M. N. Mavros1, D. K. Kalaitzoglou2, E. N. Gatsouli2, K. P. Economopoulos2,3 1MedStar Washington Hospital Center,Surgery,Washington, DC, USA 2Society Of Junior Doctors,Surgery,Athens, ATTIKI, Greece 3Massachusetts General Hospital,Surgery,Boston, MA, USA
Introduction: The perioperative use of non-steroid anti-inflammatory drugs (NSAIDs) has increased over the past decades in an attempt to limit the use of opioids and associated complications. Recent evidence however suggests that NSAIDs may be associated with higher incidence of postoperative complications.
Methods: A systematic review and meta-analysis of the existing literature was performed using the PRISMA guidelines. Both randomized and non-randomized studies were included. Sensitivity analyses were performed based on study design and class of NSAID [non-selective vs. selective cyclooxygenase 2 (COX2) inhibitor]. Calculation of pooled odds ratios (OR) and 95% confidence intervals (CI) was performed using the DerSimonian-Laird random-effects model.
Results: A total of 498 studies were identified, of which 16 (24,126 patients) were included in the meta-analysis. Six studies (502 patients) were randomized controlled trials (RCTs) and the remaining were retrospective in design. In 13 studies (23,340 patients) non-selective NSAIDs were administered, while COX2-selective NSAIDs were used in 5 studies (1,731 patients; both classes were used in 2 studies). In 14 of 16 studies, the patients underwent colorectal surgery. About 5.1% of patients (1229/24126) developed anastomotic leak postoperatively (3.8% in RCTs); incidence of leak was significantly associated with perioperative use of NSAIDs (OR=1.83; 95%CI: 1.39–2.43, p<0.001). This effect remained significant at sensitivity analysis by study design (RCTs only: OR 3.38, 95%CI: 1.23–9.28, p= 0.02) and the same trend was observed when studies were pooled by NSAID class (non-selective: OR=1.52, 95%CI: 1.21–1.91, p<0.001; COX2-selective: OR=2.48, 95%CI: 0.999–6.15, p=0.05). Based on results from retrospective studies, there was no difference in the incidence of surgical site infections (OR=0.95, 95%CI: 0.53–1.65, p=0.85); however there was a trend towards lower mortality in patients getting NSAIDs (OR=0.58, 95%CI: 0.34–1.01, p=0.05). Sensitivity analysis attributed this trend to the subgroup of patients receiving non-selective NSAIDs (OR=0.42, 95%CI: 0.32–0.55, p<0.001).
Conclusion: Perioperative use of NSAIDs may have beneficial effects in terms of pain management and length of hospital stay, however based on our analysis this comes at the cost of increased rate of anastomotic leak. Limited data based on retrospective studies suggesting lower mortality in patients receiving non-selective NSAIDs need to be reproduced in randomized studies. Further research may focus on the risk-benefit and decision-making analysis to balance the overall (beneficial and detrimental) perioperative effects of NSAIDs in gastrointestinal surgery.
