Reply to “Pancreatogastrostomy versus Pancreatojejunostomy Following Pancreatoduodenectomy: Comments on RECOPANC Study”
We thank Qiu and Chengyou for carefully evaluating our manuscript1 and for the comments addressed in their “Letter to the Editor.” In the following letter, we are addressing the questions raised.
The aim of this prospective randomized trial was to compare pancreatogastrostomy (PG) and pancreaticojejunostomy (PJ), regardless of specific details related to the anastomosis, as for example, suture material or specific suture technique. One rationale of this approach was to ensure generalizability of results by including all pancreatoduodenectomy patients from many high-volume centers in Germany performing techniques with several modifications (see supplement). Another rationale was that there is no unequivocal evidence so far that the type of suture material, specific suture technique, or stenting has any significant impact on fistula rate or other morbidity after pancreatoduodenectomy. Details of anastomotic techniques were different in each center, guaranteeing that each center is familiar with the performed anastomosis and not reflecting the learning curve within the study (see supplement). Any ex post-hoc analysis of a specific technical subtype would be underpowered; however, between all single centers also, there was no significant difference between PG and PJ as of pancreatic fistula formation, given the fact that various techniques were used. In summary, we feel that such subgroup analysis would neither make sense nor have enough statistical power in the scope of this study.
The second question of the study by Qiu et al addresses quality of life. We agree that monitoring quality of life preoperatively in comparison to postoperatively would have been beneficial. Unfortunately, we did not have these data from every participant in the study and therefore did not include this analysis in the study. Of note, we did not find significant differences in the comorbidities, ASA score, and other baseline parameters between both collectives.
Postpancreatectomy hemorrhage (PPH) was classified according to the ISGPS criteria.2 Even though there was a significant increase in bleeding events after PG, this was due to grade A and B PPH, as specified in the study. Previous studies from our group3 have shown that the increased bleeding rate after PG is mainly due to bleeding from the cut surface of the pancreas in the stomach. In his study, however, we did not ask for the location of PPH. As per definition, there was no difference in grade C PPH in this study. We do not have information in which percentage of patients endoscopic intervention in PG was superior to PJ due to accessibility, as this was not the focus of our primary and secondary endpoint.
In contrast to the study by Qiu et al, we think that further studies might not help to clarify this question in more detail. In the supplement of our study, we have added a meta-analysis that includes a total of 1629 patients from level I evidence-based studies. This evidence makes the technical question of superiority of PG or PJ to the best answered in surgical methodology. It seems that both anastomoses have comparable fistula rates. In our study this was also independent of the quality of the pancreatic remnant, soft or hard. It also seems that the PG might be easier to perform and, especially for the less experienced pancreatic surgeon, might have some benefits. Cumulative evidence for comparison of PG and PJ is stronger than for any other question in pancreatic surgery.