Abstract
Nonsteroidal anti-inflammatory drugs (NSAIDs) are widely used for pain and inflammation but are associated with gastrointestinal (GI) bleeding. While this risk is well established, most studies evaluate NSAIDs as a homogenous class, limiting clinical decision-making based on individual agent safety. This systematic review and meta-analysis aimed to quantify the risk of GI bleeding associated with individual NSAIDs. We searched PubMed from inception through January 2025 using MeSH and free-text terms for "gastrointestinal bleeding" and nine commonly used NSAIDs. A random effects meta-analysis was conducted to estimate the pooled odds ratios (ORs) for GI bleeding, with hazard ratios (HRs) and relative risks (RRs) treated as approximations of ORs due to the rare event nature of GI bleeding. Of 6,711 records screened, 25 studies met the inclusion criteria. Substantial heterogeneity in study design, populations, and outcome ascertainment was observed. Celecoxib was associated with the lowest risk of GI bleeding (OR 1.16, 95% CI: 0.84-1.61). Among non-selective NSAIDs, ibuprofen had the lowest significant risk (OR 2.28, 95% CI: 1.71-3.03), while ketorolac showed the highest risk (OR 20.67, 95% CI: 14.56-29.34). Other agents, such as piroxicam and meloxicam, also demonstrated significantly elevated risks. The risk of GI bleeding varies widely among individual NSAIDs. Celecoxib appears to have the lowest GI risk, though cardiovascular safety must also be considered. These findings highlight the need for personalized NSAID selection and suggest that NSAIDs should not be treated as a uniform class when assessing bleeding risk.