The Burden of Proof studies: assessing the evidence of risk.

Exposure to risks throughout life results in a wide variety of outcomes. Objectively judging the relative impact of these risks on personal and population health is fundamental to individual survival and societal prosperity. Existing mechanisms to quantify and rank the magnitude of these myriad effects and the uncertainty in their estimation are largely subjective, leaving room for interpretation that can fuel academic controversy and add to confusion when communicating risk. We present a new suite of meta-analyses-termed the Burden of Proof studies-designed specifically to help evaluate these methodological issues objectively and quantitatively. Through this data-driven approach that complements existing systems, including GRADE and Cochrane Reviews, we aim to aggregate evidence across multiple studies and enable a quantitative comparison of risk-outcome pairs. We introduce the burden of proof risk function (BPRF), which estimates the level of risk closest to the null hypothesis that is consistent with available data. Here we illustrate the BPRF methodology for the evaluation of four exemplar risk-outcome pairs: smoking and lung cancer, systolic blood pressure and ischemic heart disease, vegetable consumption and ischemic heart disease, and unprocessed red meat consumption and ischemic heart disease. The strength of evidence for each relationship is assessed by computing and summarizing the BPRF, and then translating the summary to a simple star rating. The Burden of Proof methodology provides a consistent way to understand, evaluate and summarize evidence of risk across different risk-outcome pairs, and informs risk analysis conducted as part of the Global Burden of Diseases, Injuries, and Risk Factors Study.

Health effects associated with vegetable consumption: a Burden of Proof study.

Previous research suggests a protective effect of vegetable consumption against chronic disease, but the quality of evidence underlying those findings remains uncertain. We applied a Bayesian meta-regression tool to estimate the mean risk function and quantify the quality of evidence for associations between vegetable consumption and ischemic heart disease (IHD), ischemic stroke, hemorrhagic stroke, type 2 diabetes and esophageal cancer. Increasing from no vegetable consumption to the theoretical minimum risk exposure level (306-372 g daily) was associated with a 23.2% decline (95% uncertainty interval, including between-study heterogeneity: 16.4-29.4) in ischemic stroke risk; a 22.9% (13.6-31.3) decline in IHD risk; a 15.9% (1.7-28.1) decline in hemorrhagic stroke risk; a 28.5% (-0.02-51.4) decline in esophageal cancer risk; and a 26.1% (-3.6-48.3) decline in type 2 diabetes risk. We found statistically significant protective effects of vegetable consumption for ischemic stroke (three stars), IHD (two stars), hemorrhagic stroke (two stars) and esophageal cancer (two stars). Including between-study heterogeneity, we did not detect a significant association with type 2 diabetes, corresponding to a one-star rating. Although current evidence supports increased efforts and policies to promote vegetable consumption, remaining uncertainties suggest the need for continued research.

Health effects associated with consumption of unprocessed red meat: a Burden of Proof study.

Characterizing the potential health effects of exposure to risk factors such as red meat consumption is essential to inform health policy and practice. Previous meta-analyses evaluating the effects of red meat intake have generated mixed findings and do not formally assess evidence strength. Here, we conducted a systematic review and implemented a meta-regression-relaxing conventional log-linearity assumptions and incorporating between-study heterogeneity-to evaluate the relationships between unprocessed red meat consumption and six potential health outcomes. We found weak evidence of association between unprocessed red meat consumption and colorectal cancer, breast cancer, type 2 diabetes and ischemic heart disease. Moreover, we found no evidence of an association between unprocessed red meat and ischemic stroke or hemorrhagic stroke. We also found that while risk for the six outcomes in our analysis combined was minimized at 0 g unprocessed red meat intake per day, the 95% uncertainty interval that incorporated between-study heterogeneity was very wide: from 0-200 g d-1. While there is some evidence that eating unprocessed red meat is associated with increased risk of disease incidence and mortality, it is weak and insufficient to make stronger or more conclusive recommendations. More rigorous, well-powered research is needed to better understand and quantify the relationship between consumption of unprocessed red meat and chronic disease.

Life Expectancy for White, Black, and Hispanic Race/Ethnicity in U.S. States: Trends and Disparities, 1990 to 2019.

BACKGROUND: Life expectancy (LE) differences within and between states by race/ethnicity have not been examined. OBJECTIVE: To estimate LE for selected race/ethnicity groups in states from 1990 to 2019. DESIGN: Cross-sectional time-series analysis. SETTING: United States. PARTICIPANTS: Deidentified death records and Census data were used to construct regression models with smoothed time series of mortality from 1990 to 2019. MEASUREMENTS: LE at birth, by sex and year, for subgroups of people reporting Hispanic, non-Hispanic Black, or non-Hispanic White race/ethnicity. RESULTS: Disparities in LE across states were 8.0 years for females and 12.2 years for males in 1990 and 7.9 years for females and 7.8 years for males in 2019. When race/ethnicity groups were accounted for, disparities across states were 20.7 years for females and 24.5 years for males in 1990, decreasing to 18.5 years for females and 23.7 years for males in 2019. Disparities across states increased within each race/ethnicity group between 1990 and 2019, with the largest increase for non-Hispanic White males and the smallest for Hispanic females. The disparity between race/ethnicity groups within states decreased for most of the 23 states with estimates for all 3 groups but increased for females in 7 states and males in 5 states. LIMITATION: Because of small sample size, LE was not estimated for 37 of 153 state-race/ethnicity groups. CONCLUSION: Disparity in LE across states was greater when race/ethnicity groups were considered. Disparities across all state-race/ethnicity groups in general have decreased over the past 3 decades. Within each race/ethnicity group, disparities across states have increased. Although racial/ethnic disparities decreased in most of the 23 states for which LE was estimated for all 3 groups, they increased for females in 7 states and males in 5 states. PRIMARY FUNDING SOURCE: National Heart, Lung, and Blood Institute.

Basis for changes in the disease burden estimates related to vitamin A and zinc deficiencies in the 2017 and 2019 Global Burden of Disease Studies.

BACKGROUND: The Global Burden of Disease (GBD) Study provides estimates of death and disability from eighty-seven risk factors, including some micronutrient deficiencies. OBJECTIVES: To review methodological changes that led to large differences in the disease burden estimates for vitamin A and Zn deficiencies between the GBD 2017 and 2019 Studies. METHODS: GBD publications were reviewed; additional information was provided by GBD researchers. RESULTS: Vitamin A deficiency prevalence is based on plasma retinol concentration, whereas the estimate for Zn deficiency prevalence uses dietary adequacy as a proxy. The estimated global prevalence of vitamin A deficiency for children aged 1-4 years in the year 2017 decreased from 0·20 (95 % CI 0·17, 0·24) in GBD 2017 to 0·16 (95 % CI 0·15, 0·19) in GBD 2019, while the global prevalence of Zn deficiency did not change between the two studies (0·09 (95 % CI 0·04, 0·17) and 0·09 (95 % CI 0·03, 0·18)). New to 2019 was that meta-analyses were performed using Meta Regression - Bayesian, Regularized, Trimmed, a method developed for GBD. Due to this and multiple other methodological changes, the estimated number of deaths due to vitamin A deficiency dropped from 233 000 (179 000-294 000) to 24 000 (3000-50 000) from GBD 2017 to 2019, and for Zn deficiency from 29 000 (1000-77 000) to 2800 (700-6500), respectively. CONCLUSION: The changes in the estimated disease burdens due to vitamin A and Zn deficiencies in the GBD reports from 2017 to 2019 are due primarily to changes in the analytical methods employed, so may not represent true changes in disease burden. Additional effort is needed to validate these results.

The estimated health impact of sodium reduction through food reformulation in Australia: A modeling study.

BACKGROUND: The Australian Government recently established sodium targets for packaged foods to encourage voluntary reformulation to reduce population sodium consumption and related diseases. We modeled the health impact of Australia's sodium reformulation targets and additional likely health gains if more ambitious, yet feasible sodium targets had been adopted instead. METHODS AND FINDINGS: Using comparative risk assessment models, we estimated the averted deaths, incidence, and disability-adjusted life years (DALYs) from cardiovascular disease (CVD), chronic kidney disease (CKD) and stomach cancer after implementation of (a) Australia's sodium targets (overall and by individual companies); (b) United Kingdom's targets (that covers more product categories); and (c) an optimistic scenario (sales-weighted 25th percentile sodium content for each food category included in the UK program). We used nationally representative data to estimate pre- and post-intervention sodium intake, and other key data sources from the Global Burden of Disease study. Full compliance with the Australian government's sodium targets could prevent approximately 510 deaths/year (95% UI, 335 to 757), corresponding to about 1% of CVD, CKD, and stomach cancer deaths, and prevent some 1,920 (1,274 to 2,600) new cases and 7,240 (5,138 to 10,008) DALYs/year attributable to these diseases. Over half (59%) of deaths prevented is attributed to reformulation by 5 market-dominant companies. Compliance with the UK and optimistic scenario could avert approximately an additional 660 (207 to 1,227) and 1,070 (511 to 1,856) deaths/year, respectively, compared to Australia's targets. The main limitation of this study (like other modeling studies) is that it does not prove that sodium reformulation programs will prevent deaths and disease events; rather, it provides the best quantitative estimates and the corresponding uncertainty of the potential effect of the different programs to guide the design of policies. CONCLUSIONS: There is significant potential to strengthen Australia's sodium reformulation targets to improve its health impact. Promoting compliance by market-dominant food companies will be critical to achieving the potential health gains.

Challenges for Estimating the Global Prevalence of Micronutrient Deficiencies and Related Disease Burden: A Case Study of the Global Burden of Disease Study.

Information on the prevalence of micronutrient deficiencies is needed to determine related disease burden; underpin evidence-based advocacy; and design, deliver, and monitor safe, effective interventions. Assessing the global prevalence of deficiency requires a valid micronutrient status biomarker with an appropriate cutoff to define deficiency and relevant data from representative surveys across multiple locations and years. The Global Burden of Disease Study includes prevalence estimates for iodine, iron, zinc, and vitamin A deficiencies, for which recommended biomarkers and appropriate deficiency cutoffs exist. Because representative survey data are lacking, only retinol concentration is used to model vitamin A deficiency, and proxy indicators are used for the other micronutrients (goiter for iodine, hemoglobin for iron, and dietary food adequacy for zinc). Because of data limitations, complex statistical modeling is required to produce current estimates, relying on assumptions and proxies that likely understate the extent of micronutrient deficiencies and the consequent global health burden.

Approaches to Defining Healthy Diets: A Background Paper for the International Expert Consultation on Sustainable Healthy Diets.

BACKGROUND: Healthy diets promote optimal growth and development and prevent malnutrition in all its forms, including undernutrition, obesity, and diet-related noncommunicable diseases (NCDs). OBJECTIVE: This background paper for the International Expert Consultation on Sustainable Healthy Diets characterizes healthy diets and their implications for food system sustainability. METHODS: Three complementary approaches to defining healthy diets are compared: World Health Organization (WHO) guidelines or recommendations developed between 1996 and 2019; 2017 Global Burden of Disease (GBD) risk factor study estimates of diet-related risk-outcome associations; and analyses associating indices of whole dietary patterns with health outcomes in population studies and clinical trials. RESULTS: World Health Organization dietary recommendations are global reference points for preventing undernutrition and reducing NCD risks; they emphasize increasing intakes of fruits, vegetables (excepting starchy root vegetables), legumes, nuts, and whole grains; limiting energy intake from free sugars and total fats; consuming unsaturated rather than saturated or trans fats; and limiting salt intake. Global Burden of Disease findings align well with WHO recommendations but include some additional risk factors such as high consumption of processed meat; this approach quantifies contributions of diet-related risks to the NCD burden. Evidence on whole dietary patterns supports WHO and GBD findings and raises concerns about potential adverse health effects of foods with high levels of industrial processing. CONCLUSIONS: Implied shifts toward plant foods and away from animal foods (excepting fish and seafood), and for changes in food production systems have direct relevance to the sustainability agenda.

The burden of non-communicable diseases attributable to high BMI in Brazil, 1990-2017: findings from the Global Burden of Disease Study.

BACKGROUND: The prevalence and burden of disease resulting from obesity have increased worldwide. In Brazil, more than half of the population is now overweight. However, the impact of this growing risk factor on disease burden remains inexact. Using the 2017 Global Burden of Disease (GBD) results, this study sought to estimate mortality and disability-adjusted life years (DALYs) lost to non-communicable diseases caused by high body mass index (BMI) in both sexes and across age categories. This study also aimed to describe the prevalence of overweight and obesity throughout the states of Brazil. METHODS: Age-standardized prevalence of overweight and obesity were estimated between 1990 and 2017. A comparative risk assessment was applied to estimate DALYs and deaths for non-communicable diseases and for all causes linked to high BMI. RESULTS: The prevalence of overweight and obesity increased during the period of analysis. Overall, age-standardized prevalence of obesity in Brazil was higher in females (29.8%) than in males (24.6%) in 2017; however, since 1990, males have presented greater rise in obesity (244.1%) than females (165.7%). Increases in prevalence burden were greatest in states from the North and Northeast regions of Brazil. Overall, burden due to high BMI also increased from 1990 to 2017. In 2017, high BMI was responsible for 12.3% (8.8-16.1%) of all deaths and 8.4% (6.3-10.7%) of total DALYs lost to non-communicable diseases, up from 7.2% (4.1-10.8%), and 4.6% (2.4-6.0%) in 1990, respectively. Change due to risk exposure is the leading contributor to the growth of BMI burden in Brazil. In 2017, high BMI was responsible for 165,954 deaths and 5,095,125 DALYs. Cardiovascular disease and diabetes have proven to be the most prevalent causes of deaths, along with DALYs caused by high BMI, regardless of sex or state. CONCLUSIONS: This study demonstrates increasing age-standardized prevalence of obesity in all Brazilian states. High BMI plays an important role in disease burdens in terms of cardiovascular diseases, diabetes, and all causes of mortality. Assessing levels and trends in exposures to high BMI and the resulting disease burden highlights the current priority for primary prevention and public health action initiatives focused on obesity.

The burden of diabetes and hyperglycemia in Brazil: a global burden of disease study 2017.

BACKGROUND: The Global Burden of Diseases (GBD) 2017 database permits an up-to-date evaluation of the frequency and burden of diabetes at the state level in Brazil and by type of diabetes. The objective of this report is to describe, using these updated GBD data, the current and projected future burden of diabetes and hyperglycemia in Brazil, as well as its variation over time and space. METHODS: We derived all estimates using the GBD 2016 and 2017 databases to characterize disease burden related to diabetes and hyperglycemia in Brazil, from 1990 to 2040, using standard GBD methodologies. RESULTS: The overall estimated prevalence of diabetes in Brazil in 2017 was 4.4% (95%UI 4.0-4.9%), with 4.0% of those with diabetes being identified as having type 1 disease. While the crude prevalence of type 1 disease has remained relatively stable from 1990, type 2 prevalence has increased 30% for males and 26% for females. In 2017, approximately 3.3% of all disability-adjusted life years lost were due to diabetes and 5.9% to hyperglycemia. Diabetes prevalence and mortality were highest in the Northeast region and growing fastest in the North, Northeast, and Center-West regions. Over this period, despite a slight decrease in age-standardized incidence of type 2 diabetes, crude overall burden due to hyperglycemia has increased 19%, with population aging being a main cause for this rise. Cardiovascular diseases, responsible for 38.3% of this burden in 1990, caused only 25.9% of it in 2017, with premature mortality attributed directly to diabetes causing 31.6% of the 2017 burden. Future projections suggest that the diabetes mortality burden will increase 144% by 2040, more than twice the expected increase in crude disease burden overall (54%). By 2040, diabetes is projected to be Brazil's third leading cause of death and hyperglycemia its third leading risk factor, in terms of deaths. CONCLUSIONS: The disease burden in Brazil attributable to diabetes and hyperglycemia, already large, is predicted by GBD estimates to more than double to 2040. Strong actions by the Ministry of Health are necessary to counterbalance the major deleterious effects of population aging.

Mortality, morbidity, and risk factors in China and its provinces, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017.

BACKGROUND: Public health is a priority for the Chinese Government. Evidence-based decision making for health at the province level in China, which is home to a fifth of the global population, is of paramount importance. This analysis uses data from the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2017 to help inform decision making and monitor progress on health at the province level. METHODS: We used the methods in GBD 2017 to analyse health patterns in the 34 province-level administrative units in China from 1990 to 2017. We estimated all-cause and cause-specific mortality, years of life lost (YLLs), years lived with disability (YLDs), disability-adjusted life-years (DALYs), summary exposure values (SEVs), and attributable risk. We compared the observed results with expected values estimated based on the Socio-demographic Index (SDI). FINDINGS: Stroke and ischaemic heart disease were the leading causes of death and DALYs at the national level in China in 2017. Age-standardised DALYs per 100 000 population decreased by 33·1% (95% uncertainty interval [UI] 29·8 to 37·4) for stroke and increased by 4·6% (-3·3 to 10·7) for ischaemic heart disease from 1990 to 2017. Age-standardised stroke, ischaemic heart disease, lung cancer, chronic obstructive pulmonary disease, and liver cancer were the five leading causes of YLLs in 2017. Musculoskeletal disorders, mental health disorders, and sense organ diseases were the three leading causes of YLDs in 2017, and high systolic blood pressure, smoking, high-sodium diet, and ambient particulate matter pollution were among the leading four risk factors contributing to deaths and DALYs. All provinces had higher than expected DALYs per 100 000 population for liver cancer, with the observed to expected ratio ranging from 2·04 to 6·88. The all-cause age-standardised DALYs per 100 000 population were lower than expected in all provinces in 2017, and among the top 20 level 3 causes were lower than expected for ischaemic heart disease, Alzheimer's disease, headache disorder, and low back pain. The largest percentage change at the national level in age-standardised SEVs among the top ten leading risk factors was in high body-mass index (185%, 95% UI 113·1 to 247·7]), followed by ambient particulate matter pollution (88·5%, 66·4 to 116·4). INTERPRETATION: China has made substantial progress in reducing the burden of many diseases and disabilities. Strategies targeting chronic diseases, particularly in the elderly, should be prioritised in the expanding Chinese health-care system. FUNDING: China National Key Research and Development Program and Bill & Melinda Gates Foundation.

Global Improvement in Dietary Quality Could Lead to Substantial Reduction in Premature Death.

BACKGROUND: The preventable premature mortality achievable by improvement in dietary quality at a global level is unclear. OBJECTIVE: The aim of this study was to assess dietary quality globally, and to quantify the potential global impact of improving dietary quality on population health. METHODS: We applied the Alternate Healthy Eating Index (AHEI, potential range 0-100) to a global dietary database to assess dietary quality among adults in 190 countries/territories. The relation of AHEI score to risks of major chronic disease was estimated from 2 large cohorts of men and women for whom many repeated dietary assessments during up to 30 years were available. We calculated the preventable premature deaths achievable by shifting from current national diets to a reference healthy diet. RESULTS: The global mean AHEI score in 2017 was 49.5 for males and 50.5 for females. Large differences between current and target intakes existed for whole grains, sodium, long-chain n-3 polyunsaturated fats, polyunsaturated fats, and fruits. From 1990 to 2017, the global mean AHEI score increased modestly from 45.4 to 50.0. Diet quality varied substantially across the world. Coastal Mediterranean nations, the Caribbean region, and Eastern Asia (except China and Mongolia) had a higher AHEI score, whereas Central Asia, the South Pacific, and Eastern and Northern Europe had a lower score. An improvement in dietary quality from the current global diet to the reference healthy diet could prevent >11 million premature deaths, ∼24% of total deaths in 2017. These included 1.6 million cancer deaths, 3.9 million coronary artery disease deaths, 1.0 million stroke deaths, 1.7 million respiratory disease deaths, 0.4 million neurodegenerative disease deaths, 0.5 million kidney disease deaths, 0.6 million diabetes deaths, and 1.2 million digestive disease deaths. CONCLUSIONS: Global dietary quality is slowly improving, but remains far from optimal and varies across countries. Improvements in dietary quality have the potential to reduce mortality rates substantially.

Assessment of the burden of diseases and injuries attributable to risk factors in Canada from 1990 to 2016: an analysis of the Global Burden of Disease Study.

BACKGROUND: An understanding of the risk factors contributing to disease burden is critical for determining research priorities and informing national health policy. We aimed to identify the risk factor trends in Canada. METHODS: As part of the Global Burden of Disease (GBD) study (1990-2016), we conducted an analysis of country-level estimates for Canada to assess the burden of diseases and injuries attributable to risk factors. For both 1990 and 2016, metabolic, environmental and behavioural risk factors were ranked according to their contribution to disability-adjusted life years (healthy years of life lost), total deaths and years lived with disability. RESULTS: In 2016, the risk factors accounting for the largest percentage of disability-adjusted life years in Canada were (1) tobacco, (2) diet, (3) high body mass index, (4) high fasting plasma glucose, (5) high systolic blood pressure, (6) alcohol and drug use, (7) occupational risks, (8) high total cholesterol, (9) impaired kidney function and (10) air pollution. Risk factor rankings remained similar from 1990 to 2016 despite some substantial declines in burden, including a 47% (± 3%) decline in the age-standardized disability-adjusted life years rate attributable to tobacco since 1990. Risk factors with an increasing contribution to disability-adjusted life years rates from 1990 to 2016 included high body mass index, high fasting plasma glucose and alcohol and drug use. INTERPRETATION: Metabolic and behavioural risk factors, including modifiable factors such as tobacco use and diet, remain the leading risk factors contributing to the burden of diseases and injuries in Canada. This work identifies priorities and targets for reducing premature death and disability burden in Canada.

A Meta-Analysis of Food Labeling Effects on Consumer Diet Behaviors and Industry Practices.

CONTEXT: The influence of food and beverage labeling (food labeling) on consumer behaviors, industry responses, and health outcomes is not well established. EVIDENCE ACQUISITION: PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines were followed. Ten databases were searched in 2014 for studies published after 1990 evaluating food labeling and consumer purchases/orders, intakes, metabolic risk factors, and industry responses. Data extractions were performed independently and in duplicate. Studies were pooled using inverse-variance random effects meta-analysis. Heterogeneity was explored with I2, stratified analyses, and meta-regression; and publication bias was assessed with funnel plots, Begg's tests, and Egger's tests. Analyses were completed in 2017. EVIDENCE SYNTHESIS: From 6,232 articles, a total of 60 studies were identified, including 2 million observations across 111 intervention arms in 11 countries. Food labeling decreased consumer intakes of energy by 6.6% (95% CI= -8.8%, -4.4%, n=31), total fat by 10.6% (95% CI= -17.7%, -3.5%, n=13), and other unhealthy dietary options by 13.0% (95% CI= -25.7%, -0.2%, n=16), while increasing vegetable consumption by 13.5% (95% CI=2.4%, 24.6%, n=5). Evaluating industry responses, labeling decreased product contents of sodium by 8.9% (95% CI= -17.3%, -0.6%, n=4) and artificial trans fat by 64.3% (95% CI= -91.1%, -37.5%, n=3). No significant heterogeneity was identified by label placement or type, duration, labeled product, region, population, voluntary or legislative approaches, combined intervention components, study design, or quality. Evidence for publication bias was not identified. CONCLUSIONS: From reviewing 60 intervention studies, food labeling reduces consumer dietary intake of selected nutrients and influences industry practices to reduce product contents of sodium and artificial trans fat.

The Global Nutrient Database: availability of macronutrients and micronutrients in 195 countries from 1980 to 2013.

BACKGROUND: Few data are available on the supply and consumption of nutrients at the country level. To address this data gap, we aimed to create a database that provides information on availability (ie, supply) of 156 nutrients across 195 countries and territories from 1980 to 2013. METHODS: We matched 394 food and agricultural commodities from the Food and Agriculture Organization of the United Nations Supply and Utilization Accounts (SUAs) to food items in the United States Department of Agriculture Food Composition Database and obtained data on nutrient composition of the SUAs' food items. Then, after adjusting for inedible portion of each food item, we added the contributions of individual food items to the availability of each nutrient and estimated the national availability of macronutrients and micronutrients in each year. We validated our estimates by comparing our results with those of national nutrition surveys from three countries (the USA, South Korea, and Ecuador). Using dietary consumption data from the Global Burden of Disease study and two popular machine learning algorithms (Random Forest and XGBoost [extreme gradient boosting]), we developed predictive models to estimate the consumption of each nutrient based on their national availability. FINDINGS: Globally 2710 kcal (95% UI 2660-2770) were available per person per day in 2013. Carbohydrates were the major contributor to energy availability (70·5%), followed by fats (17·4%), and protein (10·5%). The energy availability and the contribution of macronutrients to total energy widely varied across levels of development. Countries at the higher level of development (high Socio-demographic Index countries) had more energy available per person per day (3270 kcal, 3220-3310); greater contributions from fats (26·0%) and proteins (11·9%) to total energy availability; and lower contributions from carbohydrate (54·8%). During 1980-2013, energy availability and the contributions of protein and fats to energy availability have increased globally and across levels of development while the contribution of carbohydrates to total energy availability has decreased. The supply of the micronutrients has also increased during the same period globally and across levels of development. Our validation analysis showed that, after accounting for waste at the retail and household level, our estimates of macronutrient availability were very close to the consumption data in nationally representative surveys. Our machine-learning models closely predicted the observed intake of nutrients with the out-of-sample correlation of greater than 0·8 between predicted and observed intake for the nutrients included in the analysis. INTERPRETATION: Our global nutrient database provides a picture of the supply of various nutrients at the country level and can be useful to assess the performance of national food systems in addressing the nutritional needs of their population. FUNDING: Bill & Melinda Gates Foundation.