BACKGROUND: Persistence with lipid-lowering drug use is important in order for patients to gain full treatment benefit. The withdrawal of cerivastatin from the market may have affected persistence due to the fear of serious adverse effects. OBJECTIE: To assess failure of patients to continue lipid-lowering drug use following the withdrawal of cerivastatin. METHODS: A cohort study including 3.5 months follow-up after the withdrawal of cerivastatin in August 2001 was conducted using data from community pharmacies in The Netherlands, covering a population of approximately 600,000 subjects. Patients selected for inclusion in the index group were current users of cerivastatin on August 10, 2001 (the date that cerivastatin was withdrawn from the market). Reference patients were gender, age and pharmacy or region matched patients who were using any HMG-CoA reductase inhibitor other than cerivastatin on the same date. The main outcome measure was discontinuation of lipid-lowering drug use. To assess whether discontinuation had increased in the reference group, discontinuation rates were compared with discontinuation rates in the previous year. Data on these rates in 2000 were obtained from the population-based PHARMO record linkage system. RESULTS: A total of 31 pharmacies of the research network (response rate 86.1%) provided medication histories of 234 current users of cerivastatin and 431 matched patients using any other HMG-CoA reductase inhibitor. In addition, 352 current users of cerivastatin and 704 matched patients using any other HMG-CoA reductase inhibitor were obtained from the PHARMO database. Overall, 13.7% of subjects in the cerivastatin group (n=586) and 9.5% in the reference group (n=1135) discontinued lipid-lowering medication (adjusted odds ratio [OR] 1.44; 95% CI 1.04-2.00). The rate of discontinuation in the reference group was comparable to this rate in the previous year. Discontinuation was more prevalent in women who had been taking cerivastatin (adjusted OR 1.74; 95% CI 1.09-2.78), those receiving low doses of cerivastatin (adjusted OR 2.45; 95% CI 1.20-4.97), those who received their last cerivastatin prescription from a specialist (adjusted OR 1.92; 95% CI 1.02-3.60) and those who had recently started using cerivastatin (adjusted OR 2.80; 95% CI 0.98-7.98), although the latter was not statistically significant. CONCLUSIONS: Failure to continue lipid-lowering drug use was higher in patients using cerivastatin than in users of other HMG-CoA reductase inhibitors, especiallyin women, those using low doses of HMG-CoA reductase inhibitors and recent starters of lipid-lowering medication. The prevention of unwarranted discontinuation of drugs due to market withdrawal should be a joint task of healthcare providers, industry andregulatory bodies.