ObjectiveTo estimate the association between (1) visual impairment (VI) and (2) eye disease and 6-year mortality risk within a cohort of elderly Kenyan people.Design, setting and participantsThe baseline of the Nakuru Posterior Segment Eye Disease Study was formed from a population-based survey of 4318 participants aged ≥50 years, enrolled in 2007–2008. Ophthalmic and anthropometric examinations were undertaken on all participants at baseline, and a questionnaire was administered, including medical and ophthalmic history. Participants were retraced in 2013–2014 for a second examination. Vital status was recorded for all participants through information from community members. Cumulative incidence of mortality, and its relationship with baseline VI and types of eye disease was estimated. Inverse probability weighting was used to adjust for non-participation.Primary outcome measuresCumulative incidence of mortality in relation to VI level at baseline.ResultsOf the baseline sample, 2170 (50%) were re-examined at follow-up and 407 (10%) were known to have died (adjusted risk of 11.9% over 6 years). Compared to those with normal vision (visual acuity (VA) ≥6/12, risk=9.7%), the 6-year mortality risk was higher among people with VI (<6/18 to ≥6/60; risk=28.3%; risk ratio (RR) 1.75, 95% CI 1.28 to 2.40) or severe VI (SVI)/blindness (<6/60; risk=34.9%; RR 1.98, 95% CI 1.04 to 3.80). These associations remained after adjustment for non-communicable disease (NCD) risk factors (mortality: RR 1.56, 95% CI 1.14 to 2.15; SVI/blind: RR 1.46, 95% CI 0.80 to 2.68). Mortality risk was also associated with presence of diabetic retinopathy at baseline (RR 3.18, 95% CI 1.98 to 5.09), cataract (RR 1.26, 95% CI 0.95 to 1.66) and presence of both cataract and VI (RR 1.57, 95% CI 1.24 to 1.98). Mortality risk was higher among people with age-related macular degeneration at baseline (with or without VI), compared with those without (RR 1.42, 95% CI 0.91 to 2.22 and RR 1.34, 95% CI 0.99 to 1.81, respectively).ConclusionsVisual acuity was related to 6-year mortality risk in this cohort of elderly Kenyan people, potentially because both VI and mortality are related to ageing and risk factors for NCD.