Skip to content

Type 2 diabetes — epidemiology and burden

TL;DR — Diabetes affected an estimated 536.6 million adults aged 20–79 years in 2021 in the IDF model, while GBD estimated 529 million people of all ages; the estimates should not be averaged because their populations and methods differ (Sun 2022, PMID 34879977; GBD 2021 Diabetes Collaborators 2023, PMID 37356446). GBD attributed 96.0% of prevalent diabetes and 95.4% of diabetes disability-adjusted life-years (DALYs) to type 2 diabetes (T2D), and projected more than 1.31 billion people with diabetes by 2050 (GBD 2021 Diabetes Collaborators 2023, PMID 37356446). Burden is shifting toward younger ages, low- and middle-income settings, and disadvantaged urban populations, while early diagnosis carries greater lifetime vascular risk (Xie 2022, PMID 36740855; Nanayakkara 2021, PMID 33313987). Costs and ascertainment are unusually heterogeneous: direct annual cost estimates vary by more than an order of magnitude across settings, and undiagnosed disease remains common where diagnostic access is limited (Seuring 2015, PMID 25787932; Hall 2011, PMID 21756350).

What is being counted

Epidemiologic estimates differ because “diabetes” may include all types, a measured glucose-defined condition, a self-reported diagnosis, or a modelled type-specific residual. GBD estimated total and type 1 diabetes separately and computed T2D by subtraction; IDF combined 219 sources and used logistic regression to smooth age-specific prevalence for adults aged 20–79 years (Sun 2022, PMID 34879977; GBD 2021 Diabetes Collaborators 2023, PMID 37356446).

Source/model Population and year Estimate Methodological boundary
IDF Diabetes Atlas Ages 20–79, 2021 536.6 million; 10.5% 219 sources; extrapolation for countries without qualifying data; diabetes of all types (Sun 2022, PMID 34879977)
GBD 2021 All ages, 2021 529 million (95% UI 500–564); age-standardised 6.1% (5.8–6.5) Bayesian meta-regression plus claims, surveys, literature and mortality models (GBD 2021 Diabetes Collaborators 2023, PMID 37356446)
GBD type-specific decomposition All ages, 2021 T2D = 96.0% (95.1–96.8) of diabetes cases T2D calculated after modelling total and type 1 diabetes (GBD 2021 Diabetes Collaborators 2023, PMID 37356446)
GBD-derived T2D analysis All ages, 2021 506.0 million prevalent; 23.9 million incident; 1.6 million deaths; 75.3 million DALYs Secondary analysis of GBD 2021 (Huang 2025, PMID 40895618)
Historical WHO-member-state projection All ages, 2000 171 million; 2.8% Country prevalence extrapolated to 191 states; useful as historical baseline, not a current estimate (Wild 2004, PMID 15111519)

The IDF and GBD point estimates are close numerically but answer different questions. Their uncertainty is structural as well as statistical: fasting glucose, oral-glucose-tolerance testing, HbA1c, self-report, claims and medication algorithms do not identify identical populations.

Scale and trajectory

Measure Current estimate Projection/trend Source
Adult diabetes prevalence 10.5% (536.6 million), 2021 12.2% (783.2 million), 2045 (Sun 2022, PMID 34879977)
All-age diabetes prevalence 529 million (500–564), 2021 >1.31 billion (1.22–1.39), 2050 (GBD 2021 Diabetes Collaborators 2023, PMID 37356446)
T2D DALYs 75 million (63–90), 2021 Burden increased at an accelerated annual rate of 0.42% (0.34–0.51) over 1990–2021 (Zhang 2024, PMID 39151887)
Non-elderly T2D incidence 196.3/100,000 (145.2–257.4), 1990 361.1/100,000 (275.2–458.4), 2021 (He 2025, PMID 40348179)
Young-adult T2D Ages 15–39 Incidence increased in most of eight high-income jurisdictions during 2000–2020 (Magliano 2024, PMID 39541997)

Population growth, ageing and changes in age-specific risk all contribute. The older projection of 366 million people by 2030 was already exceeded before 2021, illustrating why projections are scenario outputs rather than fixed forecasts (Wild 2004, PMID 15111519; Sun 2022, PMID 34879977).

Geography and socioeconomic patterning

Contrast Estimate Interpretation
Urban vs rural, global adults 12.1% vs 8.3% in 2021 IDF model; does not prove urban residence itself is causal (Sun 2022, PMID 34879977)
High- vs low-income countries 11.1% vs 5.5% in 2021 Lower measured prevalence can coexist with greater case-fatality and poorer access (Sun 2022, PMID 34879977)
Highest GBD super-region North Africa/Middle East 9.3% (8.7–9.9) age-standardised Oceania was highest region at 12.3% (11.5–13.0) (GBD 2021 Diabetes Collaborators 2023, PMID 37356446)
Nepal urban vs rural 8.1% (7.3–8.9) vs 1.0% (0.7–1.3) Ten-study meta-analysis; cross-sectional heterogeneity limits causal inference (Gyawali 2015, PMID 26613684)
Urban-slum T2D Study estimates 0.9%–25.0% 62-study review; major between-country and within-country heterogeneity (Uthman 2022, PMID 35210339)
Sub-Saharan African studies 1% rural Uganda to 12% urban Kenya 1999–2011 review; >40% undiagnosed in several screening studies (Hall 2011, PMID 21756350)

The simple “urban affluent disease” description is no longer adequate. India has documented movement from urban to rural settings, affluent to less privileged groups, and older to younger ages (Pradeepa 2017, PMID 28422124). Low-SDI countries have lower modelled prevalence in some datasets but higher age-standardised mortality and DALY rates, consistent with later detection and reduced access to complication prevention (Huang 2025, PMID 40895618).

Age at onset and lifetime burden

Age at diagnosis is not just a demographic descriptor. A meta-analysis of 26 observational studies and 1,325,493 people found that, after adjustment for current age, each one-year older age at T2D diagnosis was associated with 4% lower all-cause mortality risk, 3% lower macrovascular risk and 5% lower microvascular risk (Nanayakkara 2021, PMID 33313987).

Population Quantitative signal Source
Ages 15–39 in eight high-income jurisdictions 349,591 incident diabetes cases across 346 million person-years, 2000–2020 (Magliano 2024, PMID 39541997)
Adolescents/young adults globally GBD 1990–2019 documented rising early-onset incidence and DALY burden (Xie 2022, PMID 36740855)
Adults across 30 countries Every year younger at diagnosis corresponded to higher adjusted vascular and mortality risk (Nanayakkara 2021, PMID 33313987)
Non-elderly population Male burden exceeded female burden across nearly all age groups in a GBD-derived analysis (He 2025, PMID 40348179)

Early-onset disease combines longer exposure with an adverse risk-factor phenotype. It is covered separately in youth-onset T2D.

Sex and gender

Worldwide estimates suggest approximately 17.7 million more men than women live with diabetes, and men are generally diagnosed at younger age and lower body-fat mass (Kautzky-Willer 2023, PMID 36897358). Women often have a greater cardiometabolic risk-factor burden at diagnosis and a greater relative excess risk of cardiovascular disease and death; reproductive history, gestational diabetes and menopause alter life-course risk (Kautzky-Willer 2023, PMID 36897358).

These observations mix biology, exposure, access, diagnosis and treatment. Most guidelines do not yet provide sex-specific treatment algorithms, and sex-stratified estimates should not be treated as immutable biological differences (Kautzky-Willer 2023, PMID 36897358).

Undiagnosed disease and late presentation

No single global undiagnosed fraction is reproduced here because its value depends strongly on test, sampling frame and the denominator chosen. Regional studies demonstrate the consequence: screening studies in sub-Saharan Africa found more than 40% of cases previously undiagnosed, while reported five-year mortality among diabetes cohorts ranged from 4% to 57% (Hall 2011, PMID 21756350).

Retinopathy at diagnosis is a biological marker of prediagnostic exposure. A systematic review found substantial diabetic-retinopathy prevalence among newly diagnosed T2D, supporting eye assessment at diagnosis rather than after years of known disease (Cai 2023, PMID 36286346). The global diabetic-retinopathy burden is projected to increase through 2045 as diabetes prevalence and survival rise (Teo 2021, PMID 33940045).

Mortality and disability

T2D accounted for 95.4% (94.9–95.9) of all diabetes DALYs in GBD 2021 (GBD 2021 Diabetes Collaborators 2023, PMID 37356446). In 2021, common-metabolic-disease analysis attributed 75 million DALYs (63–90) to T2D, far above the 3.67 million (2.90–4.61) estimated for metabolic dysfunction-associated steatotic liver disease in the same framework (Zhang 2024, PMID 39151887).

Outcome lens Finding Caveat/source
Cause-specific mortality Cardiovascular disease remains a major driver, but causes shift with duration and competing risks 24-year prospective cohort (Andersson 2018, PMID 29421310)
Risk-factor control Excess mortality and cardiovascular outcomes vary markedly with achieved risk-factor control Swedish registry analysis (Rawshani 2017, PMID 28402770)
Metabolic-disease trend T2D and obesity mortality rates did not fall from 2000–2019 in the GBD-derived analysis Ecological/modelled comparison (Chew 2023, PMID 36889281)
Diabetic kidney disease Model projects continued global growth without intervention GBD-derived ARIMA forecast; includes type 1 and type 2 diabetes (Ma 2025, PMID 40060381)

Attributable risks

GBD estimated that high BMI accounted for 52.2% (25.5–71.8) of global T2D DALYs in 2021; the wide uncertainty interval is important, and “attributable” does not mean that individual cases can be assigned to one exposure (GBD 2021 Diabetes Collaborators 2023, PMID 37356446).

Suboptimal diet was estimated to contribute to incident T2D across 184 countries, but comparative risk assessment depends on exposure measurement, assumed dose-response functions and counterfactual intake (O'Hearn 2023, PMID 37069363). Ambient PM2.5 is also included in GBD comparative-risk modelling; its T2D-attributable burden increased from 1990 to 2019 (GBD 2019 Diabetes and Air Pollution Collaborators 2022, PMID 35809588).

Risk domain Epidemiologic signal What it does not establish
High BMI 52.2% (25.5–71.8) of global T2D DALYs Individual inevitability or a purely behavioural cause (GBD 2021 Diabetes Collaborators 2023, PMID 37356446)
Diet Modelled incident cases across 184 countries Effect of a single policy or food on its own (O'Hearn 2023, PMID 37069363)
PM2.5 Rising attributable burden, 1990–2019 Randomised causal effect or precise individual attribution (GBD 2019 Diabetes and Air Pollution Collaborators 2022, PMID 35809588)
Weight category Incidence differs strongly by weight status That temporal change is explained by weight alone (Yu 2023, PMID 37651631)

Economic burden

IDF estimated global diabetes-related health expenditure of US$966 billion in 2021, projected to US$1.054 trillion in 2045 (Sun 2022, PMID 34879977). This is a modelled health-expenditure total, not the sum of all lost productivity, informal care and household financial harm.

Evidence set Range/finding Interpretation
Global cost-of-illness review Direct cost US$242–11,917; indirect cost US$45–16,914 per person-year in 2011 international dollars 86 cost and 23 labour-market studies; methods highly heterogeneous (Seuring 2015, PMID 25787932)
Low/lower-middle-income review Total annual per-person estimates US$29.91–237.38; hospitalisation was the largest direct-cost component Only eight eligible studies; large method variation (Afroz 2018, PMID 30558591)
Cardiovascular-disease cost review Cardiovascular disease adds substantial diabetes-related cost Estimates vary by health system and costing perspective (Einarson 2018, PMID 30005761)

Out-of-pocket payments account for a larger share of burden in many lower-income settings, so equal dollar totals would not imply equal household consequences (Seuring 2015, PMID 25787932).

What the estimates can and cannot answer

  • Prevalence is not incidence: rising prevalence can reflect more new disease, longer survival, better detection, or all three.
  • Modelled projections are conditional on demographic and exposure assumptions; the IDF 2045 and GBD 2050 figures are not competing measurements of the same future population.
  • “T2D” is frequently inferred rather than directly classified in population data, particularly where autoantibodies, C-peptide and genetic testing are unavailable.
  • Country averages conceal large gradients by income, ethnicity, migration, food environment and access to diagnosis.
  • DALYs permit comparison across diseases but inherit uncertainty from prevalence, mortality, disability weights and attribution models.

Measurement-dependent burden and the care cascade

Global totals differ because the estimands differ. NCD-RisC pooled 1,108 studies with measured fasting glucose/HbA1c and medication data and estimated 828 million adults aged ≥18 years (95% CrI 757–908 million) with diabetes in 2022; 445 million adults aged ≥30 years (401–496 million), or 59%, were untreated (NCD-RisC 2024, PMID 39549716). Its earlier 751-study analysis estimated an increase from 108 million adults in 1980 to 422 million in 2014; 28.5% of the numerical increase was attributed to prevalence change, 39.7% to population growth and 31.8% to ageing and interaction (NCD-RisC 2016, PMID 27061677). These are all-diabetes estimates and cannot be substituted for type-specific IDF or GBD figures.

Cascade level Estimate Population/method Source
Diagnosed among all diabetes 55.8% (95% UI 49.3–62.3) Age ≥15, 204 countries, GBD-linked modelling, 2023 Stafford 2025, PMID 40934935
Treated among diagnosed 91.4% (88.0–94.2) Same model Stafford 2025, PMID 40934935
Optimal glycaemia among treated 41.6% (35.7–48.5) Same model Stafford 2025, PMID 40934935
Optimal glycaemia on treatment among all diabetes 21.2% (17.4–25.6) Same model Stafford 2025, PMID 40934935
Prior diagnosis in 55 LMIC surveys 43.9% (41.9–45.9) 37,094 adults with diabetes; survey-weighted cross-section Flood 2021, PMID 35211689
All recommended treatments received 4.6% (3.9–5.4) Same LMIC analysis Flood 2021, PMID 35211689
Glucose-/BP-/lipid-lowering coverage 50.5% / 41.3% / 6.3% Same LMIC analysis Flood 2021, PMID 35211689

The cascade separates four different failures—detection, initiation, comprehensive risk-factor treatment and achieved control. A country can therefore report high medication use among diagnosed people while still controlling only a minority of the total affected population.

Heterogeneity within national averages

In 2011–2012 US NHANES, diabetes prevalence was 14.3% (95% CI 12.2%–16.8%) when HbA1c, fasting glucose and OGTT were all used, including 5.2% (4.0%–6.9%) undiagnosed; omitting OGTT reduced the estimate to 12.3% and the undiagnosed fraction from 36.4% to 25.2% (Menke 2015, PMID 26348752). Thus, diagnostic test choice changes both prevalence and apparent equity gaps.

Early-onset T2D is epidemiologically consequential because earlier diagnosis extends lifetime exposure and is associated with excess microvascular, cardiovascular, mortality and pregnancy risk relative to later onset (Misra 2023, PMID 37708901). A GBD-derived adolescent analysis estimated 14.6 million people aged 10–24 living with T2D in 2021 and attributed 32.84% of T2D DALYs in this group to high BMI; these are model outputs, not enumerated cases (Chen 2025, PMID 39876009).

Socioeconomic position is not a nuisance covariate. Neighbourhood deprivation predicted incident T2D independently of individual position in a British cohort of older men (Bush 2023, PMID 37907278), while a nationwide South Korean cohort linked deprivation to worse complication outcomes (Choi 2020, PMID 32611580). A 19-study meta-analysis found higher all-cause mortality with low socioeconomic status among people with diabetes, although indicators and diabetes type varied between studies (Konkol 2026, PMID 42433401). Swedish registry data also demonstrate that excess mortality persists but varies markedly by age, glycaemic control and renal complications (Tancredi 2015, PMID 26510021).

Controversies that should remain visible

Controversy Position A Position B Resolution needed
Which global total is “correct”? IDF/GBD offer comparable time series and projections NCD-RisC prioritises measured biomarkers across population studies Report estimand, age range, case definition and uncertainty; never average
Is treatment coverage improving? Coverage rose in many countries from 1990–2022 Untreated numbers increased to 445 million because prevalence grew faster Track absolute untreated count beside percentage coverage (NCD-RisC 2024, PMID 39549716)
Is obesity the dominant explanation? High BMI accounts for a large modelled fraction Deprivation, detection and access shape incidence and outcomes beyond BMI Causal, locally evaluated policies rather than attributable fractions alone

Open questions

  • Which surveillance design can measure diagnosed and undiagnosed T2D with the same biochemical definition across low-, middle- and high-income settings? The current models must harmonise unlike sources (Sun 2022, PMID 34879977; GBD 2021 Diabetes Collaborators 2023, PMID 37356446).
  • How much of the rising early-onset burden reflects true incidence rather than changing screening and coding? Multinational administrative data show increases but remain concentrated in high-income jurisdictions (Magliano 2024, PMID 39541997).
  • Which policies reduce incidence without widening socioeconomic disparities? Comparative-risk estimates identify high BMI and diet but do not establish policy effectiveness (GBD 2021 Diabetes Collaborators 2023, PMID 37356446; O'Hearn 2023, PMID 37069363).
  • How should economic evaluations incorporate unpaid care, treatment discontinuation and catastrophic household spending in lower-income settings? Available cost studies use incompatible perspectives (Seuring 2015, PMID 25787932; Afroz 2018, PMID 30558591).

References

  1. Sun H, et al. IDF Diabetes Atlas: Global, regional and country-level diabetes prevalence estimates for 2021 and projections for 2045. Diabetes Res Clin Pract. 2022;183:109119. PMID 34879977
  2. GBD 2021 Diabetes Collaborators. Global, regional, and national burden of diabetes from 1990 to 2021, with projections of prevalence to 2050. Lancet. 2023;402:203-234. PMID 37356446
  3. Wild S, et al. Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care. 2004;27:1047-1053. PMID 15111519
  4. Huang Q, et al. Global burden and risk factors of type 2 diabetes mellitus from 1990 to 2021, with forecasts to 2050. Front Endocrinol. 2025;16:1538143. PMID 40895618
  5. Zhang H, et al. Global burden of metabolic diseases, 1990-2021. Metabolism. 2024;160:155999. PMID 39151887
  6. He Q, et al. Global Burden of type 2 diabetes in non-elderly individuals 1990 to 2021 and projections for 2050. Diabetes Metab. 2025;51:101660. PMID 40348179
  7. Magliano DJ, et al. Trends in the incidence of young-adult-onset diabetes by diabetes type. Lancet Diabetes Endocrinol. 2024;12:915-923. PMID 39541997
  8. Nanayakkara N, et al. Impact of age at type 2 diabetes mellitus diagnosis on mortality and vascular complications. Diabetologia. 2021;64:275-287. PMID 33313987
  9. Xie J, et al. Global burden of type 2 diabetes in adolescents and young adults, 1990-2019. BMJ. 2022;379:e072385. PMID 36740855
  10. Kautzky-Willer A, et al. Sex differences in type 2 diabetes. Diabetologia. 2023;66:986-1002. PMID 36897358
  11. Gyawali B, et al. Prevalence of type 2 diabetes in Nepal. Glob Health Action. 2015;8:29088. PMID 26613684
  12. Uthman OA, et al. Global prevalence and trends in hypertension and type 2 diabetes mellitus among slum residents. BMJ Open. 2022;12:e052393. PMID 35210339
  13. Hall V, et al. Diabetes in Sub Saharan Africa 1999-2011. BMC Public Health. 2011;11:564. PMID 21756350
  14. Pradeepa R, Mohan V. Prevalence of type 2 diabetes and its complications in India and economic costs to the nation. Eur J Clin Nutr. 2017;71:816-824. PMID 28422124
  15. Cai K, et al. Prevalence of diabetic retinopathy in patients with newly diagnosed type 2 diabetes. Diabetes Metab Res Rev. 2023;39:e3586. PMID 36286346
  16. Teo ZL, et al. Global Prevalence of Diabetic Retinopathy and Projection of Burden through 2045. Ophthalmology. 2021;128:1580-1591. PMID 33940045
  17. Andersson T, et al. Mortality trends and cause of death in patients with new-onset type 2 diabetes and controls. Diabetes Res Clin Pract. 2018;138:81-89. PMID 29421310
  18. Rawshani A, et al. Mortality and Cardiovascular Disease in Type 1 and Type 2 Diabetes. N Engl J Med. 2017;376:1407-1418. PMID 28402770
  19. Chew NWS, et al. The global burden of metabolic disease: Data from 2000 to 2019. Cell Metab. 2023;35:414-428.e3. PMID 36889281
  20. Ma X, et al. Global burden of chronic kidney disease due to diabetes mellitus, 1990-2021, and projections to 2050. Front Endocrinol. 2025;16:1513008. PMID 40060381
  21. O'Hearn M, et al. Incident type 2 diabetes attributable to suboptimal diet in 184 countries. Nat Med. 2023;29:982-995. PMID 37069363
  22. GBD 2019 Diabetes and Air Pollution Collaborators. Estimates, trends, and drivers of the global burden of type 2 diabetes attributable to PM2.5 air pollution, 1990-2019. Lancet Planet Health. 2022;6:e586-e600. PMID 35809588
  23. Yu HJ, et al. Incidence and temporal trends in type 2 diabetes by weight status. J Glob Health. 2023;13:04088. PMID 37651631
  24. Seuring T, et al. The Economic Costs of Type 2 Diabetes: A Global Systematic Review. Pharmacoeconomics. 2015;33:811-831. PMID 25787932
  25. Afroz A, et al. Cost-of-illness of type 2 diabetes mellitus in low and lower-middle income countries. BMC Health Serv Res. 2018;18:972. PMID 30558591
  26. Einarson TR, et al. Economic Burden of Cardiovascular Disease in Type 2 Diabetes. Value Health. 2018;21:881-890. PMID 30005761
  27. NCD Risk Factor Collaboration. Worldwide trends in diabetes prevalence and treatment from 1990 to 2022. Lancet. 2024;404:2077-2093. PMID 39549716
  28. NCD Risk Factor Collaboration. Worldwide trends in diabetes since 1980. Lancet. 2016;387:1513-1530. PMID 27061677
  29. Stafford LK, et al. Global, regional, and national cascades of diabetes care, 2000-23. Lancet Diabetes Endocrinol. 2025;13:924-934. PMID 40934935
  30. Flood D, et al. The state of diabetes treatment coverage in 55 low-income and middle-income countries. Lancet Healthy Longev. 2021;2:e340-e351. PMID 35211689
  31. Menke A, et al. Prevalence of and Trends in Diabetes Among Adults in the United States, 1988-2012. JAMA. 2015;314:1021-1029. PMID 26348752
  32. Misra S, et al. Current insights and emerging trends in early-onset type 2 diabetes. Lancet Diabetes Endocrinol. 2023;11:768-782. PMID 37708901
  33. Chen X, et al. Global, regional, and national burdens of type 1 and type 2 diabetes mellitus in adolescents from 1990 to 2021. BMC Med. 2025;23:48. PMID 39876009
  34. Konkol J, et al. Association between low socioeconomic status and all-cause mortality in patients with diabetes. Front Public Health. 2026;14:1841893. PMID 42433401
  35. Bush KJ, et al. Neighbourhood-level socioeconomic deprivation and risk of developing type 2 diabetes. BMJ Open Diabetes Res Care. 2023;11. PMID 37907278
  36. Choi DW, et al. Socioeconomic deprivation and outcomes of diabetes complications in type 2 diabetes. BMJ Open Diabetes Res Care. 2020;8. PMID 32611580
  37. Tancredi M, et al. Excess Mortality among Persons with Type 2 Diabetes. N Engl J Med. 2015;373:1720-1732. PMID 26510021