Abstract
BACKGROUND: Postoperative pancreatic fistula (POPF) is a potentially fatal complication following pancreaticoduodenectomy. Early prediction and exclusion of POPF may be highly advantageous to enhance patient outcomes, and accelerate recovery. In this meta-analysis, we sought to assess the prediction of drain pancreatic amylase concentration on postoperative day 1 (DPA1) for POPF. METHODS: By searching online databases up to April 2018, all researches mentioned DPA1 for detecting POPF were analyzed. STATA 12.0 was used to analyze pooled predictive parameters. RESULTS: Seventeen studies were finally analyzed including 4676 patients in total. The pooled sensitivity and specificity of DPA1 were respectively 0.85 (95% CI: 0.71, 0.93), 0.80 (95% CI: 0.74, 0.85) to predict overall POPF, and 0.70 (95% CI: 0.53, 0.82), 0.88 (95% CI: 0.86, 0.90) to predict CR-POPF. If pretest probability was 50%, corresponding post-test (+) were respectively 81%, 86% for overall POPF and CR-POPF when DPA1 was above cutoffs, while the post-test (-) were respectively 16%, 26% when DPA1 was under cutoffs. In subgroup analysis, sensitivities of cutoff >5000 group, 1000< cutoff <5000 group, and cutoff <1000 group were respectively 0.65 (0.43-0.82), 0.82 (0.71-0.89), 0.87 (0.78-0.92); and specificities were respectively 0.88 (0.83-0.92), 0.83 (0.77-0.88), 0.71 (0.62-0.79). Positive LR was 5.5 (3.4-8.8), 4.8 (3.4-6.7), and 3.0 (2.3-4.0) respectively. Negative LR was 0.40 (0.22-0.72), 0.22 (0.13-0.37), and 0.19 (0.11-0.32) respectively. CONCLUSION: DPA1, which has good sensitivity and specificity, is useful for predicting overall POPF and CR-POPF, according to the present studies. Meanwhile, it should be cautious to apply because there is a wide range in cutoffs between different studies.