An examination of the effect of cytochrome P450 drug interactions of hydroxymethylglutaryl-coenzyme a reductase inhibitors on health care utilization: A Canadian population-based study

Abstract

Background: Cytochrome P450-related drug interactions can lead to adverse effects that may affect health care resource utilization. Objective: The purpose of this study was to quantify the impact of drug interactions involving hydroxymethylglutaryl-coenzyme A reductase inhibitors (statins) on health care resource utilization. Methods: Using the Manitoba Health Research database, we identified patients who had used statins between January 1, 1995, and March 31, 1998. New statin users (NSUs) were those who received a first prescription for a statin after April 30, 1995; old statin users (OSUs) were those who had a statin prescription before January 1, 1995. The number of hospitalizations, physician visits, and prescriptions, and their associated costs to the Manitoba health care system were calculated. Statin interacters were defined as users with >1 prescription for an interacting drug while receiving a statin. Interacting drugs were classified into 2 groups: group A included drugs whose levels increased as a result of the statin prescription; drugs in group B increased statin levels. The Wilcoxon rank-sum test was used to analyze differences by statin on health care resource use. Results: A total of 28,705 statin users (18, 181 NSUs and 10,524 OSUs) were identified. During the study period, 24,496 (85.3%) individuals took 1 statin, 3751 (13.1%) took 2 statins, and 458 (1.6%) took 3 to 5 statins. The most common coadministered group A interacting drugs were diclofenac (5.8%), amitriptyline (4.9%), warfarin (4.5%), and ibuprofen (1.8%). The most common group B interacting drugs were erythromycin (8.2%), omeprazole (5.5%), cimetidine (3.6%), and clarithromycin (3.5%). Statin interacters consumed significantly more health care resources than did noninteracters for both incident and prevalent analyses ( P < 0.001). In the prevalent analysis (NSUs + OSUs), pravastatin users taking interacting drugs had significantly fewer hospitalizations (mean, 1.3), fewer physician visits (mean, 24.2), and lower health care costs (mean, $5526) compared with prevalent users of lovastatin (1.7, 28.0, and $6925, respectively) and fewer physician visits than simvastatin users (25.6, P < 0.001). In the incident analysis, pravastatin users had significantly less physician visits (20.8 vs 23.5, P < 0.001) and lower health care costs ($4739 vs $6323, P < 0.001) than lovastatin users. Conclusion: Pravastatin was associated with fewer hospitalizations, physician visits, and overall health care resource utilization in prevalent users than lovastatin, possibly due to a lack of drug interaction effects.