Background and objectives: Individuals with prevalent diabetes were known to have a higher risk of dementia and lower cognitive function. However, trends of cognitive function before diabetes and in the short term after new-onset diabetes remain unclear.
Methods: This study included participants without baseline diabetes from the China Health and Retirement Longitudinal Study. Cognitive tests were conducted at baseline (wave 1) and at least one time from wave 2 (2013) to wave 4 (2018). Cognitive function was assessed using a global cognition score which was the summary measure of 4 cognitive tests. A linear mixed model was constructed to fit the trends in cognitive function before and after diabetes onset and the trends among nondiabetes. The threshold of statistical significance was p < 0.05.
Results: During the 7-year follow-up, 1,207 (9.7% of 12,422, 59.1 ± 8.6 years, 39.9% male participants) participants developed new-onset diabetes. The cognitive function of both the without diabetes group and the diabetes group declined annually during the follow-up. The annual decline rate of the diabetes group before diabetes onset was similar to that of the without diabetes group during the whole follow-up period. After diabetes onset, participants experienced statistically significant faster cognitive declines in global cognition (-0.023 SD/year; 95% CI -0.043 to -0.004; p = 0.019) and visuospatial abilities test (-0.036 SD/year; -0.061 to -0.011; p = 0.004), but not in tests of episodic memory (-0.018 SD/year; -0.041 to 0.004; p = 0.116), attention and calculation (-0.017 SD/year; -0.037 to 0.003; p = 0.090), or orientation (0.001 SD/year; -0.018 to 0.020; p = 0.894), compared with the cognitive slope before diabetes. In subgroup analysis, compared with those who developed diabetes between 45-54 years, those developing diabetes older (55-64 years, p for interaction = 0.701; 65-74 years, p for interaction = 0.996) did not demonstrate different rates of global cognitive decline after diabetes.
Discussion: Individuals experienced faster rate of cognitive decline in a few years after diabetes onset, but not during the prediabetes period. Age did not modify the effect of diabetes on postdiabetes cognitive decline. Efforts in eliminating the adverse impacts on cognition should be started on diagnosis of diabetes.