Background The joint impact of healthy lifestyle behaviours (HLBs) on incident disability among elderly populations is still uncertain. This cohort study was conducted to estimate the population-attributable fraction (PAF) of combined HLBs for disability reduction in elderly Japanese.

Methods We analysed 10-year follow-up data for 9910 community-dwelling elderly people (≥65 years) in a prospective cohort study. Information on lifestyle behaviours and food consumption was collected via a questionnaire in 2006. The exposure variable was defined as a healthy lifestyle index (HLI), which represented the summed number of HLBs (‘never or former smoker’, ‘time spent walking ≥0.5 hour/day’ and ‘vegetable and fruit consumption volume ≥median’). Data on incident disability were retrieved from the public Long term Care Insurance database. HRs and 95% CIs were estimated by Cox proportional regression and the PAFs and their 95% CIs were estimated with the multivariate-adjusted model.

Results The 10-year incidence of disability was 35.7%. An inverse dose–response relationship was observed (HR (95% CI): 0.85(0.81 to 0.90) for each one-point increase of the HLI score, p-trend <0.001). Based on multivariate-adjustment, adherence to each one additional HLB gives PAF of 10.5%(95% CI 9.0% to 12.0%) for disability reduction. The PAF would have been 25.9%(14.2% to 36.0%) if all subjects had adhered to all three HLBs.

Conclusion Combined HLBs may have a substantial impact on reducing the risk of incident disability among elderly people. Even having one more healthy lifestyle habit may bring considerable benefit.