Role of hemoglobin A1c at different cut-off points for diabetes diagnosis

Objective To assess the value of hemoglobin A1c(HbA1c) at different cut-off points for diagnosing diabetes mellitus(DM).Methods We studied the patients who were suspected of having diabetes or at high risk of diabetes and underwent 75 g-OGTT for diagnosing diabetes from January 2007 to October 2009 in the outpatient department of our hospital.A to tall of 1,427 subjects,including 645 males and 782 females,with a mean age of(53±15) years,were enrolled in the present study.All the subjects underwent oral glucose tolerance test(OGTT),and the HbA1c and glycated albumin(GA) level were measured after 10 hours fasting.Results(1) Based on the diabetes diagnosis criteria of WHO(1999),the subjects were divided into normal glucose tolerance(NGT) group(n=582),impaired glucose regulation(IGR) group(n=378),and diabetes mellitus(DM) group(n=467).HbA1c level had an increasing trend in all the three groups(P0.05).(2) Pearson correlation analysis demonstrated that HbA1c was positively correlated with GA,fasting plasma glucose(FPG),and 2h post-load plasma glucose(2hPG) level determined by OGTT(r=0.74,0.82,0.80,P0.01).(3) Receiver operator characteristic(ROC) curve was plotted to detect the undiagnosed diabetes.The analysis revealed that the area under the curve(AUC) was 0.943 of HbA1c(95% confidence interval: 0.931-0.955),which was similar to that of FPG(AUC 0.955,95% confidence interval 0.944-0.965,P0.05).The sensitivity of using HbA1c≥6.3% as the diabetes diagnosis criterion was significantly higher than using FPG ≥7.0 mmol/L(79.7% vs.51.4%,P0.01) or HbA1c ≥6.5%(79.7% vs.67.2%,P0.01).Conclusions HbA1c measurement is valid for diabetes diagnosis when OGTT is not available.An HbA1c value of 6.3% is an optimal cut-off point for diabetes diagnosis in Chinese population.(Shanghai Med J,2011,34: 341-344)