Incident type 2 diabetes attributable to suboptimal diet in 184 countries

Abstract

The global burden of diet-attributable type 2 diabetes (T2D) is not well established. This risk assessment model estimated T2D incidence among adults attributable to direct and body weight-mediated effects of 11 dietary factors in 184 countries in 1990 and 2018. In 2018, suboptimal intake of these dietary factors was estimated to be attributable to 14.1 million (95% uncertainty interval (UI), 13.8-14.4 million) incident T2D cases, representing 70.3% (68.8-71.8%) of new cases globally. Largest T2D burdens were attributable to insufficient whole-grain intake (26.1% (25.0-27.1%)), excess refined rice and wheat intake (24.6% (22.3-27.2%)) and excess processed meat intake (20.3% (18.3-23.5%)). Across regions, highest proportional burdens were in central and eastern Europe and central Asia (85.6% (83.4-87.7%)) and Latin America and the Caribbean (81.8% (80.1-83.4%)); and lowest proportional burdens were in South Asia (55.4% (52.1-60.7%)). Proportions of diet-attributable T2D were generally larger in men than in women and were inversely correlated with age. Diet-attributable T2D was generally larger among urban versus rural residents and higher versus lower educated individuals, except in high-income countries, central and eastern Europe and central Asia, where burdens were larger in rural residents and in lower educated individuals. Compared with 1990, global diet-attributable T2D increased by 2.6 absolute percentage points (8.6 million more cases) in 2018, with variation in these trends by world region and dietary factor. These findings inform nutritional priorities and clinical and public health planning to improve dietary quality and reduce T2D globally.

Fig. 1. The proportional burden of T2D attributable to suboptimal diet jointly and by each individual dietary factor globally in 2018.

Bars represent the estimated percentage of T2D incidence due to suboptimal intake of 11 dietary factors jointly (suboptimal diet) and separately at the global level in 2018. The burden due to suboptimal diet was estimated using proportional multiplication, assuming that half the benefit of whole-grain intake is mediated through replacement of refined rice and wheat intake. Refined rice and wheat were modeled separately but combined for this aggregate analysis using proportional multiplication. The attributable burden of T2D for four dietary factors (insufficient intake of fruit, nuts and seeds, non-starchy vegetables and excess intake of fruit juice) were estimated only based on effects mediated through weight gain (for example, no direct effects on T2D risk were identified in the literature). See Supplementary Table 5 for more details on the inputs for each dietary factor. Data are presented as the central estimate (median) and the corresponding 95% UI, derived from the 2.5th and 97.5th percentiles of 1,000 multiway probabilistic Monte Carlo model simulations.

Fig. 2. The burden of T2D attributable to suboptimal diet by key sociodemographic factors at the global level in 2018.

Bars represent the estimated percentage burden (a) and absolute burden per 1 million population (b) of T2D incidence due to suboptimal intake of 11 dietary factors jointly: insufficient intake of whole grains, yogurt, fruit, nuts and seeds, and non-starchy vegetables and excess intake of refined rice and wheat, processed meats, unprocessed red meat, SSBs, potatoes and fruit juice. The burden due to suboptimal diet was estimated using proportional multiplication, assuming that half the benefit of whole-grain intake is mediated through replacement of refined rice and wheat intake. See Supplementary Table 5 for more details on the inputs for each dietary factor. Data are presented as the central estimate (median) and the corresponding 95% UI, derived from the 2.5th and 97.5th percentiles of 1,000 multiway probabilistic Monte Carlo model simulations.

Fig. 3. The proportional burden of T2D attributable to suboptimal intake of eight individual risk factors by world region in 2018.

Bars represent the estimated percentage of T2D incidence due to suboptimal intake of eight individual dietary factors separately. The attributable burden of T2D for four dietary factors (insufficient intake of fruit, nuts and seeds, and non-starchy vegetables and excess intake of fruit juice) was estimated only based on effects mediated through weight gain (that is, no direct effects on T2D risk were identified in the literature) and is reported in Extended Data Fig. 1. Countries were delineated into world regions by the GDD. Data are presented as the central estimate (median) and the corresponding 95% UI, derived from the 2.5th and 97.5th percentiles of 1,000 multiway probabilistic Monte Carlo model simulations.

Fig. 4. The burden attributable to suboptimal diet at the national level in the top 30 most populous countries in 2018.

Bars represent the estimated percentage burden (a) and absolute burden per 1 million population (b) of T2D incidence due to suboptimal intake of 11 dietary factors jointly: insufficient intake of whole grains, yogurt, fruit, nuts and seeds, and non-starchy vegetables and excess intake of refined rice and wheat, processed meats, unprocessed red meat, SSBs, potatoes and fruit juice. The burden due to suboptimal diet was estimated using proportional multiplication, assuming that half the benefit of whole-grain intake is mediated through replacement of refined rice and wheat intake. Countries are ordered based on population size in 2018, from highest to lowest. See Supplementary Table 5 for more details on the inputs for each dietary factor. Data are presented as the central estimate (median) and the corresponding 95% UI, derived from the 2.5th and 97.5th percentiles of 1,000 multiway probabilistic Monte Carlo model simulations.

Fig. 5. The absolute change in the proportional burden of T2D attributable to suboptimal diet and each individual risk factor between 1990 and 2018 globally and by world region for four select dietary factors.

Bars represent the estimated absolute change in proportional burden of T2D incidence (a) globally due to suboptimal intake of 12 dietary factors jointly and individually: insufficient intake of whole grains, yogurt, fruit, nuts and seeds, and non-starchy vegetables and excess intake of refined rice, refined wheat, processed meats, unprocessed red meat, SSBs, potatoes and fruit juice. The burden due to suboptimal diet was estimated using proportional multiplication, assuming that half the benefit of whole-grain intake is mediated through replacement of refined rice and wheat intake. In addition, excess intake of four dietary factors (unprocessed red meat (b), refined rice (c), SSBs (d) and processed meat (e)) is included as illustrative examples of the estimated absolute change in percentage burden of T2D, with the remaining dietary factors included in Extended Data Figs. 5 and 6. A different x-axis range was used for b to account for the magnitude of absolute change in T2D burden attributable to excess intake of unprocessed red meat in southeast and East Asia. A negative absolute change in proportional burden indicates a reduction in the diet-attributable burden of T2D between 1990 and 2018 (for example, reduced intake of harmful dietary factors, increased intake of protective dietary factors), while a positive absolute change in percentage burden indicates an increase in the diet-attributable burden of T2D during that time frame (for example, increased intake of harmful dietary factors, decreased intake of harmful dietary factors). Countries were delineated into world regions by the GDD. Data are presented as the central estimate (median) and the corresponding 95% UI, derived from the 2.5th and 97.5th percentiles of 1,000 multiway probabilistic Monte Carlo model simulations.

Fig. 6. Correlation of national-level diet-attributable T2D burden and national SDI in 2018 and 1990.

Points represent the 184 countries included in this analysis (labeled with their ISO3 code and colored based on world region) in 2018 (a) and 1990 (b). The gray line represents overall correlation, with Pearson correlation coefficient and associated P value (two-tailed) provided. No adjustments were made for multiple comparisons. The y axis is based on estimated proportional burden of T2D incidence due to suboptimal intake of 11 dietary factors jointly: insufficient intake of whole grains, yogurt, fruit, nuts and seeds, and non-starchy vegetables and excess intake of refined rice and wheat, processed meats, unprocessed red meat, SSBs, potatoes and fruit juice. The burden due to suboptimal diet was estimated using proportional multiplication, assuming that half the benefit of whole-grain intake is mediated through replacement of refined rice and wheat intake. SDI is a measure of a nation’s development expressed on a scale of 0 to 1 sourced from the Global Burden of Disease study, based on a compositive average of the rankings of income per capita, average educational attainment and fertility rates.

Extended Data Fig. 1. Logic pathway for estimating the direct and BMI-mediated diet-attributable T2D burden due to suboptimal diet globally.

Modeling inputs and outputs provided in boxes. Blue text and arrows suggest the logical pathway through which dietary intake impacts T2D incidence. BMI, body mass index; PAF, population attributable fraction; RR, relative risk; T2D, type II diabetes.

Extended Data Fig. 2. The proportional burden of T2D attributable to suboptimal intake of four individual risk factors by world region in 2018, %.

Bars represent the estimated percentage of T2D incidence due to suboptimal intake of 4 individual dietary factors – insufficient intake of (a) fruit, (b) non-starchy vegetables, (c) nuts and seeds; and excess intake of (d) fruit juice – that were estimated only based on effects mediated through weight gain (for example, no direct effects on T2D risk were identified in the literature). Remaining dietary factors reported in Fig. 3 in the main text. Countries were delineated into world regions by the Global Dietary Database. Data are presented as the central estimate (median) and corresponding 95% uncertainty interval, derived from the 2.5th and 97.5th percentiles of 1000 multiway probabilistic Monte Carlo model simulations.

Extended Data Fig. 3. The absolute burden of T2D attributable to suboptimal diet at the national level per 1 M population in (A) 1990 and (B) 2018.

Heatmap reflects the estimated absolute burden of T2D incidence due to suboptimal intake of 11 dietary factors jointly: insufficient intake of whole grains, yogurt, fruit, nuts and seeds, and non-starchy vegetables; and excess intake of refined rice and wheat, processed meats, unprocessed red meat, sugar-sweetened beverages, potatoes, and fruit juice. The burden due to suboptimal diet was estimated using proportional multiplication, assuming that half the benefit of whole grains intake is mediated through replacement of refined rice and wheat intake. The absolute burden per 1 million population was calculated by dividing the absolute number of diet-attributable cases by the country population in that year and multiplying by 1 million. Different scales (0 to 4000 cases vs. 0 to 8000 cases) were used to better reflect the absolute case distribution globally in 1990 and 2018, respectively.

Extended Data Fig. 4. The absolute change in the proportional burden of T2D attributable to suboptimal diet by world region between 1990–2018.

Bars represent the estimated absolute change between 1990 and 2018 by world region in proportional burden of T2D incidence attributable to suboptimal intake of 11 dietary factors jointly: insufficient intake of whole grains, yogurt, fruit, nuts and seeds, and non-starchy vegetables; and excess intake of refined rice and wheat, processed meats, unprocessed red meat, sugar-sweetened beverages, potatoes, and fruit juice. The burden due to suboptimal diet was estimated using proportional multiplication, assuming that half the benefit of whole grains intake is mediated through replacement of refined rice and wheat intake. A negative absolute change in proportional burden indicates a reduction in the diet-attributable burden of T2D between 1990 and 2018, while a positive absolute change in percentage burden indicates an increase in the diet-attributable burden of T2D during that time frame. Countries were delineated into world regions by the Global Dietary Database. Data are presented as the central estimate (median) and corresponding 95% uncertainty interval, derived from the 2.5th and 97.5th percentiles of 1000 multiway probabilistic Monte Carlo model simulations.

Extended Data Fig. 5. The absolute change in the proportional burden of T2D attributable to (A) excess potatoes, (B) excess refined wheat, (C) insufficient whole grains, and (D) insufficient yogurt by world region between 1990–2018.

Bars represent the estimated absolute change in proportional burden of T2D incidence between 1990 and 2018 attributable to four dietary factors, globally and by world region – (a) excess intake of potatoes, (b) excess intake of refined wheat, (c) insufficient intake of whole grains, and (d) insufficient intake of yogurt. Note varying x-axis ranges across dietary factors. A negative absolute change in proportional burden indicates a reduction in the diet-attributable burden of T2D between 1990 and 2018 (for example, reduced intake of harmful dietary factors, increased intake of protective dietary factors), while a positive absolute change in percentage burden indicates an increase in the diet-attributable burden of T2D during that time frame (for example, increased intake of harmful dietary factors, decreased intake of harmful dietary factors). Countries were delineated into world regions by the Global Dietary Database. Data are presented as the central estimate (median) and corresponding 95% uncertainty interval, derived from the 2.5th and 97.5th percentiles of 1000 multiway probabilistic Monte Carlo model simulations.

Extended Data Fig. 6. The absolute change in the proportional burden of T2D attributable to (E) insufficient fruit, (F) insufficient non-starchy vegetables, (G) insufficient nuts & seeds, and (H) excess fruit juice by world region between 1990–2018.

Bars represent the estimated absolute change in proportional burden of T2D incidence between 1990 and 2018 attributable to eight dietary factors, globally and by world region – (e) insufficient intake of fruit, (f) insufficient intake of non-starchy vegetables, (g) insufficient intake of nuts & seeds, and (h) excess intake of fruit juice. Note varying x-axis ranges across dietary factors. A negative absolute change in proportional burden indicates a reduction in the diet-attributable burden of T2D between 1990 and 2018 (for example, reduced intake of harmful dietary factors, increased intake of protective dietary factors), while a positive absolute change in percentage burden indicates an increase in the diet-attributable burden of T2D during that time frame (for example, increased intake of harmful dietary factors, decreased intake of harmful dietary factors). Countries were delineated into world regions by the Global Dietary Database. Data are presented as the central estimate (median) and corresponding 95% uncertainty interval, derived from the 2.5th and 97.5th percentiles of 1000 multiway probabilistic Monte Carlo model simulations.

Extended Data Fig. 7. Difference in the absolute burden of T2D attributable to suboptimal diet between 1990–2018 in the top 30 most populous countries in 2018.

Bars represent the estimated absolute change in the absolute burden per 1 M population of T2D incidence due to suboptimal intake of 11 dietary factors jointly: insufficient intake of whole grains, yogurt, fruit, nuts and seeds, and non-starchy vegetables; and excess intake of refined rice and wheat, processed meats, unprocessed red meat, sugar-sweetened beverages, potatoes, and fruit juice. The burden due to suboptimal diet was estimated using proportional multiplication, assuming that half the benefit of whole grains intake is mediated through replacement of refined rice and wheat intake. Countries are ordered based on population size in 2018, from highest to lowest. See Supplementary Table 1 for more details on the inputs for each dietary factor. Data are presented as the central estimate (median) and corresponding 95% uncertainty interval, derived from the 2.5th and 97.5th percentiles of 1000 multiway probabilistic Monte Carlo model simulations.

Extended Data Fig. 8. The burden of T2D attributable to suboptimal diet by key sociodemographic factors at the global level in 1990.

Bars represent the estimated (a) percentage burden and (b) absolute burden per 1 M population of T2D incidence due to suboptimal intake of 11 dietary factors jointly: insufficient intake of whole grains, yogurt, fruit, nuts and seeds, and non-starchy vegetables; and excess intake of refined rice and wheat, processed meats, unprocessed red meat, sugar-sweetened beverages, potatoes, and fruit juice. The burden due to suboptimal diet was estimated using proportional multiplication, assuming that half the benefit of whole grains intake is mediated through replacement of refined rice and wheat intake. See Supplementary Table 1 for more details on the inputs for each dietary factor. Data are presented as the central estimate (median) and corresponding 95% uncertainty interval, derived from the 2.5th and 97.5th percentiles of 1000 multiway probabilistic Monte Carlo model simulations.

Extended Data Fig. 9. Proportional burden of T2D attributable to suboptimal diet by urbanicity and education level at the world region level in 2018.

Bars represent the estimated proportional burden of T2D incidence due to suboptimal intake of 11 dietary factors jointly: insufficient intake of whole grains, yogurt, fruit, nuts and seeds, and non-starchy vegetables; and excess intake of refined rice and wheat, processed meats, unprocessed red meat, sugar-sweetened beverages, potatoes, and fruit juice. The burden due to suboptimal diet was estimated using proportional multiplication, assuming that half the benefit of whole grains intake is mediated through replacement of refined rice and wheat intake. Education level and urban/rural residence were defined previously by the Global Dietary Database Project. Uncertainty in the T2D incidence estimates by education level and urban/rural residence were not incorporated into the population attributable fraction calculation. See Methods for further details. Data are presented as the central estimate (median) and corresponding 95% uncertainty interval, derived from the 2.5th and 97.5th percentiles of 1000 multiway probabilistic Monte Carlo model simulations.