Early drain removal safe in 60% of pancreaticoduodenectomy patients

Key clinical point: Most patients undergoing pancreaticoduodenectomy can safely have drains removed 1 or 2 days following the procedure, reducing complication risks associated with longer-term drain use.

Major finding: Drain amylase values below 600 U/L on postoperative day 1, seen in more than 60% of the cohort, correspond with a low risk of fistula development.

Data source: A prospective cohort study evaluating about 500 patients, all undergoing PD at a major surgical center.

Disclosures: Dr. Fong and colleagues disclosed no external funding or conflicts of interest.




The majority of patients undergoing pancreaticoduodenectomy will not develop pancreatic fistula and may safely have drains removed on the first postoperative day, a prospective cohort study has found.

Less than 1% of patients with drain amylase levels below 600 U/L on postoperative day 1 will develop pancreatic fistula (PF). This means that in this group – which represents about 60% of PD patients – early drain removal may be a safe management option.

For their research, published online in Annals of Surgery (2015 Jan. 12 [doi:10.1097/SLA.0000000000001038), Dr. Zhi Ven Fong and colleagues at Massachusetts General Hospital and Harvard Medical School, Boston, sought to find the threshold value for drain amylase that predicts fistula, with the goal of helping guide surgeons’ decisions on drain management.

Most surgeons place intraperitoneal drains during PD to control leakage in the event that anastomoses fail. However, some surgeons have abandoned drain placement after PD out of concerns that drains can introduce infection and cause other complications, with risks increasing the longer drains are in place. Drain placement followed by amylase measurement and early removal in low-risk patients “represents a middle ground between the two practices,” Dr. Fong and colleagues wrote.

The investigators evaluated results from two cohorts of consecutive patients undergoing pancreaticoduodenectomy at their surgical center, an initial training cohort (n = 126) and a validation cohort (n = 369). Closed-suction drains were used in all patients, and drain output and amylase levels were prospectively measured daily until drain removal or patient discharge.

Results from the first cohort showed that a drain amylase level of 612 U/L or higher showed the best accuracy (86%), sensitivity (93%), and specificity (79%) in predicting fistula, compared with other established variables.

In the larger validation cohort, the 140 patients with drain amylase values of 600 or higher on postoperative day 1 saw a PF rate of 31.4% (odds ratio = 52, P < .0001). Of the 229 patients with values lower than 600, a group comprising 62.1% of the cohort, fistula developed in only two cases (0.9%). An amylase value below 600 proved a stronger predictor of the absence of PF (OR = 0.0192, P < .0001) than pancreatic gland texture (OR = 0.193, P = .002) and duct diameter (OR = 0.861, P = .835).

“We believe that the debate [over] current intraperitoneal drain management after PD should not be ‘to drain or not to drain’ but rather ‘who and when can we stop draining,’ ” Dr. Fong and colleagues wrote.

“Risk of PF is less than 1% if POD [postoperative day] 1 drain amylase level is lower than 600 U/L. We propose that in this group, which comprises more than 60% of patients, intraperitoneal drains should be removed on PODs 1 and 2, and are currently validating this strategy in our practice,” they wrote.

The investigators cautioned, however, that theirs was a one-site study at an institution whose fistula rates after PD are low, compared with historic rates. “Utilizing POD 1 drain amylase of less than 600 U/L as an early stratification of patients to guide drain removal should not be assumed to be a universally safe practice,” they wrote, until data from larger multisite studies become available.

Dr. Fong and colleagues disclosed no external funding or conflicts of interest.

Next Article:

   Comments ()