Introduction. Non-steroidal anti-inflamatory drugs (NSAIDs) are one of the most common causes of adverse drug reactions, notably dyspepsia and bleeding. In this study NSAIDs utilization pattern was assesed during the first six months of 2000, considering its gastrointestinal (GI) toxicity. Methods. NSAIDs utilization was measured as a number of defined daily doses (DDDs) per 1000 inhab/day. To determine the prescribing quality, NSAIDs utilization accounting for 90% of the use (DU90% segment) was determined. Within the DU90% segment, the proportion of «high riskfl (ketoprofen, piroxicam), «low riskfl (diclofenac, ibuprofen) and «intermediate riskfl (or no documentation) drugs regarding GI toxicity was evaluated. Results. In Rijeka, 14 DDD/1000 inhab/day of NSAIDs were prescribed. The most often prescribed NSAID was diclofenac with 55%. Three NSAIDs were found in the DU90% segment including diclofenac and ibuprofen. Within the DU 90% segment there was 21% of «high riskfl classified NSAIDs prescribed. Concusion. This study suggests that evidence-based medicine was not the leading impact factor in prescribing NSAIDs in Rijeka. Since no important diferences regarding efficacy exist among NSAIDs, the ultimate choice should depend on their relative toxicity and cost. Resuts of this study could be used as a basis for preparing prescription guidelines.