Background <p>The increasing prevalence of diabetes mellitus (DM) and prediabetes (Pre-DM) in elderly individuals necessitates refined diagnostic tools. The oral glucose tolerance test (OGTT) is cumbersome, and age may alter glycemic parameters, suggesting a need for age-specific thresholds. This study aimed to evaluate the diagnostic performance of glycated hemoglobin (HbA1c) and glycated albumin (GA) for dysglycemia in community-dwelling older adults in Chengdu, China.</p> Methods <p>This was a community-based cross-sectional study that consecutively enrolled eligible community-dwelling residents aged 60&#xa0;years or older in Chengdu, China. A total of 1,162 participants underwent measurements of fasting plasma glucose (FPG), HbA1c, and GA. A standard 75‑g oral glucose tolerance test (OGTT) with two‑hour post-prandial plasma glucose (2hPG) measurement was performed, and glycemic status was defined according to the World Health Organization (WHO) diagnostic criteria. Receiver operating characteristic (ROC) curve analysis was used to determine optimal cutoff values for HbA1c and GA to detect DM and Pre-DM.</p> Results <p>Among the 1,162 participants (65.2% were female; median age, 67&#xa0;years), 38.2%, 34.5%, and 27.3% were classified as having normal glucose tolerance (NGT), Pre-DM, and DM, respectively. For the diagnosis of DM, HbA1c demonstrated an AUC of 0.86 (95% CI: 0.83–0.88). An optimal cutoff value of 6.0% yielded greater sensitivity (72.56%) than did the conventional threshold of 6.5% (sensitivity: 34.70%), while maintaining a specificity of 85.36%. Diagnostic performance was superior in females than in males (AUC = 0.87 vs 0.83, <i>P</i> = 0.001) and varied significantly across age and BMI groups, supporting the consideration of age‑specific cutoff values. GA demonstrated moderate diagnostic performance (AUC = 0.70), and its combination with HbA1c did not improve performance. Both biomarkers exhibited limited ability for identifying Pre-DM.</p> Conclusion <p>HbA1c at a lower cutoff value of 6.0% demonstrated high diagnostic accuracy for DM in older adults, though its performance was influenced by age, sex, and BMI. In contrast, GA did not confer a significant diagnostic advantage. These findings suggest that the current cutoff value of HbA1c warrants reassessment in elderly individuals while also highlighting its limitations for Pre-DM screening, where OGTT may still be necessary.</p> Trial registration <p>The trial was registered with the Chinese Clinical Trial Registry (<a href="https://www.chictr.org.cn">https://www.chictr.org.cn</a>, ChiCTR2300070831) on 24 April 2023.</p> Graphical Abstract <p></p>