· Lancet (London, England) 2016 · pooled cross-sectional modeling study · n=4,372,000 participants (751 studies)

Worldwide trends in diabetes since 1980: a pooled analysis of 751 population-based studies with 4.4 million participants.

Cited 4418 times in the scientific literature.

Level 3 - non-randomized controlled study

Pooled analysis and Bayesian hierarchical modeling of population-based observational surveys (by design analogy)

PubMed 27061677 · doi:10.1016/S0140-6736(16)00618-8 · record verified 2026-08-29

What was done

Researchers pooled data from 751 population-based studies comprising 4,372,000 adults across 146 countries that collected biomarker data on diabetes (fasting plasma glucose ≥7.0 mmol/L, self-reported diagnosis, or use of insulin or oral hypoglycaemic drugs). A Bayesian hierarchical model was used to estimate trends in age-standardised adult diabetes prevalence from 1980 to 2014 across 200 countries and territories by sex, and to calculate the posterior probability of meeting the global target of halting prevalence at 2010 levels by 2025.

What was found

Global age-standardised diabetes prevalence increased from 4.3% (95% credible interval 2.4–7.0) in 1980 to 9.0% (7.2–11.1) in 2014 in men, and from 5.0% (2.9–7.9) to 7.9% (6.4–9.7) in women. Total adults with diabetes increased from 108 million to 422 million (28.5% driven by rising prevalence, 39.7% by population growth and ageing, and 31.8% by their interaction). Prevalence in 2014 was lowest in northwestern Europe and highest in Polynesia and Micronesia (approaching 25%, with American Samoa >30% in both sexes). The probability of meeting the global 2025 target is <1% for men and 1% for women, with only 9 countries for men and 29 for women having a ≥50% probability.

Why it matters

This study shows that adult diabetes prevalence increased or remained unchanged in every country over 34 years, with disproportionate growth in low- and middle-income regions. The findings demonstrate that global targets to halt diabetes expansion are virtually impossible to meet without substantial policy and clinical changes.

Limits

Biomarker survey data were available for only 146 of the 200 countries estimated, requiring model-based extrapolation for 54 countries. The definition combines type 1 and type 2 diabetes without distinction. Estimates relied on fasting plasma glucose or existing diagnosis/medication use, which may miss cases identifiable only by oral glucose tolerance test or HbA1c.

Cited by