Association of prediabetes progression and regression with cognitive decline: Findings from the CHARLS

Abstract Aim The aim of the study was to describe the association of prediabetes progression and regression with change in cognitive function. Methods Data from three waves (2011, 2015 and 2018) of the China Health and Retirement Longitudinal Study (CHARLS) were analysed. Diabetic statuses in 2011 and 2015 were ascertained using the American Diabetes Association criteria. Cognitive function was assessed and standardized at all three waves, where a total score and its two components (episodic memory and metal status) were calculated. We evaluated the association of prediabetes progression and regression (from 2011 to 2015) with changes in cognitive function from 2011 to 2015 and from 2015 to 2018. Results Of 2590 participants (56% women, mean age 58.6 ± 8.4 years) with prediabetes, 12% progressed to diabetes and 41% regressed to normoglycaemia. Compared with participants who remained as prediabetes, those who progressed to diabetes showed a trend to have accelerated decline in episodic memory ( β = −0.11, 95% confidence interval −0.22 to 0.003, p = 0.057). However, participants who regressed to normoglycaemia did not have less cognitive decline. Neither prediabetes progression nor regression predicted change in cognitive function from 2015 to 2018. In a separate group of participants who remained as normoglycaemia ( n = 858), changes in cognitive function from 2011 to 2015 and from 2015 to 2018 were similar to those who remained as prediabetes. Conclusion In people with prediabetes, progression to diabetes may be associated with accelerated cognitive decline but regression to normoglycaemia does not retard cognitive decline. Prediabetes progression and regression may not be predictive of change in cognitive function.