Background
Drain amylase content in the days immediately after major pancreatic resection has been investigated previously as a predictor of postoperative pancreatic fistula (POPF). Its accuracy, however, has not been determined conclusively. The purpose of this study was to evaluate the accuracy of drain amylase content on the first day after major pancreatic resection in predicting the occurrence of POPF.

Methods
A literature search of the MEDLINE, Embase and Scopus® databases to 13 May 2015 was performed to identify studies evaluating the accuracy of drain amylase values on day 1 after surgery in predicting the occurrence of POPF. The area under the hierarchical summary receiver operating characteristic (ROC) curve (AUChSROC) was calculated as an index of accuracy, and pooled estimates of accuracy indices (sensitivity and specificity) were calculated at different cut-off levels. Subgroup and meta-regression analyses were performed to test the robustness of the results.

Results
Thirteen studies involving 4416 patients were included. The AUChSROC was 0·89 (95 per cent c.i. 0·86 to 0·92) for clinically significant POPF and 0·88 (0·85 to 0·90) for POPF of any grade. Pooled estimates of sensitivity and specificity were calculated for the different cut-offs: 90–100 units/l (0·96 and 0·54 respectively), 350 units/l (0·91 and 0·84) and 5000 units/l (0·59 and 0·91). Accuracy was independent of the type of operation, type of anastomosis performed and octreotide administration.

Conclusion
Evaluation of drain amylase content on the first day after surgery is highly accurate in predicting POPF following major pancreatic resection. It may allow early drain removal and institution of an enhanced recovery pathway.