically, the absolute value of the difference in means, mean ranks, or proportions divided by the pooled standard deviation). We used ASD rather than standard significance tests because using multiple significant tests to evaluate baseline variables can be misleading. Each test carries a 5% probability of type 1 error; with multiple statistical tests, the type I error rate becomes exaggerated, resulting in an inflated probability of a significant difference between groups based on random chance alone. The use of ASD mitigates the risk of amplifying type 1 errors. Groups were considered imbalanced for any variable that had absolute standardized differences greater than 0.392. All baseline variables had an ASD less than 0.392, suggesting that groups were not imbalanced for these variables.

To the Editor We read with interest the article of Sims et al,1 who reported that low-dose supplementation of arginine vasopressin reduced blood transfusion products and mortality in patients with traumatic hemorrhagic shock in their article in JAMA Surgery.1 Regarding the application of low-dose vasopressin for trauma patients in clinical practice, we should discuss some points referring to previous clinical studies.

To the Editor The work of Sims et al1 on low-dose supplementation of arginine vasopressin in patients with trauma and hemorrhagic shock has gained our greatest interest. We applaud the authors for their work, which we consider an important contribution to the trauma community and ongoing controversy on the use of vasopressors in hemorrhagic shock after trauma.

Despite the remarkable progress in transplant over 65 years, organ transportation has not been innovated. The current system is challenged by expense, time, and safety. Unmanned aircraft systems (UAS) capable of on-demand life-urgent delivery of organs, blood, and medications may represent a pathway toward saving lives in transplant and beyond.1 In the last 5 years, dramatic advances in UAS technologies have been realized. Unmanned aircraft system technologies allow for ranges of more than 1000 miles, speeds faster than 150 miles per hour, and vertical takeoff and landing. Further, artificial intelligence technologies have ushered in a new era of autonomous flight, allowing for pilotless travel. Our group recently showed that UAS technologies can be used to move human organs.1

The uptake of bariatric surgery remains low in the United States, with fewer than 1% of eligible patients undergoing this life-saving treatment.1 In addition to the obesity bias that delays referral for surgery, some continue to fear the short- and long-term risks of surgery. To date, most long-term outcome data have focused on Roux-en-Y gastric bypass (RYGB),2 but sleeve gastrectomy (SG) is now the most common bariatric operation in the United States.3 A better understanding of its long-term outcomes may help allay some of these fears.

In an attempt to minimize missed injury rates, potentially decrease mortality, and enhance rapid patient disposition, standard-dose whole-body computed tomographic (WBCT) imaging has become ubiquitous at trauma centers for the hemodynamically stable patient admitted with trauma.1,2 The radiation dose from WBCT ranges from 10 to 20 mGy, which results in an approximately 0.08% estimated lifetime cancer mortality for 45-year-old persons.3 Risk of mortality due to missed injury is therefore higher than the risk of future radiation-induced cancer.

In Reply The secondary analysis of European Study Group for Pancreatic Cancer 4 (ESPAC-4) demonstrated that there were no significant differences between the time to recurrence and subsequent and overall survival between local and distant recurrence.1 Patients with metastases to the lungs had a much longer survival compared with those with local recurrence or metastases to other sites such as the liver. Thus, it is clear that pancreatic cancer appears to behave as a systemic disease requiring effective systemic therapy after resection.

To the Editor In the article by Jones et al,1 the authors conclude that pancreatic ductal adenocarcinoma behaves as a systemic disease, requiring more effective systemic therapies. While we agree that systemic therapy is critical, we feel that several points need to be made about the risk of local recurrence and the important role of local therapy. First, the observed rate of local-only first failure in approximately 50% of patients warrants considerable thought because this rate is exceedingly high compared with other solid tumors in which extirpative surgery plays a major role. Given that the ability of an isolated local recurrence to potentially seed distant sites is a tangible risk, future efforts to further improve local control are certainly warranted. Second, local recurrence occurred at a median of 2.3 months later than the median time to distant recurrence, but the median survival rates associated with local-only and distant-only first recurrences were very similar. Furthermore, there were no differences in median survival measured from the time of recurrence between patients with local vs distant recurrences. Perhaps even more thought-provoking is that gemcitabine plus capecitabine led to fewer local-only first recurrences, while distant-only recurrences were higher. Finally, multivariate analysis found that positive resection margin and local invasion were significantly associated with poor survival and that local recurrence also trended to being significantly associated with worse survival (P = .054). Taken together, these findings argue that having a local-only first failure phenotype may be at least as poor a prognostic sign as having a metastatic-only first recurrence phenotype, and that lower rates of local recurrence observed in the experimental arm could be driving the observed survival improvement.

Given the technological utopianism that shapes medical headlines, surgeons (and their patients) might be excused for embracing the fiction that with sufficient National Institutes of Health funding, we might all get out of life alive.2 Yet surgical practice is rooted in the reality that despite our best efforts to preserve function and extend life, death comes to us all. And when it does, patient priorities often shift such that what matters most can only be measured through subjective perceptions of support, communication, and the care of one human being for another. Such patient-reported outcomes have been slow to arrive in surgery, but the growing focus on value-based care will only increase their relevance.

In Reply We appreciate the interest in our article and thank JAMA Surgery for the opportunity to reply to the letters to the editor.1 Shi et al advised that combining patients without postoperative pancreatic fistula (POPF) and with biochemical leakage may have biased the analysis regarding long-term survival in patients receiving neoadjuvant therapy (NAT). While biochemical leakage may reflect a more preserved gland function, which has been associated with better survival in upfront resected pancreatic cancer,2 a preliminary analysis had shown no significant differences between biochemical leakage and no pancreatic fistula in patients with NAT.

To the Editor We read with interest the study of Hank et al,1 which proposes that there is a decreased risk of developing clinically relevant pancreatic fistula (CR-POPF) when neoadjuvant chemotherapy (NAT) therapy is applied for pancreatic adenocarcinoma. They suggest induction of fibrosis of the pancreatic gland as a proposed mechanism. While the data appear intriguing, there are some methodologic issues that raise questions.

To the Editor We have read with great interest the article by Hank et al.1 The authors indicated that neoadjuvant therapy (NAT) could reduce the rate of clinically relevant postoperative pancreatic fistula (CR-POPF). In addition, they demonstrated that CR-POPF is associated with a significant reduction in long-term survival for patients with pancreatic cancer with NAT.

Mission Statement: To promote the art and science of surgery by publishing relevant peer-reviewed research to assist the surgeon in optimizing patient care. JAMA Surgery will also serve as a forum for the discussion of issues pertinent to surgery, such as the education and training of the surgical workforce, quality improvement, and the ethics and economics of health care delivery.

In the Original Investigation titled “Association Between Liver Transplant Wait-list Mortality and Frailty Based on Body Mass Index”1 that was published online September 11, 2019, there was an error in the byline. Dr Rahimi’s middle initial was missing. His name should read Robert S. Rahimi, MD. This article was corrected online.

In Reply We thank you all for taking the time to critically evaluate our work.1 With regards to Kawada’s statement, we agree that generally speaking, one would expect that the odds ratio (OR) of the number of annual colonoscopies for proximal sessile serrated polyp detection rate (pSSPDR) to be statistically significant and decreasing with increasing volume. This seems discordant with our study’s findings and may be driven by the fact that pSSPDR was only recently accorded importance as a quality indicator. Therefore, those who have trained more recently have been trained more rigorously to identify pSSPDR. However, this does not take away from the realization that with increased procedure volume, one becomes better at identifying them. Thank you for the clarification regarding the Crockett et al study2 evaluating sessile serrated polyp detection not being derived solely from proximal sessile serrated polyps. While most sessile serrated polyps are proximal, we do agree that further studies are needed with focus on pSSPDR to be an adequate comparator with our study.

To the Editor Sarvepalli et al1 found that during colorectal cancer screening, most of the differences in adenoma detection rate between endoscopists disappeared when adjusting for patient and colonoscopy characteristics. Adenoma detection rate should have been the quality metric for programs of colorectal cancer screening long ago.2 This is not yet the case, despite colonoscopy having been endorsed in the United States as the preferred strategy for colorectal cancer screening as early as 2000. One team even published arbitrary and very different distributions of values from one report to another one.3 The steering committee of the World Endoscopy Organization just issued a consensus statement regarding quality of colonoscopy but failed to reach a consensus for adenoma detection rate.4

To the Editor In their carefully conducted and well-reported study, Sarvepalli et al1 conclude that prior reports of lower polyp detection rates among nongastroenterologists were flawed by residual confounding. However, their study may be flawed by the opposite error: overadjusting for factors that operate as part of the causal pathway linking endoscopist characteristics with polyp detection. Specifically, eAppendix 1 in the Supplement lists preparation adequacy (which could be a function of quality of patient education delivered by different clinician types), cecal intubation, and withdrawal time as control variables. Prior studies have noted associations between each of these factors and polyp detection,2-4 and it is not unreasonable to assume that any or all would be associated with endoscopist specialty. It would be interesting to examine the results with these 3 variables omitted.

To the Editor I read with great interest the article by Sarvepalli et al1 to assess the association between endoscopist characteristics and detection of positive clinical findings by colonoscopy with appropriate adjustments. The authors adopted a multilevel mixed-effects logistic regression model for the analysis, and 7 endoscopist characteristic variables were used as independent variables to determine the association with adenoma detection rates and proximal sessile serrated polyp detection rates (pSSPDRs). Adenoma detection rates were not significantly associated with endoscopist characteristics. In contrast, adjusted odds ratios of years in practice per increment of 10 years and the number of annual colonoscopies performed per 50 colonoscopies per year for pSSPDR were 0.86 (95% CI, 0.83-0.89) and 1.05 (95% CI, 1.01-1.09), respectively. The authors concluded that there was a need for additional training for increasing pSSPDRs. I have some queries regarding their study.

In Reply We appreciate the insights from Lipof et al about prosthetic joint infection prevention. The authors express concern that our findings may not extrapolate well to general orthopedic patients undergoing total joint replacement (TJR). Fortunately, many previous studies, including randomized clinical trials, similarly found that extended prophylaxis regimens do not reduce surgical site infections (SSI).1 The TJRs were covered by the Surgical Care Improvement Project, which included early discontinuation of antimicrobials as a core measure; high rates of compliance with Surgical Care Improvement Project infection metrics are associated with SSI reductions, providing additional support to the notion that longer durations of prophylaxis do not improve infectious outcomes.2 In our TJR cohort (n = 38 675), risk of SSI was increased in patients who received extended regimens, rising from 1.3% after less than 24 hours to 2.7% after 48 to 72 hours.3 Because of the direction of this trend, it is highly unlikely that these findings would reverse if more or lower-risk patients were included.

To the Editor We commend the authors of “Association of Duration and Type of Surgical Prophylaxis With Antimicrobial-Associated Adverse Events” on their investigation efforts into an important topic of discussion.1 They retrospectively reviewed the Veterans Affairs Surgical Quality Improvement Project Database to assess postoperative 30-day surgical site infections (SSIs), 7-day incidence of acute kidney injury (AKI), and development of Clostridium difficile infection at 90 days after cardiac, vascular, colorectal, and orthopedic surgery. They concluded that longer duration of prophylactic antibiotics did not lead to decreased rate of SSI but rather increased rates of AKI and C difficile infection.

In Reply We thank JAMA Surgery for the opportunity to reply to the comments raised in the letters to the editor regarding our article.1 The use of resuscitative endovascular balloon occlusion of the aorta (REBOA) has gained popularity since it has become more readily available.2 However, there is limited high-grade evidence to guide its use and a considerable complication risk should this approach be used without appropriate indications. As with any developing technology, evidence-based research should be used to determine its indications, benefit, and hazards. Currently, available data are based on numerous small single-center studies of case reports or case series. Our study highlights the results of a nationwide sample that explored the outcomes of REBOA in patients with severe trauma injuries. In our propensity score–matched cohort, after adjusting for the spectrum of measurable confounding variables available in the American College of Surgeons Trauma Quality Improvement Program (ACS-TQIP), we found that REBOA is associated with a significant increase in mortality and complications. The ACS-TQIP is the largest high-quality nationally representative trauma registry that undergoes rigorous statistical adjustments.3 Ultimately, the data speak for themselves and are consistent with the findings from the Japan Trauma Data Bank. The relative infrequency of REBOA use prohibits retrospective analyses between centers to account for major differences in volume. However, it is still worthwhile to report the currently available data and to contemplate their merits.

To the Editor We read with great interest the article by Joseph et al.1 This is a 420-patient case-control study with propensity score matching that examines the value of resuscitative endovascular balloon occlusion of the aorta (REBOA). Patients undergoing REBOA were found to have worse outcomes than matched controls. However, the use of causal language surrounding the findings of the article—that REBOA use increases mortality risk—is unsupported. We posit that at least 2 biases regarding hospital confounding and time to surgery and selection bias for survivors draw this conclusion into question.

To the Editor We read the article by Joseph et al1 on balloon aortic occlusion with interest. In this study, the authors used the Trauma Quality Improvement Project to examine outcomes of primarily patients with blunt trauma who underwent resuscitative endovascular balloon occlusion of the aorta (REBOA) compared with a propensity score–matched group of patients who did not receive REBOA. The study reported higher mortality and a greater incidence of acute kidney injury and lower extremity amputation in the REBOA group. We applaud the authors for attempting to analyze REBOA outcomes, but fundamental limitations of the propensity score–matching approach applied in this study warrant additional consideration.

To the Editor As members of the trauma community who use resuscitative endovascular balloon occlusion of the aorta (REBOA), we read with great interest the article by Joseph et al.1 The authors have executed a statistically complex and important registry-based study. However, we are concerned with identifying this research design as a case-control study.

To the Editor Joseph et al1 conducted a nationwide analysis of resuscitative endovascular balloon occlusion of the aorta (REBOA) using the 2015 to 2016 American College of Surgeons Trauma Quality Improvement Program data set. The authors reported higher rates of 24-hour mortality and complications, including amputation and acute kidney injury in patients undergoing REBOA than those who were not.1 These results were consistent with those from the Japan Trauma Data Bank.2,3 The potential limitations of the article included the unavailability of data on the duration of occlusion, occluded zone of the aorta, and the use of partial REBOA; moreover, the use of a new device compatible with a 7-Fr sheath should have limited the risk of leg ischemia.

In Reply This letter is in response to the letters by Batista, Tai et al, and Kanemoto et al. We thank the editors for giving us the opportunity to rebut the issues raised about our study.1 We are grateful for Batista’s remarks because they reflect a common misinterpretation of postoperative complications. For the Japanese CCOG 1102 randomized clinical trial,2 the primary end point is disease-free survival, and secondary end points are overall survival, peritoneal recurrence-free survival, and incidence of adverse events. Thus, detailed records and observations may be lacking in “postoperative complications” such as abdominal pain. Moreover, in the trial including 27 patients by Ronellenfitsch et al,3 only half (14 cases) were pT3/4. Because peritoneal metastasis is caused by direct cancer cell dissemination from serosa-invasive tumors, the low number of patients with free peritoneal tumor cells was detected and the trial was closed early. As the author himself describes, “our study population is too small and long-term follow-up data are presently unavailable, so no sound conclusions can be drawn to that regard.”3 The study of Batista is similar, where only 8 patients were allocated to extensive intraoperative peritoneal lavage (EIPL) protocol. In our study, all surgeons have enough experience for D2 gastrectomy (>100 procedures per year) to ensure the quality of surgery. Therefore, few patients had postoperative complications, including gastrointestinal leakage. Indeed, in the CCOG 1102 study, only 2% cases had gastrointestinal leakages in non-EIPL and EIPL group, respectively.

To the Editor Guo et al1 found that extensive intraoperative peritoneal lavage (EIPL) can be the prophylactic procedure of postoperative complications after D2 resection of locally advanced gastric cancer. Postoperative abdominal pain was observed more often in the surgery alone group (48 of 271 patients [17.7%]) compared with the surgery plus EIPL group (30 of 279 patients [10.8%]) (difference, 7.0%; 95% CI, 0.8%-13.1%; P = .02). It also showed significant difference in the postoperative complication rate between the surgery alone group (46 of 271 patients [17.0%]) and the surgery plus EIPL group (31 of 279 patients [11.1%]) (difference, 5.9%; 95% CI, 0.1%-11.6%; P = .04). Although surgery plus EIPL could be a new standard strategy of advanced gastric cancer owing to the reduction of postoperative complications safely and simply at a very low cost, we have some concerns about cytologic analysis.

To the Editor We read with great interest the article by Guo et al,1 on combining surgery with extensive intraoperative peritoneal lavage (EIPL) for patients with advanced gastric cancer. This study is important because peritoneal recurrence carries a very poor prognosis with no cure. We are currently conducting a randomized clinical trial on this topic with a similar design in which the primary end point is 3-year overall survival.2 In the Guo et al1 article, the authors reported the short-term outcomes of their trial and found that the EIPL group had less postoperative complications, ileus, and wound pain; hence, they concluded that EIPL increases the safety of surgery. However, because the trial was powered to detect an absolute difference in 3-year overall survival of 11%, which is their primary end point, we wondered whether their observed findings might be owing to chance. We estimated that based on a 2-sided test of 5%, a sample size of between 800 to 1100 would be required to achieve an 80% power to detect the observed differences in the short-term outcomes of mortality, wound pain, and ileus as reported in the paper.3 Further, based on the reported sample size of 550, the trial has less than 65% power to detect significant differences at the 5% level for the outcomes concerned. Besides, several hypotheses were evaluated in this article without accounting for multiple testing.4 Had corrections been made for multiple testing, none of these findings would be significant. In addition, we suggest the authors report the American Society of Anesthesiologists classification status, the analgesics protocol, and indicate how ileus and their postoperative complications were defined in the article. We are of the opinion that this information has significant bearing on the interpretation of their conclusions.

To the Editor The article by Guo et al1 presents initial findings for the SEIPLUS study, a randomized clinical trial exploring the value of extensive intraoperative peritoneal lavage (EIPL) for treatment of locally advanced gastric cancer. As highlighted by Levine2 in an accompanying Invited Commentary, the accrual for this trial occurred during a remarkably short period, and EIPL appears to be a simple, safe, and low-cost method to decrease the peritoneal metastases rates.

Limited donor lung availability and prolonged wait times remain substantial barriers to lung transplant. According to the Scientific Registry of Transplant Recipients, 11% of patients were removed from the wait-list in 2017 because of death or clinical deterioration and 40% of patients waited longer than 1 year before transplant.1 Ex vivo lung perfusion (EVLP) provides the opportunity to expand the donor pool, but despite encouraging early posttransplant outcomes,2,3 the long-term effect of EVLP remains unclear. In this issue of JAMA Surgery, Divithotawela et al4 present 10-year outcomes of the Toronto Lung Transplant Group’s experience with normothermic EVLP for lungs from extended-criteria donors and donation after cardiac death. They report that despite being obtained from potentially higher-risk donors, the EVLP donor lungs demonstrated equivalent rates of chronic lung allograft dysfunction and allograft survival at 10-year follow-up. Additionally, at their center, EVLP was associated with a significant increase in the number of transplants during the study period. While there were inherent limitations, such as the retrospective nature of the study, debatable higher-risk status of EVLP lungs, nonrandom assignment of EVLP, and a lack of confounder-adjusted comparison, this study suggests that satisfactory long-term outcomes can be accomplished with EVLP, a valuable contribution. Notably, the system used in this study differs from the commercially available US Food and Drug Administration–approved system in its atrial drainage (closed vs open), and the data presented in this study may not apply to that system.5

Over the past several years, increasing attention has been given to the issue of workforce sex disparities, both within medicine at large and surgery specifically. Inequalities exist between male and female surgeons in academic advancement, recognition of achievements, and compensation.1-3 As a profession, surgeons need to move beyond awareness of the problem of sex disparities and toward a deeper understanding of causes.

Venous thromboembolism (VTE) remains a significant cause of postoperative morbidity, and inpatient postoperative VTE prophylaxis remains standard of care.1 For select procedures, there is a well-recognized risk of VTE that extends beyond hospital discharge,2 and numerous guidelines support extended prophylaxis for 4-6 weeks following discharge for high-risk surgical subgroups.1,3

Since Cowley et al1 introduced the concept of the golden hour of shock in the mid-1970s, there has been an ongoing discussion about what constitutes an optimized emergency medical system (EMS). Physician-staffed EMS are continuously undergoing critique for missing this golden hour because of additional actions or interventions performed on scene, which supposedly prolong prehospital rescue times.2 Vice versa: paramedic-staffed EMS that follow a so-called scoop-and-run principle are credited with producing shorter prehospital rescue times.3 With their contribution in this issue of JAMA Surgery, Gauss et al4 add more fuel to this vivid discussion. Using retrospective registry data from 2 French trauma cohorts and a range of statistical models, the authors4 assessed the association between increasing prehospital times and increasing in-hospital all-cause mortality in trauma patients treated within 2 physician-staffed EMS in central Europe; the odds of death increased by 9% for each 10-minute increase in prehospital time and, after adjustment, by 4% for each 10-minute increase.4

With great interest, we read Li and colleagues’ article1 in this issue of JAMA Surgery. The authors sought to compare the short-term outcomes of laparoscopic distal gastrectomy with those of open distal gastrectomy for the treatment of patients with locally advanced (cT2-4aN+M0) distal gastric cancer. In this study, an open-label, single center, phase 2 randomized clinical trial with a noninferiority design, the authors assessed outcomes at 3 years. All patients received neoadjuvant therapy of oxaliplatin and capecitabine before their operation. Li and colleagues1 reported significantly lower postoperative complication rates, less postoperative pain, and better adherence to and completion of adjuvant therapy among trial participants.