New diagnostic criteria for diabetes mellitus: are we any further forward?

Abstract

For several years, the scienti®c and clinical community in the UK have waited while the World Health Organization (WHO) has re®ned both the classi®cation and diagnostic criteria for diabetes mellitus. This period has now drawn to a close with the publication in 1999 of the WHO consultation report and its subsequent endorsement by the British Diabetic Association (BDA) in February 2000 (see www.diabetes.org.uk for details). It is now pertinent to review this report in the light of its forebears and ask whether progress has been made. We shall not discuss gestational diabetes, the aetiological classi®cation of diabetes or the use of the criteria to estimate the prevalence of diabetes (all of which are addressed by the WHO report), but will restrict our comments to the clinical diagnosis of diabetes in non-pregnant individuals. For clarity, all glucose concentrations referred to will be venous plasma measurements, although the equivalent values for venous and capillary whole blood and for capillary plasma can be found in Table 1. Diabetes mellitus is a group of metabolic diseases characterized by chronic elevation of the concentration of glucose in the blood, associated with increased prevalence of microvascular disease such as nephropathy and retinopathy. This is less obvious than it may seem. Although the diagnosis of diabetes has naturally been based on measures of blood glucose concentration, there is in fact no clear cut-off between normality and diabetes, and this is complicated further by the recognition of a state of impaired glucose tolerance (IGT) which lies between the two. Added to this are problems associated with the use of different sample types and analytical methods. Given these limitations, it is clear that any diagnostic classi®cation based on blood glucose concentration will be arti®cial. Until recently, the diagnosis of diabetes worldwide was based on the WHO criteria established in 1985 [which built on earlier reports of both the National Diabetes Data Group (1979) and the WHO (1980) and were reviewed by Wiener in 1992]. According to these criteria, a random (casual) glucose concentration 45 5 mmol/L rendered the diagnosis unlikely, whereas a concentration 511 1 mmol/L established the diagnosis in a symptomatic individual. In traditional units, these two concentrations are equivalent to 100 and 200 mg/dL respectively, which re ects the rather arbitrary choice of these cut-off levels. Patients with glucose concentrations between 5 5 and 11 1 mmol/L required an oral glucose tolerance test (OGTT) to establish the diagnosis. This test had two criteria for diagnosis: a fasting glucose concentration 57 8 mmol/L, or a glucose concentration 511 1 mmol/L 2 h after ingestion of 75 g of glucose. Demonstration of plasma glucose concentration in excess of these values on a single occasion con®rmed the diagnosis in a symptomatic individual, although use of the fasting value alone was considered less reliable due to the problem of ensuring a true fasting state. Over the last decade, however, it has become clear that the fasting cut-off of 7 8 mmol/L does not detect all individuals who go on to demonstrate a 2-h post-load value of 511 1 mmol/L, which is the most rigorously validated diagnostic limit for diabetes. In the mid-1990s, both the American Diabetes Association (ADA) and WHO convened expert groups to consider this issue (amongst others). The ADA published its report in 1997. This accepted a fasting plasma glucose concentration of57 0 mmol/L as more accurately predicting a 2-h post-load value of 511 1 mmol/L than a fasting concentration of 57 8 mmol/L. Having Personal View Ann Clin Biochem 2000; 37: 588±592