Model Accurately Predicts Benefit of Lipid-Lowering Therapies

Irfan Khan, PhD

A time-dependent model accurately predicted treatment benefit of lipid-lowering therapies.

The model was able to facilitate decision making and scenario analyses with a given lipid-lowering therapy strategy in various patient populations and time frames.

Irfan Khan, PhD, and investigators developed a time-dependent model from published randomized controlled trials summarizing the relationship between low-density lipoprotein cholesterol lowering and cardiovascular risk reduction. The team aimed to apply the model to investigate the effect of treatment scenarios over time.

Khan and the investigative team searched literature using references from the American Heart Association/American College of Cardiology guidelines, the Cholesterol Treatment Trialists’ meta-analysis, and additional analysis from other investigators. The sources used rigorous criteria in randomized controlled trial selection. The search was augmented using PubMed and keywords.

Articles found were manually reviewed and evaluated based on inclusion criteria. Exclusion criteria were then applied, including open-label design, reported data not suitable for model estimation, trials involving special populations, and trials with bococizumab.

The team summarized the model-predicted and trial-reported hazard ratios for individual cardiovascular end points, including nonfatal myocardial infarction, ischemic stroke, coronary heart disease death, unstable angina requiring hospitalization, and coronary revascularization. They also summarized the hazard ratios for a three-part composite end point which represented nonfatal myocardial infarction, ischemic stroke, and coronary heart disease.

Khan and the team’s final model was applied to investigate scenarios representing different at-risk populations, treatment strategies, and treatment durations. The populations had recent acute coronary syndrome, stable atherosclerotic cardiovascular disease, diabetes mellitus primary prevention, and primary prevention.

Of the literature searched, 22 randomized controlled trials met selection criteria. The trials included primary and secondary prevention populations, follow-ups of .3-6.7 years, and treatments including statins, ezetimibe, PCSK9 inhibitors, and anacetrapib. Indicator variables included individuals end point type, lipid-lowering therapy type, and high baseline high-sensitivity C-reactive protein level.

In 15 of the trials, the new time-dependent model better predicted estimations for a composite of coronary heart disease death, nonfatal myocardial infarction, and ischemic stroke compared to cholesterol treatment trialists. The investigators explored scenarios and found an absolute risk reduction of at least 2% with intensive treatment with high-intensity statin, ezetimibe, and high-dose proprotein convertase subtilisin/kexin type 9 inhibitor compared with high- or moderate-intensity statin alone. The absolute risk reductions were achieved in higher-risk populations with 2-5 years of treatment and among lower-risk populations with 9-11 years of treatment.

“Evaluation of the implications of sustained (low-density lipoprotein cholesterol) lowering on cardiovascular outcomes is important for appropriate clinician-patient shared decision making,” the study authors wrote. “It is apparent given the evidence from lipid-lowering trials that the relative risk reduction gradually improves over time for cardiovascular events with sustained (low-density lipoprotein cholesterol) lowering.”

The team’s model demonstrated the ability to capture treatment benefit accurately for statin and nonstatin treatments. The model can be used to simulate treatment benefit over time with time-varying low-density lipoprotein cholesterol caused by situations caused by trial design or real-world aspects.

The study, “Time-Dependent Cardiovascular Treatment Benefit Model for Lipid-Lowering Therapies,” was published in the Journal of the American Heart Association.