Life-Course Trajectories and Modifiable Risk Factors for Incident Walking Limitation and Mortality in 25 High-Income, Middle-Income, and Low-Income Countries (Pure): A Prospective Cohort Study Publisher Pubmed



Joundi RA ; Rangarajan S ; Bangdiwala S ; Leong D ; Smith E ; Miller V ; Bosch J ; Ismail R ; Alhabib KF ; Rosengren A ; Lopezjaramillo P ; Iqbal R ; Kelishadi R ; Seron P Show All Authors
Source: The Lancet Healthy Longevity Published:2026

Abstract

Background: Life expectancy is increasing globally, but if people are to age healthily, they must do so with fewer limitations in their daily activities. However, information on either the frequency or risk factors for limitations to walking ability or other key activities across different regions of the world is limited. Our aim was to describe the incidence, trajectories, risk factors, and population-attributable fraction of new-onset walking limitations in 25 countries at all socioeconomic levels. Methods: PURE is an ongoing, prospective cohort study. The current analysis included community-dwelling participants who lived in four high-income countries (HICs), 16 middle-income countries (MICs), and five low-income countries (LICs). Individuals aged 35–70 years at baseline who completed a baseline questionnaire about activity limitations between Jan 12, 2001, and May 6, 2019, were included in our analysis. The activity limitation screen included questions on self-reported difficulty with walking, grasping, bending, seeing close-up, seeing distance, and hearing. The primary outcome was incident walking limitation and our analytic sample comprised those with no walking limitation at baseline. We estimated the incidence rates, adjusted for age and sex, per 100 person-years in the overall PURE population, by country income level (and separately for China) and sex. We used multistate modelling to evaluate trajectories across the life course, analysed across continuous age, through three distinct sequential states: no limitation, walking limitation, and death. We used survival models to evaluate the associations of socioeconomic status, vascular and behavioural factors, community walkability, and incident adverse events, with incident walking limitations. We then calculated the population-attributable fraction of selected modifiable factors and compared the risk factors for walking limitation and mortality. Findings: 172 889 people from the PURE cohort answered questions on walking limitations at baseline, 150 221 of whom reported no walking limitation and were included in the multistate model. Of these 150 221 individuals, 122 538 had at least one follow-up assessment with walking limitations data (mean age at baseline 49·7 years [SD 9·5]; 71 424 [58·3%] female and 51 114 [41·7%] male). Mean follow-up was 14·5 years (SD 3·3). Incidence of a new walking limitation per 100 person-years was higher in LICs (3·34 [95% CI 3·27–3·41]), and lowest in China (0·58 [0·56–0·60]), compared with other MICs (1·80 [1·77–1·84]) and HICs (1·31 [1·27–1·37]). The incidence of walking limitation was higher in female participants (1·84 [1·81–1·87]) than in male participants (1·25 [1·22–1·28]). In multistate models, state transitions from no walking limitation to walking limitation and death occurred at a higher rate and earlier in LICs, where the age at which the probability of transitioning to a walking limitation was reached by an estimated one-third of people at 64 years compared with age 76 years in HICs. Female participants had a higher probability of incident walking limitation across the age spectrum compared with male participants. Many socioeconomic, vascular, and behavioural risk factors, community walkability, and incident adverse events, especially incident stroke, were associated with incident walking limitations. The population-level risk factors with the highest population-attributable fractions for walking limitation were low education (11·1% [95% CI 9·9–12·4]), obesity (5·2% [4·7–5·8]), hypertension (3·6% [2·2–5·0]), and low recreational physical activity (4·3% [2·3–6·3]), with obesity being the highest in HICs (12·9% [11·2–14·6]) and low education being the highest elsewhere. Potentially modifiable individual-level risk factors explained approximately 32·9% of the population's risk of walking limitations and approximately 47·4% of mortality, and four of the top five factors were shared for both outcomes (low education, low recreational activity, poor diet, and hypertension). Interpretation: Individuals in LICs had an accelerated transition to walking limitation, which was approximately 12 years earlier than those in HICs. Walking limitation and mortality shared a common set of modifiable risk factors, accounting for almost one-third of the population-level risk of walking limitations and highlighting opportunities for integrated prevention strategies in mid-life that simultaneously target disability and premature mortality across socioeconomic settings. Funding: Funding sources are listed at the end of the Article. © 2026 The Author(s)