Long-term variability of glycemic markers and risk of all-cause mortality in type 2 diabetes: A systematic review and meta-analysis

Background & aims This research aimed to assess how long-term fluctuations in glycemic indicators (FPG and HbA1c), measured by the coefficient of variation (CV), influence the risk of all-cause mortality in people with type 2 diabetes. Methods A review of all literature related to the variability of glycemic indicators published up to June 2025 was conducted using PubMed, Scopus, and Google Scholar. A total of 3,663 articles were found and evaluated based on the title, introduction, and full text, leading to the selection of 15 studies for the meta-analysis. Meta-analysis was performed using a random-effects model. Subgroup and meta-regression analyses were used to explore sources of heterogeneity. Results Fifteen studies ( n = 584,237) were included. Higher glycemic variability was generally linked to increased mortality risk, but this connection showed extreme statistical heterogeneity (I² >85% for both HbA1c-CV and FPG-CV). Therefore, single pooled estimates (e.g., HR: 0.39 for HbA1c-CV) were unreliable and required stratified interpretation; initial inverse associations were corrected after harmonizing effect coding (Table S1). Importantly, meta-regression revealed that cohort mean age was the strongest moderator, explaining nearly 48.3% (R² = 48.3%) of the between-study heterogeneity. The prognostic link was notably stronger in younger populations, suggesting a diminished effect in older adults due to competing risks. Sensitivity analyses confirmed the consistency of the overall effect direction. Conclusions Long-term glycemic variability is linked to all-cause mortality in type 2 diabetes. Given the extreme heterogeneity, our findings emphasize that age strongly modifies the risk estimates, suggesting that GV is a more potent prognostic marker in younger T2DM populations. Clinically, potential monitoring of GV should be highly individualized and considered hypothesis-generating, particularly in high-risk, younger subgroups. Standardization of metrics and further prospective studies are warranted.