BACKGROUND AND AIMS : Lipid lowering treatments (LLTs) are typically approved based on LDL-C lowering effects observed in randomized controlled trials (RCTs). In economic evaluations, effectiveness in preventing cardiovascular events is typically predicted using the relationship between LDL-C lowering and cardiovascular risk. Several studies have demonstrated that treatments which reduce LDL-C levels also reduce cardiovascular events, and this relationship is robust across diverse biological mechanisms of lipid lowering. We aimed to explore the importance of baseline LDL-C in the relationship between LDL-C reduction and cardiovascular risk. CONCLUSIONS: The results support the well documented association between LDL-C and cardiovascular risk reduction. Furthermore, models accounting for baseline LDL-C as well as the LDL-C reduction on treatment may provide more accurate predictions of cardiovascular risk than models based on LDL-C reduction alone. Limitations of this work include the use of aggregate trial data rather than patient-level data (which were not available to us from source studies).
METHODS: Thirty-six RCTs were identified from a systematic literature review supplemented with targeted searches. Random effects meta-regression was performed with and without the inclusion of baseline LDL-C as a covariate. Models were compared using the Akaike and Bayesian Information Criteria (AIC and BIC). Correlation (R-squared) and statistical heterogeneity (Higgins I-squared) were assessed.
RESULTS: For major cardiovascular events, myocardial infarction, cardiovascular mortality, and revascularization endpoints, AIC, BIC, and I-squared were consistently lower, and R-squared was consistently higher, for models including baseline LDL-C as a covariate than those without. P-values for baseline LDL-C were highly significant for these endpoints (P≤0.003).
