PHILADELPHIA – A randomized multicenter trial that compared patients who had distal pancreatectomy with and without routine peritoneal drainage found no appreciable difference in the complication rates between the two groups, the lead investigator reported at the annual meeting of the American Surgical Association here.
The trial randomized 202 patients to the intraperitoneal drainage group and 197 to the nondrainage group. The groups were well matched in terms of patients who dropped out because of metastatic disease or other reason, as well as demographics and operative data. “They were equally matched for laparoscopic vs. open, equally matched for vascular resection, equally matched for pancreatic texture and duct size, equally matched for method of transection, equally matched for operative time and blood loss and equally matched for surgical pathology,” Dr. Van Buren said.
The primary outcome was frequency of grade 2 complications or greater, which occurred in 44% (76) of the drainage group and 42% (72) of the nondrainage group (P = .804). “Similarly, the groups were equal for grade 3 severe complications as well as median complication severity and median number of complications,” Dr. Van Buren said. The rates of complications of grade 3 or greater were 29% (51) for the drainage group and 26% (44) for the nondrainage group (P = .477). Ninety-day mortality also was similar between the two groups: None died in the drainage group and two (1%) died in the nondrainage group (P = .24).
Drilling down into types of complications, Dr. Van Buren added that rates were similar across the board. “There was no difference in clinically relevant postoperative pancreatic fistula between the drain and no-drain group,” he said. “There was no difference in intra-abdominal abscess, no difference in the rates of postoperative imaging, no difference in the rates of postoperative percutaneous drain placements, and there was no difference in readmission or reoperation.”
One outcome that was noticeably different between the two groups was intra-abdominal fluid collections, reported in 9% (15) of the drainage group and 22% (38) of the nondrainage group (P = .0004). “However,” Dr. Van Buren said, “these were asymptomatic.”
In his discussion of the presentation,, of Thomas Jefferson University, Philadelphia, said that “the scope and the rigor of this study are impressive and quite laudable” but raised a number of questions, including concerns about the two deaths in the nondrainage group in the context of two other trials: a smaller multicenter trial, coauthored by Dr. Van Buren, that compared postoperative use of drains and nondrainage in pancreaticoduodenectomy but was halted because of eight deaths in the nondrainage group ( ); and the German reported at last year’s ASA meeting that found nondrainage after pancreaticoduodenectomy to be superior to drainage in terms of reintervention and fistula rates.
Dr. Van Buren replied that the first study from Baylor – referenced by Dr. Yeo – involved “well-balanced groups, and they were equally matched and had minimal dropout throughout.” Because of that, the finding that drain placement for the pancreaticoduodenectomy “was definitive,” whereas the PANDRA trial was subject to some criticisms. The screening and randomization processes in PANDRA have been criticized because 3,200 patients were eligible for enrollment, suggesting a screening bias, Dr. Van Buren said. In addition, drains were placed in 21% of patients who were allocated to the no-drain group, suggesting surgeons deviated from the protocol in higher-risk situations, resulting in additional selection bias. This implies PANDRA was more of a selective draining trial, he said.
Dr. Van Buren and Dr. Yeo reported having no financial disclosures.
The complete manuscript of this study and its presentation at the American Surgical Association’s 137th Annual Meeting, April 2017, in Philadelphia, is to be published in Annals of Surgery pending editorial review.