BACKGROUND AND OBJECTIVENonsteroidal anti-inflammatory drugs (NSAIDs) are prescribed widely in Italy. They include nonspecific NSAIDs (NS-NSAIDs) and the newly marketed cyclo-oxygenase (COX)-2 specific inhibitors (COXIBs) celecoxib and rofecoxib. The objective of this study was to describe the prescribing patterns for NS-NSAIDs and COXIBs in a local Italian area, analysing an administrative database.PATIENTS AND METHODSWe extracted from the database information on subjects who had received at least one reimbursed prescription of an NSAID during the period between 1 January 2001 and 31 December 2001, including age, sex, patient identification code, Anatomical Therapeutic Chemical (ATC) classification system code, strength, formulation, number of packs prescribed, prescription date, and prescription of gastroprotective agents (GPAs) on the same day as the prescription of the NSAID. On the basis of the type of NSAID received, we divided the patients into five cohorts: oral NS-NSAIDs only during the observed year, injectable NS-NSAIDs only, celecoxib only, rofecoxib only, and a combination. For descriptive purposes, we defined three age groups: <40 years, 40-64 years, and >64 years. The duration of exposure to NSAID therapy was calculated using the most commonly prescribed dose for the different drugs. Subjects receiving >/=30 doses per year were defined as "regular users". Analyses included mean age, mean duration of exposure, percentage of regular users, and percentage of GPAs co-prescribed in the different cohorts.RESULTSNSAIDs were prescribed to 62 059 subjects from a resident population of 365 321 inhabitants; 43.8% received oral NS-NSAIDs only, 22.6% injectable NS-NSAIDs only, 7.2% celecoxib only, 5.2% rofecoxib only, and 22% different regimens of different types of NSAIDs. The mean duration of treatment increased with age in all cohorts; the mean age was 56 years in the NS-NSAID cohort, 61 years in the celecoxib cohort, and 62 years in the rofecoxib cohort (p = 0.01, COXIBs vs NS-NSAIDs). The mean duration of therapy was 11.4 days/year for injectable NS-NSAIDs, 43.8 days/year for rofecoxib, 50.5 days/year for oral NS-NSAIDs, and 53.7 days/year for celecoxib. Fifty-four percent of subjects in the oral NS-NSAID cohort were regular users versus 64% in the rofecoxib and 70% in the celecoxib groups (p = 0.001, COXIBs vs NS-NSAIDs). Co-prescription with GPAs was 9.5% for NS-NSAIDs, 8.4% for rofecoxib, and 7.7% for celecoxib.CONCLUSIONSAnalysis of an administrative database in Italy showed a trend suggesting that COXIBs are prescribed to an older population and for a longer period of time than NS-NSAIDs, and that their use is less frequently associated with GPAs.