Type 2 diabetes remission trajectories and variation in risk of diabetes complications: A population-based cohort study.
Biochemical remission of type 2 diabetes is achievable through dietary changes, physical activity and subsequent weight loss. We aim to identify distinct diabetes remission trajectories in a large population-based cohort over seven-years follow-up and to examine associations between remission trajectories and diabetes complications. Group-based trajectory modelling examined longitudinal patterns of HbA1c level (adjusting for remission status) over time. Multivariable Cox models quantified the association between each remission trajectory and microvascular complications, macrovascular complications, cardiovascular (CVD) events and all-cause mortality. Four groups were assigned. Group 1 (8,112 [13.5%]; achieving HbA1c <48 mmol/mol (6.5%) followed by increasing HbA1c levels); Group 2 (6,369 [10.6%]; decreasing HbA1c levels >48 mmol/mol (6.5%)); Group 3 (36,557 [60.6%]; stable high HbA1c levels); Group 4 (9,249 [15.3%]; stable low HbA1c levels (<48mmol/mol or <6.5%)). Compared to Group 3, Groups 1 and 4 had lower risk of microvascular complications (aHRs (95% CI): 0.65 (0.61-0.70), p-value <0.001;0.59 (0.55-0.64) p-value<0.001, respectively)), macrovascular complications (aHRs (95% CI): 0.83 (0.75-0.92), p-value<0.001; 0.66 (0.61-0.71), p-value<0.001) and CVD events (aHRs (95% CI): 0.74(0.67-0.83), p-value<0.001; 0.67(0.61-0.73), p-vlaue<0.001). Risk of CVD outcomes were similar for Groups 2 and 3. Compared to Group 3, Group 1 (aHR: 0.82(95% CI: 0.76-0.89)) had lower risk of mortality, but Group 4 had higher risk of mortality (aHR: 1.11(95% CI: 1.03-1.19)). Risk of CVD outcomes vary by pattern of remission over time, with lowest risk for those in remission longer. People who achieve remission, even for shorter periods of time, continue to benefit from this lower exposure to hyperglycaemia, which may, in turn, lower the risk of CVD outcomes including mortality.
Funder: NIHR Oxford Biomedical Research Centre; funder-id: http://dx.doi.org/10.13039/501100013373