This analysis of five independent cohorts provides novel evidence that 1 h PG is superior to FPG, 2 h PG, HbA1c or the combination of FPG and HbA1c for diagnosing type 2 diabetes. The 1 h PG consistently demonstrated discriminative ability across all cohorts (AUCs 0.96–0.98) and significant improvements over the combination of FPG and HbA1c (ΔAUC 0.08–0.22, p<0.001). Sensitivity and specificity of 1 h PG were high, reaching >90% sensitivity and specificity, whereas FPG, 2 h PG and HbA1c exhibited moderate to lower AUCs (0.75–0.86, 0.79–0.91 and 0.60–0.83, respectively), with variable sensitivities (53.3–87.8%) and specificities (58.7–100%). These findings indicate that the 1 h PG outperforms both individual and combined traditional markers, providing significantly enhanced diagnostic performance and better-balanced sensitivity and specificity across diverse populations.

Selvin et al investigated the prognostic performance of combined FPG and HbA1c measurements to confirm undiagnosed diabetes at an early stage [4]. While this approach showed high specificity (98.1%), it had moderate sensitivity (54.9%) at 5 years of follow-up. Importantly, as their study did not compare this strategy with OGTT-derived variables, the question of whether alternative markers, such as the 1 h PG, might provide superior diagnostic performance remained unresolved. Our findings address this gap by demonstrating that 1 h PG consistently outperforms FPG plus HbA1c, showing substantially higher sensitivities and specificities.

It is worth noting that the 2 h PG was found not to differ from the combination of FPG and HbA1c (p for meta=0.656), strongly implying that a post-load glucose marker is important for improving the capacity to diagnose type 2 diabetes. The 1 h PG, as well as the FPG and 2 h PG, was found to be highly reproducible [15]. The 1 h PG offers a shorter, reproducible and more sensitive alternative to 2 h PG, can improve compliance with testing, and facilitates earlier intervention strategies [15]. Nevertheless, the 2 h PG remains valuable, particularly for prognostic assessment in individuals with impaired glucose tolerance or impaired fasting glucose at high risk of progression and adverse outcomes.

To the best of our knowledge, this is the first AUC-ROC analysis to elucidate the diagnostic performance of single and combined biomarkers for diagnosing type 2 diabetes. The key strength of this study lies in its analysis of large, heterogeneous cohorts from Western Europe and East Asia, enhancing the generalisability of findings. Furthermore, the consistent methodology across cohorts with varying study designs, combined with robust statistical comparisons, further strengthens the validity of our findings. Limitations of this analysis include the under-representation of certain populations, including Latin American, African, Pacific Islander, Southeast Asian and Middle Eastern cohorts and the enrichment of high-risk individuals defined by ADA criteria in PLIS/TULIP and GENFIEV. Nevertheless, evidence from South Korean [7] and Native American cohorts [8] supports the broader applicability of 1 h PG. Future research should assess the utility of 1 h PG in gestational diabetes and validate its performance in diverse youth populations.

In summary, this study reinforces the use of the 1 h PG as a criterion for diagnosing type 2 diabetes. Its superior diagnostic power and practical advantages over traditional markers positions the 1 h PG as a pivotal tool for the early diagnosis of type 2 diabetes.