Abstract P238: Evaluation of Low-Density Lipoprotein Cholesterol as a Confounder for Adverse Events
Several new lipid-lowering drugs may soon be available for secondary prevention of coronary heart disease (CHD). Claims data will likely be used for pharmacovigilance of these drugs. However, claims data frequently do not have laboratory results, such as low-density lipoprotein cholesterol (LDL-C). In order to determine whether LDL-C is an uncontrolled confounder in claims-based pharmacovigilance studies, we examined associations between LDL-C and incident cancer, serious hospitalized infection, dementia, and diabetes. These outcomes were chosen due to their possible associations with statin use and low LDL-C. LDL-C was measured at baseline (2003-2007) among participants in the REasons for Geographic and Racial Differences in Stroke (REGARDS) study. Analyses were restricted to 3,043 participants ≥ 65 years of age with a history of CHD or risk equivalents (stroke/carotid disease, peripheral artery disease, abdominal aortic aneurysm and diabetes) and Medicare fee-for-service coverage. Sensitivity analyses were performed (1) with LDL-C as a continuous variable, (2) among participants with a history of CHD or (3) among participants with a history of CHD or risk equivalents taking statins. Outcomes were identified through 2011 using validated claims-based algorithms. Multivariable models indicated LDL-C did not have any significant associations with any of the outcomes under study (Table). However, some hazard ratios indicated non-significant differences in the incidence of safety outcomes by LDL-C level. For example, compared to participants with LDL-C < 70 mg/dL, the hazard ratio (95% CI) for incident dementia during follow-up was 1.30 (0.85, 1.98), 1.41 (0.92, 2.18), and 1.29 (0.79, 2.10) for participants with LDL-C of 100 - 129, 130 - 159, and ≥ 160 mg/dL, respectively. Sensitivity analyses had similar results. This study suggests that LDL-C is not likely to be a major confounder in pharmacovigilance studies of lipid-lowering drugs with outcomes of cancer, serious hospitalized infection, dementia, or diabetes.
Author Disclosures: S.T. Kent: B. Research Grant; Significant; Amgen. E.B. Levitan: B. Research Grant; Significant; Amgen. H. Zhao: B. Research Grant; Significant; Amgen. A. Manthripragada: A. Employment; Significant; Amgen. M.M. Safford: None. J.R. Curtis: None. P. Muntner: B. Research Grant; Significant; Amgen.
- © 2015 by American Heart Association, Inc.