Higher Patient Numbers-Stronger Evidence?
High-level evidence on the safety and efficacy of treatment strategies is generated by randomized controlled trials (RCTs) and meta-analyses. Pancreatic surgeons around the world have expended considerable effort, performing many RCTs and meta-analyses on strategies for reducing the incidence of POPF, which include somatostatin analogues,2 stenting the pancreatic anastomosis,3 different types of pancreatic anastomosis,4 prophylactic peritoneal drainage,5 and others. However, the evidence on these different mitigation strategies remains unsatisfactory and somewhat inconsistent. One important factor that accounts for the lack of definitive evidence—and served as a rationale for the study by Ecker et al1—is that the majority of available RCTs did not analyze treatment strategies in a population at a high risk of POPF.
Ecker et al1 performed a retrospective multi-institutional study on the values of different strategies to avoid POPF in a high-risk cohort [fistula risk score (FRS) = 7 to 106] selected from 5323 patients who had undergone PD. A total of 522 (9.8%) cases met the criteria for high risk and were included in the study. Clinically relevant POPF occurred in 152 (29%) of these patients, and was associated with secondary complications, mortality, and increased use of resources for managing complications. By univariate analysis, surgeon career PD volume, institutional PD volume, pancreatogastrostomy, prophylactic octreotide administration, and transanastomotic stents were treatment-related factors associated with risk of POPF. By multivariate analysis, prophylactic octreotide administration was independently associated with an increased risk, and use of an external stent was associated with a decreased risk of POPF. On the basis of these results, Ecker et al1 defined the “optimal” mitigation regimen as the combination of omission of prophylactic octreotide and use of external stents. They then compared the subgroup of 167 patients in which this “optimal” regimen had been employed to a propensity-score-matched control group (controlled for sex, obesity, surgeon career PD volume and institutional volume), and found a markedly lower rate of clinically relevant POPF if the “optimal” regimen was used (13.2% vs 33.5%). Ecker et al1 concluded that externalized stents and omission of prophylactic octreotide provide optimal outcomes in the setting of pancreaticojejunostomy and intraperitoneal drainage.
At first glance, the data and conclusions provided by Ecker et al1 seem convincing, and suggest that the authors have successfully completed the quest for the optimal strategy to reduce the rate of POPF after PD for high-risk cases. However, we wonder whether such conclusions can be drawn from this type of study. The retrospective multi-institutional design contains several inherent weaknesses that might have considerably affected the key results and conclusions:
RCTs and meta-analyses that were mainly conducted in populations including low- to high-risk pancreatic anastomosis have concluded that prophylactic somatostatin use reduces the risk of complications (including POPF) and that evidence on the efficacy of internal and external stents is limited.2,3 Conflicting with this high-level evidence, Ecker et al1 concluded, based on their retrospective study, that prophylactic octreotide is harmful and external stents are beneficial in a high-risk setting.