Sigmoidal maximal effect modeling of low-density lipoprotein cholesterol concentration and annual incidence of coronary heart disease events in secondary prevention trials

• Published in: Pharmacotherapy Vol. 34; no. 5; p. 452
• Main Authors: Charland, Scott L, Stanek, Eric J
• Format: Journal Article
• Language: English
• Published: United States 01.05.2014
• Subjects: Cholesterol, LDL - blood; Clinical Trials as Topic - statistics & numerical data; Coronary Disease - blood; Coronary Disease - drug therapy; Coronary Disease - epidemiology; Coronary Disease - mortality; Humans; Hydroxymethylglutaryl-CoA Reductase Inhibitors - administration & dosage; Hydroxymethylglutaryl-CoA Reductase Inhibitors - therapeutic use; Incidence; Models, Statistical; Secondary Prevention - statistics & numerical data
• ISSN: 1875-9114
• DOI: 10.1002/phar.1368

To evaluate the relationship between low-density lipoprotein cholesterol (LDL-C) concentration and the annual incidence of combined coronary heart disease (CHD) events—death or nonfatal myocardial infarction (NFMI)—by using sigmoidal maximal effect (sEmax) modeling of published data in various populations at risk for CHD events, and to use the best performing sEmax model generated to calculate the number needed to treat (NNT) to prevent a single CHD death or NFMI event across a range of LDL-C concentrations. Literature-based modeling analysis. A total of 95,955 patients from 22 published cardiovascular secondary prevention trials. Four distinct sEmax models were created based on intervention approach and CHD event risk for each trial population. Model outputs included the following: Emax (maximum CHD death/NFMI rate), E0 (minimum CHD death/NFMI rate), and fit parameters. The best-fitting sEmax model was compared with linear, log-linear, and logit models, and it was used for calculation of annualized NNT to prevent one CHD death or NFMI event with statins. The best fitting sEmax model was constructed from nine statin intervention trials in 60,483 clinically stable patients with CHD or CHD risk equivalents (Emax = 4.84%/year [95% confidence interval (CI) 4.11–5.41%/year], E0 = 1.24%/year [95% CI 0.64–1.83%/year]) and was superior to linear, log-linear, and logit models. Reduction of CHD death/NFMI incidence diminished at an LDL-C level near 90 mg/dl and became near static at an LDL-C level of 60–70 mg/dl. Annual NNT for LDL-C reduction from a baseline of 130–100 mg/dl, 90, and 70 mg/dl was 129, 104, and 83, respectively, and from a baseline of 100–70 mg/dl was 232. An sEmax model fully characterized the relationship between LDL-C concentration and incidence of CHD death or NFMI in a high-risk population receiving statins, with diminishing event reduction at an LDL-C level less than 90 mg/dl, and limited projected event reduction beyond an LDL-C level of ~60–70 mg/dl. As baseline LDL-C level declines, the NNT sharply increases.