Several nonsteroidal anti-inflammatory drugs (NSAIDs) have been withdrawn from the market because of hepatic adverse drug reactions (ADRs). Moreover, some cases of liver diseases have been reported in patients taking NSAIDs (arylcarboxylic NSAIDs, piroxicam, sulindac, nimesulide, etc.). Pharmacoepidemiological studies have shown a risk of hepatic ADRs with NSAIDs used in association with other hepatotoxic drugs. In contrast, other studies performed in hospitalized patients did not found any association. The aim of this study was to assess the hepatic risk associated with the use of NSAID in the setting of primary care. The study design was a case-control study where cases and controls were all recruited among patients seen in the context of medical community care. Eighty-eight cases and 178 controls were included between January 1998 and December 2000. Cases used more drugs than controls in the 15 days before index day (2.9 +/- 2.2 vs. 1.8 +/- 1.8 different consumed drugs; P < 10(-4)). After adjustment, we found a significant association between liver injury and NSAID exposure in women [odds ratio (OR) = 6.49 (1.67-25.16)] but not in men [OR = 1.06 (0.36-3.12)]. A total of 22 cases were exposed to NSAIDs. Of them, seven patients were exposed to salicylates, five to diclofenac, four to ibuprofen, four to ketoprofen, two to niflumic acid, one to flurbiprofen and one to meloxicam (two patients were simultaneously exposed to two different NSAIDs: salicylate + niflumic acid and salicylate + diclofenac). These patients suffered from hepatocellular (53.3%), cholestatic (20%) or mixed (26.7%) injury. In 18 cases, liver enzymes returned to normal values after discontinuation of drug. No case had a fatal outcome. This study shows the existence of a significant association between liver disturbances and NSAID use in women.