Percentage of people aged 18 years and over with fasting plasma glucose (FPG) ≥7 mmol/l (126 mg/dl), or glycated haemoglobin (HbA1c) ≥ 6.5% (48 mmol/mol), or on glucose-lowering medication for diabetes.

Age and sex

Estimated from population-based surveys as follows:

Numerator: number of respondents with FPG ≥7.0 mmol/L, or HbA1c ≥6.5%, or taking medication for diabetes
Denominator: total number of respondents

Population-representative studies with measurements of fasting glucose and/or glycated haemoglobin (HbA1c), and information on diabetes treatment were compiled. In studies with data on both FPG and HbA1c, treated diabetes was added to the prevalence of untreated diabetes based on the definition above. In studies with data on one biomarker only, participants whose measured biomarker was elevated were considered to have diabetes. For the remainder of the sample, who neither used treatment nor had elevated level of the measured biomarker, regressions were used to estimate the probability of having an elevated level of the second (unmeasured) biomarker.

Data were pooled using a Bayesian hierarchical meta-regression model to estimate diabetes prevalence.

Age-standardized estimates are produced by applying the age-specific estimates to the WHO Standard Population.

Full details of input data and methods are available at:

NCD Risk Factor Collaboration (NCD-RisC). Worldwide trends in diabetes prevalence and treatment from 1990 to 2022: a pooled analysis of 1108 population-representative studies with 141 million participants. Lancet 2024. Available online at: https://doi.org/10.1016/S0140-6736(24)02317-1

For any comparison over time or with other populations, age standardization is recommended.

3356

Data are not available for every country and year, and not all available data are nationally representative. Some studies only measured FPG or only measured HbA1c, and regressions were used to estimate how many people would have been identified with the unmeasured biomarker. A Bayesian hierarchical model is used to estimate prevalence for all country-years.