The diagnostic accuracy of HbA(1c), compared to the oral glucose tolerance test, for screening for type 2 diabetes mellitus in Africa-A systematic review and meta-analysis

Objective To assess the diagnostic accuracy of glycated haemoglobin A(1c) (HbA(1c)), compared to fasting plasma glucose (FPG) and the oral glucose tolerance test (OGTT), in screening for type 2 diabetes (T2D) in Africa. Methods We systematically searched databases for studies that compared the HbA(1c) to either the OGTT, or the FPG for T2D diagnosis were included. The QUADAS 2 tool was used for assessing the quality of included studies. We used the split component synthesis (SCS) method for the meta-analysis of diagnostic accuracy studies to pool the studies for meta-analysis of sensitivity and specificity, primarily at the HbA(1c) >= 48 mmol/mol (6.5%) cut-off and at other cut-offs. We assessed heterogeneity using the I-2 statistic and publication bias using Doi plots. Results Eleven studies, from seven African countries, with 12,925 participants, were included. Against the OGTT, HbA(1c) >= 48 mmol/mol (6.5%) had a pooled sensitivity of 57.7% (95% confidence interval [CI] 43.4-70.9) and specificity of 92.3% (95% CI 83.9-96.5). Against the FPG, HbA(1c) >= 48 mmol/mol (6.5%) had a pooled sensitivity of 64.5% (95% CI 50.5-76.4) and specificity of 94.3% (95% CI 87.9-97.5). The highest sensitivity for HbA(1c), against the OGTT, was at the 42 mmol/mol (6.0%) cut-off. Conclusion In Africa, the HbA(1c) >= 48 mmol/mol (6.5%) cut-off may miss almost half of the individuals with T2D based on blood glucose measures.