Contributions of 2-h post-load glucose, fasting blood glucose and glycosylated haemoglobin elevations to the prevalence of diabetes and pre-diabetes in adults: A systematic analysis of global data.

Abstract

Aims: Some studies found that the association of fasting blood glucose (FPG) or glycosylated haemoglobin (HbA1c) elevation with diabetic complications was not statistically significant after controlling for the confounding caused by 2-h post-load glucose (2hPG) elevation. Furthermore, inclusion of HbA1c as a diagnostic measure for diabetes has raised some concerns about over-diagnosis and over-treatment. This study aimed to quantify the contributions of 2hPG, FPG and HbA1c elevations to the prevalence of diabetes and pre-diabetes respectively, by synthesising global data, which can serve as essential parameters in cost-effectiveness analysis and inform updates of practice guidelines about prevention and control of diabetes.

Materials and methods: The levels of 2hPG, FPG and HbA1c were classified as either 'elevated' or 'normal.' Each distinct combination of these markers (e.g., 'elevated 2hPG, normal FPG and normal HbA1c') defined a unique subgroup, resulting in a total of seven subgroups for both diabetes and pre-diabetes. The contribution of 2hPG elevation to diabetes prevalence was calculated as its proportion among all diabetes cases; the same approach was applied to FPG and HbA1c elevations, as well as to pre-diabetes prevalence. To retrieve global data, five electronic databases (i.e., PubMed, EMBASE, Web of Science, China National Knowledge Infrastructure and Wan Fang) were searched from their inception to February 2024. Studies that fulfilled the following criteria were considered eligible: (1) were conducted in adults without previously diagnosed diabetes; (2) were cross-sectional studies or baseline surveys of cohort studies (which can be regarded as a special type of cross-sectional studies); and (3) reported directly or allowed for calculation of the proportion of each subgroup out of all diabetes and/or the proportion of each subgroup out of all pre-diabetes. A 10-item tool selected from literature was used to appraise the quality of included data. The proportions of each subgroup among all diabetes cases were meta-analysed across eligible studies with the random-effects model using the MetaXL software. Similar meta-analyses were conducted for pre-diabetes.

Results: Thirty-two eligible studies were identified, with 25 reporting on newly detected diabetes (n = 289 094) and 15 on pre-diabetes (n = 221 988) and the majority of them using identical diagnostic cutoffs proposed by the American Diabetes Association (ADA). The mean age of participants ranged from 24 to 68 years (median of mean ages: 51). Twenty-four studies (75%) were assessed as at low risk for ≥7 out of the 10 quality items. In the general population, based on the ADA criteria, the weighted prevalence of diabetes and pre-diabetes was 15% and 69%, respectively. Among those with newly detected diabetes (n = 24 214), 69% had elevated 2hPG, 44% elevated FPG and 61% elevated HbA1c; 7% had isolated FPG elevation; 20% had isolated HbA1c elevation. Among those with newly detected pre-diabetes (n = 133 621), 33% had elevated 2hPG, 51% elevated FPG and 68% elevated HbA1c; 17% had isolated FPG elevation; 34% had isolated HbA1c elevation. Sensitivity analyses stratified by participant comorbidities and study quality produced results consistent with the main findings.

Conclusions: The largest contributors to the prevalence of diabetes and pre-diabetes are 2hPG and HbA1c, respectively. Isolated FPG and HbA1c elevations account for over a quarter of all diabetes and more than half of all pre-diabetes.