Postoperative nonsteroidal anti-inflammatory drugs and risk of anastomotic leak: meta-analysis of clinical and experimental studies.

BACKGROUND: Enhanced recovery programs following colorectal resection recommend the use of nonsteroidal anti-inflammatory drugs (NSAIDs) as part of multimodal analgesia. The present study aimed to assess whether postoperative NSAID use increased the risk of anastomotic leak.
METHODS: A systematic review of published literature was performed for studies comparing anastomotic leak following NSAID administration versus control. Meta-analysis was conducted for studies in human patients and experimental animal models. The primary endpoint was anastomotic leak.
RESULTS: The final analysis included 8 studies in humans and 12 experimental animal studies. Use of NSAIDs was significantly associated with anastomotic leak in humans (8 studies, 4,464 patients, odds ratio [OR] 2.14; p < 0.001). This effect was seen with nonselective NSAIDs (6 studies, 3,074 patients, OR 2.37; p < 0.001), but not with selective NSAIDs (4 studies, 1,223 patients, OR 2.32; p = 0.170). There was strong evidence of selection bias from all clinical studies, with additional inconsistent definitions and outcomes assessment. From experimental animal models, anastomotic leak was more likely with NSAID use (ten studies, 575 animals, OR 9.51; p < 0.001). Bursting pressures at day 7 were significantly lower in NSAID versus controls (7 studies, 168 animals, weighted mean difference -35.7 mmHg; p < 0.001).
CONCLUSIONS: Emerging data strongly suggest that postoperative NSAIDs are linked to anastomotic leak, although most studies are flawed and may be describing pre-existing selection bias. However, when combined with experimental data, these increasing concerns suggest caution is needed when prescribing NSAIDs to patients with pre-existing risk factors for leak, until more definitive evidence emerges.