Abstract
BACKGROUND: Frailty is a critical geriatric syndrome associated with modifiable lifestyle factors, yet their population-level contributions remain unclear. This study aimed to quantify the proportion of frailty incidence attributable to modifiable lifestyle factors and assess temporal trends in the US. METHODS: We analyzed data from 26,247 participants in the Health and Retirement Study (2004–2020) involving adults aged ≥ 50 years. Frailty was assessed using the Paulson-Lichtenberg Frailty Index. Lifestyle exposures included smoking, drinking, physical inactivity, and sleep disturbance. Associations between lifestyle factors and frailty were examined using Cox models. Population attributable fractions (PAFs) were calculated for each factor, with temporal trends assessed using generalized estimating equations with splines. RESULTS: From 2004 to 2020, frailty incidence declined from 55.2 to 46.6 per 1,000 person-years. Current smoking (HR = 1.46, 95% CI: 1.38, 1.55), physical inactivity (HR = 1.56, 95% CI: 1.48, 1.65), and sleep disturbance (HR = 1.42, 95% CI: 1.36, 1.47) were positively associated with frailty, while current drinking (HR = 0.79, 95% CI: 0.76, 0.83) was negatively associated. The PAFs analyses revealed that sleep disturbance (PAF = 14.43%) and non-current drinking (13.91%) were the leading contributors of frailty, followed by physical inactivity (5.45%) and smoking (5.08%). Only the contribution of physical inactivity increased significantly over time (P(trend) < 0.01). CONCLUSION: Four lifestyle factors collectively accounted for nearly 40% of frailty cases among US adults aged ≥ 50 years. The apparent protective role of drinking warrants caution due to potential “sick quitter” bias. The increasing contribution of physical inactivity highlights the need for prioritizing exercise promotion in frailty prevention. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12889-026-26605-z.