Prediabetes, diabetes and loss of disability-free survival in a community-based older cohort: a post-hoc analysis of the ASPirin in Reducing Events in the Elderly trial

Abstract

Evidence for the prognostic implications of hyperglycaemia in older adults is inconsistent. To evaluate disability-free survival (DFS) in older individuals by glycaemic status. This analysis used data from a randomised trial recruiting 19,114 community-based participants aged ≥70years, who had no prior cardiovascular events, dementia and physical disability. Participants with sufficient information to ascertain their baseline diabetes status were categorised as having normoglycaemia (fasting plasma glucose [FPG] < 5.6mmol/l, 64%), prediabetes (FPG 5.6 to <7.0mmol/l, 26%) and diabetes (self-report or FPG ≥ 7.0mmol/l or use of glucose-lowering agents, 11%). The primary outcome was loss of disability-free survival (DFS), a composite of all-cause mortality, persistent physical disability or dementia. Other outcomes included the three individual components of the DFS loss, as well as cognitive impairment-no dementia (CIND), major adverse cardiovascular events (MACE) and any cardiovascular event. Cox models were used for outcome analyses, with covariate adjustment using inverse-probability weighting. We included 18,816 participants (median follow-up: 6.9 years). Compared to normoglycaemia, participants with diabetes had greater risks of DFS loss (weighted HR: 1.39, 95% CI 1.21-1.60), all-cause mortality (1.45, 1.23-1.72), persistent physical disability (1.73, 1.35-2.22), CIND (1.22, 1.08-1.38), MACE (1.30, 1.04-1.63) and cardiovascular events (1.25, 1.02-1.54) but not dementia (1.13, 0.87-1.47). The prediabetes group did not have an excess risk for DFS loss (1.02, 0.93-1.12) or other outcomes. Among older people, diabetes was associated with reduced DFS, and higher risk of CIND and cardiovascular outcomes, whereas prediabetes was not. The impact of preventing or treating diabetes in this age group deserves closer attention.