Patterns of lipid lowering drug use after the withdrawal of cerivastatin

Abstract

In August 2001, the lipid lowering drug cerivastatin was withdrawn from the market after its use was associated with the deaths of 52 patients due to rhabdomyolysis. Our aim was to assess the patterns of lipid lowering drug use after the withdrawal of cerivastatin. A total of 36 Dutch community pharmacies were invited to extract medication histories from all patients who were prevalent users of cerivastatin (index group) on August 10, 2001 (index date), and from two age and gender matched controls. The controls had to be a prevalent user of any other statin at the index date. Medication histories were complete for a follow up period of 3.5 months (November 26, 2001: last day of the study). Logistic regression models were used to assess the association between patient and medication characteristics and discontinuation of lipid lowering drug use, defined as no prescription refilled for any lipid lowering drug after the index date and before end of the study. Of the 36 community pharmacies, 31 participated in the study (response rate 86.1%). Medication histories were available for 234 index patients and 431 matched controls. On the last day of the study, 37 patients in the index group (15.8%) and 41 controls (9.5%) had not refilled a prescription for any lipid lowering drug, and were therefore assumed to have discontinued lipid lowering drug use (OR = 1.8; 95% CI 1.1, 3.0). In men, the risk of discontinuation was similar for both users of cerivastatin and users of any other statin (OR = 0.9; 95% CI 0.5, 1.9), whereas in women the risk of discontinuation was higher for those using cerivastatin (OR = 3.4; 95% CI 1.7, 6.8). Stratification by age showed that discontinuation was more pronounced in those younger than 50 years of age (OR = 3.8; 95% CI 0.7, 22.4) than in older age categories (ORs ranging from 1.6 to 1.8, none statistically significant). Stratification by prescribed daily dose at the index date (expressed as defined daily doses (DDD)) showed that discontinuation was also more prevalent in those with the lowest doses of statins, < 1.00 DDDs day−1 (OR = 5.3; 95% CI 1.8, 15.6), compared with 1.00–1.99 DDDs day−1 (OR = 1.1; 95% CI 0.5, 2.5) and 2.00 DDDs day−1 (OR = 1.8; 95% CI 0.8, 3.7). In the cerivastatin group, 130 patients (66.0%) filled a new prescription for a lipid lowering drug before the theoretical end date of the prevalent prescription at the index date compared with 139 (35.6%) controls (OR = 3.5; 95% CI 2.5,5.1). Most of the patients who were on cerivastatin therapy switched to other lipid lowering drugs before they ran out of tablets. However, discontinuation of lipid lowering drug use was more prevalent in these patients than in users of other statins, especially in women, younger patients and those with low doses of statin therapy.