Fig. 2

Dose–response associations between sedentary time and all-cause mortality among people with prediabetes (A) and diabetes (B). A restricted cubic spline regression model with four knots (at the 5th, 35th, 65th, and 95th percentiles) was used to estimate the dose–response association of sedentary time with all-cause mortality. The reference was the median of the lowest tertile of sedentary time. The solid line and gray shading represent hazard ratios and 95% CIs, respectively. Models were adjusted for age, sex, race/ethnicity, education level, family income-to-poverty ratio, smoking status, alcohol consumption, body mass index, glycosylated hemoglobin A1c level, total cholesterol level, high-density lipoprotein cholesterol level, prevalent cardiovascular disease, prevalent hypertension, prevalent cancer, total wear time, and moderate- to vigorous-intensity physical activity. Diabetes duration and glucose-lowering medications were additionally adjusted for patients with diabetes. P-values for nonlinear associations are all ≥ 0.05