Objective: Guidelines and expert opinion regard the relationship between low-density lipoprotein cholesterol (LDL-C) and coronary heart disease (CHD) event rates from prevention trials as linear or log-linear. However, these relationships require key assumptions that are typically invalid in biological systems, and may be more fully described by a sigmoidal maximum effect function. Methods: Data were extracted from statin secondary prevention trials of at least 4-years duration (CARE, LIPID, 4S, HPS, TNT, IDEAL; N=57,042; average duration 5.2 yrs, range 4.8-6.1 yrs). Linear and modified nonlinear sigmoid maximum effect (sEmax) models were constructed using WinNONLIN (v.1.5, Pharsight Corporation, Mountain View, CA) to evaluate the relationship of annualized absolute rates of CHD death plus nonfatal myocardial infarction (NFMI) versus average on-treatment LDL-C. Model output included E0 (CHD death+NFMI %/yr at LDL-C=0 mg/dL) and fit parameters [r2 and Akaike's Information Criteria (AIC)]. The model-dependent number needed to treat (NNT) for one year to prevent one CHD death+NFMI event with LDL-C reduction from 100 mg/dL to 70 mg/dL was also calculated. Results: Fit parameters indicated that the sEmax was the more correct model (r2=0.906, AIC=8.40; linear r2=0.876, AIC=15.99). The sEmax model yielded an E0 of 1.37%/yr, whereas the linear model E0 was biologically implausible at –0.76%/yr. The CHD death+NFMI rate at LDL-C=100 mg/dL (sEmax 1.91%/yr; linear 2.13%/yr) and LDL-C=70 mg/dL (sEmax 1.58%/yr; linear 1.14%/yr) resulted in NNT of 303 and 101 based on sEmax and linear models, respectively. Conclusion: The relationship between LDL-C and annualized rate of CHD death + NFMI is sigmoidal and best described by a nonlinear maximum effect model (sEmax). This model demonstrates a marked diminishing rate of return with aggressive LDL-C lowering, strongly suggesting alternative risk modification measures be explored at LDL-C <100 mg/dL. These findings have clinical trial design, treatment guideline, managed care, economic, and public health implications.