Glycated hemoglobin thresholds for detecting 2-hour plasma glucose-defined abnormal glucose metabolism in Vietnamese women with polyendocrine metabolic ovarian syndrome

Objective To determine glycated hemoglobin (HbA1c) thresholds for detecting 2-hour plasma glucose (2hPG)-defined post-load dysglycemia and diabetes-range hyperglycemia, and to assess whether adding fasting plasma glucose (FPG) improves discrimination in Vietnamese women with polyendocrine metabolic ovarian syndrome (PMOS). Methods We performed a secondary diagnostic-accuracy analysis of a prospective cohort of 1,149 women with PMOS seeking fertility care. FPG was measured immediately before glucose administration at the OGTT visit after an overnight fast of at least 8 hours. The reference standard was 2hPG during a 75-g oral glucose tolerance test (OGTT): post-load dysglycemia was defined as 2hPG ≥140 mg/dL and diabetes-range 2hPG hyperglycemia was defined as a single 2hPG ≥200 mg/dL. Discrimination was assessed using receiver operating characteristic curves, area under the curve (AUC), Youden-optimal thresholds, paired DeLong tests, and bootstrap validation. Results Of the 1,149 women, 342 (29.8%) had 2hPG-defined post-load dysglycemia, including 61 (5.3% of the total cohort) with 2hPG-defined diabetes-range hyperglycemia. HbA1c outperformed FPG for both outcomes (AUC, 0.711 vs 0.634 and 0.799 vs 0.719, respectively). The Youden-optimal HbA1c threshold was 5.58% (37 mmol/mol) for 2hPG-defined post-load dysglycemia (sensitivity, 53.8%; specificity, 80.5%) and 5.62% (38 mmol/mol) for 2hPG-defined diabetes-range hyperglycemia (sensitivity, 68.9%; specificity, 80.0%). The Youden-derived cutoffs are exploratory statistical reference points. For the diabetes-range endpoint, an exploratory HbA1c cutoff of 5.3% (34 mmol/mol) achieved a point-estimate sensitivity of 93.4% and specificity of 34.5%, but would lead to OGTT in 670 per 1,000 women screened. Adding FPG produced small numerical increases in AUC, but there was no statistically significant evidence of improved discrimination over HbA1c alone; bootstrap validation of the combined models showed minimal optimism. Conclusion Among Vietnamese women with PMOS seeking fertility care, HbA1c may help prioritize confirmatory OGTT, but it cannot rule out 2hPG-defined post-load dysglycemia or replace OGTT for definitive glycemic classification. The Youden-derived thresholds should be considered exploratory, and external validation is required before clinical use.