ABSTRACT
Cyclooxygenase (COX)exists as two isoforms. Cox-1 in present under physiological conditions and COX-2 is induced by inflammatory stimuli. Previous studies in vitro have suggested that the side-effects of non-steroidal anti-inflammatory drugs (NSAIDs) correlate with their ability to inhibit COX-1, while the anti-inflammatory effects are due to their ability to inhibit COX-2 In order to strengthen this hypothesis, we examined the correlation between the inhibitory effects of eight NSAIDs (ibuprofen, aspirin, diclofenac, sulindac, naproxen, indomethacin, tolmetin and piroxicam) on the activity of COX isoforms in vitro and the range of relative risks of gastrointestinal (GI) complications that have been reported with individual NSAID in a meta-analysis. Analysis by Spearman rank correlation test demonstrated that IC50 complications (rs = 0.74, p < 0.05). A similar analysis using the IC50 of NSAIDs on the activity of COX-2 did not demonstrate any significant correlation (rs =- 0.59, p = 0.12). The ratio of COX-2/COX-1 did not appear to have any significant correlation with the risks of clinical GI side-effects (rs = 0.5, p = 0.21). Further analyses of in vitro data from other sources are encouraged to validate the usefulness of this model in predicting the GI side effects of each particular NSAID.