Medication Discontinuation in the IMPROVE-IT Trial

BACKGROUND: Although cholesterol-lowering medications can reduce the risk of recurrent cardiovascular events, premature discontinuation limits effectiveness. Discontinuation rates have not been systematically reported for lipid-lowering trials. METHODS AND RESULTS: We evaluated medication discontinuation in IMPROVE-IT (Improved Reduction of Outcomes: Vytorin Efficacy International Trial), which evaluated placebo+ simvastatin versus ezetimibe+ simvastatin in patients hospitalized with the acute coronary syndrome and followed longitudinally postdischarge. Reasons for discontinuation were evaluated from randomization through study end (median 71.9 [ interquartile range 51.8-85.8] months). Kaplan-Meier (KM) discontinuation rates were evaluated at 30 days, 1 year, and through year 7, and compared by treatment arm and region, with Cox proportional hazards modeling used to evaluate predictors of discontinuation. Overall, 46.7% of subjects discontinued study medication (KM rate by study end 50.9% [95% CI, 50.1%-51.7%]). The risk of discontinuation was highest early in the trial but decreased with increasing time, with a terminal KM rate per 100 person-years of 8.4 (8.2-8.6) from years 1 to 7. Discontinuation was higher in the placebo+ simvastatin versus ezetimibe+ simvastatin arm (KM rate 52.0% versus 49.8%, P=0.049) and was highest in the United States (7-year KM rate 57.4%). In multivariable modeling, smoking, prior revascularization, hypertension, unstable angina, female sex, nonwhite race, and US location were associated with higher discontinuation rates. CONCLUSIONS: Although discontinuation was highest early and stabilized to 8% per year, because of prolonged follow-up, most discontinuation occurred after year 1. Adding ezetimibe to statin therapy did not increase discontinuation risk. Geographic differences and patient-level factors should be considered in trial design and analysis.