Figure 2.

Dose-response relationships of dietary cholesterol intake with cardiometabolic mortality. HRs (solid line) and 95% CIs (dashed line) were adjusted for age, sex, education, annual income, marital status, total energy intake, smoking status, smoking pack-years, alcohol consumption, physical activity level, body mass index, healthy eating index, history of diabetes, hypertension, dyslipidemia, coronary heart disease, and stroke, and hormone replacement therapy (for women only). Intake of refined carbohydrate was further adjusted for in the Chinese population. To minimize the potential effects of extreme values, participants with the top 1% of cholesterol intake were excluded from the analysis; 150 mg/day was set as the reference, and three knot positions were fitted at the 5th, 50th, and 95th percentiles. P-nonlinearity for cardiometabolic mortality were 0.07, 0.06, and 0.001, respectively. Abbreviations: CI, confidence interval; HR, hazard ratio.