Fasting glucose plus HbA1c testing optimum for diagnosing diabetes
MedWire News: A combination of glycated hemoglobin (HbA1c) and fasting plasma glucose (FPG) testing is needed for optimum detection of diabetes in people at high risk for cardiovascular disease (CVD), say researchers.
They say that an algorithm including both HbA1c and FPG could be easily implemented in clinical practice to improve patient diagnoses.
Erik Badings (Deventer Hospital, The Netherlands) and colleagues carried out a cross-sectional study of 2907 people at high risk for CVD (prior myocardial infarction or stroke, angina pectoris, micro- or macroalbuminuria, or significant coronary or peripheral arterial stenosis) and with no prior history of diabetes, aged 63.6 years on average.
FPG and HbA1c were measured for all the participants. The accuracy of these factors alone and in combination for diagnosing Type 2 diabetes was then assessed using 2-hour glucose results from the oral glucose tolerance test.
The team found that the sensitivity of the normal FPG cutpoint of 7.0 mmol/l for diagnosing diabetes was low, at 59%. The optimal cutpoint for FPG was 6.4 mmol/l, with a sensitivity of 75.7%, a specificity of 77.5%, and a likelihood ratio of 3.37.
For HbA1c, the optimal cutpoint was 5.9%, with a sensitivity of 68.7%, a specificity of 67.1%, and a likelihood ratio of 2.09.
Both FPG and HbA1c were less accurate for diagnosing impaired fasting glucose (IFG; according to 2-hour glucose measures) than diabetes, with IFG cutpoints of 6.1 mmol/l for FPG (sensitivity 57.1%, specificity 57.9%) and 5.7% for HbA1c (sensitivity 63.8%, specificity 60.3%) and likelihood ratios ranging from 1.36 to 1.61.
When HbA1c and FPG were used together for diabetes prediction, the likelihood ratios ranged from 0.21 if both tests were negative (below the cutpoints) to 5.59 if they were both positive (above the cutpoints).
For IFG prediction, the combination of FPG and HbA1c resulted in a likelihood ratio ranging from 0.56 if both were below the cutpoints to 2.05 if both were above the cutpoints.
These findings demonstrate that HbA1c and FPG are not particularly useful for diagnosing IFG, say the researchers.
"In a population of patients at high risk of CVD, the use of the proposed algorithm with both FPG and HcA1c leads to a more accurate diagnosis of diabetes," write Badings et al in the Journal of Diabetes.
"Our findings should… be validated in another study, which also should address cost-effectiveness and feasibility of the algorithm," they suggest.
MedWire (www.medwire-news.md) is an independent clinical news service provided by Springer Healthcare Limited. © Springer Healthcare Ltd; 2011
By Helen Albert