Early-onset type 2 diabetes is rising fastest where incomes are lowest

Diabetes burden figures are usually quoted in aggregate, which obscures the part of the trend that matters most for lifetime risk. Two analyses published in November do something more specific: they isolate type 2 diabetes diagnosed in early adulthood, and they follow where it is concentrating.

Both are modelling studies built on Global Burden of Disease data, and both are best read as estimates with wide uncertainty rather than counts. Within those limits, they agree on direction.

The global picture

The global analysis, published 10 November in Diabetes, Obesity and Metabolism, defined early-onset type 2 diabetes as diagnosis between ages 15 and 39, and assessed incidence, prevalence, mortality, and disability-adjusted life years across 204 countries using GBD 2021 data [s2].

Between 1990 and 2021, global prevalent cases rose by 222.32% [s2]. Age-standardised incidence, prevalence, and DALY rates all rose significantly, with estimated annual percentage changes of 2.45, 2.85, and 2.10 respectively [s2].

The distribution is uneven in a specific way. The burden peaked in middle-sociodemographic-index regions and in South and East Asia, while Oceania had the highest age-standardised rates [s2]. The 35–39 age group and males bore the highest burden [s2]. Socioeconomic inequalities widened over the period, with a 35% increase in the mortality disparity affecting low-income populations [s2].

High BMI accounted for between 44.7% and 69.2% of attributable DALYs [s2] — a range wide enough to be worth noticing, but with a floor high enough that no account of early-onset type 2 diabetes can route around body weight.

The projection: age-standardised incidence rising to 529.52 per 100,000 by 2050, with 17.14 million new cases [s2].

The Americas, in more detail

The regional analysis, published 24 November in Diabetes Research and Clinical Practice, narrows to the Americas and to a slightly different age band — 25 to 39 — using GBD 2023 data, and restricts itself to the share of mortality and DALYs attributable to modifiable risk factors [s1]. It used Joinpoint regression for trends, slope and concentration indices for inequality, and a Bayesian age-period-cohort model for projections [s1].

In 2023, Central Latin America recorded the highest age-standardised rates in the region, while high-income regions reported the lowest [s1]. Between 1990 and 2023, the age-standardised DALY rate trended upward in Andean Latin America and in high-income North America [s1].

That pairing is the most interesting line in either paper. Two regions at opposite ends of the income distribution moving in the same direction is not what a simple story about development and diet would predict, and the paper does not resolve why.

The inequality analysis found a slight concentration of burden toward lower socioeconomic groups [s1]. Projections forecast continued rises in absolute deaths and DALYs through 2050 [s1]. Males consistently bore a higher burden than females across all regions [s1] — the same sex pattern the global analysis reports [s2].

What burden modelling is and is not

Neither of these papers measured anyone. GBD outputs are estimates synthesised from surveys, registries, verbal autopsy, and statistical modelling, and their quality tracks the quality of underlying data — which is systematically thinner in exactly the low- and middle-income settings where the burden is estimated to be concentrating. Projections to 2050 compound that: a Bayesian age-period-cohort model extrapolates observed trends and cannot anticipate a change in them.

The two papers also use different GBD rounds (2021 and 2023), different age bands (15–39 and 25–39), and different outcome framings (all early-onset type 2 diabetes versus the fraction attributable to modifiable risk factors) [s1][s2]. They are not interchangeable, and the numbers in one should not be read into the other.

What holds across both

Three things survive the caveats. Early-onset type 2 diabetes is increasing rather than plateauing [s1][s2]. Men carry more of the burden than women in both analyses [s1][s2]. And the gradient runs against lower-income populations, widening globally [s2] and slightly concentrated in the Americas [s1].

Diabetes diagnosed at 30 rather than 60 means three additional decades of exposure to complications. That is the reason this age band gets analysed separately, and the reason a projection of 17.14 million new cases a year by 2050 [s2] is a different kind of number from a projection of total diabetes prevalence.

This article is informational and does not constitute medical advice.

Sources

• [s1] National and regional burden of early-onset type 2 diabetes mellitus in the Americas from 1990 to 2023, attributable to modifiable risk factors, and projections to 2050: a systematic analysis for the global burden of disease study 2023. Diabetes Research and Clinical Practice, 24 November 2025. https://doi.org/10.1016/j.diabres.2025.113007
• [s2] The global burden of early-onset type 2 diabetes (1990–2050): An age–period–cohort analysis of incidence, disparities, and projections. Diabetes, Obesity and Metabolism, 10 November 2025. https://doi.org/10.1111/dom.70287