Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Global burden of 87 risk factors in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019

Global burden of 87 risk factors in 204 countries and territories, 1990–2019: a systematic... Global Health Metrics Global burden of 87 risk factors in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019 GBD 2019 Risk Factors Collaborators* Summary Background Rigorous analysis of levels and trends in exposure to leading risk factors and quantification of their effect on Lancet 2020; 396: 1223–49 human health are important to identify where public health is making progress and in which cases current efforts are *For the list of Collaborators see Viewpoint Lancet 2020; inadequate. The Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2019 provides a standardised and 396: 1135–59 comprehensive assessment of the magnitude of risk factor exposure, relative risk, and attributable burden of disease. Correspondence to: Prof Christopher J L Murray, Methods GBD 2019 estimated attributable mortality, years of life lost (YLLs), years of life lived with disability (YLDs), Institute for Health Metrics and and disability-adjusted life-years (DALYs) for 87 risk factors and combinations of risk factors, at the global level, Evaluation, University of Washington, Seattle, WA 98195, regionally, and for 204 countries and territories. GBD uses a hierarchical list of risk factors so that specific risk factors USA (eg, sodium intake), and related aggregates (eg, diet quality), are both evaluated. This method has six analytical steps. [email protected] (1) We included 560 risk–outcome pairs that met criteria for convincing or probable evidence on the basis of research studies. 12 risk–outcome pairs included in GBD 2017 no longer met inclusion criteria and 47 risk–outcome pairs for risks already included in GBD 2017 were added based on new evidence. (2) Relative risks were estimated as a function of exposure based on published systematic reviews, 81 systematic reviews done for GBD 2019, and meta-regression. (3) Levels of exposure in each age-sex-location-year included in the study were estimated based on all available data sources using spatiotemporal Gaussian process regression, DisMod-MR 2.1, a Bayesian meta-regression method, or alternative methods. (4) We determined, from published trials or cohort studies, the level of exposure associated with minimum risk, called the theoretical minimum risk exposure level. (5) Attributable deaths, YLLs, YLDs, and DALYs were computed by multiplying population attributable fractions (PAFs) by the relevant outcome quantity for each age- sex-location-year. (6) PAFs and attributable burden for combinations of risk factors were estimated taking into account mediation of different risk factors through other risk factors. Across all six analytical steps, 30 652 distinct data sources were used in the analysis. Uncertainty in each step of the analysis was propagated into the final estimates of attributable burden. Exposure levels for dichotomous, polytomous, and continuous risk factors were summarised with use of the summary exposure value to facilitate comparisons over time, across location, and across risks. Because the entire time series from 1990 to 2019 has been re-estimated with use of consistent data and methods, these results supersede previously published GBD estimates of attributable burden. Findings The largest declines in risk exposure from 2010 to 2019 were among a set of risks that are strongly linked to social and economic development, including household air pollution; unsafe water, sanitation, and handwashing; and child growth failure. Global declines also occurred for tobacco smoking and lead exposure. The largest increases in risk exposure were for ambient particulate matter pollution, drug use, high fasting plasma glucose, and high body- mass index. In 2019, the leading Level 2 risk factor globally for attributable deaths was high systolic blood pressure, which accounted for 10·8 million (95% uncertainty interval [UI] 9·51–12·1) deaths (19·2% [16·9–21·3] of all deaths in 2019), followed by tobacco (smoked, second-hand, and chewing), which accounted for 8·71 million (8·12–9·31) deaths (15·4% [14·6–16·2] of all deaths in 2019). The leading Level 2 risk factor for attributable DALYs globally in 2019 was child and maternal malnutrition, which largely affects health in the youngest age groups and accounted for 295 million (253–350) DALYs (11·6% [10·3–13·1] of all global DALYs that year). The risk factor burden varied considerably in 2019 between age groups and locations. Among children aged 0–9 years, the three leading detailed risk factors for attributable DALYs were all related to malnutrition. Iron deficiency was the leading risk factor for those aged 10–24 years, alcohol use for those aged 25–49 years, and high systolic blood pressure for those aged 50–74 years and 75 years and older. Interpretation Overall, the record for reducing exposure to harmful risks over the past three decades is poor. Success with reducing smoking and lead exposure through regulatory policy might point the way for a stronger role for public policy on other risks in addition to continued efforts to provide information on risk factor harm to the general public. Funding Bill & Melinda Gates Foundation. Copyright © 2020 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. www.thelancet.com Vol 396 October 17, 2020 1223 Global Health Metrics Research in context Evidence before this study the systematic reviews and meta-regressions, 47 new risk– The Global Burden of Diseases, Injuries, and Risk Factors Study outcome pairs have been included for risks that were previously (GBD) 2017 provided the most recent assessment of deaths, included. This includes outcomes linked to low birthweight and years of life lost because of premature mortality, years of life short gestational age as intermediate outcomes linked to lived with disability, and disability-adjusted life-years particulate matter with a diameter smaller than 2·5 μm (PM ), 2·5 attributable to metabolic, environmental and occupational, which has increased the burden attributable to PM . (8) New 2·5 and behavioural risk factors. GBD 2017 provided estimates from cohorts, trials, and case-control studies have been added for the 1990 to 2017 for 195 countries and territories. Many reports assessment of risk functions. (9) New sources have been added explore the burden of disease that can be attributed to a specific to the analysis of risk factor exposure by age, sex, and location. risk factor in a specific country or territory, region, or globally, (10) Corrections for non-reference method exposure but none attempts to assess an extensive list of risk factors in all measurements have been revised using network or related countries and regions. meta-regression. (11) For dietary risks, the theoretical minimum risk exposure level (TMREL) has been revised based on the new Added value of this study systematic reviews. (12) The distribution of alcohol use across GBD 2019 advances the technical quantification of attributable individuals has been revised to better capture the asymmetric burden in 12 ways. (1) In support of the agreement between GBD nature of the distribution. In addition to the technical and WHO, nine new countries have been added to the analysis: improvements in each step of the quantification of risk factor Cook Islands, Monaco, San Marino, Nauru, Niue, Palau, Saint Kitts exposure, relative risk, TMREL, and attributable burden, in this and Nevis, Tokelau, and Tuvalu. (2) Subnational assessments for study we have focused attention on the broad trends in risk Italy, Nigeria, Pakistan, the Philippines, and Poland have been exposure by computing summary exposure values for added to GBD 2019. (3) High and low non-optimal temperatures aggregations of risk factors. Isolating the long-term global and have been added as risk factors (54 new risk–outcome pairs). national trends in risk exposure reveals in which cases the world (4) For 81 risk–outcome pairs, new systematic reviews have been has been successful in reducing exposure to harmful risks. done as part of GBD 2019. (5) For 139 risk–outcome pairs, dose–response meta-regressions have been done to evaluate if Implications of all the available evidence the relationship between exposure and relative risk might not be Improved analysis of risk exposure and burden attributable to adequately captured by assuming a log-linear relationship risk factors at the national, regional, and global level can help to between risk and per unit increase in exposure. (6) On the basis of focus attention on risks for which exposure is increasing and in the systematic reviews and dose–response meta-regression, which locations. This quantification is an essential input into 12 risk–outcome pairs have been excluded from GBD 2019 public health prioritisation and evaluation of programme because they no longer met inclusion criteria. (7) On the basis of success. Introduction To help track risk exposures and the burden The mechanism for much of disease and injury preven­ attributable to these exposures, many studies are pub­ tion is through modifying environmental, occupational, lished each year on the burden of specific risks, often 4–7 behavioural, and metabolic risk factors. Other pathways, in specific countries or regions. To our knowledge, such as vaccination or addressing social determinants the only effort to examine attributable burden with of health, are crucially important, but a substantial standardised methods across a wide set of risk factors component of public health has targeted modifying the spanning all countries is the Global Burden of Diseases, 8–12 aforementioned risk factors. Whether the risk factor is Injuries, and Risk Factors Study (GBD). Many choices targeted through public policy such as taxation or go into the comparable quantification of the burden of regulation, through programmes such as water supply risk factors; GBD provides a rules­ based approach to improvement, or primary care advice and pharmacological evidence synthesis that follows the Guidelines on intervention, it is essential to track progress on risk Accurate and Transparent Health Estimates Reporting. exposure. Which risk factors are declining, stagnating, or Comparable quantification across risks over time even increasing gives insights into where current efforts and across populations facilitates identi fying relative are working or are insufficient. Understanding where the importance and how population health risks are promise of prevention is being realised might generate evolving over time. GBD also provides a framework lessons that can be applied to other risks in which progress to understand both the trends in risk exposure and is slow. Tracking the burden attributable to risk exposure, the trends in burden attributable to risks. Quantifying measured by deaths, years of life lost (YLLs), years lived and reporting both exposure and attributable burden is with disability (YLDs), or disability­ adjusted life­ years important because exposure might be increasing and (DALYs), can also help governments, donor agencies, attributable burden decreasing if other drivers of the international organisations, and civil society organisations underlying outcomes are declining at a fast enough 1–3 to identify new priorities. rate. 1224 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics In this study, we present new or updated results for Islands, Monaco, San Marino, Nauru, Niue, Palau, the quantification of 560 risk–outcome pairs including Saint Kitts and Nevis, Tokelau, and Tuvalu) were added, updated data for exposure, updated data for relative risks, such that the GBD location hierarchy now includes methods innovation in evaluating risk­ exposure functions, all WHO member states. These new locations were and the addition of two new risk factors—high and previously included in regional totals by assuming low non­ optimal temperatures. In addition to providing that age­ specific rates were equal to the regional rates. quantification of exposure and attributable burden in At the most detailed level, we generated estimates 204 locations over the period 1990–2019, we used summary for 990 locations. The GBD diseases and injuries exposure values (SEVs) for aggregates of risk factors to analytical framework generated estimates for every year understand where public health is making progress from 1990 to 2019. tackling the major environmental, occupational, beha­ vioural, and metabolic risk factors, and where it is not. GBD risk factor hierarchy Individual risk factors such as low birthweight or Methods ambient ozone pollution are evaluated in the GBD CRA. Overview In addition, there has been policy interest in groups of The GBD 2019 estimation of attributable burden fol­ risk factors such as household air pollution combined lowed the general framework established for comparative with ambient particulate matter. To accommodate these 14,15 risk assessment (CRA) used in GBD since 2002. Here, diverse interests, the GBD CRA has a risk factor hier­ we provide a general overview and details on major archy. Level 1 risk factors are behavioural, environmental innovations since GBD 2017. More detailed methods are and occupational, and metabolic; Level 2 risk factors available in appendix 1. CRA can be divided into six key include 20 risks or clusters of risks; Level 3 includes See Online for appendix 1 steps: inclusion of risk–outcome pairs in the analysis; 52 risk factors or clusters of risks; and Level 4 includes estimation of relative risk as a function of exposure; 69 spe cific risk factors. Counting all specific risk factors estimation of exposure levels and distributions; deter­ and aggregates computed in GBD 2019 yields 87 risks or mination of the counterfactual level of exposure, the level clusters of risks. For a full list of risk factors by level, see of exposure with minimum risk called the theoretical appendix 1 (section 5, table S2). minimum risk exposure level (TMREL); computation of population attributable fractions and attributable burden; Determining the inclusion of risk–outcome pairs in GBD and estimation of mediation of different risk factors Since GBD 2010, we have used the World Cancer Research through other risk factors such as high body­ mass index Fund criteria for convincing or probable evidence of risk– (BMI) and ischaemic heart disease, mediated through outcome pairs. For GBD 2019, we completely updated our elevated systolic blood pressure (SBP), elevated fasting systematic reviews for 81 risk–outcome pairs. Preferred plasma glucose (FPG), and elevated LDL cholesterol, to Reporting Items for Systematic Reviews and Meta­ compute the burden attributable to various combinations Analyses flowcharts on these reviews are available in of risk factors. appendix 1 (section 4). Convincing evidence requires more than one study type, at least two cohorts, no substantial Geographical units, age groups, and time periods unexplained heterogeneity across studies, good­ quality GBD 2019 estimated prevalence of exposure and attri­ studies to exclude the risk of confounding and selection butable deaths, YLLs, YLDs, and DALYs for 23 age bias, and biologically plausible dose–response gradients. groups; males, females, and both sexes combined; For GBD, for a newly proposed or evaluated risk–outcome and 204 countries and territories that were grouped pair, we additionally required that there was a significant into 21 regions and seven super­ regions. GBD 2019 association (p<0·05) after taking into account sources of includes subnational analyses for Italy, Nigeria, Pakistan, potential bias. To avoid risk–outcome pairs repetitively the Philippines, and Poland, and 16 countries previously entering and leaving the analysis with each cycle of GBD, estimated at subnational levels (Brazil, China, Ethiopia, the criteria for exclusion requires that with the available India, Indonesia, Iran, Japan, Kenya, Mexico, New studies the association has a p value greater than 0·1. Zealand, Norway, Russia, South Africa, Sweden, the UK, On the basis of these reviews and meta­ regressions, and the USA). All subnational analyses are at the 12 risk–outcome pairs included in GBD 2017 were excluded first level of administrative organi sation within each from GBD 2019: vitamin A deficiency and lower respiratory country except for New Zealand (by Māori ethnicity), infections; zinc deficiency and lower respiratory infections; Sweden (by Stockholm and non­ Stockholm), the UK diet low in fruits and four outcomes: lip and oral cavity (by local government authorities), and the Philippines cancer, nasopharynx cancer, other pharynx cancer, and (by province). In this publication, we present subnational larynx cancer; diet low in whole grains and two outcomes: estimates for Brazil, India, Indonesia, Japan, Kenya, intracerebral haemorrhage and subarachnoid haemor­ Mexico, Sweden, the UK, and the USA; given space rhage; intimate partner violence and maternal abortion constraints, these results are presented in appendix 2. and miscarriage; and high FPG and three outcomes: See Online for appendix 2 For this cycle, nine countries and territories (Cook chronic kidney disease due to hyper tension, chronic www.thelancet.com Vol 396 October 17, 2020 1225 Global Health Metrics kidney disease due to glomerulonephritis, and chronic sections 2.1, 4). For many risk factors, exposure data were kidney disease due to other and unspecified causes. modelled using either spatiotemporal Gaussian process 17,18 In addition, on the basis of multiple requests to begin regression or DisMod­ MR 2.1, which are Bayesian sta­ capturing important dimensions of climate change into tistical models developed over the past 12 years for GBD GBD, we evaluated the direct relationship between high analyses. For most risk factors, the distribution of exposure and low non­ optimal temperatures on all GBD disease and across individuals was estimated by modelling a measure injury outcomes. Rather than rely on a heterogeneous of dispersion, usually the SD, and fitting an ensemble of literature with a small number of studies examining parametric distributions to the predicted mean and SD. relationships with specific diseases and injuries, we ana­ Ensemble distributions for each risk were estimated lysed individual­ level cause of death data for all locations based on individual­ level data. Details for each risk factor with available information on daily temperature, location, modelling for mean, SD, and ensemble distribution are and International Classification of Diseases ­ coded cause of available in appendix 1 (section  4). Because of the strong death. These data totalled 58·9 million deaths covering dependency between birthweight and gestational age, eight countries. On the basis of this analysis, 27 GBD exposure for these risks was modelled as a joint distribution cause Level 3 outcomes met the inclusion criteria for each using the copula method. non­ optimal risk factor (appendix 1 section 2.2.1) and were In many cases, exposure data were available for the included in this analysis. Other climate­ related relation­ reference method of ascertainment and for alternative ships, such as between precipitation or humidity and methods, such as tobacco surveys reporting daily smoking health outcomes, have not yet been evaluated. versus total smoking; in these cases, we estimated the statistical relationship between the reference and alter­ Estimating relative risk as a function of exposure for native methods of ascertainment using network meta­ each risk–outcome pair regression and corrected the alternative data using this In GBD, we use published systematic reviews and for relationship. GBD 2019, we updated these where necessary to include any new studies that became available before Dec 31, 2019. Determining the TMREL We did meta­ analyses of relative risks from these studies For harmful risk factors with monotonically increasing as a function of exposure (appendix 1 sections 2.2.2, 4). For risk functions, the theoretical minimum risk level was GBD 2019, 81 new systematic reviews were done, including set to 0. For risk factors with J­ shaped or U­ shaped risk for 44 diet risk–outcome pairs. To allow for risk functions functions, such as for sodium and ischaemic heart that might not be log­ linear, we relaxed the meta­ regres­ disease or BMI and ischaemic heart disease, the TMREL sion assumptions to allow for monotonically increasing was determined as the low point of the risk function. or decreasing but potentially non­ linear functions for When the bottom of the risk function was flat or 147 risk–outcome pairs. Appendix 1 (section 2) provides the poorly determined, the TMREL uncertainty interval (UI) mathematical and computational details for how we captured the range over which risks are indistinguishable. implemented this approach for meta­ regression. 218 risk– For protective risks with monotonically declining risk outcome pairs were estimated assuming log­ linear functions with exposure, namely risk factors where relationships. For 126 risk–out come pairs, exposure was exposure lowers the risk of an outcome, the challenge is dichotomous or polytomous. For 37 risk–outcome pairs, selecting the level of exposure with the lowest level of the population attributable fractions were assumed by risk strongly supported by the available data. Projecting definition to be 100% (eg, 100% of diabetes is assumed to beyond the level of exposure supported by the available be, by definition, related to elevated FPG). For 32 risk– studies could exaggerate the attributable burden for a outcome pairs, other approaches were used that reflected risk factor. In these cases, for each risk–outcome pair, we the nature of the evidence that has been collected for determined the exposure level at the 85th percentile of those risks (appendix 1 section 4). For risks that affect exposure in the cohorts or trials used in the risk meta­ cardiovascular outcomes, we adjusted relative risks by age regression. We then generated the TMREL by weighting such that they follow the empirical pattern of attenuation each risk–outcome pair by the relative global magnitude seen in published studies for elevated SBP, FPG, and LDL of each outcome. Appendix 1 (section 2.4 and 4) provides cholesterol. details on the TMREL estimation for each risk. Estimation of the distribution of exposure for each risk Estimation of the population attributable fraction and by age-sex-location-year attributable burden For each risk factor, we systematically searched for For each risk factor j, we computed the population published studies, household surveys, censuses, admin­ attributable fraction (PAF) by age­ sex­ location­ year using istrative data, ground monitor data, or remote sensing the following general formula for a continuous risk: data that could inform estimates of risk exposure. To u ∫ RR (x)P (x)dx – RR (TMREL ) x=l joasg jasgt joasg jas estimate mean levels of exposure by age­ sex­ location­ year, PAF = joasgt specific methods varied across risk factors (appendix 1 ∫ RR (x)P (x)dx x = l joasg jasgt 1226 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics where PAF is the PAF for cause o, for age group a, distribution of exposure. We then averaged across joasgt sex s, location g, and year t; RR (x) is the relative risk as outcomes to compute the SEV for a given risk as joasg a function of exposure level x for risk factor j, for cause o controlled for confounding, age group a, sex s, and SEV = SEV r rc N(c) location g with the lowest level of observed exposure as l c and the highest as u; P (x) is the distribution of exposure jasgt at x for age group a, sex s, location g, and year t; and where N(c) is the total number of outcomes for a risk. TMREL is the TMREL for risk factor j, age group a, and The SEV is effectively excess risk ­ weighted prevalence, jas sex s. Where risk exposure is dichotomous or polytomous, which allows for comparisons across different types of this formula simplifies to the discrete form of the exposures. Maximum risk in the denominator of the SEV equation. is determined by the relative risk at the 99th percentile of Estimation of the PAF took into account the risk the global distribution of exposure. The SEV is on a function and the distribution of exposure across 0–100 scale where 100 means the entire population is at individuals in each age­ sex­ location­ year. By drawing maximum risk and 0 means everyone in the population 1000 samples from the risk function, 1000 distributions is at minimum risk. We computed age­ standardised of exposure for each age­ sex­ location­ year, and SEVs by age­ standardising age­ specific SEVs across the 1000 samples from the TMREL, we propagated all of age groups in which that risk factor was evaluated; this these sources of uncertainty into the PAF distributions. method is a change from GBD 2017 in which age­ PAFs were also applied at the draw level to the uncer­ standardisation included age groups in which the risk tainty distributions of each associated outcome for that was not evaluated. For example, the SEV for low age­ sex­ location­ year. birthweight is now age­ standardised across age groups 0–6 days to 7–27 days. Estimating the PAF and attributable burden for To estimate SEVs for groups of risk factors, we first combinations of risk factors estimated the value of RR without mediation through For the estimation of each specific risk factor, the risk 1 (RR ). 2/1 counterfactual distribution of exposure is the TMREL for that specific risk with no change in other risk factors. RR =MF (RR − 1) + 1 2/1 2/1 2 Thus, the sum of these risk­ specific estimates of attri­ butable burden can exceed 100% for some causes, such where RR is the relative risk of risk factor 2 and MF is 2 2/1 as cardiovascular diseases. It is also useful to assess the the mediation factor, or the proportion of the risk of risk PAF and attributable burden for combinations of factor 2 that is mediated through risk factor 1. We then risk factors, such as all diet components together or computed the PAF using the non­ mediated relative risk household air and ambient particulate matter pollution. (RR ) and computed the joint PAF as 1/2 To estimate the combined effects of risk factors, we should take into account how one risk factor might be PAF =1 −∏ (1 – PAF ). 1..j j mediated through another (eg, the effect of fruit intake j=1 might be partly mediated through fibre intake). We used the mediation matrix as developed in GBD 2017 to try We cannot simply multiply RR values used for the max to correct for overestimation of the PAF and the attri­ SEV of each component risk as this would exaggerate butable burden for combinations of risks if we were to the joint RR . We approximated the 99th percentile of max simply assume independence without any mediation. risk for the combination of risk factors by taking the Appendix 1 (section 5, table S6) provides the estimated geometric mean of the ratio between the individual mediation matrix. risk maximum risk and the individual risk global mean risk and multiplied that by the global mean joint risk. Summary exposure value Formally, As in previous rounds of GBD, we summarised expo­ sure distributions for dichotomous, polytomous, and RR N(r) max ∏RR ∏ global mean — continuous risk factors using the SEV. The SEV r r RR global mean compares the distribution of excess risk times exposure level to a population where everyone is at maximum risk. where N(r) is the total number of risks. ∫ P(x)RR(x)dx – 1 x=l Risk-deleted death rates SEV = rc RR – 1 We computed risk­ deleted death rates as the death max rates that would be observed if all risk factors were set For a given risk r and outcome c pair where RR is to their respective TMRELs. This was calculated as max the relative risk at the 99th percentile of the global the death rate in each age­ sex group multiplied by www.thelancet.com Vol 396 October 17, 2020 1227 Global Health Metrics 1 minus the all­ risk PAF for that age­ sex group in each (annual rate of change larger than –0·5%), substantial location. increases (annual rate of change greater than 0·5%) and the remainder of risks with either non­ significant rates Role of the funding source of change or significant rates of change between –0 ·5% The funders of the study had no role in study design, and 0·5%. The declining risks fall into two categories. data collection, data analysis, data interpretation, or First, a set of risks that are strongly linked to social writing of the report. The corresponding author had full and economic development, measured by the Socio­ access to all the data in the study and had final demographic Index (SDI): household air pollution; responsibility for the decision to submit for publication. unsafe water, sanitation, and handwashing; child growth failure; vitamin A deficiency; and zinc deficiency. The Results second set of declining risks includes tobacco smoking Global exposure to risks and lead, which historically have not been negatively The table shows the trends in risk exposure for each risk correlated with SDI. These risks could in fact increase as factor at the global level over two time intervals: the full countries and territories increase SDI, at least for a phase duration of the study, 1990–2019, and the past decade, in the development process. For a long list of risk factors, 2010–19. On the basis of this table, we can divide risks including some large risks, the annual rate of change into three groups based on the percentage change in was either statistically insignificant (p>0·05) or the the global SEV from 2010 to 2019: substantial declines annual rate of change was between –0·5% and 0·5% per SEV 1990 SEV 2010 SEV 2019 ARC 1990–2019 ARC 2010–19 All risk factors 23·09 (20·22 to 25·67) 21·21 (18·04 to 24·26) 21·22 (18·05 to 24·42) –0·29% (-0·46 to -0·15)* 0·00% (–0·18 to 0·20) Environmental and occupational risks 52·55 (48·66 to 55·92) 48·50 (44·44 to 52·15) 45·36 (41·16 to 49·19) –0·51% (–0·62 to –0·40)* –0·74% (–0·88 to –0·61)* Unsafe water, sanitation, and 55·40 (54·39 to 56·61) 49·70 (48·99 to 50·47) 47·13 (46·51 to 47·84) –0·56% (–0·61 to –0·51)* –0·59% (–0·67 to –0·52)* handwashing Unsafe water source 42·78 (41·06 to 44·39) 36·29 (34·57 to 37·92) 32·74 (30·82 to 34·41) –0·92% (–1·08 to –0·76)* –1·14% (–1·52 to –0·77)* Unsafe sanitation 56·28 (54·14 to 58·38) 38·21 (35·98 to 40·80) 28·93 (26·81 to 31·24) –2·29% (–2·52 to –2·07)* –3·09% (–3·68 to –2·47)* No access to handwashing facility 36·77 (36·54 to 37·03) 34·05 (33·80 to 34·32) 32·19 (31·92 to 32·48) –0·46% (–0·50 to –0·42)* –0·63% (–0·70 to –0·56)* Air pollution 45·11 (32·85 to 56·03) 38·36 (28·33 to 48·55) 34·68 (25·76 to 44·37) –0·91% (–1·21 to –0·60)* –1·12% (–1·48 to –0·81)* Particulate matter pollution 44·22 (31·97 to 55·06) 37·56 (27·57 to 47·75) 33·94 (25·11 to 43·56) –0·91% (–1·24 to –0·61)* –1·13% (–1·48 to –0·81)* Ambient particulate matter pollution 15·65 (10·62 to 21·58) 22·98 (18·28 to 27·62) 26·22 (21·57 to 30·50) 1·78% (0·95 to 2·71)* 1·46% (0·81 to 2·10)* Household air pollution from solid fuels 27·08 (16·20 to 38·13) 16·33 (9·59 to 24·52) 11·71 (6·64 to 18·27) –2·89% (–3·60 to –2·25)* –3·70% (–4·64 to –2·88)* Ambient ozone pollution 47·56 (22·76 to 60·54) 54·34 (29·48 to 65·36) 55·06 (32·21 to 67·16) 0·51% (0·27 to 1·24)* 0·15% (–0·10 to 1·08) Non-optimal temperature 29·57 (26·06 to 33·72) 30·21 (26·17 to 34·83) 29·53 (25·41 to 34·26) 0·00% (–0·13 to 0·11) –0·25% (–0·39 to –0·13)* High temperature 25·98 (22·07 to 30·21) 29·25 (24·92 to 33·82) 29·59 (25·16 to 34·26) 0·45% (0·29 to 0·59)* 0·13% (–0·01 to 0·26) Low temperature 33·21 (29·24 to 37·58) 33·47 (29·06 to 38·25) 32·92 (28·44 to 37·82) –0·03% (–0·13 to 0·06) –0·18% (–0·31 to –0·07)* Other environmental risks 50·81 (40·53 to 59·86) 45·11 (34·46 to 55·29) 39·67 (29·01 to 50·86) –0·85% (–1·18 to –0·55)* –1·43% (–1·95 to –0·93)* Residential radon 18·54 (12·37 to 25·82) 18·20 (12·23 to 25·41) 18·12 (12·17 to 25·43) –0·08% (–0·27 to 0·10) –0·05% (–0·25 to 0·14) Lead exposure 68·52 (53·18 to 80·97) 59·82 (43·52 to 74·40) 51·26 (35·09 to 67·32) –1·00% (–1·43 to –0·63)* –1·72% (–2·40 to –1·09)* Occupational risks 3·36 (2·99 to 3·90) 3·33 (2·97 to 3·89) 3·32 (2·96 to 3·87) –0·05% (–0·15 to 0·05) –0·05% (–0·22 to 0·13) Behavioural risks 16·80 (14·82 to 19·05) 15·38 (13·28 to 17·72) 15·09 (12·96 to 17·43) –0·37% (–0·50 to –0·25)* –0·21% (–0·36 to –0·07)* Child and maternal malnutrition 20·05 (19·06 to 21·19) 17·77 (16·61 to 19·07) 17·23 (15·98 to 18·55) –0·52% (–0·67 to –0·40)* –0·34% (–0·51 to –0·18)* Suboptimal breastfeeding 21·66 (20·28 to 22·96) 20·05 (18·26 to 21·34) 19·34 (17·42 to 20·68) –0·39% (–0·55 to –0·31)* –0·40% (–0·61 to –0·21)* Non-exclusive breastfeeding 21·34 (14·67 to 29·82) 19·40 (13·38 to 27·18) 18·39 (12·91 to 25·53) –0·51% (–0·61 to –0·40)* –0·59% (–0·83 to –0·31)* Discontinued breastfeeding 12·33 (12·04 to 12·65) 10·73 (10·50 to 10·99) 10·24 (9·96 to 10·54) –0·64% (–0·77 to –0·52)* –0·52% (–0·87 to –0·17)* Child growth failure 4·93 (4·41 to 5·57) 4·21 (3·70 to 4·78) 3·53 (3·01 to 4·10) –1·15% (–1·43 to –0·83)* –1·95% (–2·37 to –1·50)* Child underweight 13·32 (11·73 to 14·71) 10·51 (8·98 to 11·97) 8·13 (6·50 to 9·68) –1·70% (–2·05 to –1·45)* –2·86% (–3·54 to –2·37)* Child wasting 5·28 (4·50 to 5·98) 5·23 (4·41 to 5·97) 4·89 (4·08 to 5·61) –0·26% (–0·34 to –0·21)* –0·74% (–0·88 to –0·64)* Child stunting 24·07 (16·71 to 26·41) 19·65 (13·76 to 22·01) 16·24 (11·45 to 18·72) –1·36% (–1·63 to –1·17)* –2·11% (–2·68 to –1·74)* Low birthweight and short gestation 11·92 (10·66 to 13·44) 11·32 (10·15 to 12·67) 11·10 (9·99 to 12·42) –0·25% (–0·46 to –0·10)* –0·21% (–0·49 to 0·02) Short gestation 13·88 (12·81 to 15·20) 13·04 (12·19 to 13·96) 13·17 (12·30 to 14·13) –0·18% (–0·43 to –0·01)* 0·11% (–0·22 to 0·39) Low birthweight 11·03 (10·41 to 11·81) 10·11 (9·68 to 10·52) 9·69 (9·28 to 10·14) –0·45% (–0·69 to –0·28)* –0·47% (–0·76 to –0·21)* Iron deficiency 22·65 (21·51 to 23·98) 20·11 (18·78 to 21·59) 19·57 (18·11 to 21·12) –0·50% (–0·65 to –0·38)* –0·30% (–0·47 to –0·14)* Vitamin A deficiency 33·42 (30·78 to 36·10) 22·00 (19·70 to 24·45) 15·01 (13·55 to 16·86) –2·76% (–3·13 to –2·30)* –4·25% (–5·02 to –3·47)* Zinc deficiency 13·84 (5·91 to 24·06) 11·88 (4·96 to 21·34) 8·78 (2·89 to 17·60) –1·57% (–2·57 to –1·07)* –3·35% (–6·44 to –2·04)* (Table continues on next page) 1228 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics SEV 1990 SEV 2010 SEV 2019 ARC 1990–2019 ARC 2010–19 (Continued from previous page) Tobacco 30·54 (29·08 to 32·10) 25·32 (24·00 to 26·80) 24·03 (22·75 to 25·44) –0·83% (–0·89 to –0·77)* –0·58% (–0·69 to –0·47)* Smoking 14·85 (13·27 to 16·56) 12·41 (11·08 to 13·94) 11·14 (9·93 to 12·54) –0·99% (–1·04 to –0·94)* –1·20% (–1·29 to –1·11)* Chewing tobacco 4·58 (4·18 to 4·98) 4·95 (4·71 to 5·20) 5·11 (4·80 to 5·44) 0·37% (0·03 to 0·76)* 0·36% (–0·32 to 1·05) Secondhand smoke 43·20 (42·80 to 43·62) 37·76 (37·32 to 38·19) 37·51 (37·00 to 38·09) –0·49% (–0·54 to –0·43)* –0·07% (–0·20 to 0·06) Alcohol use 6·50 (4·62 to 8·84) 6·68 (4·81 to 9·02) 6·99 (4·98 to 9·41) 0·25% (0·00 to 0·56) 0·50% (0·05 to 0·95)* Drug use 0·18 (0·12 to 0·28) 0·18 (0·13 to 0·27) 0·19 (0·14 to 0·27) 0·28% (–0·19 to 0·69) 0·53% (0·06 to 0·97)* Dietary risks 51·31 (40·44 to 62·42) 48·28 (36·60 to 60·37) 47·10 (35·39 to 59·62) –0·30% (–0·50 to –0·15)* –0·28% (–0·50 to –0·10)* Diet low in fruits 66·70 (59·36 to 75·08) 59·09 (51·17 to 67·81) 56·86 (49·36 to 65·37) –0·55% (–0·71 to –0·42)* –0·43% (–0·58 to –0·29)* Diet low in vegetables 51·32 (38·33 to 65·78) 40·29 (29·88 to 52·52) 40·24 (29·59 to 52·46) –0·84% (–0·93 to –0·74)* –0·02% (–0·14 to 0·10) Diet low in legumes 69·46 (36·73 to 91·69) 61·20 (28·89 to 84·10) 59·67 (27·55 to 83·28) –0·52% (–1·08 to –0·32)* –0·28% (–0·67 to 0·00) Diet low in whole grains 79·92 (72·52 to 87·44) 79·57 (72·09 to 87·12) 78·81 (71·06 to 86·78) –0·05% (–0·07 to –0·03)* –0·11% (–0·17 to –0·06)* Diet low in nuts and seeds 57·76 (29·48 to 73·08) 50·13 (25·10 to 68·03) 47·47 (23·73 to 66·35) –0·68% (–0·92 to –0·29)* –0·61% (–0·91 to –0·26)* Diet low in milk 80·09 (68·47 to 89·10) 80·81 (70·31 to 89·37) 82·54 (71·88 to 91·12) 0·10% (0·05 to 0·18)* 0·23% (0·16 to 0·33)* Diet high in red meat 40·50 (33·75 to 47·06) 43·15 (36·95 to 49·10) 43·94 (38·03 to 49·58) 0·28% (0·15 to 0·47)* 0·20% (–0·04 to 0·50) Diet high in processed meat 30·95 (20·80 to 42·39) 30·56 (20·13 to 43·05) 29·81 (19·04 to 43·32) –0·13% (–0·39 to 0·12) –0·27% (–0·69 to 0·10) Diet high in sugar-sweetened beverages 29·97 (22·97 to 42·54) 29·35 (21·94 to 41·88) 30·36 (22·71 to 43·05) 0·04% (–0·43 to 0·37) 0·38% (–0·22 to 0·76) Diet low in fiber 36·87 (25·93 to 47·86) 31·43 (21·20 to 41·62) 27·62 (18·60 to 36·95) –1·00% (–1·23 to –0·81)* –1·43% (–1·78 to –1·11)* Diet low in calcium 52·64 (43·62 to 64·79) 48·63 (38·79 to 62·22) 46·02 (35·93 to 60·32) –0·46% (–0·68 to –0·23)* –0·61% (–0·89 to –0·31)* Diet low in seafood omega-3 fatty acids 96·35 (93·21 to 99·89) 93·13 (89·11 to 98·47) 93·52 (88·71 to 99·41) –0·10% (–0·18 to –0·01)* 0·05% (–0·07 to 0·15) Diet low in polyunsaturated fatty acids 69·53 (49·68 to 82·70) 62·66 (37·55 to 79·83) 61·86 (35·56 to 80·13) –0·40% (–1·08 to –0·08)* –0·14% (–0·50 to 0·14) Diet high in trans fatty acids 50·54 (43·82 to 63·48) 45·22 (38·20 to 58·98) 44·67 (37·57 to 58·75) –0·43% (–0·58 to –0·17)* –0·14% (–0·41 to 0·08) Diet high in sodium 48·42 (32·26 to 64·13) 46·04 (28·63 to 62·81) 44·97 (27·44 to 62·14) –0·25% (–0·59 to –0·09)* –0·26% (–0·60 to –0·07)* Intimate partner violence 22·48 (13·03 to 30·15) 22·17 (13·13 to 29·08) 22·98 (13·31 to 30·37) 0·07% (0·00 to 0·16) 0·40% (0·00 to 0·73) Childhood sexual abuse and bullying 7·55 (4·99 to 11·23) 8·46 (5·63 to 12·84) 9·10 (6·04 to 13·85) 0·65% (0·49 to 0·79)* 0·81% (0·64 to 0·96)* Childhood sexual abuse 8·68 (6·85 to 10·90) 8·65 (6·89 to 10·78) 9·36 (7·40 to 11·79) 0·26% (0·18 to 0·33)* 0·87% (0·60 to 1·15)* Bullying victimisation 5·51 (2·34 to 11·04) 6·83 (3·04 to 13·36) 7·31 (3·25 to 14·34) 0·98% (0·82 to 1·28)* 0·76% (0·54 to 0·90)* Unsafe sex ·· ·· ·· ·· ·· Low physical activity 3·34 (1·79 to 6·00) 3·43 (1·90 to 6·08) 3·54 (1·95 to 6·26) 0·20% (0·06 to 0·41)* 0·37% (–0·13 to 0·87) Metabolic risks 14·90 (12·02 to 18·55) 19·40 (16·12 to 23·38) 22·14 (18·63 to 26·36) 1·37% (1·17 to 1·56)* 1·46% (1·26 to 1·69)* High fasting plasma glucose 7·88 (6·96 to 8·85) 10·41 (9·43 to 11·42) 11·72 (10·56 to 12·94) 1·37% (1·27 to 1·46)* 1·32% (1·01 to 1·64)* High LDL cholesterol 35·68 (32·92 to 38·73) 32·67 (29·73 to 35·84) 32·44 (29·49 to 35·57) –0·33% (–0·38 to –0·28)* –0·08% (–0·12 to –0·05)* High systolic blood pressure 27·12 (25·51 to 28·87) 26·50 (24·51 to 28·46) 27·74 (25·70 to 29·72) 0·08% (–0·12 to 0·28) 0·51% (0·04 to 1·00)* High body-mass index 11·09 (7·96 to 15·23) 16·46 (12·79 to 21·04) 19·45 (15·57 to 24·39) 1·94% (1·56 to 2·35)* 1·86% (1·55 to 2·19)* Low bone mineral density 17·06 (12·11 to 23·39) 16·42 (11·66 to 22·72) 16·26 (11·41 to 22·60) –0·16% (–0·25 to –0·10)* –0·10% (–0·34 to 0·09) Kidney dysfunction 20·56 (14·29 to 27·97) 22·35 (15·82 to 29·79) 22·74 (16·24 to 30·25) 0·35% (0·26 to 0·47)* 0·19% (0·13 to 0·28)* Data in parentheses are 95% uncertainty intervals. SEVs are measured on a 0 to 100 scale, in which 100 is when the entire population is exposed to maximum risk and 0 is when the entire population is at minimum risk. SEVs are shown for all levels of the risk factor hierarchy. ARC=annualised rate of change. SEVs=summary exposure values. *Statistically significant increase or decrease. Table: Global age-standardised SEVs for both sexes combined in 1990, 2010, and 2019, and annualised rate of change between 1990 and 2019 and 2010 and 2019 year: ambient ozone pollution, high temperature, low pollution, alcohol use, drug use, childhood sexual temperature, residential radon, occupational risks, sub­ abuse, bullying victimisation, high FPG, high SBP, and optimal breastfeeding, short gestation, low birthweight, high BMI. Many of the increasing risks are metabolic iron deficiency, chewing tobacco, dietary risks as a group, risk factors; in fact, taken together, the exposure to intimate partner violence, low physical activity, high metabolic risks increased 1·37% per year (95% UI LDL cholesterol, low bone mineral density, and kidney 1·17–1·56) from 1990 to 2019 and 1·46% per year dysfunction. Many of these stagnating risks have been or (1·26–1·69) from 2010 to 2019. Figure 1A, which shows are targets of concerted public health efforts spanning the trends in the age­ standardised SEV for each risk public policy, targeted programmes, and primary care factor compared with the fraction of global DALYs intervention. attributable to each risk factor, further emphasises Concerning for both current and future health are these patterns. In 2019, there were three risks that the exposures that are increasing at more than 0·5% per accounted for more than 1% of DALYs and were year. This list includes ambient particulate matter increasing in age­ standardised SEVs by more than www.thelancet.com Vol 396 October 17, 2020 1229 Global Health Metrics Behavioural Environmental or occupational Metabolic Global High SDI 2·0 3·0 High body-mass index Drug use Ambient particulate matter pollution High fasting plasma glucose 1·0 2·0 Alcohol use High systolic blood pressure High fasting plasma glucose Drug use Kidney dysfunction Secondhand Short gestation Low temperature smoke High systolic blood pressure Dietary risks Occupational risks High LDL cholesterol Iron deficiency 1·0 High body-mass index Low birthweight Kidney dysfunction No access to handwashing facility Low bone –1·0 Smoking mineral density Alcohol use Unsafe water source Dietary risks Occupational risks Child growth failure High LDL cholesterol Low temperature –2·0 Ambient particulate matter pollution –1·0 –3·0 Unsafe sanitation Household air pollution from solid fuels Smoking –4·0 –2·0 0 2·5 5·0 7·5 10·0 0 5·0 10·0 15·0 High–middle SDI Middle SDI 2·0 2·5 High body-mass index Alcohol use High body-mass index High fasting plasma glucose Kidney Drug use Short gestation Ambient particulate matter pollution High systolic blood pressure dysfunction Secondhand smoke High LDL cholesterol Dietary risks Low birthweight Low temperature –1·0 Occupational risks Smoking Child growth failure Alcohol use Lead exposure –2·5 Short gestation High fasting plasma glucose Low birthweight Kidney dysfunction Dietary risks Occupational risks High LDL cholesterol Drug use High systolic blood pressure Secondhand smoke Ambient particulate matter pollution –5·0 Low temperature Smoking Household air pollution from solid fuels –1·0 –7·5 0 5·0 10·0 15·0 0 5·0 10·0 15·0 Low–middle SDI Low SDI 5·0 5·0 Ambient particulate matter pollution Ambient particulate matter pollution High body-mass index High body-mass index 2·5 Alcohol use High fasting plasma glucose 2·5 High fasting plasma glucose High systolic blood pressure Kidney Secondhand smoke dysfunction High LDL cholesterol Alcohol use Kidney Short gestation High systolic blood pressure Dietary risks dysfunction Iron deficiency Occupational risks Short gestation Occupational risks Smoking Dietary risks Low birthweight No access to Unsafe water source High LDL cholesterol Lead No access to handwashing facility Non-exclusive breastfeeding handwashing exposure Iron deficiency facility Child growth failure Low birthweight Unsafe water source –2·5 Smoking Unsafe sanitation Child growth failure Unsafe sanitation –2·5 Household air pollution from solid fuels –5·0 Household air pollution from solid fuels –7·5 –5·0 0 2·5 5·0 7·5 10·0 0 5·0 10·0 15·0 DALYs attributable to each risk, 2019 (%) DALYs attributable to each risk, 2019 (%) (Figure 1 continues on next page) 1230 www.thelancet.com Vol 396 October 17, 2020 ARC in age-standardised SEVs, 2010–19 (%) ARC in age-standardised SEVs, 2010–19 (%) ARC in age-standardised SEVs, 2010–19 (%) Global Health Metrics 5·0 Global Low SDI Low–middle SDI Middle SDI High–middle SDI 4·0 High SDI 3·0 2·0 1·0 –1·0 –2·0 High systolic Smoking High fasting Low birthweight High body-mass Short gestation Ambient particulate High LDL Alcohol use blood pressure plasma glucose index matter pollution cholesterol Figure 1: ARC in age-standardised SEVs, globally and by SDI quintile, 2010–19 (A) Level 4 risks and occupational risks, dietary risks, and child growth failure, compared with percentage of DALYs attributable to each risk. (B) Top nine Level 4 risks by attributable DALYs. Only risk factors causing more than 1% of DALYs are shown in panel A. SEVs are measured on a 0–100 scale in which 100 is when the entire population is exposed to maximum risk and 0 is when the entire population is at minimum risk. ARC=annualised rate of change. DALYs=disability-adjusted life-years. SDI=Socio-demographic Index. SEVs=summary exposure values. 1% per year, dominating the figure: high FPG, high in this study leads to large percentage reductions in BMI, and ambient particulate matter pollution. Reductions mortality in those younger than 5 years and in the middle in risks that currently still have large attributable burden and older age groups. Risk reduction can have a slightly are almost exclusively those inversely associated with larger effect on male mortality than female mortality; in rising SDI, except smoking. It might be assumed that other words, some of the difference between male and effective efforts to reduce risk exposure have been female life expectancy can be explained by risk exposures. concentrated on the world’s largest risk factors, but we The percentage of age­ specific mortality explained by all see no discernible pattern between trends in exposure risk factors combined in 1990 is very similar to the and attributable burden. The global trends shown in share shown in figure 2A (appendix 2 table S3). Figure 2B the table and figure 1 give a high­ level view of how well shows the annualised rate of decline in risk­ deleted the world is managing exposure to an extensive list age­ specific mortality from 1990 to 2019. Risk ­ deleted of harmful risks, but regional and country trends can mortality rates declined from 1990 to 2019 in all age be markedly variable. Figure 1B shows trends for groups other than in those aged 95 years and older, the largest risks in terms of global attributable age­ declining between 1·0% and 3·3% per year for all the age standardised DALY rates for countries grouped into groups younger than 75 years, and at lower rates for quintiles of SDI in 2019. There is considerable variation those aged 75 years and older. The substantial declines in across quintiles in trends in exposure. Notably, ambient risk­ deleted mortality rates are likely to be related to particulate matter pollution exposure is increasing in reductions in risks not included in our assessment, the low SDI up to middle SDI quintiles but decreasing reductions in case­ fatality rates, or other factors. The in the high SDI quintile. High FPG and high BMI are observed rates of decline for all­ cause mortality for ages increasing in all quintiles, as is alcohol use. Smoking is younger than 10 years and older than 65 years have been declining in all SDI quintiles. Regional and national faster than the risk­ deleted rates, suggesting reduction trends in SEVs are available in appendix 2 (table S1). of risks included in our analysis has played a role in Figure 2A provides an alternative way to consider progress in these age groups, particularly in those the link between risk exposures and overall trends in younger than 5 years. Notably, risk­ deleted death rates mortality. Removing the effect of all risk factors included have declined faster than observed rates, particularly for www.thelancet.com Vol 396 October 17, 2020 1231 ARC in age-standardised SEVs, 2010–19 (%) Global Health Metrics five risks for attri butable deaths for females were high SBP (5·25 million [95% UI 4·49–6·00] deaths, or 20·3% [17·5–22·9] of all female deaths in 2019), dietary risks (3·48 million [2·78–4·37] deaths, or 13·5% [10·8–16·7] of –10 all female deaths in 2019), high FPG (3·09 million [2·40–3·98] deaths, or 11·9% [9·4–15·3] of all female –20 deaths in 2019), air pollution (2·92 million [2·53–3·33] deaths or 11·3% [10·0–12·6] of all female deaths in 2019), –30 and high BMI (2·54 million [1·68–3·56] deaths or 9·8% [6·5–13·7] of all female deaths in 2019). For males, the top five risks differed slightly. In 2019, the leading Level 2 risk –40 factor for attributable deaths globally in males was tobacco (smoked, second­ hand, and chewing), which accounted –50 for 6·56 million (95% UI 6·02–7·10) deaths (21·4% [20·5–22·3] of all male deaths in 2019), followed by high –60 SBP, which accounted for 5·60 million (4·90–6·29) deaths (18·2% [16·2–20·1] of all male deaths in 2019). The third –70 largest Level 2 risk factor for attributable deaths among Females males in 2019 was dietary risks (4·47 million [3·65–5·45] Males deaths, or 14·6% [12·0–17·6] of all male deaths in 2019) –80 followed by air pollution (ambient particulate matter and ambient ozone pollution, accounting for 3·75 million 0·5 [3·31–4·24] deaths (12·2% [11·0–13·4] of all male deaths in 2019), and then high FPG (3·14 million [2·70–4·34] deaths, or 11·1% [8·9–14·1] of all male deaths in 2019). Outside of the top five, there were large differences between attributable deaths in males and females due to –0·5 alcohol use, which accounted for 2·07 million (1·79–2·37) deaths in males and 0·374 million (0·298–0·461) deaths in –1·0 females in 2019. Newly included in GBD 2019, non­ optimal temperature accounted for 1·01 million (0·880–1·15) –1·5 deaths in males and 0·946 million (0·812–1·09) deaths in females. For both sexes combined, the leading Level 2 risk –2·0 factor for deaths was high SBP, accounting for 10·8 million (9·51–12·1) deaths in 2019 (19·2% [16·9–21·3] of all deaths that year), followed by tobacco, with 8·71 million –2·5 (8·12–9·31) deaths (15·4% [14·6–16·2] of all deaths that year). –3·0 When viewed in terms of DALYs (figure 3C, D), the ranking of Level 2 risk factors reflects the differential ages –3·5 of death and the contribution of non­ fatal disease burden. Most notably, child and maternal malnutrition –4·0 (including low birthweight, short gestation, child growth failure, non­ optimal breastfeeding, and low intake of micronutrients), which has predominant health effects among the young, was the second leading Level 2 risk Age (neonatal stage or years) factor for males and leading risk factor for females in Figure 2: Change in global mortality rates after risk deletion, by age group and sex 2019, accounting for 11·5% (95% UI 10·1–13·1) of DALYs (A) Percentage change in age-specific mortality rates in 2019 after removing the effect of all risk factors in this for males and 11·7% (10·5–13·2) of DALYs for females. study. (B) ARC in risk-deleted mortality rates from 1990 to 2019. ARC=annualised rate of change. Tobacco was ranked first for males and seventh for women aged between 25 and 59 years, implying that risk females in terms of attributable DALYs. For both sexes exposure has increased in those age groups. combined, the leading Level 2 risk factor globally for attributable DALYs was child and maternal malnutri­ Global attributable burden tion, at 295 million (95% UI 253–350) DALYs in 2019 Figure 3A and 3B show global attributable deaths for (11·6% [10·3–13·1] of all DALYs that year). females and males in 2019 for the 20 risk factors at Level 2 Figure 4 shows the ranking of Level 2 risk factors by of the risk factor hierarchy (appendix 2 table S3) The top attributable DALYs, both for SDI quintiles and the 21 GBD 1232 www.thelancet.com Vol 396 October 17, 2020 Early neonatal Late neonatal Post neonatal 1–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 80–84 85–89 90–94 ≥95 ARC in risk-deleted mortality rates, 1990–2019 (%) Change in mortality rate due to risk deletion, 2019 (%) Global Health Metrics regions. Risk factors are shaded by the trend in the handwashing. In the low­ middle SDI quintile, malnutri­ attributable DALY rates over the past decade. In the tion and air pollution were still the leading risk factors for low SDI quintile, the most important risk factors were attributable DALYs, but high SBP rose to third. In the malnutrition; air pollution; and water, sanitation, and middle to high SDI quintiles, the leading risks transitioned A Global attributable deaths from Level 2 risk factors for females in 2019 High systolic blood pressure Dietary risks High fasting plasma glucose Air pollution High body-mass index Tobacco High LDL cholesterol Kidney dysfunction Cardiovascular diseases Child and maternal malnutrition Chronic respiratory diseases Non-optimal temperature Diabetes and kidney diseases Digestive diseases Unsafe water, sanitation, and handwashing Enteric infections Unsafe sex HIV/AIDS and sexually transmitted infections Maternal and neonatal disorders Low physical activity Musculoskeletal disorders Alcohol use Neoplasms Neurological disorders Other environmental risks Nutritional deficiencies Occupational risks Other infectious diseases Other non-communicable diseases Low bone mineral density Respiratory infections and tuberculosis Self-harm and interpersonal violence Drug use Substance use disorders Intimate partner violence Transport injuries Unintentional injuries Childhood sexual abuse and bullying B Global attributable deaths from Level 2 risk factors for males in 2019 Tobacco High systolic blood pressure Dietary risks Air pollution High fasting plasma glucose High body-mass index High LDL cholesterol Alcohol use Kidney dysfunction Child and maternal malnutrition Non-optimal temperature Occupational risks Unsafe water, sanitation, and handwashing Other environmental risks Low physical activity Drug use Unsafe sex Low bone mineral density Childhood sexual abuse and bullying Intimate partner violence 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 Number of deaths (in 1000s) (Figure 3 continues on next page) www.thelancet.com Vol 396 October 17, 2020 1233 Global Health Metrics C Global attributable DALYs from Level 2 risk factors for females in 2019 Child and maternal malnutrition High systolic blood pressure Air pollution High fasting plasma glucose High body-mass index Dietary risks Tobacco Unsafe water, sanitation, and handwashing Cardiovascular diseases Chronic respiratory diseases High LDL cholesterol Diabetes and kidney diseases Kidney dysfunction Digestive diseases Enteric infections Unsafe sex HIV/AIDS and sexually transmitted infections Occupational risks Maternal and neonatal disorders Musculoskeletal disorders Non-optimal temperature Neoplasms Alcohol use Neurological disorders Nutritional deficiencies Drug use Other infectious diseases Low bone mineral density Other non-communicable diseases Respiratory infections and tuberculosis Other environmental risks Self-harm and interpersonal violence Intimate partner violence Sense organ diseases Substance use disorders Low physical activity Transport injuries Unintentional injuries Childhood sexual abuse and bullying D Global attributable DALYs from Level 2 risk factors for males in 2019 Tobacco Child and maternal malnutrition High systolic blood pressure Air pollution Dietary risks High fasting plasma glucose High body-mass index Alcohol use High LDL cholesterol Occupational risks Unsafe water, sanitation, and handwashing Kidney dysfunction Non-optimal temperature Drug use Unsafe sex Other environmental risks Low bone mineral density Low physical activity Childhood sexual abuse and bullying 0 2 4 6 8 10 12 14 DALYs (%) Figure 3: Global number of deaths and percentage of DALYs attributable to Level 2 risk factors, by cause and sex, 2019 DALYs=disability-adjusted life-years. to tobacco, high SBP, dietary risks, high BMI, and high dysfunction, dietary risks, and high LDL cholesterol. In FPG. The risk transition is evident across quintiles. Select seven regions, child and maternal malnutrition is the regional patterns stand out. In the Caribbean and central leading risk factor, and in another seven regions, tobacco is Latin America, large increases were seen in attributable the leading risk factor. In the remainder of regions, the burden for high FPG, high BMI, high SBP, kidney leading risk factor is high SBP (four regions), high FPG 1234 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics Annual rate of change in all-age DALYs from 2010 to 2019 –9·7% to <–3·3% –3·3% to <–1·4% –1·4% to <–0·7% –0·7% to <–0·3% –0·3% to <0·0% 0·0% to <0·4% 0·4% to <0·8% 0·8% to <1·1% 1·1% to <1·6% 1·6% to <5·0% 12 345678 910 High fasting High body- High systolic Tobacco Low SDI Malnutrition Air pollution WaSH Dietary risks Unsafe sex Alcohol use blood pressure plasma glucose mass index High fasting High body- High LDL High systolic Dietary risks WaSH Low-middle SDI Malnutrition Air pollution Tobacco Alcohol use plasma glucose mass index cholesterol blood pressure High systolic High body- High LDL High fasting Kidney Middle SDI Tobacco Dietary risks Air pollution Malnutrition Alcohol use blood pressure mass index cholesterol plasma glucose dysfunction High fasting Kidney High systolic High LDL Occupational High body- High-middle SDI Tobacco Dietary risks Air pollution Alcohol use plasma glucose dysfunction risks blood pressure cholesterol mass index High body- High systolic High LDL High fasting Kidney Occupational Dietary risks Alcohol use Drug use High SDI Tobacco cholesterol mass index blood pressure plasma glucose dysfunction risks High systolic High body- High fasting Kidney High LDL Air pollution Andean Latin America Malnutrition Dietary risks Alcohol use Tobacco blood pressure cholesterol mass index plasma glucose dysfunction High body- Kidney High systolic High fasting Occuptional High LDL Alcohol use Australasia Tobacco Dietary risks Drug use mass index blood pressure risks dysfunction plasma glucose cholesterol High fasting High systolic High body- High LDL Kidney Tobacco Dietary risks Air pollution Alcohol use Caribbean Malnutrition plasma glucose blood pressure mass index cholesterol dysfunction Kidney High systolic High body- High fasting High LDL Malnutrition Central Asia Dietary risks Tobacco Air pollution Alcohol use cholesterol dysfunction blood pressure mass index plasma glucose High systolic High body- High fasting High LDL Kidney Occupational Alcohol use Air pollution Central Europe Tobacco Dietary risks blood pressure mass index plasma glucose cholesterol dysfunction risks High fasting High body- Kidney High LDL High systolic Dietary risks Alcohol use Malnutrition Tobacco Central Latin America Air pollution plasma glucose mass index blood pressure dysfunction cholesterol Occupational High systolic High fasting High body- Central sub-Saharan Africa Malnutrition WaSH Air pollution Unsafe sex Alcohol use Dietary risks risks blood pressure plasma glucose mass index High systolic High fasting High body- High LDL Occupational Kidney Dietary risks Alcohol use East Asia Tobacco Air pollution blood pressure plasma glucose mass index cholesterol risks dysfunction High fasting High systolic High body- High LDL Kidney Tobacco Dietary risks Alcohol use Drug use Eastern Europe Air pollution blood pressure plasma glucose mass index cholesterol dysfunction High systolic High fasting High body- Alcohol use Eastern sub-Saharan Africa Malnutrition Air pollution WaSH Unsafe sex Dietary risks Tobacco blood pressure plasma glucose mass index High systolic High fasting High body- Kidney Occupational High LDL Alcohol use High-income Asia Pacific Tobacco Dietary risks Air pollution plasma glucose cholesterol blood pressure mass index dysfunction risks High fasting Kidney High body- High systolic High LDL Occupational Drug use Dietary risks Alcohol use High-income North America Tobacco plasma glucose dysfunction mass index blood pressure cholesterol risks High body- High fasting Kidney Occupational High systolic High LDL Air pollution North Africa and Middle East Malnutrition Tobacco Dietary risks blood pressure mass index plasma glucose dysfunction risks cholesterol High fasting High body- High systolic High LDL Dietary risks WaSH Oceania Malnutrition Air pollution Tobacco Unsafe sex plasma glucose mass index blood pressure cholesterol Kidney High systolic High fasting High body- High LDL Malnutrition Air pollution Tobacco Dietary risks WaSH South Asia plasma glucose dysfunction blood pressure mass index cholesterol High systolic High fasting High body- Kidney High LDL Southeast Asia Tobacco Dietary risks Air pollution Malnutrition Alcohol use blood pressure plasma glucose mass index dysfunction cholesterol High systolic High fasting Kidney High LDL High body- Occupational Dietary risks Alcohol use Southern Latin America Tobacco Malnutrition blood pressure plasma glucose dysfunction cholesterol risks mass index High systolic High body- High fasting Air pollution Southern sub-Saharan Africa Unsafe sex Malnutrition Tobacco Alcohol use WaSH Dietary risks blood pressure mass index plasma glucose High fasting High body- High systolic High LDL Kidney Alcohol use Tropical Latin America Tobacco Dietary risks Malnutrition Air pollution mass index plasma glucose blood pressure cholesterol dysfunction High systolic High fasting High body- High LDL Occupational Kidney Dietary risks Alcohol use Western Europe Tobacco Air pollution blood pressure plasma glucose mass index cholesterol risks dysfunction High systolic High body- High fasting Western sub-Saharan Africa Malnutrition WaSH Air pollution Unsafe sex Dietary risks Alcohol use Tobacco blood pressure mass index plasma glucose Figure 4: Leading ten Level 2 risk factors for attributable DALYs by GBD region and SDI quintile, 2019 For each region and SDI quintile, Level 2 risks are ranked by attributable DALYs from left (first) to right (tenth). Risks are coloured by their annualised rate of change in all-age DALY rates from 2010 to 2019. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study. Malnutrition=child and maternal malnutrition. SDI=Socio-demographic Index. WaSH=water, sanitation, and handwashing. (one region), high BMI (one region), and unsafe sex (one Attributable burden by age group region). The largest rates of increase in attributable DALYs The pattern of risk­ factor­ attributable burden varied have been seen for high FPG in ten of 21 regions and for considerably by age and over time, as shown in figure 5, high BMI in ten of 21 regions. which includes arrows plots for all age groups combined www.thelancet.com Vol 396 October 17, 2020 1235 Global Health Metrics and for five broad age groups (0–9, 10–24, 25–49, 50–74, led to marked changes in risk rankings from 1990 to 2019. and 75 years and older). These figures show specific risk In 1990, the leading risk factors were child wasting, factors at Level 4 of the risk factor hierarchy. Figure 5A low birthweight, short gestation, and household air shows how risk exposure trends, underlying changes in pollution, all of which have dropped substantially in disease rates, and rising mean age of populations have magnitude in terms of percentage of attributable DALYs A All ages Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 Child wasting 11·4 (9·5 to 13·6) 1 High systolic blood pressure 9·3 (8·2 to 10·5) 53·1 (43·0 to 62·7) –27·0 (–31·7 to –22·6) 2 Low birthweight 10·6 (9·9 to 11·4)7 2 Smoking ·9 (7·2 to 8·6) 24·3 (15·9 to 33·9) –39·0 (–43·1 to –34·4) 3 Short gestation 8·7 (8·1 to 9·5) 3 High fasting plasma glucose 6·8 (5·8 to 8·0) 122·9 (110·0 to 135·7) 7·4 (1·5 to 13·8) 4 Household air pollution 8·0 (6·2 to 10·0) 4 Low birthweight 6·3 (5·5 to 7·3) –41·4 (–49·7 to –31·0) –41·3 (–49·6 to –30·8) 5 Smoking 6·2 (5·8 to 6·6) 5 High body-mass index 6·3 (4·2 to 8·6) 138·2 (106·1 to 186·9) 18·0 (2·2 to 42·3) 6 Unsafe water 6·2 (4·7 to 7·6) 6 Short gestation 5·5 (4·7 to 6·3) –38·9 (–47·3 to –28·0) –38·9 (–47·4 to –27·9) 0·3 (–21·2 to 30·7) 7 High systolic blood pressure 5·9 (5·3 to 6·5) 7 Ambient particulate matter 4·7 (3·8 to 5·5) 67·7 (27·9 to 126·1) 8 Child underweight 4·9 (3·9 to 6·3) 8 High LDL cholesterol 3·9 (3·2 to 4·7) 41·5 (31·1 to 50·4) –32·2 (–36·7 to –27·8) 9 Unsafe sanitation 4·6 (3·7 to 5·6) 9 Alcohol use 3·7 (3·3 to 4·1) 37·1 (27·3 to 47·9) –23·7 (–29·2 to –17·7) 10 Handwashing 3·2 (2·3 to 4·0) 10 Household air pollution 3·6 (2·7 to 4·6) –56·1 (–64·7 to –46·0) –68·2 (–74·0 to –61·6) 11 High fasting plasma glucose 3·0 (2·5 to 3·5) 11 Child wasting 3·3 (2·6 to 4·1) –71·7 (–77·4 to –65·2) –72·9 (–78·4 to –66·6) 13 Ambient particulate matter 2·7 (1·8 to 3·8) 13 Unsafe water 2·6 (1·9 to 3·3) –59·3 (–68·1 to –46·7) –65·9 (–73·0 to –55·4) 14 High LDL cholesterol 2·7 (2·2 to 3·2) 17 Unsafe sanitation 1·6 (1·3 to 2·1) 65·5 (–72·9 to –54·8) –71·0 (–77·0 to –61·8) 15 Alcohol use 2·6 (2·3 to 2·9) 19 Handwashing 1·3 (0·9 to 1·8) –58·7 (–65·9 to –49·8) –64·2 (–70·5 to –56·3) 16 High body-mass index 2·6 (1·5 to 4·0) 22 Child underweight 1·1 (0·9 to 1·4) –77·8 (–82·7 to –71·7) –79·5 (–84·0 to –73·8) B 0–9 years Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 Child wasting 24·7 (20·7 to 28·9) 1 Low birthweight 28·9 (27·3 to 30·4) –43·3 (–51·8 to –33·0) –42·6 (–51·2 to –32·2) 2 Low birthweight 23·1 (22·1 to 24·1) 2 Short gestation 24·7 (23·3 to 26·1) –41·2 (–49·6 to –30·2) –40·4 (–49·0 to –29·3) 3 Short gestation 19·0 (18·1 to 19·9) 3 Child wasting 14·8 (12·3 to 17·3) –72·9 (–78·4 to –66·3) –73·6 (–79·1 to –67·3) 4 Household air pollution 11·2 (8·7 to 14·2) 4 Household air pollution 7·7 (6·0 to 9·5) –68·8 (–75·2 to –60·6) –68·9 (–75·4 to –60·9) 5 Unsafe water 11·0 (8·5 to 13·3) 5 Unsafe water 7·7 (5·9 to 9·4) –68·3 (–75·8 to –57·4) –68·9 (–76·4 to –58·6) 6 Child underweight 10·4 (8·2 to 13·3) 6 Unsafe sanitation 5·1 (4·3 to 6·0) –72·0 (–78·7 to –62·0) –72·5 (–79·3 to –63·0) 7 Unsafe sanitation 8·2 (6·8 to 9·7) 7 Handwashing 4·5 (3·2 to 5·8) –66·0 (–72·9 to –57·0) –66·7 (–73·6 to –58·0) 8 Child stunting 6·2 (3·2 to 10·5) 8 Child underweight 4·4 (3·6 to 5·4) –80·8 (–85·2 to –75·3) –81·4 (–85·7 to –76·1) 9 Handwashing 6·0 (4·3 to 7·6) 9 Ambient particulate matter 4·0 (2·8 to 5·2) –23·3 (–45·9 to 11·5) –20·5 (–46·3 to 10·8) 10 Non-exclusive breastfeeding 3·8 (2·8 to 4·9) 10 Child stunting 2·7 (1·3 to 4·8) –80·3 (–85·8 to –74·5) –81·1 (–86·4 to –75·5) 11 Ambient particulate matter 2·3 (1·3 to 3·9) 11 Non-exclusive breastfeeding 2·4 (1·8 to 3·0) –72·1 (–77·8 to –65·3) –72·1 (–77·8 to –65·3) C 10–24 years Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 Occupational injury 3·2 (2·8 to 3·7) 1 Iron deficiency 3·0 (2·3 to 3·8) –0·9 (–11·4 to 9·5) –17·6 (–26·4 to –8·8) 2 Iron deficiency 2·8 (2·1 to 3·6) 2 Alcohol use 2·6 (2·1 to 3·1)– –6·3 (–12·9 to 0·3) 22·6 (–28·0 to –17·2) 3 Unsafe water 2·7 (1·7 to 4·2) 3 Unsafe sex 2·1 (1·5 to 2·9) 108·3 (78·5 to 140·5) 73·4 (47·4 to 98·5) 4 Alcohol use 2·6 (2·1 to 3·0) 4 Unsafe water 2·0 (1·3 to 3·0) –29·8 (–43·6 to –4·4) –40·5 (–53·1 to –20·4) 5 Unsafe sanitation 2·0 (1·3 to 3·1) 5 Occupational injury 1·8 (1·6 to 2·1) –47·7 (–54·0 to –40·8) –56·6 (–61·9 to –51·0) 6 Drug use 1·4 (1·1 to 1·7) 6 Drug use 1·8 (1·4 to 2·3) 20·4 (13·7 to 27·1) –0·6 (–6·2 to 4·9) 7 Handwashing 1·0 (0·7 to 1·5) 7 Short gestation 1·3 (1·0 to 1·6) 84·6 (68·2 to 99·4) 54·1 (40·0 to 66·0) 8 Unsafe sex 1·0 (0·7 to 1·4) 8 Low birthweight 1·3 (1·0 to 1·6) 84·6 (68·2 to 99·4) 54·1 (40·0 to 66·0) 9 Kidney dysfunction 0·9 (0·8 to 1·0) 9 Unsafe sanitation 1·2 (0·9 to 1·8) –42·4 (–53·9 to –21·6) –51·1 (–61·7 to –34·7) 10 Bullying 0·7 (0·2 to 1·4) 10 Bullying 1·1 (0·4 to 2·2) 50·7 (41·3 to 69·4) 26·9 (17·5 to 41·3) 11 Short gestation 0·6 (0·5 to 0·8) 11 Kidney dysfunction 1·1 (0·9 to 1·3) 19·0 (9·1 to 28·6) –1·2 (–9·4 to 6·7) 12 Low birthweight 0·6 (0·5 to 0·8) 12 Handwashing 0·8 (0·6 to 1·1) –28·8 (–41·4 to –8·1) –40·0 (-51·2 to –23·6) Environmental and occupational risks Behavioural risks Metabolic risks (Figure 5 continues on next page) 1236 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics D 25–49 years Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 Alcohol use 6·7 (5·9 to 7·5) 1 Alcohol use 6·3 (5·5 to 7·3) 26·7 (18·0 to 35·7) –23·5 (–28·8 to –18·0) 2 Smoking 6·6 (5·9 to 7·2) 2 High systolic blood pressure 6·0 (4·9 to 7·1) –15·1 (–23·0 to –7·4) 48·4 (34·4 to 61·8) 5·9 (4·2 to 7·8) 3 High systolic blood pressure 5·4 (4·4 to 6·4) 3 High body-mass index 136·1 (95·0 to 203·5) 40·5 (12·1 to 73·9) 4 Occupational injury 3·9 (3·5 to 4·3) 4 Smoking 5·0 (4·5 to 5·6) 1·9 (–5·7 to 9·7) –42·8 (–47·1 to –38·4) 5 High LDL cholesterol 3·5 (3·0 to 4·1) 5 Unsafe sex 4·9 (4·1 to 6·0) 45·1 (26·9 to 67·2) 131·3 (102·8 to 171·1) 4·0 (3·4 to 4·6) 6 Household air pollution6 3·4 (2·6 to 4·3) High fasting plasma glucose 90·9 (76·6 to 104·3) 10·0 (1·9 to 17·8) 7 High body-mass index7 3·3 (1·9 to 5·2) High LDL cholesterol 3·8 (3·1 to 4·5) 41·4 (28·4 to 54·5) –18·9 (–26·2 to –11·6) 8 Unsafe sex 2·8 (2·1 to 3·7) 8 Drug use 2·9 (2·5 to 3·3) 22·9 (16·6 to 30·4) 94·4 (84·5 to 106·7) 2·9 (2·4 to 3·5) 9 High fasting plasma glucose9 2·8 (2·4 to 3·2) Ambient particulate matter 120·4 (76·2 to 180·7) 29·4 (1·8 to 62·5) 10 Drug use 2·0 (1·7 to 2·3) 10 Kidney dysfunction 2·4 (2·0 to 2·7) 62·4 (49·9 to 75·0) –4·2 (–11·8 to 3·1) 11 Kidney dysfunction 1·9 (1·7 to 2·2) 11 Occupational injury 2·3 (2·1 to 2·6) –21·3 (–30·0 to –11·8) –50·4 (–55·9 to –44·5) 1·7 (1·2 to 2·3) –34·2 (–47·2 to –21·0) –61·9 (–69·4 to –54·2) 12 Ambient particulate matter 1·8 (1·2 to 2·3) 12 Household air pollution E 50–74 years Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 Smoking 19·4 (18·2 to 20·6) 1 High systolic blood pressure 16·1 (14·2 to 18·0) 47·7 (36·9 to 58·0) –28·3 (–33·6 to –23·3) 2 High systolic blood pressure 16·8 (14·9 to 18·7) 2 Smoking 15·5 (14·1 to 16·7) 22·6 (13·9 to 32·6) –40·3 (–44·6 to –35·5) 3 Household air pollution 8·5 (6·3 to 10·7) 3 High fasting plasma glucose 12·2 (10·4 to 14·4) 127·2 (113·4 to 141·5) 10·2 (3·4 to 17·0) 4 High fasting plasma glucose 8·3 (7·0 to 9·8) 4 High body-mass index 11·8 (7·9 to 16·0) 138·4 (106·5 to 186·2) 19·1 (0·7 to 39·5) 5 High body-mass index 7·6 (4·3 to 11·6) 5 Ambient particulate matter 6·8 (5·7 to 8·0) 122·5 (78·2 to 185·1) 9·8 (–13·6 to 38·3) 6 High LDL cholesterol 7·0 (5·6 to 8·5) 6 High LDL cholesterol 6·2 (4·9 to 7·7) 37·8 (27·8 to 47·5) –32·6 (–37·5 to –27·8) 7 Alcohol use 5·1 (4·5 to 5·7) 7 Alcohol use 5·0 (4·4 to 5·7) 51·2 (37·6 to 65·1) –25·8 (–32·6 to –19·1) 8 Ambient particulate matter 4·7 (3·3 to 6·3) 8 Kidney dysfunction 4·7 (4·0 to 5·3) 92·8 (80·4 to 105·3) –6·4 (–12·6 to –0·5) –36·7 (–50·4 to –21·6) –69·3 (–76·0 to –62·0) 9 High sodium 4·0 (1·4 to 8·0) 9 Household air pollution 3·5 (2·4 to 4·8) 10 Kidney dysfunction 3·7 (3·2 to 4·2) 10 High sodium 3·4 (1·1 to 7·1) 31·9 (–1·6 to 51·0) –37·1 (–52·1 to –26·5) F ≥75 years Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 High systolic blood pressure 22·0 (18·6 to 25·3) 1 High systolic blood pressure 19·5 (16·3 to 22·7) 69·6 (58·6 to 80·5) –30·0 (–34·3 to –25·7) 2 Smoking 14·8 (13·9 to 15·7) 2 High fasting plasma glucose 13·5 (10·2 to 18·0) 144·5 (130·1 to 158·7) 1·8 (–4·8 to 7·9) 3 High fasting plasma glucose 10·5 (7·8 to 14·4) 3 Smoking 12·3 (11·4 to 13·0) 58·2 (48·9 to 69·1) –31·9 (–35·8 to –27·3) 4 High LDL cholesterol 9·2 (6·0 to 13·2) 4 High body-mass index 7·3 (4·3 to 11·1) 145·1 (123·1 to 180·2) 4·7 (–6·0 to 17·9) 5 Household air pollution 7·8 (5·7 to 10·2) 5 High LDL cholesterol 7·2 (4·5 to 10·6) 50·0 (39·2 to 58·7) –40·2 (–43·5 to –37·1) 6 High body-mass index 5·7 (3·0 to 9·2) 6 Ambient particulate matter 6·7 (5·6 to 7·8) 143·7 (94·6 to 211·9) 4·1 (–18·2 to 31·0) 7 Ambient particulate matter 5·2 (3·7 to 6·8) 7 Kidney dysfunction 5·9 (4·9 to 6·9) 121·7 (108·6 to 134·1) –8·6 (–14·2 to –3·6) 8 Kidney dysfunction 5·1 (4·1 to 6·1) 8 Low temperature 3·4 (2·9 to 3·9) 42·2 (32·5 to 53·1) –41·8 (–45·7 to –37·6) 9 Low temperature 4·6 (3·9 to 5·3) 9 Household air pollution 3·1 (2·1 to 4·3) –24·5 (–41·1 to –4·8) –67·7 (–74·9 to –59·4) 10 Low whole grains 3·5 (1·8 to 4·4) 10 Low whole grains 3·0 (1·6 to 3·9) 66·2 (57·6 to 74·6) –32·3 (–35·7 to –28·8) Environmental and occupational risks Behavioural risks Metabolic risks Figure 5: Leading ten Level 4 risks by attributable DALYs, 1990–2019 For all ages (A), 0–9 years (B), 10–24 years (C), 25–49 years (D), 50–74 years (E), and 75 years and older (F). DALYs=disability-adjusted life-years. and rank by 2019. The leading risks in 2019 were high The largest declines among the leading ten risks were SBP, smoking, high FPG, low birthweight, and high for child growth failure (child underweight, stunting, BMI. Other notable shifts include the large increase in and wasting); water, sanitation, and handwashing; percentage of attributable DALYs and rank for ambient and house hold air pollution. Large but more moderate particulate matter pollution, high LDL cholesterol, and declines in attributable burden occurred for short alcohol use. Among the youngest age group (0–9 years), gestation and low birthweight, with the smallest reduc­ shown in figure 5B, the leading Level 4 risk factors tion observed for ambient particulate matter pollution. were composed exclusively of malnutrition and environ­ Among adolescents and young adults (aged 10–24 years; mental risk factors. Over the 1990–2019 period, there figure 5C), the pattern of risk factor burden was notably were substantial reductions in the burden attributable to different from the 0–9 years age group, with iron these risk factors in both absolute numbers and rates. deficiency, alcohol use, and unsafe sex ranking first to www.thelancet.com Vol 396 October 17, 2020 1237 Global Health Metrics A Child and maternal malnutrition DALYs attributable 0% to <2% 10% to <12·5% 2% to <4% 12·5% to <15% 4% to <6% 15% to <17·5% 6% to <8% 17·5% to <20% 8% to <10% ≥20% Caribbean and central America Persian Gulf Balkan Peninsula Southeast Asia West Africa Eastern Mediterranean Northern Europe B High systolic blood pressure DALYs attributable 0% to <2% 10% to <12·5% 2% to <4% 12·5% to <15% 4% to <6% 15% to <17·5% 6% to <8% 17·5% to <20% 8% to <10% ≥20% Caribbean and central America Persian Gulf Balkan Peninsula Southeast Asia West Africa Eastern Mediterranean Northern Europe (Figure 6 continues on next page) 1238 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics C Tobacco DALYs attributable 0% to <2% 10% to <12·5% 2% to <4% 12·5% to <15% 4% to <6% 15% to <17·5% 6% to <8% 17·5% to <20% 8% to <10% ≥20% Eastern Caribbean and central America Persian Gulf Balkan Peninsula Southeast Asia West Africa Mediterranean Northern Europe D Air pollution DALYs attributable 0% to <2% 10% to <12·5% 2% to <4% 12·5% to <15% 4% to <6% 15% to <17·5% 6% to <8% 17·5% to <20% 8% to <10% ≥20% Eastern Caribbean and central America Persian Gulf Balkan Peninsula Southeast Asia West Africa Mediterranean Northern Europe (Figure 6 continues on next page) www.thelancet.com Vol 396 October 17, 2020 1239 Global Health Metrics E Dietary risks DALYs attributable 0% to <2% 10% to <12·5% 2% to <4% 12·5% to <15% 4% to <6% 15% to <17·5% 6% to <8% 17·5% to <20% 8% to <10% ≥20% Eastern Caribbean and central America Persian Gulf Balkan Peninsula Southeast Asia West Africa Mediterranean Northern Europe Figure 6: Percentage of all DALYs attributable to the five leading Level 2 risk factors, 2019 DALYs attributable to child and maternal malnutrition (A), high systolic blood pressure (B), tobacco (C), air pollution (D), and dietary risks (E). DALYs=disability-adjusted life-years. third for attributable DALYs in this age group in 2019. National findings There were declines in unsafe sex DALYs in the second The leading risk factors for attributable DALYs had half of the study period, but due to rapid increases highly varied geographical patterns, as shown in figure 6, from 1990 to 2004, there was still a 108·3% (95% UI which presents maps of the percentage of burden 78·5–140·5) increase in unsafe sex DALYs from 1990 to attributable to the top five Level 2 risk factors globally 2019. The long­ term consequences of short gestation and in 2019. The highest proportions (greater than 20%) of low birthweight also increased in importance. burden attributable to the leading Level 2 risk factor In the 25–49 years age group (figure 5D), alcohol use was in 2019, child and maternal malnutrition, were seen the leading Level 4 risk factor for attributable burden, in most of western, central, and eastern sub­ Saharan followed by high SBP and then high BMI, smoking, unsafe African regions (figure 6A). In addition, rates greater sex, and high FPG. The number of DALYs increased for all than 20% were seen in Afghanistan, Pakistan, states in the top ten risks, but age­ standardised attributable DALY northern India, Yemen, and Papua New Guinea. Rates rates increased only for high BMI, unsafe sex, high FPG, between 10% and 20% were seen in a diverse set of drug use, and ambient particulate matter pollution. In the central American countries, states in Brazil, Tajikistan, two oldest age groups, the set of leading risks are quite Uzbekistan, Myanmar, regions of the Philippines, and similar to one another, dominated by high SBP at the top, some Indonesian provinces. and followed by other metabolic risk factors including Figure 6B shows the burden attributable to the second high FPG, high BMI, high LDL cholesterol, and kidney leading Level 2 risk factor in 2019, high SBP. Locations dysfunction. Smoking also contributed substantially to with more than 20% of DALYs attributable to high SBP the risk attributable burden in these age groups, ranked included Georgia and most of central and eastern second in ages 50–74 years (figure 5E) and third in ages Europe. Most countries in north Africa and the Middle 75 years and older (figure 5F). In the oldest age group, East had between 10% and 20% of DALYs attributable to low temperature was also one of the top ten risks, high SBP as did states in southern India and many parts although age­standardised attributable DALY rates of southeast Asia. The only countries with less than declined from 1990 to 2019. Sex­ specific rankings by age 2% of all­ age DALYs attributable to high SBP were in group are available in appendix 2 (figures S4, S5). western sub­ Saharan Africa. 1240 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics The third leading Level 2 risk factor, tobacco, is shown DALYs were attributed to present and past exposure for in figure 6C. Locations with more than 20% of DALYs the 87 environmental, occupational, behavioural, and attributable in 2019 include countries in the Balkan metabolic risk factors and combinations of risk factors Peninsula and two provinces in China—Liaoning and included in this analysis. Overall, combined global expo­ Heilongjiang. Most countries in Europe had between sure to the risks included in this study has remained 10% and 20% of DALYs attributable to smoking; Canada, remarkably constant over the past 30 years. Risk­ deleted most states in the USA, Russia, the rest of China, and mortality rates over the same period have declined, many parts of southeast Asia were also in this category. ranging from a 3·3% decline per year in females aged Attributable burden remains less than 6% in most of 1–4 years to a 0·3% decline per year in males aged Mexico, central America, and Andean Latin America. 90–94 years. Despite this overall pattern, reductions in The burden attributable to tobacco is less than 2% in key risks highly correlated with SDI—unsafe water, much of western and eastern sub­ Saharan Africa. sanitation, and handwashing; household air pollution; Figure 6D shows the burden attributable to air pollution child growth failure; and vitamin A and zinc defi­ (ambient particulate matter, household air pollution, and ciencies—have contributed to reductions in global child ambient ozone pollution). No location had more than death rates. Among the most detailed major non­ commu­ 20% of DALYs attributable to air pollution. But a wide nicable disease risks, only tobacco smoking has declined range of countries in western and eastern sub­ Saharan steadily. At the global level in 2019, there were three risk Africa had attributable burden percentages between 10% factors that accounted for more than 1% of DALYs and and 15%. Similarly, nearly all locations in south Asia, were increasing in exposure by more than 1% per year: many parts of southeast Asia, and most provinces in high BMI, ambient particulate matter pollution, and China also had the same levels of attributable burden. The high FPG. There is large scope for public regulatory spatial patterns of the constituent risks included in air policy, community programmes, and primary care inter­ pollution—particularly ambient particulate matter and ventions on risks to have a greater effect on prevention. household air pollution—were quite different (see GBD These broad global patterns mask considerable hetero­ For GBD Compare see https://vizhub.healthdata.org/ Compare for data), with ambient particulate matter geneity in risk levels and trends at the country level, gbd-compare/ pollution playing a much greater role in Asia than in reinforcing the need for country assessments and Africa. country­ specific prevention planning. The fifth most important Level 2 risk factor was the dietary risks that are based on the joint effects of 15 diet Important risk factor trends quality components (figure 6E). In Bulgaria, dietary risks Some risk exposures are highly correlated with social accounted for more than 20% of attributable DALYs. and economic development, as measured by SDI. However, diet accounts for more than 10% of DALYs in As countries and territories increase SDI through many locations in central and eastern Europe, central higher levels of education, particularly among women; Asia, and most of China. The lowest shares of DALYs increased GDP per capita; and improved access to attributed to dietary risks are in sub­ Saharan Africa, modern contraceptives, we should expect progress on particularly countries in the Sahel. these risks. The incremental effect of campaigns, policies, and programmes on top of this social and Risk-specific trends economic development process is yet to be established Two­ page risk­ specific summaries provide detailed in this analysis. Bending the development curve is results on attri butable deaths, YLLs, YLDs, and DALYs possible, as evidenced by the abrupt accelerations in the for a selection of the 87 risk factors in the GBD risk decline in some risk factors, such as the recent decline hierarchy. These summaries include 2019 counts, age­ in unsafe sanitation in India. Even for risks that standardised rates, and rankings for attributable burden; are historically highly correlated with SDI, intervention the com position of attributable burden for leading can accelerate progress. The range of policy initiatives to causes; patterns of attributable burden over time and accelerate the transition to cleaner cooking fuels is 21,22 age; and age­ standardised SEVs by location and SDI. another example of this effort. Analysis of exemplars, They were written to increase the accessibility to and countries with lower SEVs for these risks for their level transparency of GBD estimates for each risk factor. of SDI or faster progress than expected for the change in Summaries for select risk factors are highlighted in print SDI, could yield further insights. For all two-page summaries see (pp S216–319); summaries for all risk factors can be Two risk factors that have not been highly correlated https://www.thelancet.com/ found online. with SDI in the past have also seen declines in exposure gbd/summaries at nearly 1% per year over the study period: tobacco Discussion smoking and lead exposure. Progress on exposure to Main findings these risks stands out compared with the increases in Our analysis of risk­attributable burden using exposure to many metabolic risks and no substantial 30 652 sources for exposure, relative risk, and the TMREL change for others such as diet quality. In both of these showed that in 2019, 47·8% (95% UI 45·3–50·1) of global cases, government action through taxation and regulatory www.thelancet.com Vol 396 October 17, 2020 1241 Global Health Metrics policy for tobacco smoking, including advertising bans intake above energy requirements. Some studies sug­ 24,25 and clean air legislation, and regulation of lead content, gest that certain diet components are more likely to have had a major effect. Tobacco interventions highlight contribute to increased BMI than others; the mechanism how regulatory policy can lead to behaviour change. of these effects can be complex and include effects on International efforts for tobacco control have also been appetite, absorption, and displacement of other foods. It bolstered by the Framework Convention on Tobacco is currently hard to understand the role of physical Control. Despite the more than 1% per year decline in inactivity, excess caloric intake, and diet quality in driving age­ standardised tobacco smoking exposure between the increase in BMI. The large combined burden of 2010 and 2019, tobacco remains the third leading risk diet quality, physical inactivity, and high BMI (11·9% factor for attributable DALYs among Level 2 risks. For [95% UI 9·6–14·5] of all DALYs in 2019) indicates just the three major and rapidly increasing risks, the role of how profoundly important the nexus of diet and physical taxation and regulatory policy should be examined. For activity can be to current and future health. The setting ambient particulate matter pollution, regulation can for understanding the potential of changes in overall diet 27,28 clearly have a direct impact. For the nexus of high FPG is to use future health scenarios to trace how public and high BMI, regulatory strategies are less clear. We are policies such as subsidies, taxes, information campaigns, failing to deal with these risks, and concerted research and improving accessibility can affect health in each and policy efforts are needed to reverse the trends. country. In this study, no country or territory has had a The marked rise of metabolic risks as a group, in significant decline in the proportion of the population particular high FPG and high BMI, and their large with high BMI between 1990 to 2019 or in the past contribution to attributable burden is perhaps most decade. The complete failure to reduce BMI at the disturbing. During this period of rising metabolic risk national level implies that efforts to modify the nexus of exposure, global cardiovascular disease age­ standardised physical inactivity, diet quality, and excess energy intake mortality has been declining as documented in GBD 2019 might be very challenging. Tackling this diet quality and for diseases and injuries. The seeming paradox could, to excess energy intake will not only be important for a large extent, be explained by the effect of access to care, human health but has important ramifications for social determinants of health, cohort effects, and other environmental sustainability. behavioural, occupational, and environmental risks not The two types of exposure to particulate matter with a quantified here. Rising metabolic risks might at some diameter of less than 2·5 μm (PM ) have profoundly 2·5 point overwhelm these other drivers and eventually lead to different relationships with socio ­ demographic develop­ rising cardiovascular mortality in the future. This situation ment: household air pollution is strongly related to SDI might have arrived in some high­ income countries in and tends to decrease steadily with socio­ demographic which age­ standardised cardio vascular disease mortality development. By contrast, ambient particulate matter has plateaued or increased since 2017. If year­ on­ year pollution tends to increase with industrialisation and declines in cardiovascular disease mortality come to an then decline with air­ quality management at higher levels 37,38 end, the effect on mortality and longevity at the global level of SDI. The global increases in ambient particulate could be massive. While high BMI and high FPG have matter pollution exposure are being driven by the middle steadily increased, high LDL cholesterol has remained SDI quintiles, as seen in figure 1B. Studies have shown constant over the past decade despite the expected that for ambient particulate matter pollution (ambient correlation with BMI; this finding warrants further PM ), the main sources of exposure are residential 2·5 investigation and could be related to changes in diet energy use, industry, and power generation. The con­ quality, pharmacological intervention, or other factors. centration of PM burden in south Asia highlights how 2·5 Although not increasing at the rate of high BMI or high the absence of national policy actions can have a major FPG, high SBP has become the leading risk factor for effect. Among the large risk factors in which exposure is disease burden at the global level, among the most increasing, ambient particulate matter pollution stands detailed risks in this analysis. A range of strategies out because exposure is declining in countries with a including primary care management and reductions in higher SDI. Like tobacco and lead, regulation can have a sodium intake are known to be potentially effective in profound effect on exposure to and health effects of 32,33 reducing the burden of this critical risk factor. ambient particulate matter pollution and does not require 40,41 The rise of high BMI and its probable role in increasing individual action. There is a clear role for global high FPG needs further examination. Increased BMI can organisations to encourage regulatory change in middle be traced to the combination of physical inactivity, excess SDI countries with large and increasing exposure to caloric intake, and diet quality. At the global level, we ambient particulate matter pollution. This agenda is all find that high BMI is rising considerably faster than low the more urgent because of the direct linkage to global physical activity and poor diet quality. Diet quality on its climate change. own is the fifth leading Level 2 risk factor for attributable Because of profound global interest in the potential DALYs. The effect of diet on human health goes beyond health effects of climate change, we have included high diet quality and should include the contribution of diet and low non­ optimal temperatures in GBD 2019. Climate 1242 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics change will have impacts on human health through between avoidable and attributable burden can vary many mechanisms: direct effects of temperature rise, across countries; in high mortality settings, competing humidity changes, sea­ level rise, extreme weather events, risks might mean that avoidable burden will be sys­ and reduced agricultural yields and increased rural tematically smaller than attributable burden. 42,43 poverty. We have so far included only one of these For GBD 2019 and all previous GBD CRA efforts, our pathways in the GBD analysis, namely the direct effects inclusion criteria for a risk–outcome pair were based on of ambient temperature on different disease outcomes. the World Cancer Research Fund criteria for convincing Our analysis showed that the TMREL varies as a function or probable evidence. We also required that published of mean annual temperature. Locations where mean studies, when meta­ analysed together, yielded a sig­ temperature is higher tend to have higher optimal nificant (p<0·05) relative risk for any risk–outcome pair temperatures, probably through physical and social meeting these criteria. To avoid risk–outcome pairs on adaptation. In 2019, the burden (as measured by the cusp of statistical significance coming in and out of percentage of total DALYs) attributable to low temperature GBD with different cycles, we introduced a threshold of was 2·2 times greater than the burden attributable to p>0·1 to exclude a risk–outcome pair that has previously high temperature. This balance does not, however, hold been included in GBD. Among the included risk– true when looking at specific locations or regions. While outcome pairs, the consistency of the evidence and risk for high SDI countries, the cold­ related burden is of bias varies considerably. The evidence linking smoking 15·4 times greater than the heat­ related burden, this to lung cancer is clearly far stronger than the evidence on relationship is switched for other regions, such as south omega­ 3 and ischaemic heart disease. The UI of the Asia where we observed a 1·7 times greater heat­ related mean effect does not fully capture this difference in the burden and sub­ Saharan Africa where we observed consistency of evidence or the risk of bias. A more robust a  3·6 times greater heat­ related burden. Rising tem­ measure needs to take into account various risks of bias perature will probably have a substantial effect in and the unexplained variation in effect after taking into locations with less capacity to adapt to increased account these risks of bias as well as the magnitude of temperature, potentially exacerbating health inequalities the effect size. The relative risk of lung cancer from across countries. The social capacity to adapt is also smoking is very high across levels of exposure, with a probably tied to economic development: for example, air relative risk of 3·4 at ten pack­ years of smoking and a conditioners in the USA have mitigated the impact of relative risk of 6·5 at 20 pack­ years of smoking, which heat waves over the past 50 years. In terms of trends, makes it far more likely to be causal than an exposure there was a marked increase in exposure to high with a relative risk of 1·1. If we included the unexplained temperature from 1990 to 2010 and then a slight decline heterogeneity across studies after adjusting for risk of from 2010 to 2019; there are major annual fluctuations in bias in the UI of the relative risks, several risk–outcome temperature exposure on top of long­ term warming pairs might not meet inclusion criteria. We are working trends, and 2010 stood out as a year with high tempe­ on developing an evidence scoring system that quantifies ratures in many regions. Our analysis does not provide a consistency and risk of bias for GBD and would allow basis for understanding the full effects of future climate readers to understand that not all risk–outcome pairs change, which will operate through many different have the same evidence base. It would, however, be pathways in addition to the direct effects of temperature. misleading and potentially harmful to argue that we In the GBD CRA work to date, we have estimated the should only examine the GBD risk–outcome pairs burden attributable to past exposure in a given year. The with the highest grade of evidence. The precautionary CRA framework also laid out the important utility for principle for public policy implies that governments have policy making of estimating how changes in current and a duty to act on risk factors that are probably or potentially future exposure can change future levels of health; this harmful and not only those that have overwhelming concept is called avoidable burden. Most CRA work to evidence. Relying only on effects on the basis of the date has focused on estimating attributable burden, highest degree of evidence will very seriously delay public even though avoidable burden is arguably more relevant recognition and proactive policies, which in turn would to policy prioritisation. The dominance of work on attri­ result in perpetuating preventable burden. We hope to butable burden is founded on two premises: it is very inform both individuals and public policy makers with difficult to estimate avoidable burden as this estimate quantification of burden and strength of evidence so they requires a comprehensive future health scenarios frame­ are empowered to make sense of the available data. work; and attributable burden is likely to be highly Analysing relative risks to inform choices by individuals correlated with avoidable burden. With the availability of and choices by population health decision makers might 45,46 a GBD­ informed future health scenarios platform, the legitimately have different perspectives. Some guidelines 48,49 possibility of estimating avoidable burden is much more on systematic reviews recommend reporting absolute tractable. Future work on avoidable burden for each GBD risk levels related to exposure; this method is perhaps risk factor might allow us to examine the true relationship appropriate for informing individual choice. For public between the two approaches to CRA. The relationships policy, attributable burden might be more relevant. If a www.thelancet.com Vol 396 October 17, 2020 1243 Global Health Metrics risk factor is related to 1000 deaths, from a public policy cardiovascular diseases, particularly stroke, due to newly perspective, the concentration of the risk in a smaller or added data and changes in fitting the exposure ­ response larger group of individuals might matter less than it does curves. Given the very large burden of low birthweight to individuals. To provide a synthesis of the evidence for and short gestation on neonatal mortality, the inclusion different users, we included estimation of the all ­ cause of these intermediates has been an important change in mortality relative risk associated with exposure levels of our assessment. The burden of stroke and ischaemic each risk factor using the global distribution of burden heart disease attributable to kidney dysfunction increased across outcomes. However, risks that cause relatively between GBD 2017 and GBD 2019 after updating the modest increases for individuals but are highly prevalent, relative risks with new data from 44 cohorts. For instance, such as air pollution, are, nevertheless, legitimate targets the comparable estimate of the proportion of cardio­ for public policy. vascular DALYs due to kidney dysfunction in 2010 increased from 6·8% (95% UI 6·0–7·6) to 8·5% Substantial changes compared with GBD 2017 (6·8–10·3). Compared with GBD 2017, our GBD 2019 estimates of the burden (as measured by percentage of total DALYs) Limitations attributable to diet quality in 2017 were 29·7% lower. In GBD 2019, we undertook a reassessment of dose– These reductions stem from three major sources: response relationships and relaxed previous assumptions changes in the crosswalks between alternative and that the risk curve is log­ linear. This reassessment was reference methods for estimating diet intake, new limited, however, to dietary risks, kidney dysfunction, and systematic reviews and meta­ regressions, and more air pollution. Future reassess ments of other continuous risk empirical standardised methods for selecting the TMREL factors that currently assume a log­ linear relationship could for protective factors. Although there were changes in materially change risk factor rankings in the future and the overall burden of diet, there were larger changes could also lead to exclusion of other risk–outcome pairs. in the diet components themselves, particularly the Assessment of the joint effects of risk factors depends substantial increase in the attributable burden from on two critical factors: the correlation of risk exposure red meat and the decline in the burden attributable to and the estimation of the joint effects of groups of risks low vegetable intake. The sources of the changes were together. For exposure, we assumed that for each age­ sex­ the same as for diet quality overall. One of the most location­ year, the estimates of the prevalence of exposure important insights from this enriched analysis is that for were independent. Previous simulation analyses under­ many harmful and protective factors, the relative risk taken for GBD 2010 with use of US data from the functions tend to flatten out at higher exposure levels; National Health and Nutrition Examination Survey the previous practice of imposing a log­ linear functional suggested this assumption did not materially bias our form on the risk equation—widely used in the scientific findings. To assess the joint effects of risk factors, we literature—might have led to overestimation. For protec­ assumed in general that relative risks are multiplicative. tive diet components (whole grains, fruit, fibre, nuts This simple assumption has been modified to take into and seeds, omega­ 3, polyunsaturated fatty acids, veg­ account known pathways in which one risk factor, such etables, milk, and calcium), we set the TMREL to the as fruit consumption, is mediated through another risk 85th percentile of levels of exposure included in the factor such as fibre intake. To avoid over ­ estimation of published cohort studies or randomised controlled trials. the joint effects, we computed the non ­ mediated relative With further study of individuals with higher levels of risks and then assumed that non­ mediated relative intake, it is possible that the level of intake associated risks are multiplicative. This approach does not capture with the lowest risk is in fact higher than the TMREL potential synergy between relative risks in which some set for protective diet components in GBD 2019. 12 diet combinations might be super­ multiplicative. For some risk–outcome pairs from GBD 2017 were excluded areas such as diet, the joint estimation is very important from GBD 2019 because our re­ analysis with updated for public policy. Further, more detailed work is needed data suggested that the effects were no longer signifi­ to strengthen the evidence base for understanding cant. Some risk–outcome pairs, such as omega­ 3 and mediation. In particular, mediation implies necessarily ischaemic heart disease, which remained in the analysis that exposure between mediated risks is correlated. as the result of the new meta­ regression of 21 trials and Factoring in that implied correlation into risk exposure 27 cohort studies, met inclusion criteria but future estimation could strengthen estimates in the future. studies could shift the balance of the evidence to be The main limitation of our estimates of risk­ excluded. attributable burden is the availability and quality of Particulate matter pollution burden in 2017 was 44·6% primary data that underpin the analysis. Data for risk higher in GBD 2019 than in GBD 2017. The increase was relationships of several risk factors, such as ambient due to the inclusion of low birthweight and short ozone pollution, residential radon, occupational risks, gestation as risk factors that are themselves affected by child hood sexual abuse, intimate partner violence, PM , as well as increases in the relative risk curve for bullying victimisation, and child growth failure are 2·5 1244 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics sparse. For exposure measurement, patterns of data Conclusion availability are non­ uniform across geography and over Using the most up­ to­ date assessment of the data for time and, where available, might be based on less exposure and relative risk, we found that global exposure reliable modes of data collection such as self­ report. In to harmful environmental risks has been declining, with GBD 2019, we implemented more explicit corrections the notable exception of ambient particulate matter for bias associated with non­ reference methods of pollution. Environmental risk reduction is making an exposure measurement that improved the estimation of important contribution to reductions in child mortality. In risk exposure. Furthermore, these assessments can be aggregate, there has been no real progress reducing used to guide future data collection efforts by identifying exposure to behavioural risks, while metabolic risks are, those populations with not only sparse but low­ quality on average, increasing every year. As a world, we are data based on the collection mode. failing to change some behaviours, particularly those Our analysis, particularly the overall assessment of related to diet quality, caloric intake, and physical activity. burden attributable to all risks combined and risk­ Progress on reducing harm from one crucial behaviour, deleted mortality, is limited by several potentially tobacco smoking, shows the power of taxation and important risk factors not included in this analysis. The regulation. The promise of prevention through risk most important set is likely to be social determinants of modification is not being realised in adult populations health such as educational attainment, poverty, or social around the world. Urgent attention on more successful exclusion. We are currently doing systematic reviews on strategies to reduce risks is needed. educational attainment, which will be the first social Contributors Please see appendix 1 for more detailed information about individual determinant to be incorporated into future rounds of author contributions to the research, divided into the following the GBD CRA. There is also a wide range of other categories: managing the estimation process; writing the first draft of the risk factors not yet included such as nitrous oxide, heavy manuscript; providing data or critical feedback on data sources; metals, environ mental noise, sleep, stress, UV radiation, developing methods or computational machinery; applying analytical methods to produce estimates; providing critical feedback on methods or among others. Future rounds of GBD might evaluate results; drafting the work or revising it critically for important intellectual whether these risk factors meet inclusion criteria. content; extracting, cleaning, or cataloguing data; designing or coding To date, GBD has not included Mendelian ran­ figures and tables; and managing the overall research enterprise. domisation studies in meta­ regression. These studies Declaration of interests could provide new insights on the causal connections C A T Antonio reports personal fees from Johnson & Johnson between risks and outcomes. Not all Mendelian (Philippines), outside of the submitted work. E Beghi reports grants from 51–53 Italian Ministry of Health and SOBI, and personal fees from Arvelle randomisation studies are appropriate for inclusion. Therapeutics, outside of the submitted work. Y Béjot reports personal Future rounds of GBD will give careful consideration to fees from AstraZeneca, Bristol­ Myers Squibb, Pfizer, Medtronic, Merck including these studies for some risk–outcome pairs. Sharpe & Dohme, and Amgen; grants and personal fees from Boehringer For harmful risks with monotonically increasing risk Ingelheim; personal fees and non­ financial support from Servier; and non­ financial support from Biogen, outside of the submitted work. functions, we have generally assumed that the TMREL M L Bell reports grants from US Environmental Protection Agency, is 0. For protective risks such as fruit or whole grain National Institutes of Health (NIH), and from Wellcome Trust intake, selecting the level of exposure that is minimum Foundation, during the conduct of the study; and Honorarium or travel risk is more challenging. Extrapolating the risk function reimbursement from the NIH (for review of grant proposals), American Journal of Public Health (participation as editor), Global Research beyond where the available cohort studies or trials Laboratory and Seoul National University, Royal Society, Ohio University, support the protective effect could easily lead to both Atmospheric Chemistry Gordon Research Conference, Johns Hopkins exaggerated estimates of attributable burden and implau­ Bloomberg School of Public Health, Arizona State University, Ministry of sible recom mendations on consumption. To avoid this the Environment Japan, Hong Kong Polytechnic University, University of Illinois–Champaign, and the University of Tennessee–Knoxville, outside exaggeration, we set the TMREL for protective risks to be of the submitted work. H Christensen reports personal fees from Bristol equal to the 85th percentile of exposure in the available Myers Squibb, Bayer, and Boehringer Ingelheim, outside of the cohorts and trials. The 85th percentile is arbitrary, but submitted work. S­ C Chung reports grants from GlaxoSmithKline, sensitivity analysis did not suggest major changes if we outside of the submitted work. L Degenhardt reports grants from Indivior and Seqirus, outside of the submitted work. S M S Islam reports selected the 90th or 80th percentiles. grants from National Heart Foundation of Australia and Deakin Lastly, in most cases, we assume that relative risks as University, during the conduct of the study. S L James reports grants a function of exposure are universal and apply in from Sanofi Pasteur and employment from Genentech, outside of the all locations and time periods. Exceptions include submitted work. V Jha reports grants from Baxter Healthcare, GlaxoSmithKline, Zydus Cadilla, and Biocon and personal fees from temperature, in which the risk functions clearly depend NephroPlus, outside of the submitted work. J J Jozwiak reports personal on the annual mean temperature, and the relative risks fees from Amgen, ALAB Laboratoria, Teva, Synexus, and Boehringer for high BMI for breast cancer that differ in Asian and Ingelheim, outside of the submitted work. S V Katikireddi reports grants from NRS Senior Clinical Fellowship, UK Medical Research Council, and non­ Asian populations. Our rules require that there is the Scottish Government Chief Scientist Office, during the conduct of evidence of significant differences in the relative risk for the study. M Kivimäki reports grants from the Medical Research Council, different subgroups; to date, few cases have met this UK (MR/R024227/1), during the conduct of the study. S Lorkowski standard. As evidence accumulates, more location­ reports personal fees from Akcea Therapeutics, Amedes, Amgen, Berlin­ Chemie, Boehringer Ingelheim Pharma, Daiichi Sankyo, Merck Sharp & specific or subgroup relative risks might be identified. www.thelancet.com Vol 396 October 17, 2020 1245 Global Health Metrics Dohme, Novo Nordisk, Sanofi ­ Aventis, Synlab, Unilever, and Upfield, The authors are grateful to the Ministry of Health, the survey copyright and non­ financial support from Preventicus, outside of the submitted owner, for allowing them to have the database. All results of the study work. R V Martin reports grants from the Natural Science and are those of the authors and in no way committed to the Ministry. Engineering Research Council of Canada, during the conduct of the This research used data from China Family Panel Studies, funded by study. T R Miller reports having a contract from the AB InBev 985 Program of Peking University and carried out by the Institute of Foundation, outside of the submitted work. J F Mosser reports grants Social Science Survey of Peking University. The Costa Rican Longevity from the Bill & Melinda Gates Foundation, during the conduct of the and Healthy Aging Study project is a longitudinal study by the study. S Nomura reports grants from the Ministry of Education, Culture, University of Costa Rica’s Centro Centroamericano de Población and Sports, Science, and Technology. S B Patten reports funding from the Instituto de Investigaciones en Salud, in collaboration with the Cuthbertson & Fischer Chair in Pediatric Mental Health at the University University of California at Berkeley. The original pre­ 1945 cohort was of Calgary, during the conduct of the study. C D Pond reports personal funded by the Wellcome Trust (grant 072406), and the 1945­ 1955 fees from Nutricia, outside of the submitted work; and grants from the Retirement Cohort was funded by the US National Institute on Aging National Medical Research council in relation to dementia, and travel (grant R01AG031716). The Principal Investigators are Luis Rosero­ Bixby grants and remuneration related to education of primary care and William H Dow, and co­ Principal Investigators Xinia Fernández and professionals in relation to dementia. M J Postma reports grants from Gilbert Brenes. This paper uses data from Eurostat. The responsibility BioMerieux, WHO, European Union, FIND, Antilope, DIKTI, LPDP, for all conclusions drawn from the data lies entirely with the authors. Bayer, and Budi; personal fees from Quintiles, Novartis, and Pharmerit; The Health Behaviour in School­ Aged Children study is an international grants and personal fees from IQVIA, Bristol­ Myers Squibb, Astra study carried out in collaboration with WHO/EURO. The International Zeneca, Seqirus, Sanofi, Merck Sharpe & Dohme, GlaxoSmithKline, Coordinator of the 1997–98, 2001–02, 2005–06, and 2009–10 surveys was Pfizer, Boehringer Ingelheim, and Novavax; stocks from Ingress Health, Candace Currie and the Data Bank Manager for the 1997–98 survey was and PAG; and is acting as adviser to Asc Academics, all outside of the Bente Wold, whereas for the following survey Oddrun Samdal was the submitted work. I Rakovac reports grants from WHO, during the Databank Manager. A list of principal investigators in each country can conduct of the study. A E Schutte reports personal fees from Omron, be found at http://www.hbsc.org. This paper uses data from the WHO Servier, Takeda, Novartis, and Abbott, outside of the submitted work. Study on global AGEing and adult health. Researchers interested in J A Singh reports personal fees from Crealta/Horizon, Medisys, Fidia, using Irish Longitudinal Study on Ageing data can access the data for UBM LLC, Trio Health, Medscape, WebMD, Clinical Care Options, free from the following sites: Irish Social Science Data Archive at Clearview Healthcare Partners, Putnam Associates, Spherix, Practice University College Dublin (http://www.ucd.ie/issda/data/tilda/); Point Communications, the NIH, and the American College of Interuniversity Consortium for Political and Social Research at the Rheumatology; personal fees from Simply Speaking; stock options in University of Michigan (http://www.icpsr.umich.edu/icpsrweb/ICPSR/ Amarin Pharmaceuticals and Viking Pharmaceuticals; membership in studies/34315). Data used in this paper come from the 2009–10 Ghana the steering committee of OMERACT (an international organisation that Socioeconomic Panel Study Survey, which is a nationally representative develops measures for clinical trials and receives arm’s length funding survey of more than 5000 households in Ghana. The survey is a joint from 12 pharmaceutical companies), the FDA Arthritis Advisory effort undertaken by the Institute of Statistical, Social and Economic Committee, and the Veterans Affairs Rheumatology Field Advisory Research (ISSER) at the University of Ghana, and the Economic Growth Committee; and non­ financial support from UAB Cochrane Centre (EGC) at Yale University. It was funded by EGC. At the same Musculoskeletal Group Satellite Center on Network Meta­ analysis, time, ISSER and the EGC are not responsible for the estimations outside of the submitted work. S T S Skou reports personal fees from reported by the analysts. The Palestinian Central Bureau of Statistics Journal of Orthopaedic & Sports Physical Therapy and Munksgaard and granted the researchers access to relevant data in accordance with grants from The Lundbeck Foundation, outside of the submitted work; license number SLN2014­ 3­ 170, after subjecting data to processing and being co­ founder of GLA:D. GLA:D is a non­ profit initiative hosted at aiming to preserve the confidentiality of individual data in accordance University of Southern Denmark aimed at implementing clinical with the General Statistics Law, 2000. The researchers are solely guidelines for osteoarthritis in clinical practice. J D Stanaway reports responsible for the conclusions and inferences drawn upon available grants from the Bill & Melinda Gates Foundation, during the conduct of data. The authors thank the Russia Longitudinal Monitoring Survey, the study. D J Stein reports personal fees from Lundbeck, and Sun, RLMS­ HSE, conducted by the National Research University Higher outside of the submitted work. F Topouzis reports grants from Pfizer, School of Economics and ZAO Demoscope together with Carolina Thea, Novartis, Rheon, Omikron, Pharmaten, Bayer, and Population Center, University of North Carolina at Chapel Hill, and the Bausch & Lomb; and personal fees from Novartis, and Omikron, outside Institute of Sociology, RAS for making data available. This paper uses of the submitted work. R Uddin reports travel and accommodation data from the Armenia 2016, Bangladesh 2009–10, Belarus 2016–17, reimbursement from Deakin University Institute for Physical Activity Benin 2015, Bhutan 2014, Iraq 2015, Kuwait 2006 and 2014, Libya 2009, and Nutrition, outside of the submitted work. All other authors declare Malawi 2009, Moldova 2013, and Sudan 2016 STEP Surveys, all no competing interests. implemented with the support of WHO. This paper uses data from Survey of Health, Ageing and Retirement in Europe (SHARE) Waves 1 For more on SHARE see Data sharing (DOI:10.6103/SHARE.w1.700), 2 (10.6103/SHARE.w2.700), 3 (10.6103/ http://www.share-project.org To download the data used in these analyses, please visit the Global SHARE.w3.700), 4 (10.6103/SHARE.w4.700), 5 (10.6103/SHARE.w5.700), For the Global Health Data Health Data Exchange GBD 2019 website. 6 (10.6103/SHARE.w6.700), and 7 (10.6103/SHARE.w7.700); see Börsch­ Exchange GBD 2019 website Acknowledgments Supan and colleagues (2013) for methodological details. The SHARE see http://ghdx.healthdata.org/ Research reported in this publication was supported by the Bill & data collection has been funded by the European Commission through gbd−2019 Melinda Gates Foundation; Bloomberg Philanthropies; the University of FP5 (QLK6­ CT­ 2001­ 00360), FP6 (SHARE­ I3: RII­ CT­ 2006­ 062193, Melbourne; Queensland Department of Health, Australia; the National COMPARE: CIT5­ CT­ 2005­ 028857, SHARELIFE: CIT4­ CT­ 2006­ 028812), Health and Medical Research Council, Australia; Public Health England; FP7 (SHARE­ PREP: GA number 211909, SHARE­ LEAP: GA number the Norwegian Institute of Public Health; St Jude Children’s Research 227822, SHARE M4: GA number 261982), and Horizon 2020 Hospital; the Cardiovascular Medical Research and Education Fund; (SHARE­ DEV3: GA number 676536, SERISS: GA number 654221) and the National Institute on Ageing of the National Institutes of Health by DG Employment, Social Affairs & Inclusion. Additional funding from (NIH; award P30AG047845); and the National Institute of Mental Health the German Ministry of Education and Research, the Max Planck Society of the NIH (award R01MH110163). The content is solely the for the Advancement of Science, the US National Institute on Aging responsibility of the authors and does not necessary represent the official (U01_AG09740­ 13S2, P01_AG005842, P01_AG08291, P30_AG12815, views of the funders. Data for this research was provided by MEASURE R21_AG025169, Y1­ AG­ 4553­ 01, IAG_BSR06­ 11, OGHA_04­ 064, Evaluation, funded by the United States Agency for International HHSN271201300071C) and from various national funding sources is Development (USAID). Views expressed do not necessarily reflect those gratefully acknowledged (www.share­ project.org). The United States of USAID, the US Government, or MEASURE Evaluation. This research Aging, Demographics, and Memory Study is a supplement to the Health used data from the Chile National Health Survey 2003 and 2009–10. and Retirement Study, which is sponsored by the National Institute of 1246 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics Aging (grant number NIA U01AG009740). It was conducted jointly by Research, Kalaghatgi. B­ F Hwang acknowledges support from China Duke University and the University of Michigan. This paper uses data Medical University (107­ Z­ 04), Taichung, Taiwan. N Ikeda acknowledges from Add Health, a programme project designed by J Richard Udry, support from The Japan Society for the Promotion of Science (grant Peter S Bearman, and Kathleen Mullan Harris, and funded by a grant number 15K08762 and 18H03063). S M S Islam acknowledges support P01­ HD31921 from the Eunice Kennedy Shriver National Institute of from the National Heart Foundation of Australia and Deakin University. Child Health and Human Development, with cooperative funding from M Jakovljevic acknowledges grant OI175014 of the Ministry of Education 17 other agencies. Special acknowledgment is due to Ronald R Rindfuss Science and Technological Development of the Republic of Serbia for and Barbara Entwisle for assistance in the original design. Information supporting the Serbian portion of this GBD study. P Jeemon supported on how to obtain the Add Health data files is available on the Add Health by a Clinical and Public Health intermediate fellowship (grant number For the Add Health website see website. No direct support was received from grant P01­ HD31921 for this IA/CPHI/14/1/501497) from the Wellcome Trust­ Department of http://www.cpc.unc.edu/ analysis. This paper uses data from the NIH. The content is solely the Biotechnology, India Alliance (2015–20). O John acknowledges receiving addhealth responsibility of the authors and does not necessarily represent the the UIPA scholarship from UNSW, Sydney. S V Katikireddi official views of the NIH. This analysis uses data or information from acknowledges funding from a NRS Senior Clinical Fellowship the Longitudinal Aging Study in India (LASI) Pilot microdata and (SCAF/15/02), the Medical Research Council (MC_UU_12017/13 & documentation. The development and release of the LASI Pilot Study MC_UU_12017/15), and the Scottish Government Chief Scientist Office was funded by the National Institute on Aging, a division of the NIH (SPHSU13 & SPHSU15). C Kieling acknowledges employment with the (R21AG032572, R03AG043052, and R01 AG030153). L G Abreu Conselho Nacional de Desenvolvimento Científico e Tecnológico acknowledges support from Coordenação de Aperfeiçoamento de (a Brazilian public funding agency) and being a UK Academy of Medical Pessoal de Nível Superior, Brazil (finance Code 001) and Conselho Sciences Newton Advanced Fellow. Y J Kim acknowledges support from Nacional de Desenvolvimento Científico e Tecnológico. L Abu­ Raddad the Research Management Centre, Xiamen University Malaysia, acknowledges the support of Qatar National Research Fund XMUMRF/2018­ C2/ITCM/0001. K Krishan acknowledges support from (NPRP 9­ 040­ 3­ 008). O Adetokunboh acknowledges the South African UGC Centre of Advanced Study (CAS II) awarded to the Department of Department of Science and Innovation, and National Research Anthropology, Panjab University, Chandigarh, India. B Lacey Foundation. A Agrawal acknowledges support from the Wellcome Trust acknowledges support from the NIHR Oxford Biomedical Research DBT India Alliance Senior Fellowship. R Akinyemi acknowledges Centre and the BHF Centre of Research Excellence, Oxford. T Lallukka support by Grant U01HG010273 from the NIH as part of the H3Africa acknowledges support from the Academy of Finland (grant Consortium and support by the FLAIR fellowship funded by the number 319200). I Landires is member of the Sistema Nacional de UK Royal Society and the African Academy of Sciences. S Aljunid Investigación, which is supported by the Secretaría Nacional de Ciencia, acknowledges the Department of Health Policy and Management, Tecnología e Innovación, Panamá. S Langan acknowledges support from Faculty of Public Health, Kuwait University and International Centre for a Wellcome Trust Senior Clinical Fellowship (205039/Z/16/Z). J Lazarus Casemix and Clinical Coding, Faculty of Medicine, and the National acknowledges support from a Spanish Ministry of Science, Innovation University of Malaysia for the approval and support to participate in this and Universities Miguel Servet grant (Instituto de Salud Carlos III/ESF, research project. T Bärnighausen acknowledges support from the EU [CP18/00074]). S Lorkowski acknowledges support from the German Alexander von Humboldt Foundation through the Alexander von Federal Ministry of Education and Research (nutriCARD, grant Humboldt Professor award, funded by the German Federal Ministry of agreement number 01EA1808A). A M Mantilla­ Herrera is affiliated with Education and Research. A Badawi acknowledges support from the the Queensland Centre for Mental Health Research which receives core Public Health Agency of Canada. G Britton acknowledges support from funding from the Department of Health, Queensland Government. the Sistema Nacional de Investigación of SENACYT, Panamá. J J Carrero R Martin is grateful to Washington University for additional institutional acknowledges support from the Swedish Research Council (2019­ 01059). support. J McGrath acknowledges support from the Danish National F Caravlho acknowledges support from UID/MULTI/04378/2019 and Research Foundation (Niels Bohr Professorship), and the Queensland UID/QUI/50006/2019 with funding from FCT/MCTES through national Health Department (via West Moreton HHS). W Mendoza acknowledges funds. A Cohen acknowledges support from Health Effects Institute, Population and Development department at the United Nations Boston. V M Costa acknowledges support from Grant Population Fund Country Office in Peru, which does not necessarily SFRH/BHD/110001/2015, received by Portuguese national funds endorse this study. U Mueller acknowledges support from the German through Fundação para a Ciência e Tecnologia, IP, under the Norma National Cohort Study BMBF grant number 01ER1801D. S Nomura Transitória DL57/2016/CP1334/CT0006. M DeLang acknowledges the acknowledges support from the Ministry of Education, Culture, Sports, NASA grant number NNX16AQ30G and NIOSH grant number Science, and Technology of Japan (18K10082). A Ortiz acknowledges T42­ OH008673 for supporting the work to estimate global ozone. support from ISCIII PI19/00815, DTS18/00032, ISCIII­ RETIC H Erskine acknowledges support from the Australian National Health REDinREN RD016/0009 Fondos FEDER, FRIAT, Comunidad de Madrid and Medical Research Council Early Career Fellowship (APP1137969) B2017/BMD­ 3686 CIFRA2­ CM; these funding sources had no role in the and is employed by the Queensland Centre for Mental Health Research, writing of the manuscript or the decision to submit it for publication. which receives core funding from Queensland Health. E Fernandes S Patten acknowledges support from the Cuthbertson and Fischer Chair acknowledges support from UID/MULTI/04378/2019 and in Pediatric Mental Health at the University of Calgary. G Patton UID/QUI/50006/2019 with funding from FCT/MCTES through national acknowledges support from a National Health & Medical Research funds. A Ferrari acknowledges support from a National Health and Council Fellowship. M R Phillips acknowledges support in part by the Medical Research Council Early Career Fellowship Grant (APP1121516) National Natural Science Foundation of China (number 81371502 and and the Queensland Centre for Mental Health Research, which receives 81761128031). A Raggi, D Sattin, and S Schiavolin acknowledge support funding from the Queensland Department of Health. M Ferreira by a grant from the Italian Ministry of Health (Ricerca Corrente, acknowledges support from a National Health and Medical Council of Fondazione Istituto Neurologico C Besta, Linea 4 ­ Outcome Research: Australia fellowship. M Freitas acknowledges financial support from the dagli Indicatori alle Raccomandazioni Cliniche). D Ribeiro acknowledges EU (FEDER funds through COMPETE POCI­ 01­ 0145­ FEDER­ 029248), the financial support from the EU (FEDER funds through the and National Funds (FCT, Fundação para a Ciência e Tecnologia) Operational Competitiveness Program [COMPETE]; POCI­ 01­ 0145­ through project PTDC/NAN­ MAT/29248/2017. M Ausloos, C Herteliu, FEDER­ 029253). D C Ribeiro acknowledges support from The and A Pana acknowledge partial support by a grant of the Romanian Sir Charles Hercus Health Research Fellowship (number 18/111) Health National Authority for Scientific Research and Innovation, Research Council of New Zealand. P Sachdev acknowledges support CNDS­ UEFISCDI, project number PN­ III­ P4­ ID­ PCCF­ 2016­ 0084. from the National Health and Medical Research Council of Australia C Herteliu acknowledges partial support by a grant co­ funded by Program Grant. A M Samy acknowledges support from a fellowship European Fund for Regional Development through Operational from the Egyptian Fulbright Mission Program. D Santomauro Program for Competitiveness, Project ID P_40_382. P Hoogar acknowledges affiliation with the Queensland Centre for Mental Health acknowledges the Bio Cultural Studies, Manipal Academy of Higher Research, which receives core funding from the Department of Health. Education, and Manipal and Centre for Holistic Development and M Santric­ Milicevic acknowledges support from the Ministry of www.thelancet.com Vol 396 October 17, 2020 1247 Global Health Metrics Education, Science and Technological Development of the Republic of 11 Gakidou E, Afshin A, Abajobir AA, et al. Global, regional, and national comparative risk assessment of 84 behavioural, Serbia (contract number 175087). R Sarmiento acknowledges support environmental and occupational, and metabolic risks or clusters of from the Applied and Environmental Sciences University in Bogota, risks, 1990–2016: a systematic analysis for the Global Burden of Colombia and Carlos III Institute of Health Madrid Spain. A Schutte Disease Study 2016. Lancet 2017; 390: 1345–422. acknowledges support from the South African National Research 12 Stanaway JD, Afshin A, Gakidou E, et al. Global, regional, and national Foundation SARChI Chair initiative (GUN 86895) and the South African comparative risk assessment of 84 behavioural, environmental and Medical Research Council. M Serre acknowledges the NASA grant occupational, and metabolic risks or clusters of risks for 195 countries number NNX16AQ30G for supporting the work to estimate global and territories, 1990–2017: a systematic analysis for the Global Burden ozone. S T Skou acknowledges support from a grant from Region of Disease Study 2017. Lancet 2018; 392: 1923–94. Zealand (Exercise First) and a grant from the European Research 13 Stevens GA, Alkema L, Black RE, et al. Guidelines for accurate and Council under the EU’s Horizon 2020 research and innovation transparent health estimates reporting: the GATHER statement. programme (grant agreement number 801790). J B Soriano Lancet 2016; 388: e19–23. acknowledges support by Centro de Investigación en Red de 14 Murray CJ, Lopez AD. Global mortality, disability, and the Enfermedades Respiratorias, Instituto de Salud Carlos III, Madrid, contribution of risk factors: Global Burden of Disease Study. Lancet Spain. R Tabarés­ Seisdedos acknowledges support in part by the 1997; 349: 1436–42. national grant PI17/00719 from ISCIII­ FEDER. N Taveira acknowledges 15 Murray CJ, Ezzati M, Lopez AD, Rodgers A, Vander Hoorn S. partial support from the European & Developing Countries Clinical Comparative quantification of health risks conceptual framework Trials Partnership, EU (LIFE project, reference RIA2016MC­ 1615). and methodological issues. Popul Health Metr 2003; 1: 1. M Tonelli acknowledges support from the David Freeze chair in health 16 World Cancer Research Fund, American Institute for Cancer services research, University of Calgary. S Tyrovolas acknowledges Research. Food, nutrition, and physical activity, and the prevention support from the Foundation for Education and European Culture, of cancer: a global perspective. Washington DC: AICR, 2007. the Sara Borrell postdoctoral program (reference number CD15/00019 17 James SL, Abate D, Abate KH, et al. Global, regional, and national from the Instituto de Salud Carlos III), and the Fondos Europeo de incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, Desarrollo Regional. M Wei acknowledges support from the National 1990–2017: a systematic analysis for the Global Burden of Disease Institute on Aging at the NIH (K23AG056638). J J West acknowledges Study 2017. Lancet 2018; 392: 1789–858. support from NASA grant number NNX16AQ30G and NIOSH number 18 Griswold MG, Fullman N, Hawley C, et al. Alcohol use and burden T42­ OH008673. for 195 countries and territories, 1990–2016: a systematic analysis Editorial note: the Lancet Group takes a neutral position with respect to for the Global Burden of Disease Study 2016. Lancet 2018; 392: 1015–35. territorial claims in published maps and institutional affiliations. 19 Kurowicka D, Cooke R. Sampling algorithms for generating joint References uniform distributions using the vine­ copula method. 1 WHO. Global action plan for the prevention and control of Comput Stat Data Anal 2007; 51: 2889–906. noncommunicable diseases. 2013–2020. Geneva: World Health 20 Curtis V. Explaining the outcomes of the ‘Clean India’ campaign: Organization, 2013. institutional behaviour and sanitation transformation in India. 2 WHO. UN Task Force on NCDs: The government of Ethiopia is BMJ Glob Health 2019; 4: e001892. working jointly with the UN system to strengthen the national 21 Quinn A, Bruce N, Puzzolo E, et al. An analysis of efforts to scale NCD response. 2017. http://www.who.int/ncds/un­ task­ force/joint­ up clean household energy for cooking around the world. programming­ missions/ethiopia­ november­ 2017/en/ (accessed Energy Sustain Dev 2018; 46: 1–10. Feb 25, 2020). 22 Dabadge A, Sreenivas A, Josey A. What has the pradhan mantri 3 Centers for Disease Control and Prevention. The National ujjwala yojana achieved so far? Econ Polit Wkly 2018; 53: 69–75. Environmental Public Health Tracking Program. 2019. https://www. 23 Yach D, McKee M, Lopez AD, Novotny T. Improving diet and cdc.gov/nceh/information/tracking_network.htm (accessed physical activity: 12 lessons from controlling tobacco smoking. Feb 25, 2020). BMJ 2005; 330: 898–900. 4 WHO. Global health risks: mortality and burden of disease 24 Davis JM, Elias RW, Grant LD. Current issues in human lead attributable to selected major risks. Geneva: World Health exposure and regulation of lead. Neurotoxicology 1993; 14: 15–27. Organization, 2009. 25 Dignam T, Kaufmann RB, LeStourgeon L, Brown MJ. Control of 5 Nowbar AN, Gitto M, Howard JP, Francis DP, Al­ Lamee R. Mortality lead sources in the United States, 1970–2017: public health progress from ischemic heart disease. Circ Cardiovasc Qual Outcomes 2019; and current challenges to eliminating lead expsoure. 12: e005375. J Public Health Manag Pract 2019; 25: S13–22. 6 Zhou B, Bentham J, Di Cesare M, et al. Worldwide trends in blood 26 Chung­ Hall J, Craig L, Gravely S, Sansone N, Fong GT. Impact of pressure from 1975 to 2015: a pooled analysis of 1479 population­ the WHO FCTC over the first decade: a global evidence review based measurement studies with 19·1 million participants. Lancet prepared for the Impact Assessment Expert Group. Tob Control 2017; 389: 37–55. 2019; 28 (suppl 2): 119–28. 7 NCD Risk Factor Collaboration (NCD­ RisC). Trends in adult body­ 27 United States Environmental Protection Agency. National air mass index in 200 countries from 1975 to 2014: a pooled analysis of quality status and trends through 2007. Research Triangle Park, 1698 population­ based measurement studies with 19·2 million North Carolina: U.S. Environmental Protection Agency. Office of participants. Lancet 2016; 387: 1377–96. Air Quality Planning and Standards, 2008. https://www.epa.gov/ 8 Lim SS, Vos T, Flaxman AD, et al. A comparative risk assessment sites/production/files/2017 ­ 11/documents/trends_brochure_2007. of burden of disease and injury attributable to 67 risk factors and pdf (accessed Feb 25, 2020). risk factor clusters in 21 regions, 1990–2010: a systematic analysis 28 Zalakeviciute R, Rybarczyk Y, López­ Villada J, Diaz Suarez MV. for the Global Burden of Disease Study 2010. Lancet 2012; Quantifying decade­ long effects of fuel and traffic regulations on 380: 2224–60. urban ambient PM pollution in a mid­ size South American city. 2·5 9 Forouzanfar MH, Alexander L, Anderson HR, et al. Atmos Pollut Res 2018; 9: 66–75. Global, regional, and national comparative risk assessment of 29 GBD 2019 Diseases and Injuries Collaborators. Global burden of 79 behavioural, environmental and occupational, and metabolic 369 diseases and injuries in 204 countries and territories, risks or clusters of risks in 188 countries, 1990–2013: a systematic 1990–2019: a systematic analysis for the Global Burden of Disease analysis for the Global Burden of Disease Study 2013. Lancet 2015; Study 2019. Lancet 2020; 396: 1204–22. 386: 2287–323. 30 Fullman N, Yearwood J, Abay SM, et al. Measuring performance on 10 Forouzanfar MH, Afshin A, Alexander LT, et al. Global, regional, the Healthcare Access and Quality Index for 195 countries and and national comparative risk assessment of 79 behavioural, territories and selected subnational locations: a systematic analysis environmental and occupational, and metabolic risks or clusters of from the Global Burden of Disease Study 2016. Lancet 2018; risks, 1990–2015: a systematic analysis for the Global Burden of 391: 2236–71. Disease Study 2015. Lancet 2016; 388: 1659–724. 1248 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics 31 Lopez AD, Adair T. Is the long­ term decline in cardiovascular­ 43 IPCC. Global warming of 1·5°C. An IPCC special report on the disease mortality in high­ income countries over? Evidence from impacts of global warming of 1·5°C above pre­ industrial levels national vital statistics. Int J Epidemiol 2019; 48: 1815–23. and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of 32 Saiz LC, Gorricho J, Garjón J, Celaya MC, Erviti J, Leache L. Blood climate change, sustainable development, and efforts to eradicate pressure targets for the treatment of people with hypertension and poverty. Geneva, Switzerland: World Meteorological cardiovascular disease. Cochrane Database Syst Rev 2018; 7: CD010315. Organization, 2018. 33 He FJ, Li J, MacGregor GA. Effect of longer ­ term modest salt 44 Barreca A, Clay K, Deschenes O, Greenstone M, Shapiro JS. Adapting reduction on blood pressure. Cochrane Database Syst Rev 2013; to climate change: the remarkable decline in the US temperature­ 4: CD004937. mortality relationship over the twentieth century. J Polit Econ 2016; 34 The Lancet. The global syndemic of obesity, undernutrition, 124: 105–59. and climate change: The Lancet Commission report. Lancet 2019; 45 Foreman KJ, Marquez N, Dolgert A, et al. Forecasting life 393: 791–846. expectancy, years of life lost, and all­ cause and cause­ specific 35 Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB. Changes in mortality for 250 causes of death: reference and alternative diet and lifestyle and long­ term weight gain in women and men. scenarios for 2016–40 for 195 countries and territories. Lancet 2018; N Engl J Med 2011; 364: 2392–404. 392: 2052–90. 36 Willett W, Rockström J, Loken B, et al. Food in the Anthropocene: 46 Vollset S, Goren E, Yuan C­ W, et al. Fertility, mortality, migration, the EAT–Lancet Commission on healthy diets from sustainable and population scenarios for 195 countries and territories from 2017 food systems. Lancet 2019; 393: 447–92. to 2100: a forecasting analysis for the Global Burden of Disease 37 Adesina JA, Piketh SJ, Qhekwana M, Burger R, Language B, Study. Lancet 2020; 396: 1285–306. Mkhatshwa G. Contrasting indoor and ambient particulate matter 47 Saltelli A, Funtowicz S. The Precautionary Principle: implications concentrations and thermal comfort in coal and non­ coal burning for risk management strategies. Int J Occup Med Environ Health households at South Africa Highveld. Sci Total Environ 2020; 2004; 17: 47–57. 699: 134403. 48 Alonso­ Coello P, Carrasco­ Labra A, Brignardello­ Petersen R, et al. 38 HEI International Scientific Oversight Committee. Section 11. A methodological survey of the analysis, reporting and Development, air pollution expsoure, and population health. In: interpretation of Absolute Risk ReductiOn in systematic revieWs Outdoor air pollution and health in developing countries of Asia: (ARROW): a study protocol. Syst Rev 2013; 2: 113. a comprehensive review. Special Report 18. Boston, MA: Health 49 Nuovo J, Melnikow J, Chang D. Reporting number needed to treat Effects Institute, 2010. https://www.healtheffects.org/system/files/ and absolute risk reduction in randomized controlled trials. JAMA SR18AsianLitReview.pdf (accessed Nov 27, 2019). 2002; 287: 2813–14. 39 Weagle CL, Snider G, Li C, et al. Global sources of fine particulate 50 Millwood IY, Walters RG, Mei XW, et al. Conventional and genetic matter: interpretation of PM chemical composition observed by 2·5 evidence on alcohol and vascular disease aetiology: a prospective SPARTAN using a global chemical transport model. study of 500 000 men and women in China. Lancet 2019; Environ Sci Technol 2018; 52: 11670–81. 393: 1831–42. 40 Correia AW, Pope CA, Dockery DW, Wang Y, Ezzati M, Dominici F. 51 Adam D. The gene­ based hack that is revolutionizing epidemiology. Effect of air pollution control on life expectancy in the United States: Nature 2019; 576: 196–99. an analysis of 545 U.S. counties for the period from 2000 to 2007. Epidemiology 2013; 24: 23–31. 52 Burgess S, Davey Smith G. How humans can contribute to Mendelian randomization analyses. Int J Epidemiol 2019; 48: 661–64. 41 Dedoussi IC, Eastham SD, Monier E, Barrett SRH. Premature mortality related to United States cross­ state air pollution. Nature 53 Burgess S, Timpson NJ, Ebrahim S, Davey Smith G. Mendelian 2020; 578: 261–65. randomization: where are we now and where are we going? Int J Epidemiol 2015; 44: 379–88. 42 Smith, KR, Woodward A, Campbell­ Lendrum D, et al. Human health: impacts, adaptation, and co­ benefits. In: Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Field CB, Barros VR, Dokken DJ, et al (eds)]. Cambridge, UK and New York, NY: Cambridge University Press. 2014. www.thelancet.com Vol 396 October 17, 2020 1249 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The Lancet Unpaywall

Global burden of 87 risk factors in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019

Murray, Christopher J L; Aravkin, Aleksandr Y; Zheng, Peng; Abbafati, Cristiana; Abbas, Kaja M; Abbasi-Kangevari, Mohsen; Abd-Allah, Foad; Abdelalim, Ahmed; Abdollahi, Mohammad; Abdollahpour, Ibrahim; Abegaz, Kedir Hussein; Abolhassani, Hassan; Aboyans, Victor; Abreu, Lucas Guimarães; Abrigo, Michael R M; Abualhasan, Ahmed; Abu-Raddad, Laith Jamal; Abushouk, Abdelrahman I; Adabi, Maryam; Adekanmbi, Victor; Adeoye, Abiodun Moshood; Adetokunboh, Olatunji O; Adham, Davoud; Advani, Shailesh M; Agarwal, Gina; Aghamir, Seyed Mohammad Kazem; Agrawal, Anurag; Ahmad, Tauseef; Ahmadi, Keivan; Ahmadi, Mehdi; Ahmadieh, Hamid; Ahmed, Muktar Beshir; Akalu, Temesgen Yihunie; Akinyemi, Rufus Olusola; Akinyemiju, Tomi; Akombi, Blessing; Akunna, Chisom Joyqueenet; Alahdab, Fares; Al-Aly, Ziyad; Alam, Khurshid; Alam, Samiah; Alam, Tahiya; Alanezi, Fahad Mashhour; Alanzi, Turki M; Alemu, Biresaw wassihun; Alhabib, Khalid F; Ali, Muhammad; Ali, Saqib; Alicandro, Gianfranco; Alinia, Cyrus; Alipour, Vahid; Alizade, Hesam; Aljunid, Syed Mohamed; Alla, François; Allebeck, Peter; Almasi-Hashiani, Amir; Al-Mekhlafi, Hesham M; Alonso, Jordi; Altirkawi, Khalid A; Amini-Rarani, Mostafa; Amiri, Fatemeh; Amugsi, Dickson A; Ancuceanu, Robert; Anderlini, Deanna; Anderson, Jason A; Andrei, Catalina Liliana; Andrei, Tudorel; Angus, Colin; Anjomshoa, Mina; Ansari, Fereshteh; Ansari-Moghaddam, Alireza; Antonazzo, Ippazio Cosimo; Antonio, Carl Abelardo T; Antony, Catherine M; Antriyandarti, Ernoiz; Anvari, Davood; Anwer, Razique; Appiah, Seth Christopher Yaw; Arabloo, Jalal; Arab-Zozani, Morteza; Ariani, Filippo; Armoon, Bahram; Ärnlöv, Johan; Arzani, Afsaneh; Asadi-Aliabadi, Mehran; Asadi-Pooya, Ali A; Ashbaugh, Charlie; Assmus, Michael; Atafar, Zahra; Atnafu, Desta Debalkie; Atout, Maha Moh'd Wahbi; Ausloos, Floriane; Ausloos, Marcel; Ayala Quintanilla, Beatriz Paulina; Ayano, Getinet; Ayanore, Martin Amogre; Azari, Samad; Azarian, Ghasem; Azene, Zelalem Nigussie; Badawi, Alaa; Badiye, Ashish D; Bahrami, Mohammad Amin; Bakhshaei, Mohammad Hossein; Bakhtiari, Ahad; Bakkannavar, Shankar M; Baldasseroni, Alberto; Ball, Kylie; Ballew, Shoshana H; Balzi, Daniela; Banach, Maciej; Banerjee, Srikanta K; Bante, Agegnehu Bante; Baraki, Adhanom Gebreegziabher; Barker-Collo, Suzanne Lyn; Bärnighausen, Till Winfried; Barrero, Lope H; Barthelemy, Celine M; Barua, Lingkan; Basu, Sanjay; Baune, Bernhard T; Bayati, Mohsen; Becker, Jacob S; Bedi, Neeraj; Beghi, Ettore; Béjot, Yannick; Bell, Michellr L; Bennitt, Fiona B; Bensenor, Isabela M; Berhe, Kidanemaryam; Berman, Adam E; Bhagavathula, Akshaya Srikanth; Bhageerathy, Reshmi; Bhala, Neeraj; Bhandari, Dinesh; Bhattacharyya, Krittika; Bhutta, Zulfiqar A; Bijani, Ali; Bikbov, Boris; Bin Sayeed, Muhammad Shahdaat; Biondi, Antonio; Birihane, Binyam Minuye; Bisignano, Catherine; Biswas, Raaj Kishore; Bitew, Helen; Bohlouli, Somayeh; Bohluli, Mehdi; Boon-Dooley, Alexandra S; Borges, Guilherme; Borzì, Antonio Maria; Borzouei, Shiva; Bosetti, Cristina; Boufous, Soufiane; Braithwaite, Dejana; Breitborde, Nicholas J K; Breitner, Susanne; Brenner, Hermann; Briant, Paul Svitil; Briko, Andrey Nikolaevich; Briko, Nikolay Ivanovich; Britton, Gabrielle B; Bryazka, Dana; Bumgarner, Blair R; Burkart, Katrin; Burnett, Richard Thomas; Burugina Nagaraja, Sharath; Butt, Zahid A; Caetano dos Santos, Florentino Luciano; Cahill, Leah E; Cámera, Luis LA Alberto; Campos-Nonato, Ismael R; Cárdenas, Rosario; Carreras, Giulia; Carrero, Juan J; Carvalho, Felix; Castaldelli-Maia, Joao Mauricio; Castañeda-Orjuela, Carlos A; Castelpietra, Giulio; Castro, Franz; Causey, Kate; Cederroth, Christopher R; Cercy, Kelly M; Cerin, Ester; Chandan, Joht Singh; Chang, Kai-Lan; Charlson, Fiona J; Chattu, Vijay Kumar; Chaturvedi, Sarika; Cherbuin, Nicolas; Chimed-Ochir, Odgerel; Cho, Daniel Youngwhan; Choi, Jee-Young Jasmine; Christensen, Hanne; Chu, Dinh-Toi; Chung, Michael T; Chung, Sheng-Chia; Cicuttini, Flavia M; Ciobanu, Liliana G; Cirillo, Massimo; Classen, Thomas Khaled Dwayne; Cohen, Aaron J; Compton, Kelly; Cooper, Owen R; Costa, Vera Marisa; Cousin, Ewerton; Cowden, Richard G; Cross, Di H; Cruz, Jessica A; Dahlawi, Saad M A; Damasceno, Albertino Antonio Moura; Damiani, Giovanni; Dandona, Lalit; Dandona, Rakhi; Dangel, William James; Danielsson, Anna-Karin; Dargan, Paul I; Darwesh, Aso Mohammad; Daryani, Ahmad; Das, Jai K; Das Gupta, Rajat; das Neves, José; Dávila-Cervantes, Claudio Alberto; Davitoiu, Dragos Virgil; De Leo, Diego; Degenhardt, Louisa; DeLang, Marissa; Dellavalle, Robert Paul; Demeke, Feleke Mekonnen; Demoz, Gebre Teklemariam; Demsie, Desalegn Getnet; Denova-Gutiérrez, Edgar; Dervenis, Nikolaos; Dhungana, Govinda Prasad; Dianatinasab, Mostafa; Dias da Silva, Diana; Diaz, Daniel; Dibaji Forooshani, Zahra Sadat; Djalalinia, Shirin; Do, Hoa Thi; Dokova, Klara; Dorostkar, Fariba; Doshmangir, Leila; Driscoll, Tim Robert; Duncan, Bruce B; Duraes, Andre Rodrigues; Eagan, Arielle Wilder; Edvardsson, David; El Nahas, Nevine; El Sayed, Iman; El Tantawi, Maha; Elbarazi, Iffat; Elgendy, Islam Y; El-Jaafary, Shaimaa I; Elyazar, Iqbal RF; Emmons-Bell, Sophia; Erskine, Holly E; Eskandarieh, Sharareh; Esmaeilnejad, Saman; Esteghamati, Alireza; Estep, Kara; Etemadi, Arash; Etisso, Atkilt Esaiyas; Fanzo, Jessica; Farahmand, Mohammad; Fareed, Mohammad; Faridnia, Roghiyeh; Farioli, Andrea; Faro, Andre; Faruque, Mithila; Farzadfar, Farshad; Fattahi, Nazir; Fazlzadeh, Mehdi; Feigin, Valery L; Feldman, Rachel; Fereshtehnejad, Seyed-Mohammad; Fernandes, Eduarda; Ferrara, Giannina; Ferrari, Alize J; Ferreira, Manuela L; Filip, Irina; Fischer, Florian; Fisher, James L; Flor, Luisa Sorio; Foigt, Nataliya A; Folayan, Morenike Oluwatoyin; Fomenkov, Artem Alekseevich; Force, Lisa M; Foroutan, Masoud; Franklin, Richard Charles; Freitas, Marisa; Fu, Weijia; Fukumoto, Takeshi; Furtado, João M; Gad, Mohamed M; Gakidou, Emmanuela; Gallus, Silvano; Garcia-Basteiro, Alberto L; Gardner, William M; Geberemariyam, Biniyam Sahiledengle; Gebreslassie, Assefa Ayalew Ayalew Ayalew; Geremew, Abraham; Gershberg Hayoon, Anna; Gething, Peter W; Ghadimi, Maryam; Ghadiri, Keyghobad; Ghaffarifar, Fatemeh; Ghafourifard, Mansour; Ghamari, Farhad; Ghashghaee, Ahmad; Ghiasvand, Hesam; Ghith, Nermin; Gholamian, Asadollah; Ghosh, Rakesh; Gill, Paramjit Singh; Ginindza, Themba G G; Giussani, Giorgia; Gnedovskaya, Elena V; Goharinezhad, Salime; Gopalani, Sameer Vali; Gorini, Giuseppe; Goudarzi, Houman; Goulart, Alessandra C; Greaves, Felix; Grivna, Michal; Grosso, Giuseppe; Gubari, Mohammed Ibrahim Mohialdeen; Gugnani, Harish Chander; Guimarães, Rafael Alves; Guled, Rashid Abdi; Guo, Gaorui; Guo, Yuming; Gupta, Rajeev; Gupta, Tarun; Haddock, Beatrix; Hafezi-Nejad, Nima; Hafiz, Abdul; Haj-Mirzaian, Arvin; Haj-Mirzaian, Arya; Hall, Brian J; Halvaei, Iman; Hamadeh, Randah R; Hamidi, Samer; Hammer, Melanie S; Hankey, Graeme J; Haririan, Hamidreza; Haro, Josep Maria; Hasaballah, Ahmed I; Hasan, Md Mehedi; Hasanpoor, Edris; Hashi, Abdiwahab; Hassanipour, Soheil; Hassankhani, Hadi; Havmoeller, Rasmus J; Hay, Simon I; Hayat, Khezar; Heidari, Golnaz; Heidari-Soureshjani, Reza; Henrikson, Hannah J; Herbert, Molly E; Herteliu, Claudiu; Heydarpour, Fatemeh; Hird, Thomas R; Hoek, Hans W; Holla, Ramesh; Hoogar, Praveen; Hosgood, H Dean; Hossain, Naznin; Hosseini, Mostafa; Hosseinzadeh, Mehdi; Hostiuc, Mihaela; Hostiuc, Sorin; Househ, Mowafa; Hsairi, Mohamed; Hsieh, Vivian Chia-rong; Hu, Guoqing; Hu, Kejia; Huda, Tanvir M; Humayun, Ayesha; Huynh, Chantal K; Hwang, Bing-Fang; Iannucci, Vincent C; Ibitoye, Segun Emmanuel; Ikeda, Nayu; Ikuta, Kevin S; Ilesanmi, Olayinka Stephen; Ilic, Irena M; Ilic, Milena D; Inbaraj, Leeberk Raja; Ippolito, Helen; Iqbal, Usman; Irvani, Seyed Sina Naghibi; Irvine, Caleb Mackay Salpeter; Islam, M Mofizul; Islam, Sheikh Mohammed Shariful; Iso, Hiroyasu; Ivers, Rebecca Q; Iwu, Chidozie C D; Iwu, Chinwe Juliana; Iyamu, Ihoghosa Osamuyi; Jaafari, Jalil; Jacobsen, Kathryn H; Jafari, Hussain; Jafarinia, Morteza; Jahani, Mohammad Ali; Jakovljevic, Mihajlo; Jalilian, Farzad; James, Spencer L; Janjani, Hosna; Javaheri, Tahereh; Javidnia, Javad; Jeemon, Panniyammakal; Jenabi, Ensiyeh; Jha, Ravi Prakash; Jha, Vivekanand; Ji, John S; Johansson, Lars; John, Oommen; John-Akinola, Yetunde O; Johnson, Catherine Owens; Jonas, Jost B; Joukar, Farahnaz; Jozwiak, Jacek Jerzy; Jürisson, Mikk; Kabir, Ali; Kabir, Zubair; Kalani, Hamed; Kalani, Rizwan; Kalankesh, Leila R; Kalhor, Rohollah; Kanchan, Tanuj; Kapoor, Neeti; Karami Matin, Behzad; Karch, André; Karim, Mohd Anisul; Kassa, Getachew Mullu; Katikireddi, Srinivasa Vittal; Kayode, Gbenga A; Kazemi Karyani, Ali; Keiyoro, Peter Njenga; Keller, Cathleen; Kemmer, Laura; Kendrick, Parkes J; Khalid, Nauman; Khammarnia, Mohammad; Khan, Ejaz Ahmad; Khan, Maseer; Khatab, Khaled; Khater, Mona M; Khatib, Mahalaqua Nazli; Khayamzadeh, Maryam; Khazaei, Salman; Kieling, Christian; Kim, Yun Jin; Kimokoti, Ruth W; Kisa, Adnan; Kisa, Sezer; Kivimäki, Mika; Knibbs, Luke D; Knudsen, Ann Kristin Skrindo; Kocarnik, Jonathan M; Kochhar, Sonali; Kopec, Jacek A; Korshunov, Vladimir Andreevich; Koul, Parvaiz A; Koyanagi, Ai; Kraemer, Moritz U G; Krishan, Kewal; Krohn, Kris J; Kromhout, Hans; Kuate Defo, Barthelemy; Kumar, G Anil; Kumar, Vivek; Kurmi, Om P; Kusuma, Dian; La Vecchia, Carlo; Lacey, Ben; Lal, Dharmesh Kumar; Lalloo, Ratilal; Lallukka, Tea; Lami, Faris Hasan; Landires, Iván; Lang, Justin J; Langan, Sinéad M; Larsson, Anders O; Lasrado, Savita; Lauriola, Paolo; Lazarus, Jeffrey V; Lee, Paul H; Lee, Shaun Wen Huey; LeGrand, Kate E; Leigh, James; Leonardi, Matilde; Lescinsky, Haley; Leung, Janni; Levi, Miriam; Li, Shanshan; Lim, Lee-Ling; Linn, Shai; Liu, Shiwei; Liu, Simin; Liu, Yang; Lo, Justin; Lopez, Alan D; Lopez, Jaifred Christian F; Lopukhov, Platon D; Lorkowski, Stefan; Lotufo, Paulo A; Lu, Alton; Lugo, Alessandra; Maddison, Emilie R; Mahasha, Phetole Walter; Mahdavi, Mokhtar Mahdavi; Mahmoudi, Morteza; Majeed, Azeem; Maleki, Afshin; Maleki, Shokofeh; Malekzadeh, Reza; Malta, Deborah Carvalho; Mamun, Abdullah A; Manda, Ana Laura; Manguerra, Helena; Mansour-Ghanaei, Fariborz; Mansouri, Borhan; Mansournia, Mohammad Ali; Mantilla Herrera, Ana M; Maravilla, Joemer C; Marks, Ashley; Martin, Randall V; Martini, Santi; Martins-Melo, Francisco Rogerlândio; Masaka, Anthony; Masoumi, Seyedeh Zahra; Mathur, Manu Raj; Matsushita, Kunihiro; Maulik, Pallab K; McAlinden, Colm; McGrath, John J; McKee, Martin; Mehndiratta, Man Mohan; Mehri, Fereshteh; Mehta, Kala M; Memish, Ziad A; Mendoza, Walter; Menezes, Ritesh G; Mengesha, Endalkachew Worku; Mereke, Alibek; Mereta, Seid Tiku; Meretoja, Atte; Meretoja, Tuomo J; Mestrovic, Tomislav; Miazgowski, Bartosz; Miazgowski, Tomasz; Michalek, Irmina Maria; Miller, Ted R; Mills, Edward J; Mini, GK; Miri, Mohammad; Mirica, Andreea; Mirrakhimov, Erkin M; Mirzaei, Hamed; Mirzaei, Maryam; Mirzaei, Roya; Mirzaei-Alavijeh, Mehdi; Misganaw, Awoke Temesgen; Mithra, Prasanna; Moazen, Babak; Mohammad, Dara K; Mohammad, Yousef; Mohammad Gholi Mezerji, Naser; Mohammadian-Hafshejani, Abdollah; Mohammadifard, Noushin; Mohammadpourhodki, Reza; Mohammed, Ammas Siraj; Mohammed, Hussen; Mohammed, Jemal Abdu; Mohammed, Shafiu; Mokdad, Ali H; Molokhia, Mariam; Monasta, Lorenzo; Mooney, Meghan D; Moradi, Ghobad; Moradi, Masoud; Moradi-Lakeh, Maziar; Moradzadeh, Rahmatollah; Moraga, Paula; Morawska, Lidia; Morgado-da-Costa, Joana; Morrison, Shane Douglas; Mosapour, Abbas; Mosser, Jonathan F; Mouodi, Simin; Mousavi, Seyyed Meysam; Mousavi Khaneghah, Amin; Mueller, Ulrich Otto; Mukhopadhyay, Satinath; Mullany, Erin C; Musa, Kamarul Imran; Muthupandian, Saravanan; Nabhan, Ashraf F; Naderi, Mehdi; Nagarajan, Ahamarshan Jayaraman; Nagel, Gabriele; Naghavi, Mohsen; Naghshtabrizi, Behshad; Naimzada, Mukhammad David; Najafi, Farid; Nangia, Vinay; Nansseu, Jobert Richie; Naserbakht, Morteza; Nayak, Vinod C; Negoi, Ionut; Ngunjiri, Josephine W; Nguyen, Cuong Tat; Nguyen, Huong Lan Thi; Nguyen, Minh; Nigatu, Yeshambel T; Nikbakhsh, Rajan; Nixon, Molly R; Nnaji, Chukwudi A; Nomura, Shuhei; Norrving, Bo; Noubiap, Jean Jacques; Nowak, Christoph; Nunez-Samudio, Virginia; Oţoiu, Adrian; Oancea, Bogdan; Odell, Christopher M; Ogbo, Felix Akpojene; Oh, In-Hwan; Okunga, Emmanuel Wandera; Oladnabi, Morteza; Olagunju, Andrew T; Olusanya, Bolajoko Olubukunola; Olusanya, Jacob Olusegun; Omer, Muktar Omer; Ong, Kanyin L; Onwujekwe, Obinna E; Orpana, Heather M; Ortiz, Alberto; Osarenotor, Osayomwanbo; Osei, Frank B; Ostroff, Samuel M; Otstavnov, Nikita; Otstavnov, Stanislav S; Øverland, Simon; Owolabi, Mayowa O; P A, Mahesh; Padubidri, Jagadish Rao; Palladino, Raffaele; Panda-Jonas, Songhomitra; Pandey, Anamika; Parry, Charles D H; Pasovic, Maja; Pasupula, Deepak Kumar; Patel, Sangram Kishor; Pathak, Mona; Patten, Scott B; Patton, George C; Pazoki Toroudi, Hamidreza; Peden, Amy E; Pennini, Alyssa; Pepito, Veincent Christian Filipino; Peprah, Emmanuel K; Pereira, David M; Pesudovs, Konrad; Pham, Hai Quang; Phillips, Michael R; Piccinelli, Cristiano; Pilz, Tessa M; Piradov, Michael A; Pirsaheb, Meghdad; Plass, Dietrich; Polinder, Suzanne; Polkinghorne, Kevan R; Pond, Constance Dimity; Postma, Maarten J; Pourjafar, Hadi; Pourmalek, Farshad; Poznańska, Anna; Prada, Sergio I; Prakash, V; Pribadi, Dimas Ria Angga; Pupillo, Elisabetta; Quazi Syed, Zahiruddin; Rabiee, Mohammad; Rabiee, Navid; Radfar, Amir; Rafiee, Ata; Raggi, Alberto; Rahman, Muhammad Aziz; Rajabpour-Sanati, Ali; Rajati, Fatemeh; Rakovac, Ivo; Ram, Pradhum; Ramezanzadeh, Kiana; Ranabhat, Chhabi Lal; Rao, Puja C; Rao, Sowmya J; Rashedi, Vahid; Rathi, Priya; Rawaf, David Laith; Rawaf, Salman; Rawal, Lal; Rawassizadeh, Reza; Rawat, Ramu; Razo, Christian; Redford, Sofia Boston; Reiner, Robert C; Reitsma, Marissa Bettay; Remuzzi, Giuseppe; Renjith, Vishnu; Renzaho, Andre M N; Resnikoff, Serge; Rezaei, Negar; Rezaei, Nima; Rezapour, Aziz; Rhinehart, Phoebe-Anne; Riahi, Seyed Mohammad; Ribeiro, Daniel Cury; Ribeiro, Daniela; Rickard, Jennifer; Rivera, Juan A; Roberts, Nicholas L S; Rodríguez-Ramírez, Sonia; Roever, Leonardo; Ronfani, Luca; Room, Robin; Roshandel, Gholamreza; Roth, Gregory A; Rothenbacher, Dietrich; Rubagotti, Enrico; Rwegerera, Godfrey M; Sabour, Siamak; Sachdev, Perminder S; Saddik, Basema; Sadeghi, Ehsan; Sadeghi, Masoumeh; Saeedi, Reza; Saeedi Moghaddam, Sahar; Safari, Yahya; Safi, Sare; Safiri, Saeid; Sagar, Rajesh; Sahebkar, Amirhossein; Sajadi, S Mohammad; Salam, Nasir; Salamati, Payman; Salem, Hosni; Salem, Marwa R Rashad; Salimzadeh, Hamideh; Salman, Omar Mukhtar; Salomon, Joshua A; Samad, Zainab; Samadi Kafil, Hossein; Sambala, Evanson Zondani; Samy, Abdallah M; Sanabria, Juan; Sánchez-Pimienta, Tania G; Santomauro, Damian Francesco; Santos, Itamar S; Santos, João Vasco; Santric-Milicevic, Milena M; Saraswathy, Sivan Yegnanarayana Iyer; Sarmiento-Suárez, Rodrigo; Sarrafzadegan, Nizal; Sartorius, Benn; Sarveazad, Arash; Sathian, Brijesh; Sathish, Thirunavukkarasu; Sattin, Davide; Saxena, Sonia; Schaeffer, Lauren E; Schiavolin, Silvia; Schlaich, Markus P; Schmidt, Maria Inês; Schutte, Aletta Elisabeth; Schwebel, David C; Schwendicke, Falk; Senbeta, Anbissa Muleta; Senthilkumaran, Subramanian; Sepanlou, Sadaf G; Serdar, Berrin; Serre, Marc L; Shadid, Jamileh; Shafaat, Omid; Shahabi, Saeed; Shaheen, Amira A; Shaikh, Masood Ali; Shalash, Ali S; Shams-Beyranvand, Mehran; Shamsizadeh, Morteza; Sharafi, Kiomars; Sheikh, Aziz; Sheikhtaheri, Abbas; Shibuya, Kenji; Shield, Kevin David; Shigematsu, Mika; Shin, Jae Il; Shin, Min-Jeong; Shiri, Rahman; Shirkoohi, Reza; Shuval, Kerem; Siabani, Soraya; Sierpinski, Radoslaw; Sigfusdottir, Inga Dora; Sigurvinsdottir, Rannveig; Silva, João Pedro; Simpson, Kyle E; Singh, Jasvinder A; Singh, Pushpendra; Skiadaresi, Eirini; Skou, Søren T; Skryabin, Valentin Yurievich; Smith, Emma U R; Soheili, Amin; Soltani, Shahin; Soofi, Moslem; Sorensen, Reed J D; Soriano, Joan B; Sorrie, Muluken Bekele; Soshnikov, Sergey; Soyiri, Ireneous N; Spencer, Cory N; Spotin, Adel; Sreeramareddy, Chandrashekhar T; Srinivasan, Vinay; Stanaway, Jeffrey D; Stein, Caroline; Stein, Dan J; Steiner, Caitlyn; Stockfelt, Leo; Stokes, Mark A; Straif, Kurt; Stubbs, Jacob L; Sufiyan, Mu'awiyyah Babale; Suleria, Hafiz Ansar Rasul; Suliankatchi Abdulkader, Rizwan; Sulo, Gerhard; Sultan, Iyad; Szumowski, Łukasz; Tabarés-Seisdedos, Rafael; Tabb, Karen M; Tabuchi, Takahiro; Taherkhani, Amir; Tajdini, Masih; Takahashi, Ken; Takala, Jukka S; Tamiru, Animut Tagele; Taveira, Nuno; Tehrani-Banihashemi, Arash; Temsah, Mohamad-Hani; Tesema, Getayeneh Antehunegn; Tessema, Zemenu Tadesse; Thurston, George D; Titova, Mariya Vladimirovna; Tohidinik, Hamid Reza; Tonelli, Marcello; Topor-Madry, Roman; Topouzis, Fotis; Torre, Anna E; Touvier, Mathilde; Tovani-Palone, Marcos Roberto Roberto; Tran, Bach Xuan; Travillian, Ravensara; Tsatsakis, Aristidis; Tudor Car, Lorainne; Tyrovolas, Stefanos; Uddin, Riaz; Umeokonkwo, Chukwuma David; Unnikrishnan, Bhaskaran; Upadhyay, Era; Vacante, Marco; Valdez, Pascual R; van Donkelaar, Aaron; Vasankari, Tommi Juhani; Vasseghian, Yasser; Veisani, Yousef; Venketasubramanian, Narayanaswamy; Violante, Francesco S; Vlassov, Vasily; Vollset, Stein Emil; Vos, Theo; Vukovic, Rade; Waheed, Yasir; Wallin, Mitchell Taylor; Wang, Yafeng; Wang, Yuan-Pang; Watson, Alexandrea; Wei, Jingkai; Wei, Melissa Y Wei; Weintraub, Robert G; Weiss, Jordan; Werdecker, Andrea; West, J Jason; Westerman, Ronny; Whisnant, Joanna L; Whiteford, Harvey A; Wiens, Kirsten E; Wolfe, Charles D A; Wozniak, Sarah S; Wu, Ai-Min; Wu, Junjie; Wulf Hanson, Sarah; Xu, Gelin; Xu, Rixing; Yadgir, Simon; Yahyazadeh Jabbari, Seyed Hossein; Yamagishi, Kazumasa; Yaminfirooz, Mousa; Yano, Yuichiro; Yaya, Sanni; Yazdi-Feyzabadi, Vahid; Yeheyis, Tomas Y; Yilgwan, Christopher Sabo; Yilma, Mekdes Tigistu; Yip, Paul; Yonemoto, Naohiro; Younis, Mustafa Z; Younker, Theodore Patrick; Yousefi, Bahman; Yousefi, Zabihollah; Yousefinezhadi, Taraneh; Yousuf, Abdilahi Yousuf; Yu, Chuanhua; Yusefzadeh, Hasan; Zahirian Moghadam, Telma; Zamani, Mohammad; Zamanian, Maryam; Zandian, Hamed; Zastrozhin, Mikhail Sergeevich; Zhang, Yunquan; Zhang, Zhi-Jiang; Zhao, Jeff T; Zhao, Xiu-Ju George; Zhao, Yingxi; Zhou, Maigeng; Ziapour, Arash; Zimsen, Stephanie R M; Brauer, Michael; Afshin, Ashkan; Lim, Stephen S
The Lancet – Oct 1, 2020
Download PDF
27 pages

Loading next page...
 
/lp/unpaywall/global-burden-of-87-risk-factors-in-204-countries-and-territories-1990-19JQIFFq4H

References (64)

  • Bin Zhou, J. Bentham, M. Cesare, Honor Bixby, G. Danaei, Melanie Cowan, C. Paciorek, Gitanjali Singh, Kaveh Hajifathalian, J. Bennett, C. Taddei, V. Bilano, R. Carrillo-Larco, Shirin Djalalinia, S. Khatibzadeh, C. Lugero, N. Peykari, Wan Zhang, Yuan Lu, Gretchen Stevens, L. Riley, P. Bovet, P. Elliott, D. Gu, Nayu Ikeda, R. Jackson, M. Joffres, A. Kengne, T. Laatikainen, T. Lam, A. Laxmaiah, Jing Liu, J. Miranda, C. Mondo, H. Neuhauser, J. Sundström, L. Smeeth, M. Sorić, M. Woodward, M. Ezzati, Leandra Abarca-Gómez, Z. Abdeen, H. Rahim, N. Abu-Rmeileh, Benjamín Acosta-Cázares, R. Adams, W. Aekplakorn, K. Afsana, C. Aguilar-Salinas, C. Agyemang, A. Ahmadvand, W. Ahrens, R. Raddadi, Rihab Woyatan, Mohamed Ali, A. Alkerwi, Eman Aly, P. Amouyel, A. Amuzu, L. Andersen, S. Anderssen, L. Ängquist, R. Anjana, D. Ansong, H. Aounallah-Skhiri, J. Araújo, I. Ariansen, T. Aris, N. Arlappa, K. Aryal, D. Arveiler, F. Assah, M. Assunção, M. Avdičová, A. Azevedo, F. Azizi, B. Babu, S. Bahijri, N. Balakrishna, P. Bandosz, J. Banegas, C. Barbagallo, A. Barceló, A. Barkat, A. Barros, M. Barros, I. Bata, A. Batieha, L. Baur, R. Beaglehole, H. Romdhane, M. Benet, L. Benson, A. Bernabé-Ortiz, G. Bernotiene, H. Bettiol, A. Bhagyalaxmi, Sumit Bharadwaj, S. Bhargava, Y. Bi, M. Bikbov, P. Bjerregaard, E. Bjertness, C. Björkelund, A. Blokstra, Simona Bo, M. Bobák, H. Boeing, J. Boggia, C. Boissonnet, V. Bongard, L. Braeckman, I. Brajkovich, F. Branca, Juergen Breckenkamp, H. Brenner, L. Brewster, G. Bruno, H. Bueno-de-Mesquita, Anna Bugge, Con Burns, M. Bursztyn, A. León, J. Cacciottolo, C. Cameron, G. Can, A. Cândido, V. Capuano, V. Cardoso, A. Carlsson, M. Carvalho, F. Casanueva, J. Casas, C. Caserta, S. Chamukuttan, A. Chan, Q. Chan, H. Chaturvedi, N. Chaturvedi, Chien-Jen Chen, Fangfang Chen, Huashuai Chen, Shuohua Chen, Zhengming Chen, Ching-Yu Cheng, Imane Dekkaki, A. Chetrit, A. Chiolero, S. Chiou, A. Chiriță-Emandi, B. Cho, Yumi Cho, J. Chudek, R. Cífková, F. Claessens, E. Clays, H. Concin, C. Cooper, R. Cooper, T. Coppinger, S. Costanzo, D. Cottel, C. Cowell, C. Craig, A. Crujeiras, J. Cruz, G. D’Arrigo, E. d’Orsi, J. Dallongeville, A. Damasceno, R. Dankner, T. Dantoft, L. Dauchet, G. Backer, D. Bacquer, G. Gaetano, S. Henauw, D. Smedt, M. Deepa, A. Dehghan, H. Delisle, V. Deschamps, K. Dhana, A. Castelnuovo, J. Dias-da-Costa, Alejandro Díaz, Ty Dickerson, Ha Do, A. Dobson, C. Donfrancesco, Silvana Donoso, A. Döring, K. Doua, W. Drygas, V. Dulskienė, A. Džakula, V. Dzērve, E. Dziankowska-Zaborszczyk, R. Eggertsen, U. Ekelund, J. Ati, U. Ellert, R. Elosúa, R. Erasmus, C. Erem, L. Eriksen, J. Peña, A. Evans, D. Faeh, C. Fall, F. Farzadfar, F. Félix-Redondo, T. Ferguson, D. Fernández-Bergés, D. Ferrante, M. Ferrari, C. Ferreccio, J. Ferrières, J. Finn, K. Fischer, B. Föger, L. Foo, Ann‐Sofie Forslund, M. Forsner, S. Fortmann, H. Fouad, Damian Francis, M. Franco, O. Franco, G. Frontera, F. Fuchs, S. Fuchs, Y. Fujita, Takuro Furusawa, Z. Gaciong, D. Gareta, S. Garnett, J. Gaspoz, Magda Gasull, Louise Gates, D. Gavrila, J. Geleijnse, Anoosheh Ghasemian, Anup Ghimire, S. Giampaoli, F. Gianfagna, J. Giovannelli, R. Goldsmith, H. Gonçalves, M. Gross, J. Rivas, F. Gottrand, S. Graff-Iversen, D. Grafnetter, Aneta Grajda, R. Gregor, T. Grodzicki, A. Grøntved, G. Gruden, V. Grujić, Ong Guan, V. Gudnason, Ramiro Guerrero, I. Guessous, A. Guimarães, M. Gulliford, Johanna Gunnlaugsdottir, M. Gunter, P. Gupta, O. Gureje, Beata Gurzkowska, L. Gutierrez, F. Gutzwiller, F. Hadaegh, J. Halkjær, I. Hambleton, R. Hardy, Rachakulla Harikumar, J. Hata, A. Hayes, Jiang He, M. Hendriks, A. Henriques, L. Cadena, S. Herrala, R. Heshmat, Ilpo Hihtaniemi, S. Ho, S. Ho, M. Hobbs, A. Hofman, G. Dinç, Claudia Hormiga, B. Horta, L. Houti, C. Howitt, T. Htay, A. Htet, Yonghua Hu, J. Huerta, A. Husseini, I. Huybrechts, N. Hwalla, L. Iacoviello, A. Iannone, M. Ibrahim, M. Ikram, V. Irazola, Muhammad Islam, V. Ivković, M. Iwasaki, Jeremy Jacobs, T. Jafar, K. Jamrozik, I. Janszky, G. Jasienska, B. Jelaković, C. Jiang, M. Johansson, J. Jonas, T. Jørgensen, Pradeep Joshi, A. Juolevi, G. Jurak, V. Jureša, R. Kaaks, A. Kafatos, O. Kalter‐Leibovici, N. Kamaruddin, A. Kasaeian, J. Katz, J. Kauhanen, P. Kaur, M. Kavousi, G. Kazakbaeva, U. Keil, L. Boker, S. Keinänen-Kiukaanniemi, R. Kelishadi, H. Kemper, M. Kersting, T. Key, Y. Khader, D. Khalili, Y. Khang, K. Khaw, S. Kiechl, J. Killewo, Jeongseon Kim, J. Klumbienė, E. Kolle, P. Kolsteren, P. Korrovits, S. Koskinen, K. Kouda, S. Kozieł, P. Kristensen, S. Krokstad, D. Kromhout, H. Kruger, R. Kubínová, R. Kuciene, D. Kuh, U. Kujala, K. Kula, Z. Kułaga, R. Kumar, P. Kurjata, Y. Kusuma, K. Kuulasmaa, C. Kyobutungi, C. Lachat, O. Landrove, V. Lánská, G. Lappas, B. Larijani, L. Laugsand, Khanh Bao, T. Le, C. Leclercq, Jongseok Lee, Jeonghee Lee, T. Lehtimäki, Rampal Lekhraj, L. León-Muñoz, N. Levitt, Yanping Li, C. Lilly, W. Lim, M. Lima-Costa, Hsien-Ho Lin, Xu Lin, A. Linneberg, L. Lissner, M. Litwin, R. Lorbeer, P. Lotufo, J. Lozano, D. Lukšienė, A. Lundqvist, N. Lunet, Per Lytsy, G. Ma, Jun Ma, G. Machado-Coelho, S. Machi, S. Maggi, D. Magliano, Marjeta Majer, Marcia Makdisse, R. Malekzadeh, R. Malhotra, Kodavanti Rao, S. Malyutina, Y. Manios, J. Mann, E. Manzato, P. Margozzini, P. Marques-Vidal, J. Marrugat, R. Martorell, E. Mathiesen, A. Matijasevich, T. Matsha, J. Mbanya, A. Posso, S. McFarlane, S. Mcgarvey, S. McLachlan, R. McLean, B. McNulty, A. Khir, S. Médiène-Benchekor, J. Medzioniene, A. Meirhaeghe, C. Meisinger, A. Menezes, G. Menon, I. Meshram, A. Metspalu, J. Mi, Kairit Mikkel, Jody Miller, J. Miquel, M. Misigoj-Durakovic, M. Mohamed, K. Mohammad, N. Mohammadifard, V. Mohan, M. Yusoff, N. Møller, D. Molnár, A. Momenan, K. Monyeki, L. Moreira, A. Morejón, L. Moreno, K. Morgan, G. Moschonis, M. Mossakowska, A. Mostafa, J. Mota, M. Motlagh, Jorge Motta, M. Muiesan, M. Müller-Nurasyid, N. Murphy, J. Mursu, V. Musil, G. Nagel, B. Naidu, Harunobu Nakamura, Jana Námešná, E. Nang, V. Nangia, S. Narake, E. Navarrete-Muñoz, N. Ndiaye, W. Neal, Ilona Nenko, F. Nervi, N. Nguyen, Q. Nguyen, R. Nieto-Martínez, T. Niiranen, G. Ning, T. Ninomiya, S. Nishtar, M. Noale, O. Noboa, A. Noorbala, Teresa Noorbala, D. Noto, M. Nsour, D. O’Reilly, Kyungwon Oh, M. Olinto, I. Oliveira, M. Omar, A. Onat (2017)

    Worldwide trends in blood pressure from 1975 to 2015: a pooled analysis of 1479 population-based measurement studies with 19·1 million participants

    Lancet (London, England), 389

  • (2018)

    Alcohol use and burden for 195 countries and territories, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016

    Lancet (London, England), 392

  • (2014)

    www.thelancet.com Vol 396 October 17, 2020 42

  • Ipcc (2022)

    Global Warming of 1.5°C

  • D. Mozaffarian, Tao Hao, E. Rimm, W. Willett, F. Hu (2011)

    Changes in diet and lifestyle and long-term weight gain in women and men.

    The New England journal of medicine, 364 25

  • 廣畑 富雄 (2012)

    Food, Nutrition, Physical Activity, and the Prevention of Cancer : a Global Perspective : 食物、栄養、身体活動とがんの予防 : 世界的展望(前篇)

    , 4

  • F. He, G. MacGregor (2000)

    Effect of longer-term modest salt reduction on blood pressure.

    The Cochrane database of systematic reviews, 3

  • Andrea Saltelli, S. Funtowicz (2005)

    The Precautionary Principle: Implications for Risk Management Strategies

    Human and Ecological Risk Assessment: An International Journal, 11

  • Gretchen Stevens, L. Alkema, R. Black, G. Collins, M. Ezzati, John Grove, D. Hogan, M. Hogan, R. Horton, A. Marušić, C. Mathers, C. Murray, I. Rudan, J. Salomon, P. Simpson, T. Vos (2017)

    [Guidelines for Accurate and Transparent Health Estimates Reporting: the GATHER Statement].

    Epidemiologia e servicos de saude : revista do Sistema Unico de Saude do Brasil, 26 1

  • A. Dabadge (2018)

    What has the Pradhan Mantri Ujjwala Yojana Achieved so Far

  • J. Nuovo, J. Melnikow, Denise Chang (2002)

    Reporting number needed to treat and absolute risk reduction in randomized controlled trials.

    JAMA, 287 21

  • Irene Dedoussi, S. Eastham, E. Monier, S. Barrett (2020)

    Premature mortality related to United States cross-state air pollution

    Nature, 578

  • V. Curtis (2019)

    Explaining the outcomes of the 'Clean India' campaign: institutional behaviour and sanitation transformation in India

    BMJ Global Health, 4

  • Alan Barreca, Karen Clay, O. Deschenes, Michael Greenstone, Joseph Shapiro (2016)

    Adapting to Climate Change: The Remarkable Decline in the US Temperature-Mortality Relationship over the Twentieth Century

    Journal of Political Economy, 124

  • Andrew Correia, C. Pope, D. Dockery, Yun Wang, Majid Ezzati, Francesca Dominici (2013)

    Effect of Air Pollution Control on Life Expectancy in the United States: An Analysis of 545 U.S. Counties for the Period from 2000 to 2007

    Epidemiology, 24

  • P. Alonso-Coello, A. Carrasco-Labra, R. Brignardello-Petersen, I. Neumann, E. Akl, Xin Sun, B. Johnston, M. Briel, J. Busse, D. Glujovsky, C. Granados, A. Iorio, Affan Irfan, L. García, R. Mustafa, Anggie Ramírez-Morera, I. Solá, K. Tikkinen, S. Ebrahim, P. Vandvik, Yuqing Zhang, A. Selva, A. Sanabria, Oscar Zazueta, R. Vernooij, H. Schünemann, G. Guyatt (2013)

    A methodological survey of the analysis, reporting and interpretation of Absolute Risk ReductiOn in systematic revieWs (ARROW): a study protocol

    Systematic Reviews, 2

  • Kyle Foreman, N. Marquez, A. Dolgert, K. Fukutaki, N. Fullman, Madeline McGaughey, Martin Pletcher, Amanda Smith, Kendrick Tang, Chun-Wei Yuan, Jonathan Brown, Joseph Friedman, Jiawei He, K. Heuton, Mollie Holmberg, Disha Patel, P. Reidy, A. Carter, Kelly Cercy, Abigail Chapin, D. Douwes-Schultz, Tahvi Frank, Falko Goettsch, Patrick Liu, Vishnu Nandakumar, M. Reitsma, Vincent Reuter, Nafis Sadat, Reed Sorensen, V. Srinivasan, R. Updike, H. York, Alan Lopez, R. Lozano, Stephen Lim, A. Mokdad, S. Vollset, C. Murray (2018)

    Forecasting life expectancy, years of life lost, and all-cause and cause-specific mortality for 250 causes of death: reference and alternative scenarios for 2016–40 for 195 countries and territories

    Lancet (London, England), 392

  • S. Burgess, G. Smith (2019)

    How humans can contribute to Mendelian randomization analyses.

    International journal of epidemiology

  • (2014)

    eds)]. Cambridge, UK and New York, NY

  • Section 11. Development, air pollution expsoure, and population health

  • J. Yang (2016)

    Of markets and rights: Discourse in the 2008 and 2013 Global Action Plan for the Prevention and Control of Noncommunicable Diseases

    Annals of global health, 82

  • R. Zalakeviciute, Y. Rybarczyk, Jesús López-Villada, M. Suarez (2018)

    Quantifying decade-long effects of fuel and traffic regulations on urban ambient PM2.5 pollution in a mid-size South American city

    Atmospheric Pollution Research, 9

  • Operational Competitiveness Program [COMPETE]; POCI­01­0145­ FEDER­029253)

  • dagli Indicatori alle Raccomandazioni Cliniche

  • C. Weagle, G. Snider, Chi Li, A. Donkelaar, Sajeev Philip, Sajeev Philip, P. Bissonnette, Jaqueline Burke, John Jackson, R. Latimer, Emily Stone, I. Abboud, C. Akoshile, N. Anh, J. Brook, A. Cohen, Jinlu Dong, M. Gibson, D. Griffith, K. He, B. Holben, R. Kahn, C. Keller, C. Keller, J. Kim, N. Lagrosas, P. Lestari, Yeo Khian, Yang Liu, E. Marais, J. Martins, A. Misra, Ulfi Muliane, R. Pratiwi, E. Quel, A. Salam, L. Segev, S. Tripathi, Chien Wang, Qiang Zhang, M. Brauer, Y. Rudich, R. Martin, R. Martin (2018)

    Global Sources of Fine Particulate Matter: Interpretation of PM2.5 Chemical Composition Observed by SPARTAN using a Global Chemical Transport Model.

    Environmental science & technology, 52 20

  • Stephen Lim, T. Vos, A. Flaxman, G. Danaei, K. Shibuya, Heather Adair-Rohani, Mohammad Almazroa, M. Amann, H. Anderson, K. Andrews, M. Aryee, C. Atkinson, L. Bacchus, A. Bahalim, K. Balakrishnan, J. Balmes, S. Barker-Collo, A. Baxter, M. Bell, J. Blore, F. Blyth, C. Bonner, G. Borges, R. Bourne, M. Boussinesq, M. Brauer, P. Brooks, N. Bruce, B. Brunekreef, C. Bryan-Hancock, C. Bucello, R. Buchbinder, F. Bull, R. Burnett, T. Byers, Bianca Calabria, J. Carapetis, Emily Carnahan, Z. Chafe, F. Charlson, Honglei Chen, J. Chen, A. Cheng, J. Child, A. Cohen, K. Colson, B. Cowie, Sarah Darby, S. Darling, A. Davis, L. Degenhardt, F. Dentener, D. Jarlais, K. Devries, M. Dherani, E. Ding, E. Dorsey, T. Driscoll, K. Edmond, Suad Ali, Rebecca Engell, P. Erwin, S. Fahimi, Gail Falder, F. Farzadfar, A. Ferrari, M. Finucane, S. Flaxman, F. Fowkes, Greg Freedman, Michael Freeman, E. Gakidou, Santu Ghosh, E. Giovannucci, G. Gmel, Kathryn Graham, R. Grainger, B. Grant, D. Gunnell, H. Gutierrez, Wayne Hall, H. Hoek, A. Hogan, H. Hosgood, D. Hoy, Howard Hu, B. Hubbell, S. Hutchings, Sydney Ibeanusi, G. Jacklyn, Rashmi Jasrasaria, J. Jonas, H. Kan, J. Kanis, N. Kassebaum, N. Kawakami, Y. Khang, S. Khatibzadeh, J. Khoo, Cindy Kok, F. Laden, R. Lalloo, Q. Lan, T. Lathlean, J. Leasher, J. Leigh, Yang Li, J. Lin, S. Lipshultz, S. London, R. Lozano, Yuan Lu, J. Mak, R. Malekzadeh, Leslie Mallinger, W. Marcenes, L. March, Robert Marks, R. Martin, P. Mcgale, J. Mcgrath, S. Mehta, Z. Memish, G. Mensah, T. Merriman, R. Micha, C. Michaud, V. Mishra, Khayriyyah Hanafiah, A. Mokdad, L. Morawska, D. Mozaffarian, Tasha Murphy, M. Naghavi, B. Neal, P. Nelson, J. Nolla, Rosana Norman, Casey Olives, S. Omer, J. Orchard, R. Osborne, B. Ostro, A. Page, K. Pandey, C. Parry, E. Passmore, Jayadeep Patra, N. Pearce, P. Pelizzari, M. Petzold, M. Phillips, D. Pope, C. Pope, J. Powles, M. Rao, H. Razavi, E. Rehfuess, J. Rehm, B. Ritz, F. Rivara, Thomas Roberts, Carolyn Robinson, J. Rodriguez-Portales, I. Romieu, R. Room, Lisa Rosenfeld, Ananya Roy, L. Rushton, J. Salomon, Uchechukwu Sampson, L. Sànchez-Riera, Ella Sanman, A. Sapkota, S. Seedat, P. Shi, K. Shield, R. Shivakoti, Gitanjali Singh, D. Sleet, Emma Smith, Kirk Smith, Nicolas Stapelberg, K. Steenland, H. Stöckl, L. Stovner, K. Straif, L. Straney, G. Thurston, Jimmy Tran, R. Dingenen, A. Donkelaar, J. Veerman, L. Vijayakumar, Robert Weintraub, M. Weissman, Richard White, H. Whiteford, S. Wiersma, J. Wilkinson, H. Williams, Warwick Williams, N. Wilson, A. Woolf, P. Yip, J. Zielinski, Alan Lopez, C. Murray, M. Ezzati (2012)

    A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010

    The Lancet, 380

  • L. Saiz, J. Gorricho, J. Garjón, María Celaya, J. Erviti, Leire Leache (2020)

    Blood pressure targets for the treatment of people with hypertension and cardiovascular disease.

    The Cochrane database of systematic reviews, 9

  • (2009)

    Global health risks: mortality and burden of disease attributable to selected major risks. Geneva: World Health Organization

  • (2010)

    In: Outdoor air pollution and health in developing countries of Asia: a comprehensive review

  • S. Vollset, Emily Goren, Chun-Wei Yuan, Jackie Cao, Amanda Smith, Thomas Hsiao, C. Bisignano, G. Azhar, Emma Castro, Julian Chalek, A. Dolgert, Tahvi Frank, K. Fukutaki, S. Hay, R. Lozano, A. Mokdad, Vishnu Nandakumar, Maxwell Pierce, Martin Pletcher, Toshana Robalik, Krista Steuben, H. Wunrow, Bianca Zlavog, C. Murray (2020)

    Fertility, mortality, migration, and population scenarios for 195 countries and territories from 2017 to 2100: a forecasting analysis for the Global Burden of Disease Study

    Lancet (London, England), 396

  • N. Fullman, J. Yearwood, S. Abay, C. Abbafati, F. Abd-Allah, Jemal Abdela, A. Abdelalim, Zegeye Abebe, Teshome Abebo, V. Aboyans, H. Abraha, D. Abreu, L. Abu-Raddad, A. Adane, R. Adedoyin, O. Adetokunboh, Tara Adhikari, M. Afarideh, A. Afshin, G. Agarwal, Dominic Agius, Anurag Agrawal, S. Agrawal, Aliasghar Kiadaliri, Miloud Aichour, M. Akibu, R. Akinyemi, T. Akinyemiju, N. Akseer, F. Lami, Fares Alahdab, Z. Al‐Aly, K. Alam, Tahiya Alam, D. Alasfoor, Mohammed Albittar, K. Alene, A. Al-Eyadhy, Syed Ali, M. Alijanzadeh, S. Aljunid, A. Alkerwi, F. Alla, P. Allebeck, Christine Allen, M. Alomari, Rajaa Al-Raddadi, U. Alsharif, K. Altirkawi, N. Alvis-Guzmán, Azmeraw Amare, Kebede Amenu, W. Ammar, Y. Amoako, N. Anber, C. Andrei, S. Androudi, C. Antonio, V. Araújo, O. Aremu, J. Ärnlöv, A. Artaman, K. Aryal, H. Asayesh, E. Asfaw, S. Asgedom, R. Asghar, M. Ashebir, N. Asseffa, T. Atey, S. Atre, M. Atteraya, L. Ávila-Burgos, Euripide Avokpaho, A. Awasthi, B. Quintanilla, A. Ayalew, H. Ayele, R. Ayer, T. Ayuk, P. Azzopardi, N. Azzopardi-Muscat, T. Babalola, H. Badali, A. Badawi, M. Banach, A. Banerjee, A. Banstola, Ryan Barber, M. Barboza, S. Barker-Collo, T. Bärnighausen, S. Barquera, L. Barrero, Q. Bassat, S. Basu, B. Baune, S. Bazargan-Hejazi, Neeraj Bedi, E. Beghi, Masoud Behzadifar, M. Behzadifar, Bayu Bekele, A. Belachew, S. Belay, Yihalem Belay, M. Bell, A. Bello, D. Bennett, James Bennett, I. Benseñor, D. Berhe, E. Bernabé, R. Bernstein, M. Beuran, A. Bhalla, P. Bhatt, S. Bhaumik, Z. Bhutta, B. Biadgo, Ali Bijani, B. Bikbov, C. Birungi, S. Biryukov, Hailemichael Bizuneh, I. Bolliger, K. Bolt, I. Bou-Orm, K. Bozorgmehr, O. Brady, A. Brazinova, N. Breitborde, H. Brenner, G. Britton, T. Brugha, Z. Butt, Lucero Cahuana-Hurtado, Ismael Campos-Nonato, J. Campuzano, J. Car, M. Car, Rosario Cárdenas, J. Carrero, F. Carvalho, C. Castañeda-Orjuela, Jacqueline Rivas, F. Catalá-López, Kelly Cercy, Julian Chalek, Hsing-Yi Chang, Jung-Chen Chang, A. Chattopadhyay, P. Chaturvedi, P. Chiang, V. Chisumpa, J. Choi, H. Christensen, D. Christopher, Sheng-Chia Chung, Liliana Ciobanu, M. Cirillo, D. Colombara, S. Conti, C. Cooper, Leslie Cornaby, P. Cortesi, Monica Cortinovis, Alexandre Pereira, Ewerton Cousin, M. Criqui, E. Cromwell, C. Crowe, J. Crump, A. Daba, B. Dachew, A. Dadi, L. Dandona, R. Dandona, P. Dargan, A. Daryani, M. Daryani, Jai Das, S. Das, J. Neves, Nicole Weaver, K. Davletov, B. Courten, D. Leo, Jan‐Walter Neve, R. Dellavalle, Gebre Demoz, Kebede Deribe, D. Jarlais, Subhojit Dey, S. Dharmaratne, M. Dhimal, Shirin Djalalinia, D. Doku, K. Dolan, E. Dorsey, Kadine Santos, Kerrie Doyle, T. Driscoll, M. Dubey, E. Dubljanin, B. Duncan, M. Echko, Dumessa Edessa, D. Edvardsson, Joshua Ehrlich, Erika Eldrenkamp, Z. El-Khatib, M. Endres, A. Endries, B. Eshrati, S. Eskandarieh, A. Esteghamati, M. Fakhar, T. Farag, M. Faramarzi, E. Faraon, Andre Faro, F. Farzadfar, A. Fatusi, M. Fazeli, V. Feigin, A. Feigl, Netsanet Fentahun, S. Fereshtehnejad, E. Fernandes, João Fernandes, D. Fijabi, I. Filip, F. Fischer, C. Fitzmaurice, A. Flaxman, L. Flor, N. Foigt, Kyle Foreman, J. Frostad, Thomas Fürst, N. Futran, E. Gakidou, S. Gallus, K. Gambashidze, A. Gamkrelidze, M. Ganji, A. Gebre, T. Gebrehiwot, A. Gebremedhin, Y. Gelaw, J. Geleijnse, D. Geremew, P. Gething, Reza Ghadimi, Khalil Falavarjani, Maryam Ghasemi-Kasman, P. Gill, A. Giref, M. Giroud, Melkamu Gishu, G. Giussani, W. Godwin, Srinivas Goli, H. Gómez-Dantés, P. Gona, A. Goodridge, S. Gopalani, Y. Goryakin, A. Goulart, A. Grada, Max Griswold, G. Grosso, H. Gugnani, Yuming Guo, Rahul Gupta, Rajeev Gupta, Tanush Gupta, T. Gupta, Vipin Gupta, J. Haagsma, V. Hachinski, N. Hafezi-Nejad, G. Hailu, R. Hamadeh, S. Hamidi, G. Hankey, H. Harb, Heather Harewood, S. Harikrishnan, J. Haro, H. Hassen, Rasmus Havmoeller, Caitlin Hawley, S. Hay, Jiawei He, S. Hearps, M. Hegazy, B. Heibati, M. Heidari, D. Hendrie, N. Henry, Víctor Ballesteros, C. Herteliu, D. Hibstu, Molla Hiluf, H. Hoek, Enayatollah Rad, N. Horita, H. Hosgood, M. Hosseini, S. Hosseini, M. Hostiuc, S. Hostiuc, D. Hoy, M. Hsairi, A. Htet, G. Hu, John Huang, K. Iburg, Fachmi Idris, E. Igumbor, C. Ikeda, B. Ileanu, O Ilesanmi, K. Innos, S. Irvani, C. Irvine, F. Islami, T. Jacobs, K. Jacobsen, N. Jahanmehr, R. Jain, Sudhir Jain, M. Jakovljevic, Moti Jalu, A. Jamal, Mehdi Javanbakht, A. Jayatilleke, P. Jeemon, Ravi Jha, V. Jha, Jacek Jóúwiak, Oommen John, Sarah Johnson, J. Jonas, Vasna Joshua, Mikk Jürisson, Z. Kabir, R. Kadel, Amaha Kahsay, Rizwan Kalani, C. Kar, M. Karanikolos, A. Karch, C. Karema, Seyed Karimi, A. Kasaeian, D. Kassa, G. Kassa, T. Kassa, N. Kassebaum, S. Katikireddi, Anil Kaul, N. Kawakami, K. Kazanjan, S. Kebede, P. Keiyoro, G. Kemp, A. Kengne, M. Kereselidze, Ezra Ketema, Y. Khader, M. Khafaie, A. Khajavi, I. Khalil, E. Khan, G. Khan, Md Khan, M. Khan, M. Khanal, Y. Khang, M. Khater, Abdullah Khoja, A. Khosravi, J. Khubchandani, G. Kibret, Daniel Kiirithio, Daniel Kim, Y. Kim, R. Kimokoti, Y. Kinfu, S. Kinra, A. Kisa, N. Kissoon, S. Kochhar, Y. Kokubo, J. Kopec, S. Kosen, P. Koul, A. Koyanagi, M. Kravchenko, K. Krishan, Kristopher Krohn, B. Defo, G. Kumar, Pushpendra Kumar, Michael Kutz, I. Kuzin, H. Kyu, D. Lad, A. Lafranconi, D. Lal, R. Lalloo, H. Lam, Q. Lan, J. Lang, V. Lansingh, S. Lansky, A. Larsson, A. Latifi, J. Lazarus, J. Leasher, P. Lee, Yirga Legesse, J. Leigh, C. Leshargie, S. Leta, J. Leung, R. Leung, M. Levi, Yongmei Li, Juan Liang, Misgan Liben, L. Lim, Stephen Lim, M. Lind, S. Linn, S. Listl, Patrick Liu, Shiwei Liu, R. Lodha, Alan Lopez, S. Lorch, S. Lorkowski, P. Lotufo, T. Lucas, R. Lunevicius, G. Lurton, R. Lyons, F. Maalouf, Erlyn Macarayan, M. Mackay, Emilie Maddison, F. Madotto, H. Razek, M. Razek, M. Majdan, R. Majdzadeh, A. Majeed, R. Malekzadeh, R. Malhotra, D. Malta, A. Mamun, Trey Manhertz, Helena Manguerra, M. Mansournia, L. Mantovani, Tsegahun Manyazewal, C. Mapoma, Christopher Margono, J. Martínez-Raga, S. Martins, F. Martins-Melo, Ira Martopullo, W. März, B. Massenburg, M. Mathur, P. Maulik, M. Mazidi, C. McAlinden, J. Mcgrath, M. Mckee, S. Mehata, R. Mehrotra, K. Mehta, Varshil Mehta, T. Meier, Fabiola Mejía-Rodríguez, K. Meles, M. Melku, P. Memiah, Z. Memish, W. Mendoza, D. Mengiste, D. Mengistu, Bereket Menota, G. Mensah, A. Meretoja, T. Meretoja (2018)

    Measuring performance on the Healthcare Access and Quality Index for 195 countries and territories and selected subnational locations: a systematic analysis from the Global Burden of Disease Study 2016

    Lancet (London, England), 391

  • W. Willett, J. Rockström, J. Rockström, B. Loken, M. Springmann, T. Lang, S. Vermeulen, S. Vermeulen, T. Garnett, D. Tilman, D. Tilman, F. DeClerck, F. DeClerck, A. Wood, M. Jonell, Michael Clark, L. Gordon, J. Fanzo, C. Hawkes, R. Zurayk, Juan Rivera, W. Vries, Lindiwe Sibanda, A. Afshin, A. Chaudhary, A. Chaudhary, M. Herrero, R. Agustina, F. Branca, A. Lartey, S. Fan, B. Crona, Elizabeth Fox, Victoria Bignet, M. Troell, M. Troell, T. Lindahl, T. Lindahl, Sudhvir Singh, S. Cornell, K. Reddy, S. Narain, S. Nishtar, Christopher Murray (2019)

    Food in the Anthropocene: the EAT–Lancet Commission on healthy diets from sustainable food systems

    The Lancet, 393

  • M. Cesare, J. Bentham, Gretchen Stevens, Bin Zhou, G. Danaei, Yuan Lu, Honor Bixby, Melanie Cowan, L. Riley, Kaveh Hajifathalian, L. Fortunato, C. Taddei, J. Bennett, Nayu Ikeda, Y. Khang, C. Kyobutungi, A. Laxmaiah, Yanping Li, Hsien-Ho Lin, J. Miranda, A. Mostafa, M. Turley, C. Paciorek, M. Gunter, M. Ezzati, Z. Abdeen, Zargar Hamid, N. Abu-Rmeileh, Benjamín Acosta-Cázares, R. Adams, W. Aekplakorn, C. Aguilar-Salinas, A. Ahmadvand, W. Ahrens, Mohamed Ali, A. Alkerwi, M. Álvarez-Pedrerol, Eman Aly, P. Amouyel, A. Amuzu, L. Andersen, S. Anderssen, Dolores Andrade, R. Anjana, H. Aounallah-Skhiri, I. Ariansen, T. Aris, N. Arlappa, D. Arveiler, F. Assah, M. Avdičová, F. Azizi, B. Babu, N. Balakrishna, P. Bandosz, J. Banegas, C. Barbagallo, A. Barceló, A. Barkat, M. Barros, I. Bata, A. Batieha, R. Batista, L. Baur, R. Beaglehole, H. Romdhane, M. Benet, A. Bernabé-Ortiz, G. Bernotiene, H. Bettiol, A. Bhagyalaxmi, Sumit Bharadwaj, S. Bhargava, Zaid Bhatti, Z. Bhutta, Hongsheng Bi, Y. Bi, P. Bjerregaard, E. Bjertness, M. Bjertness, C. Björkelund, M. Blake, A. Blokstra, Simona Bo, M. Bobák, L. Boddy, B. Boehm, H. Boeing, C. Boissonnet, V. Bongard, P. Bovet, L. Braeckman, M. Bragt, I. Brajkovich, F. Branca, Juergen Breckenkamp, H. Brenner, L. Brewster, G. Brian, G. Bruno, H. Bueno-de-Mesquita, Anna Bugge, Con Burns, A. León, J. Cacciottolo, Tilema Cama, C. Cameron, J. Camolas, G. Can, A. Cândido, V. Capuano, V. Cardoso, M. Carvalho, F. Casanueva, J. Casas, C. Caserta, K. Castetbon, S. Chamukuttan, A. Chan, Q. Chan, H. Chaturvedi, N. Chaturvedi, Chien-Jen Chen, Fangfang Chen, Huashuai Chen, Shuohua Chen, Zhengming Chen, Ching-Yu Cheng, A. Chetrit, A. Chiolero, S. Chiou, A. Chiriță-Emandi, Yumi Cho, K. Christensen, J. Chudek, R. Cífková, F. Claessens, E. Clays, H. Concin, C. Cooper, R. Cooper, T. Coppinger, S. Costanzo, D. Cottel, C. Cowell, C. Craig, A. Crujeiras, G. D’Arrigo, E. d’Orsi, J. Dallongeville, A. Damasceno, C. Damsgaard, R. Dankner, L. Dauchet, G. Backer, D. Bacquer, G. Gaetano, S. Henauw, D. Smedt, M. Deepa, A. Deev, A. Dehghan, H. Delisle, F. Delpeuch, K. Dhana, A. Castelnuovo, J. Dias-da-Costa, Alejandro Díaz, Shirin Djalalinia, Ha Do, A. Dobson, C. Donfrancesco, A. Doering, K. Doua, W. Drygas, E. Egbagbe, R. Eggertsen, U. Ekelund, J. Ati, P. Elliott, R. Engle-Stone, R. Erasmus, C. Erem, L. Eriksen, J. Peña, A. Evans, D. Faeh, C. Fall, F. Farzadfar, F. Félix-Redondo, T. Ferguson, D. Fernández-Bergés, D. Ferrante, M. Ferrari, C. Ferreccio, J. Ferrières, J. Finn, K. Fischer, Eric Flores, B. Foeger, L. Foo, Ann‐Sofie Forslund, S. Fortmann, H. Fouad, Damian Francis, M. Franco, O. Franco, G. Frontera, F. Fuchs, S. Fuchs, Y. Fujita, Takuro Furusawa, Z. Gaciong, M. Gafencu, D. Gareta, S. Garnett, J. Gaspoz, Magda Gasull, Louise Gates, J. Geleijnse, Anoosheh Ghasemian, S. Giampaoli, F. Gianfagna, J. Giovannelli, A. Giwercman, R. Goldsmith, M. Gross, J. Rivas, M. Gorbea, F. Gottrand, Sidsel Iversen, D. Grafnetter, Aneta Grajda, M. Grammatikopoulou, R. Gregor, T. Grodzicki, A. Grøntved, G. Gruden, V. Grujić, D. Gu, Ong Guan, V. Gudnason, Ramiro Guerrero, I. Guessous, A. Guimarães, M. Gulliford, Johanna Gunnlaugsdottir, X. Guo, Yin Guo, P. Gupta, O. Gureje, Beata Gurzkowska, L. Gutierrez, F. Gutzwiller, J. Halkjær, R. Hardy, R. Kumar, A. Hayes, Jiang He, M. Hendriks, L. Cadena, R. Heshmat, Ilpo Hihtaniemi, S. Ho, S. Ho, M. Hobbs, A. Hofman, Claudia Hormiga, B. Horta, L. Houti, T. Htay, A. Htet, M. Htike, Yonghua Hu, A. Hussieni, Chinh Huu, I. Huybrechts, N. Hwalla, L. Iacoviello, A. Iannone, M. Ibrahim, M. Ikram, V. Irazola, Muhammad Islam, M. Iwasaki, R. Jackson, Jeremy Jacobs, T. Jafar, K. Jamil, K. Jamrozik, G. Jasienska, C. Jiang, M. Joffres, M. Johansson, J. Jonas, T. Jørgensen, Pradeep Joshi, A. Juolevi, G. Jurak, V. Jureša, R. Kaaks, A. Kafatos, O. Kalter‐Leibovici, E. Kapantais, A. Kasaeian, J. Katz, P. Kaur, M. Kavousi, U. Keil, L. Boker, R. Kelishadi, H. Kemper, A. Kengne, M. Kersting, T. Key, Y. Khader, D. Khalili, K. Khaw, Ilse Khouw, S. Kiechl, J. Killewo, Jeongseon Kim, Y. Kiyohara, Jeannette Klimont, E. Kolle, P. Kolsteren, P. Korrovits, S. Koskinen, K. Kouda, S. Kozieł, W. Kratzer, S. Krokstad, D. Kromhout, H. Kruger, K. Kula, Z. Kułaga, R. Kumar, Y. Kusuma, K. Kuulasmaa, F. Laamiri, T. Laatikainen, C. Lachat, Y. Laid, T. Lam, O. Landrove, V. Lánská, G. Lappas, L. Laugsand, Khanh Bao, T. Le, C. Leclercq, Jongseok Lee, Jeonghee Lee, T. Lehtimäki, Rampal Lekhraj, L. León-Muñoz, W. Lim, M. Lima-Costa, Xu Lin, A. Linneberg, L. Lissner, M. Litwin, Jing-Jie Liu, R. Lorbeer, P. Lotufo, J. Lozano, D. Lukšienė, A. Lundqvist, N. Lunet, Per Lytsy, G. Ma, S. Machi, S. Maggi, D. Magliano, Marcia Makdisse, R. Malekzadeh, R. Malhotra, Kodavanti Rao, Y. Manios, J. Mann, E. Manzato, P. Margozzini, O. Markey, P. Marques-Vidal, J. Marrugat, Y. Martin-Prével, R. Martorell, S. Masoodi, T. Matsha, A. Mazur, J. Mbanya, S. McFarlane, S. Mcgarvey, M. Mckee, S. McLachlan, R. McLean, B. McNulty, S. Yusof, S. Médiène-Benchekor, A. Meirhaeghe, C. Meisinger, L. Mendes, A. Menezes, Gert Mensink, I. Meshram, A. Metspalu, J. Mi, K. Michaelsen, Kairit Mikkel, Jody Miller, J. Miquel, M. Misigoj-Durakovic, M. Mohamed, K. Mohammad, N. Mohammadifard, V. Mohan, M. Yusoff, Drude Molbo, N. Møller, D. Molnár, C. Mondo, E. Monterrubio, K. Monyeki, L. Moreira, A. Morejón, L. Moreno, K. Morgan, Erik Mortensen, G. Moschonis, M. Mossakowska, J. Mota, M. Motlagh, Jorge Motta, Thet Mu, M. Muiesan, Martina Mueller-Nurasyid, N. Murphy, J. Mursu, E. Murtagh, K. Musa, V. Musil, G. Nagel, Harunobu Nakamura, Jana Námešná, E. Nang, V. Nangia, M. Nankap, S. Narake, E. Navarrete-Muñoz, Ilona Nenko, M. Neovius, F. Nervi, H. Neuhauser, N. Nguyen, Q. Nguyen, R. Nieto-Martínez, G. Ning, T. Ninomiya, S. Nishtar, M. Noale, T. Norat, D. Noto, M. Nsour, D. O’Reilly, Angélica Ochoa-Avilés, Kyungwon Oh, Iman Olayan, M. Olinto, M. Ołtarzewski, M. Omar, P. Ordunez, Ana Ortíz, M. Osler, C. Osmond, S. Ostojić, J. Otero, K. Overvad, F. Paccaud, C. Padez, A. Pająk, D. Palli, A. Palloni, L. Palmieri, S. Panda-Jonas, F. Panza, W. Parnell, M. Parsaeian, M. Pednekar, P. Peeters, S. Peixoto, A. Pereira, C. Pérez, A. Peters, N. Peykari, Son Pham, I. Pigeot, H. Pikhart, A. Pilav, L. Pilotto, F. Pistelli (2016)

    Trends in adult body-mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population-based measurement studies with 19·2 million participants

    Lancet (London, England), 387

  • (2008)

    National air quality status and trends through 2007. Research Triangle Park, North Carolina: U.S. Environmental Protection Agency

  • J. Davis, R. Elias, L. Grant (1993)

    Current issues in human lead exposure and regulation of lead.

    Neurotoxicology, 14 2-3

  • Alan Lopez, T. Adair (2019)

    Is the long-term decline in cardiovascular-disease mortality in high-income countries over? Evidence from national vital statistics.

    International journal of epidemiology

  • B. Swinburn, V. Kraak, S. Allender, Vincent Atkins, Phillip Baker, J. Bogard, H. Brinsden, A. Calvillo, O. Schutter, R. Devarajan, M. Ezzati, S. Friel, S. Goenka, Ross Hammond, G. Hastings, C. Hawkes, M. Herrero, P. Hovmand, M. Howden, L. Jaacks, Ariadne Kapetanaki, Matt Kasman, H. Kuhnlein, S. Kumanyika, B. Larijani, T. Lobstein, M. Long, V. Matsudo, S. Mills, Gareth Morgan, A. Morshed, Patricia Nece, An Pan, David Patterson, G. Sacks, M. Shekar, Geoff Simmons, W. Smit, A. Tootee, S. Vandevijvere, W. Waterlander, L. Wolfenden, W. Dietz (2019)

    The Global Syndemic of Obesity, Undernutrition, and Climate Change: The Lancet Commission report

    The Lancet, 393

  • I. Millwood, R. Walters, X. Mei, Yu Guo, Ling Yang, Z. Bian, D. Bennett, Yiping Chen, C. Dong, R. Hu, G. Zhou, Bo Yu, Weifang Jia, S. Parish, R. Clarke, G. Smith, R. Collins, M. Holmes, Liming Li, R. Peto, Zhengming Chen (2019)

    Conventional and genetic evidence on alcohol and vascular disease aetiology: a prospective study of 500 000 men and women in China

    Lancet (London, England), 393

  • E. Gakidou, A. Afshin, A. Abajobir, K. Abate, C. Abbafati, K. Abbas, F. Abd-Allah, A. Abdulle, S. Abera, V. Aboyans, L. Abu-Raddad, N. Abu-Rmeileh, G. Abyu, I. Adedeji, O. Adetokunboh, M. Afarideh, Anurag Agrawal, S. Agrawal, A. Kiadaliri, H. Ahmadieh, M. Ahmed, Amani Aichour, Ibtihel Aichour, Miloud Aichour, R. Akinyemi, N. Akseer, Fares Alahdab, Z. Al‐Aly, K. Alam, N. Alam, Tahiya Alam, D. Alasfoor, K. Alene, Komal Ali, Reza Alizadeh-Navaei, A. Alkerwi, F. Alla, P. Allebeck, Rajaa Al-Raddadi, U. Alsharif, K. Altirkawi, N. Alvis-Guzmán, Azmeraw Amare, E. Amini, W. Ammar, Y. Amoako, H. Ansari, J. Antó, C. Antonio, P. Anwari, N. Arian, J. Ärnlöv, A. Artaman, K. Aryal, H. Asayesh, S. Asgedom, T. Atey, L. Ávila-Burgos, Euripide Avokpaho, A. Awasthi, P. Azzopardi, U. Bacha, A. Badawi, K. Balakrishnan, S. Ballew, A. Barać, Ryan Barber, S. Barker-Collo, T. Bärnighausen, S. Barquera, L. Barregard, L. Barrero, C. Batis, K. Battle, B. Baune, J. Beardsley, Neeraj Bedi, E. Beghi, M. Bell, D. Bennett, James Bennett, I. Benseñor, A. Berhane, D. Berhe, E. Bernabé, B. Betsu, M. Beuran, A. Beyene, A. Bhansali, Z. Bhutta, B. Bikbov, C. Birungi, S. Biryukov, C. Blosser, D. Boneya, I. Bou-Orm, M. Brauer, N. Breitborde, H. Brenner, T. Brugha, Lemma Bulto, Bl Baumgarner, Z. Butt, Lucero Cahuana-Hurtado, Rosario Cárdenas, J. Carrero, C. Castañeda-Orjuela, F. Catalá-López, Kelly Cercy, Hsing-Yi Chang, F. Charlson, Odgerel Chimed-Ochir, V. Chisumpa, Abdulaal Chitheer, H. Christensen, D. Christopher, M. Cirillo, A. Cohen, H. Comfort, C. Cooper, J. Coresh, Leslie Cornaby, P. Cortesi, M. Criqui, J. Crump, L. Dandona, R. Dandona, J. Neves, G. Davey, D. Davițoiu, K. Davletov, B. Courten, L. Degenhardt, Selina Deiparine, R. Dellavalle, Kebede Deribe, A. Deshpande, S. Dharmaratne, E. Ding, Shirin Djalalinia, H. Do, K. Dokova, D. Doku, E. Dorsey, T. Driscoll, M. Dubey, B. Duncan, Sarah Duncan, Natalie Ebert, H. Ebrahimi, Z. El-Khatib, A. Enayati, A. Endries, S. Ermakov, H. Erskine, B. Eshrati, S. Eskandarieh, A. Esteghamati, Kara Estep, E. Faraon, C. Farinha, Andre Faro, F. Farzadfar, K. Fay, V. Feigin, S. Fereshtehnejad, João Fernandes, A. Ferrari, T. Feyissa, I. Filip, F. Fischer, C. Fitzmaurice, A. Flaxman, N. Foigt, Kyle Foreman, J. Frostad, N. Fullman, Thomas Fürst, J. Furtado, M. Ganji, A. García-Basteiro, T. Gebrehiwot, J. Geleijnse, Ayele Geleto, Bikila Gemechu, H. Gesesew, P. Gething, A. Ghajar, K. Gibney, P. Gill, R. Gillum, A. Giref, Melkamu Gishu, G. Giussani, W. Godwin, P. Gona, A. Goodridge, S. Gopalani, Y. Goryakin, A. Goulart, Nicholas Graetz, H. Gugnani, Jingwen Guo, Rajeev Gupta, Tanush Gupta, Vipin Gupta, R. Gutiérrez, V. Hachinski, N. Hafezi-Nejad, G. Hailu, R. Hamadeh, S. Hamidi, M. Hammami, A. Handal, G. Hankey, H. Harb, H. Hareri, Mohammad Hassanvand, Rasmus Havmoeller, Caitlin Hawley, S. Hay, M. Hedayati, D. Hendrie, I. Heredia-Pi, H. Hoek, N. Horita, H. Hosgood, S. Hostiuc, D. Hoy, M. Hsairi, G. Hu, Hsiang Huang, John Huang, K. Iburg, C. Ikeda, M. Inoue, C. Irvine, M. Jackson, K. Jacobsen, N. Jahanmehr, M. Jakovljevic, Alejandra Jáuregui, Mehdi Javanbakht, P. Jeemon, L. Johansson, C. Johnson, J. Jonas, Mikk Jürisson, Z. Kabir, R. Kadel, Amaha Kahsay, R. Kamal, A. Karch, C. Karema, A. Kasaeian, N. Kassebaum, A. Kastor, S. Katikireddi, N. Kawakami, P. Keiyoro, Sefonias Kelbore, L. Kemmer, A. Kengne, C. Kesavachandran, Y. Khader, I. Khalil, E. Khan, Y. Khang, A. Khosravi, J. Khubchandani, C. Kieling, Daniel Kim, J. Kim, Y. Kim, R. Kimokoti, Y. Kinfu, A. Kisa, K. Kissimova-Skarbek, M. Kivimäki, L. Knibbs, A. Knudsen, J. Kopec, S. Kosen, P. Koul, A. Koyanagi, M. Kravchenko, Kristopher Krohn, H. Kromhout, B. Defo, B. Bicer, G. Kumar, Michael Kutz, H. Kyu, D. Lal, R. Lalloo, T. Lallukka, Q. Lan, V. Lansingh, A. Larsson, Alexander Lee, P. Lee, J. Leigh, J. Leung, M. Levi, Yichong Li, Yongmei Li, Xiaofeng Liang, Misgan Liben, S. Linn, Patrick Liu, R. Lodha, G. Logroscino, Katherine Looker, Alan Lopez, S. Lorkowski, P. Lotufo, R. Lozano, R. Lunevicius, Erlyn Macarayan, H. Razek, M. Razek, M. Majdan, R. Majdzadeh, A. Majeed, R. Malekzadeh, R. Malhotra, D. Malta, A. Mamun, Helena Manguerra, L. Mantovani, C. Mapoma, R. Martin, J. Martínez-Raga, F. Martins-Melo, M. Mathur, K. Matsushita, R. Matzopoulos, M. Mazidi, C. McAlinden, J. Mcgrath, S. Mehata, M. Mehndiratta, T. Meier, Y. Melaku, P. Memiah, Z. Memish, W. Mendoza, M. Mengesha, G. Mensah, Gert Mensink, S. Mereta, A. Meretoja, T. Meretoja, H. Mezgebe, R. Micha, Anoushka Millear, T. Miller, S. Minnig, M. Mirarefin, E. Mirrakhimov, Awoke Misganaw, S. Mishra, K. Mohammad, K. Mohammed, S. Mohammed, N. Ibrahim, Murali Mohan, A. Mokdad, L. Monasta, J. Hernandez, M. Montico, M. Moradi-Lakeh, P. Moraga, L. Morawska, S. Morrison, Cliff Mountjoy-Venning, U. Mueller, Erin Mullany, Kate Muller, G. Murthy, K. Musa, M. Naghavi, A. Naheed, V. Nangia, G. Natarajan, Ionut Negoi, R. Negoi, C. Nguyen, Grant Nguyen, Minh Nguyen, Q. Nguyen, T. Nguyen, E. Nichols, D. Ningrum, Marika Nomura, V. Nong, O. Norheim, B. Norrving, J. Noubiap, C. Obermeyer, F. Ogbo, Hwan-Jung Oh, O. Oladimeji, A. Olagunju, T. Olagunju, Pedro Olivares, H. Olsen, B. Olusanya, J. Olusanya, John Opio, Eyal Oren, Alberto Ortiz, E. Ota, M. Owolabi, Mahesh Pa, R. Pacella, A. Pana, B. Panda, S. Panda-Jonas, J. Pandian, C. Papachristou, Eun‐Kee Park, C. Parry, S. Patten, G. Patton, David Pereira, N. Perico, K. Pesudovs, M. Petzold, M. Phillips, J. Pillay, M. Piradov, F. Pishgar, D. Plass, Martin Pletcher, S. Polinder, S. Popova, R. Poulton, F. Pourmalek, N. Prasad, Carrie Purcell, M. Qorbani, A. Radfar, Anwar Rafay, A. Rahimi-Movaghar, V. Rahimi-Movaghar, Mahfuzar Rahman, Mohammad Rahman, M. Rahman, R. Rai, S. Rajšić, U. Ram, S. Rawaf, C. Rehm, J. Rehm, R. Reiner, M. Reitsma, L. Reynales-Shigematsu, G. Remuzzi, A. Renzaho, S. Resnikoff, S. Rezaei, A. Ribeiro, J. Rivera, K. Roba, D. Rojas-Rueda, Yesenia Román, R. Room, G. Roshandel, Gregory Roth, D. Rothenbacher, Enrico Rubagotti, L. Rushton, Nafis Sadat, M. Safdarian, S. Safi, S. Safiri, R. Sahathevan, Joseph Salama, J. Salomon, A. Samy, J. Sanabria, M. Sánchez-Niño, Tania Sánchez-Pimienta, D. Santomauro, I. Santos, M. Milicevic, B. Sartorius, Maheswar Satpathy, M. Sawhney, S. Saxena, E. Schaeffner, M. Schmidt, I. Schneider, A. Schutte, D. Schwebel, F. Schwendicke, S. Seedat, S. Sepanlou, Berrin Serdar (2017)

    Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016

    Lancet (London, England), 390

  • (2007)

    Research Triangle Park, North Carolina: U.S. Environmental Protection Agency. Oice of Air Quality Planning and Standards

  • S. Burgess, N. Timpson, S. Ebrahim, G. Smith (2015)

    Mendelian randomization: where are we now and where are we going?

    International journal of epidemiology, 44 2

  • S. Fisch (1973)

    Mortality from ischemic heart disease.

    JAMA, 224 11

  • C. Murray, M. Ezzati, Alan Lopez, A. Rodgers, S. Hoorn (2003)

    Comparative quantification of health risks: Conceptual framework and methodological issues

    Population Health Metrics, 1

  • J. Stanaway, A. Afshin, E. Gakidou, Stephen Lim, D. Abate, K. Abate, C. Abbafati, N. Abbasi, Hedayat Abbastabar, F. Abd-Allah, Jemal Abdela, A. Abdelalim, Ibrahim Abdollahpour, R. Abdulkader, Molla Abebe, Zegeye Abebe, S. Abera, Olifan Abil, H. Abraha, Aklilu Abrham, L. Abu-Raddad, Niveen Abu-Rmeileh, Manfred Accrombessi, Dilaram Acharya, Pawan Acharya, Abdullahi Adamu, A. Adane, O. Adebayo, R. Adedoyin, Victor Adekanmbi, Z. Ademi, O. Adetokunboh, Mina Adib, Amha Admasie, J. Adsuar, K. Afanvi, M. Afarideh, Gina Agarwal, Anju Aggarwal, S. Aghayan, Anurag Agrawal, S. Agrawal, A. Ahmadi, Mehdi Ahmadi, H. Ahmadieh, M. Ahmed, Amani Aichour, Ibtihel Aichour, Miloud Aichour, M. Akbari, T. Akinyemiju, N. Akseer, Z. Al‐Aly, A. Al-Eyadhy, H. Al-Mekhlafi, Fares Alahdab, K. Alam, Samiah Alam, Tahiya Alam, Alaa Alashi, S. Alavian, K. Alene, Komal Ali, Syed Ali, M. Alijanzadeh, Reza Alizadeh-Navaei, S. Aljunid, A. Alkerwi, François Alla, U. Alsharif, K. Altirkawi, N. Alvis-Guzmán, Azmeraw Amare, W. Ammar, N. Anber, Jason Anderson, C. Andrei, S. Androudi, Megbaru Animut, Mina Anjomshoa, Mustafa Ansha, Josep Antó, C. Antonio, P. Anwari, L. Appiah, Seth Appiah, J. Arabloo, O. Aremu, J. Ärnlöv, A. Artaman, K. Aryal, H. Asayesh, Z. Ataro, M. Ausloos, Euripide Avokpaho, A. Awasthi, B. Quintanilla, R. Ayer, T. Ayuk, P. Azzopardi, A. Babazadeh, H. Badali, A. Badawi, Kalpana Balakrishnan, A. Bali, Kylie Ball, S. Ballew, Maciej Banach, J. Banoub, A. Barać, S. Barker-Collo, T. Bärnighausen, L. Barrero, Sanjay Basu, B. Baune, S. Bazargan-Hejazi, Neeraj Bedi, Ettore Beghi, Masoud Behzadifar, M. Behzadifar, Y. Béjot, Bayu Bekele, Eyasu Bekru, Ezra Belay, Yihalem Belay, Morag Bell, A. Bello, Derrick Bennett, I. Benseñor, Gilles Bergeron, A. Berhane, Eduardo Bernabé, Robert Bernstein, M. Beuran, Tina Beyranvand, N. Bhala, Ashish Bhalla, S. Bhattarai, Z. Bhutta, B. Biadgo, Ali Bijani, B. Bikbov, V. Bilano, Nigus Bililign, M. Sayeed, D. Bisanzio, Tuhin Biswas, T. Bjørge, B. Blacker, A. Bleyer, R. Borschmann, I. Bou-Orm, S. Boufous, R. Bourne, O. Brady, Michael Brauer, A. Brazinova, N. Breitborde, H. Brenner, A. Briko, Gabrielle Britton, T. Brugha, Rachelle Buchbinder, Richard Burnett, Reinhard Busse, Z. Butt, Leah Cahill, Lucero Cahuana-Hurtado, Ismael Campos-Nonato, Rosario Cárdenas, Giulia Carreras, J. Carrero, F. Carvalho, C. Castañeda-Orjuela, Jacqueline Rivas, Franz Castro, Ferrán Catalá-López, K. Causey, Kelly Cercy, Ester Cerin, Y. Chaiah, Hsing-Yi Chang, Jung-Chen Chang, Kai-Lan Chang, F. Charlson, Aparajita Chattopadhyay, Vijay Chattu, M. Chee, Ching-Yu Cheng, Adrienne Chew, P. Chiang, Odgerel Chimed-Ochir, K. Chin, Abdulaal Chitheer, J. Choi, Rajiv Chowdhury, H. Christensen, D. Christopher, Sheng-Chia Chung, F. Cicuttini, Massimo Cirillo, Aaron Cohen, D. Collado-Mateo, C. Cooper, Owen Cooper, J. Coresh, Leslie Cornaby, P. Cortesi, Monica Cortinovis, Megan Costa, Ewerton Cousin, M. Criqui, E. Cromwell, D. Cundiff, A. Daba, B. Dachew, A. Dadi, Albertino Damasceno, L. Dandona, R. Dandona, Sarah Darby, P. Dargan, A. Daryani, R. Gupta, J. Neves, T. Dasa, Aditya Dash, D. Davițoiu, K. Davletov, Vanessa Cruz-Góngora, Fernando Hoz, Diego Leo, Jan‐Walter Neve, L. Degenhardt, Selina Deiparine, R. Dellavalle, Gebre Demoz, E. Denova-Gutiérrez, Kebede Deribe, N. Dervenis, Aniruddha Deshpande, Don Jarlais, Getenet Dessie, G. deVeber, Subhojit Dey, S. Dharmaratne, M. Dhimal, Mesfin Dinberu, Eric Ding, Helen Diro, Shirin Djalalinia, H. Do, K. Dokova, D. Doku, Kerrie Doyle, T. Driscoll, Manisha Dubey, E. Dubljanin, Eyasu Duken, B. Duncan, A. Durães, Natalie Ebert, H. Ebrahimi, S. Ebrahimpour, D. Edvardsson, A. Effiong, A. Eggen, C. Bcheraoui, Z. El-Khatib, I. Elyazar, A. Enayati, A. Endries, Benjamin Er, H. Erskine, S. Eskandarieh, A. Esteghamati, Kara Estep, H. Fakhim, Mahbobeh Faramarzi, M. Fareed, Talha Farid, C. Farinha, A. Farioli, Andre Faro, M. Farvid, M. Farzaei, Batool Fatima, K. Fay, A. Fazaeli, Valery Feigin, A. Feigl, S. Fereshtehnejad, Eduarda Fernandes, João Catalão, Giannina Ferrara, A. Ferrari, Manuela Ferreira, I. Filip, J. Finger, F. Fischer, N. Foigt, Kyle Foreman, Takeshi Fukumoto, N. Fullman, Thomas Fürst, João Furtado, Neal Futran, S. Gall, Silvano Gallus, A. Gamkrelidze, M. Ganji, A. García-Basteiro, William Gardner, A. Gebre, A. Gebremedhin, Teklu Gebremichael, Tilayie Gelano, J. Geleijnse, Y. Geramo, P. Gething, Kebede Gezae, Reza Ghadimi, K. Ghadiri, Khalil Falavarjani, Maryam Ghasemi-Kasman, M. Ghimire, Rakesh Ghosh, A. Ghoshal, Simona Giampaoli, P. Gill, Tiffany Gill, R. Gillum, I. Ginawi, G. Giussani, E. Gnedovskaya, W. Godwin, Srinivas Goli, H. Gómez-Dantés, P. Gona, S. Gopalani, A. Goulart, A. Grada, M. Grams, G. Grosso, H. Gugnani, Yuming Guo, Rahul Gupta, R. Gupta, Tanush Gupta, R. Gutiérrez, Daniela Gutiérrez-Torres, J. Haagsma, T. Habtewold, Vladimir Hachinski, N. Hafezi-Nejad, Tekleberhan Hagos, Tewodros Hailegiyorgis, G. Hailu, Arvin Haj-Mirzaian, A. Haj-Mirzaian, R. Hamadeh, S. Hamidi, A. Handal, Graeme Hankey, Yuantao Hao, H. Harb, S. Harikrishnan, J. Haro, H. Hassankhani, H. Hassen, Rasmus Havmoeller, Caitlin Hawley, S. Hay, A. Hedayatizadeh-Omran, B. Heibati, B. Heidari, Mohsen Heidari, D. Hendrie, A. Henok, I. Heredia-Pi, C. Herteliu, F. Heydarpour, S. Heydarpour, D. Hibstu, T. Higazi, E. Hilawe, Hans Hoek, Howard Hoffman, M. Hole, Enayatollah Rad, Praveen Hoogar, H. Hosgood, Seyed Hosseini, M. Hosseinzadeh, M. Hostiuc, S. Hostiuc, D. Hoy, M. Hsairi, Thomas Hsiao, Guoqing Hu, Howard Hu, John Huang, Mamusha Hussen, Chantal Huynh, K. Iburg, Nayu Ikeda, O Ilesanmi, Usman Iqbal, S. Irvani, C. Irvine, Sheikh Islam, F. Islami, Maria Jackson, K. Jacobsen, L. Jahangiry, N. Jahanmehr, Sudhir Jain, Mihajlo Jakovljevic, S. James, S. Jassal, A. Jayatilleke, P. Jeemon, Ravi Jha, Vivekanand Jha, John Ji, J. Jonas, J. Jonnagaddala, Z. Shushtari, Ankur Joshi, J. Jozwiak, Mikk Jürisson, Zubair Kabir, Amaha Kahsay, Rizwan Kalani, T. Kanchan, Surya Kant, Chittaranjan Kar, Manoochehr Karami, B. Matin, A. Karch, C. Karema, Narges Karimi, Seyed Karimi, A. Kasaeian, D. Kassa, G. Kassa, T. Kassa, N. Kassebaum, S. Katikireddi, Anil Kaul, Norito Kawakami, Zhila Kazemi, Ali Karyani, A. Kefale, P. Keiyoro, Grant Kemp, A. Kengne, Andre Keren, C. Kesavachandran, Yousef Khader, Behzad Khafaei, M. Khafaie, Alireza Khajavi, Nauman Khalid, I. Khalil, Gulfaraz Khan, M. Khan, Muhammad Khan, Y. Khang, M. Khater, Mohammad Khazaei, Habibolah Khazaie, Abdullah Khoja, A. Khosravi, Mohammad Khosravi, A. Kiadaliri, Daniel Kiirithio, Cho-il Kim, Daniel Kim, Young-Eun Kim, Y. Kim, R. Kimokoti, Y. Kinfu, A. Kisa, K. Kissimova-Skarbek, Mika Kivimäki, L. Knibbs, A. Knudsen, S. Kochhar, Yoshihiro Kokubo, T. Kolola, J. Kopec, S. Kosen, P. Koul, Ai Koyanagi, M. Kravchenko, K. Krishan, Kristopher Krohn, H. Kromhout, B. Defo (2018)

    Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017

    Lancet (London, England), 392

  • T. Dignam, R. Kaufmann, Lauren LeStourgeon, M. Brown (2019)

    Control of Lead Sources in the United States, 1970-2017: Public Health Progress and Current Challenges to Eliminating Lead Exposure

    Journal of Public Health Management and Practice, 25

  • acknowledges funding from a NRS Senior Clinical Fellowship

  • (2000)

    at aiming to preserve the confidentiality of individual data in accordance with the General Statistics Law

  • (2017)

    UN Task Force on NCDs: The government of Ethiopia is working jointly with the UN system to strengthen the national NCD response

  • The National Environmental Public Health Tracking Program

  • Prof Christopher, J. Murray, Navid Rabiee (2020)

    Global burden of 369 diseases and injuries in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019

    Lancet (London, England), 396

  • J. Adesina, S. Piketh, M. Qhekwana, R. Burger, B. Language, G. Mkhatshwa (2020)

    Contrasting indoor and ambient particulate matter concentrations and thermal comfort in coal and non-coal burning households at South Africa Highveld.

    The Science of the total environment, 699

  • Janet Chung-Hall, L. Craig, Shannon Gravely, Natalie Sansone, G. Fong (2018)

    Impact of the WHO FCTC over the first decade: a global evidence review prepared for the Impact Assessment Expert Group

    Tobacco Control, 28

  • acknowledges support from the NIHR

  • M. Forouzanfar, A. Afshin, Lily Alexander, H. Anderson, Z. Bhutta, S. Biryukov, M. Brauer, R. Burnett, Kelly Cercy, F. Charlson, A. Cohen, L. Dandona, Kara Estep, A. Ferrari, J. Frostad, N. Fullman, P. Gething, W. Godwin, Max Griswold, S. Hay, Y. Kinfu, H. Kyu, H. Larson, Xiaofeng Liang, Stephen Lim, Patrick Liu, Alan Lopez, R. Lozano, L. Marczak, G. Mensah, A. Mokdad, M. Moradi-Lakeh, M. Naghavi, B. Neal, M. Reitsma, Gregory Roth, J. Salomon, P. Sur, T. Vos, Joseph Wagner, Haidong Wang, Yi Zhao, Maigeng Zhou, Gunn Aasvang, A. Abajobir, K. Abate, C. Abbafati, K. Abbas, F. Abd-Allah, A. Abdulle, S. Abera, Biju Abraham, L. Abu-Raddad, G. Abyu, A. Adebiyi, I. Adedeji, Z. Ademi, A. Adou, J. Adsuar, E. Agardh, A. Agarwal, Anurag Agrawal, A. Kiadaliri, O. Ajala, T. Akinyemiju, Z. Al‐Aly, K. Alam, N. Alam, S. Aldhahri, R. Aldridge, Z. Alemu, R. Ali, A. Alkerwi, F. Alla, P. Allebeck, U. Alsharif, K. Altirkawi, E. Martin, N. Alvis-Guzmán, Azmeraw Amare, A. Amberbir, A. Amegah, H. Amini, W. Ammar, S. Amrock, H. Andersen, B. Anderson, C. Antonio, P. Anwari, J. Ärnlöv, A. Artaman, H. Asayesh, R. Asghar, Reza Assadi, S. Atique, Euripide Avokpaho, A. Awasthi, B. Quintanilla, P. Azzopardi, U. Bacha, A. Badawi, M. Bahit, K. Balakrishnan, A. Barać, Ryan Barber, S. Barker-Collo, T. Bärnighausen, S. Barquera, L. Barregard, L. Barrero, S. Basu, C. Batis, S. Bazargan-Hejazi, J. Beardsley, Neeraj Bedi, E. Beghi, B. Bell, M. Bell, A. Bello, D. Bennett, I. Benseñor, A. Berhane, E. Bernabé, B. Betsu, A. Beyene, N. Bhala, A. Bhansali, S. Bhatt, S. Biadgilign, B. Bikbov, D. Bisanzio, E. Bjertness, J. Blore, R. Borschmann, S. Boufous, R. Bourne, M. Brainin, A. Brazinova, N. Breitborde, H. Brenner, D. Broday, T. Brugha, B. Brunekreef, Z. Butt, L. Cahill, Bianca Calabria, Ismael Campos-Nonato, Rosario Cárdenas, D. Carpenter, J. Carrero, Daniel Casey, C. Castañeda-Orjuela, Jacqueline Rivas, R. Castro, F. Catalá-López, Jung-Chen Chang, P. Chiang, Mirriam Chibalabala, Odgerel Chimed-Ochir, V. Chisumpa, Abdulaal Chitheer, J. Choi, H. Christensen, D. Christopher, Liliana Ciobanu, M. Coates, Samantha Colquhoun, A. Manzano, Leslie Cooper, Kimberly Cooperrider, Leslie Cornaby, Monica Cortinovis, J. Crump, L. Cuevas-Nasu, A. Damasceno, R. Dandona, S. Darby, P. Dargan, J. Neves, A. Davis, K. Davletov, E. Castro, Vanessa Cruz-Góngora, D. Leo, L. Degenhardt, Liana Gobbo, B. Pozo-Cruz, R. Dellavalle, A. Deribew, D. Jarlais, S. Dharmaratne, P. Dhillon, C. Díaz-Torné, D. Dicker, E. Ding, E. Dorsey, Kerrie Doyle, T. Driscoll, L. Duan, M. Dubey, B. Duncan, I. Elyazar, A. Endries, S. Ermakov, H. Erskine, B. Eshrati, A. Esteghamati, S. Fahimi, E. Faraon, T. Farid, C. Farinha, Andre Faro, M. Farvid, F. Farzadfar, V. Feigin, S. Fereshtehnejad, J. Fernandes, F. Fischer, J. Fitchett, T. Fleming, N. Foigt, Kyle Foreman, F. Fowkes, R. Franklin, Thomas Fürst, N. Futran, E. Gakidou, A. García-Basteiro, T. Gebrehiwot, A. Gebremedhin, J. Geleijnse, B. Gessner, A. Giref, M. Giroud, Melkamu Gishu, G. Giussani, S. Goenka, M. Gómez-Cabrera, H. Gómez-Dantés, P. Gona, A. Goodridge, S. Gopalani, C. Gotay, A. Goto, H. Gouda, H. Gugnani, F. Guillemin, Yuming Guo, Rahul Gupta, Rajeev Gupta, R. Gutiérrez, J. Haagsma, N. Hafezi-Nejad, D. Haile, G. Hailu, Y. Halasa, R. Hamadeh, S. Hamidi, A. Handal, G. Hankey, Y. Hao, H. Harb, S. Harikrishnan, J. Haro, Mohammad Hassanvand, T. Hassen, Rasmus Havmoeller, I. Heredia-Pi, N. Hernández-Llanes, P. Heydarpour, H. Hoek, H. Hoffman, M. Horino, N. Horita, H. Hosgood, D. Hoy, M. Hsairi, A. Htet, G. Hu, John Huang, A. Husseini, S. Hutchings, I. Huybrechts, K. Iburg, B. Idrisov, B. Ileanu, M. Inoue, T. Jacobs, K. Jacobsen, N. Jahanmehr, M. Jakovljevic, H. Jansen, S. Jassal, Mehdi Javanbakht, S. Jayaraman, A. Jayatilleke, S. Jee, P. Jeemon, V. Jha, Ying Jiang, T. Jibat, Ye Jin, C. Johnson, J. Jonas, Z. Kabir, Y. Kalkonde, R. Kamal, H. Kan, A. Karch, C. Karema, C. Karimkhani, A. Kasaeian, Anil Kaul, N. Kawakami, Dhruv Kazi, P. Keiyoro, L. Kemmer, A. Kemp, A. Kengne, A. Keren, C. Kesavachandran, Y. Khader, Abdur Khan, E. Khan, G. Khan, Y. Khang, S. Khatibzadeh, S. Khera, T. Khoja, J. Khubchandani, C. Kieling, Cho-il Kim, Daniel Kim, R. Kimokoti, N. Kissoon, M. Kivipelto, L. Knibbs, Y. Kokubo, J. Kopec, P. Koul, A. Koyanagi, M. Kravchenko, H. Kromhout, H. Krueger, Tiffany Ku, B. Defo, R. Kuchenbecker, B. Bicer, E. Kuipers, G. Kumar, G. Kwan, D. Lal, R. Lalloo, T. Lallukka, Q. Lan, A. Larsson, A. Latif, A. Lawrynowicz, J. Leasher, J. Leigh, J. Leung, M. Levi, Xiaohong Li, Yichong Li, Juan Liang, Shiwei Liu, B. Lloyd, G. Logroscino, P. Lotufo, R. Lunevicius, Michael MacIntyre, M. Mahdavi, M. Majdan, A. Majeed, R. Malekzadeh, D. Malta, Wondimu Manamo, C. Mapoma, W. Marcenes, R. Martin, J. Martínez-Raga, F. Masiye, K. Matsushita, R. Matzopoulos, B. Mayosi, J. Mcgrath, M. Mckee, P. Meaney, C. Medina, A. Mehari, Fabiola Mejía-Rodríguez, A. Mekonnen, Y. Melaku, Z. Memish, W. Mendoza, Gert Mensink, A. Meretoja, T. Meretoja, Y. Mesfin, F. Mhimbira, Anoushka Millear, T. Miller, E. Mills, M. Mirarefin, Awoke Misganaw, C. Mock, A. Mohammadi, S. Mohammed, G. Mola, L. Monasta, J. Hernandez, M. Montico, L. Morawska, R. Mori, D. Mozaffarian, U. Mueller, Erin Mullany, J. Mumford, G. Murthy, J. Nachega, A. Naheed, V. Nangia, N. Nassiri, J. Newton, Marie Ng, Q. Nguyen, M. Nisar, P. Pete, O. Norheim, Rosana Norman, B. Norrving, L. Nyakarahuka, C. Obermeyer, F. Ogbo, I. Oh, O. Oladimeji, Pedro Olivares, H. Olsen, B. Olusanya, J. Olusanya, John Opio, Eyal Oren, R. Orozco, Alberto Ortiz, E. Ota, Mahesh Pa, A. Pana, Eun‐Kee Park, C. Parry, M. Parsaeian, Tejas Patel, A. Caicedo, Snehal Patil, S. Patten, G. Patton, N. Pearce, David Pereira, N. Perico, K. Pesudovs, M. Petzold, M. Phillips, F. Piel, J. Pillay, D. Plass, S. Polinder, C. Pond, C. Pope, D. Pope, S. Popova, R. Poulton, F. Pourmalek, N. Prasad, M. Qorbani, Rynaz Rabiee, A. Radfar, Anwar Rafay, V. Rahimi-Movaghar, Mahfuzar Rahman, Mohammad Rahman, S. Rahman, R. Rai, S. Rajšić, M. Raju, U. Ram, S. Rana, K. Ranganathan, P. Rao, Christian García, A. Refaat, C. Rehm, J. Rehm, Nikolas Reinig, G. Remuzzi, S. Resnikoff, A. Ribeiro, J. Rivera, H. Roba, Alina Rodriguez, S. Rodríguez-Ramírez, D. Rojas-Rueda (2016)

    Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015

    Lancet (London, England), 388

  • D. Yach, M. Mckee, Alan Lopez, T. Novotny (2005)

    Improving diet and physical activity: 12 lessons from controlling tobacco smoking

    BMJ : British Medical Journal, 330

  • D. Adam (2019)

    The gene-based hack that is revolutionizing epidemiology

    Nature, 576

  • Spencer Abbastabar, S. James, D. Abate, K. Abate, Solomon Abay, C. Abbafati, Nooshin Abbasi, Hedayat Abbastabar, F. Abd-Allah, Jemal Abdela, A. Abdelalim, Ibrahim Abdollahpour, R. Abdulkader, Zegeye Abebe, S. Abera, Olifan Abil, H. Abraha, L. Abu-Raddad, Niveen Abu-Rmeileh, Manfred Accrombessi, Dilaram Acharya, Pawan Acharya, Ilana Ackerman, Abdullahi Adamu, O. Adebayo, Victor Adekanmbi, O. Adetokunboh, Mina Adib, Jose Adsuar, K. Afanvi, Mohsen Afarideh, A. Afshin, Gina Agarwal, Kareha Agesa, Rakesh Aggarwal, S. Aghayan, S. Agrawal, Alireza Ahmadi, Mehdi Ahmadi, H. Ahmadieh, M. Ahmed, Amani Aichour, Ibtihel Aichour, Miloud Aichour, Tomi Akinyemiju, N. Akseer, Z. Al‐Aly, A. Al-Eyadhy, Hesham Al-Mekhlafi, Rajaa Al-Raddadi, Fares Alahdab, K. Alam, Tahiya Alam, Alaa Alashi, S. Alavian, K. Alene, M. Alijanzadeh, Reza Alizadeh-Navaei, S. Aljunid, A. Alkerwi, François Alla, Peter Allebeck, Mohamed Alouani, K. Altirkawi, N. Alvis-Guzmán, Azmeraw Amare, L. Aminde, W. Ammar, Y. Amoako, N. Anber, C. Andrei, S. Androudi, Megbaru Animut, Mina Anjomshoa, Mustafa Ansha, C. Antonio, P. Anwari, J. Arabloo, Antonio Arauz, O. Aremu, F. Ariani, Bahroom Armoon, Johan Ärnlöv, Amit Arora, A. Artaman, K. Aryal, H. Asayesh, R. Asghar, Z. Ataro, Sachin Atre, M. Ausloos, L. Ávila-Burgos, Euripide Avokpaho, A. Awasthi, B. Quintanilla, R. Ayer, P. Azzopardi, A. Babazadeh, Hamid Badali, A. Badawi, A. Bali, Katherine Ballesteros, S. Ballew, Maciej Banach, J. Banoub, A. Banstola, A. Barać, Miguel Barboza, S. Barker-Collo, T. Bärnighausen, L. Barrero, B. Baune, S. Bazargan-Hejazi, Neeraj Bedi, Ettore Beghi, Masoud Behzadifar, Meysam Behzadifar, Yannick Béjot, A. Belachew, Yihalem Belay, Michelle Bell, A. Bello, I. Benseñor, Eduardo Bernabé, Robert Bernstein, M. Beuran, Tina Beyranvand, N. Bhala, S. Bhattarai, Soumyadeep Bhaumik, Z. Bhutta, B. Biadgo, A. Bijani, B. Bikbov, V. Bilano, Nigus Bililign, M. Sayeed, D. Bisanzio, B. Blacker, Fiona Blyth, I. Bou-Orm, S. Boufous, R. Bourne, Oliver Brady, Michael Brainin, L. Brant, A. Brazinova, N. Breitborde, H. Brenner, P. Briant, Andrew Briggs, A. Briko, Gabrielle Britton, T. Brugha, Rachelle Buchbinder, Reinhard Busse, Z. Butt, Lucero Cahuana-Hurtado, Jorge Cano, Rosario Cárdenas, J. Carrero, A. Carter, F. Carvalho, C. Castañeda-Orjuela, Jacqueline Rivas, Franz Castro, Ferrán Catalá-López, Kelly Cercy, Ester Cerin, Y. Chaiah, Alex Chang, Hsing-Yi Chang, Jung-Chen Chang, F. Charlson, Aparajita Chattopadhyay, Vijay Chattu, Pankaj Chaturvedi, P. Chiang, Ken Chin, Abdulaal Chitheer, J. Choi, Rajiv Chowdhury, H. Christensen, D. Christopher, F. Cicuttini, Liliana Ciobanu, Massimo Cirillo, R. Claro, D. Collado-Mateo, Cyrus Cooper, Josef Coresh, P. Cortesi, Monica Cortinovis, Megan Costa, Ewerton Cousin, M. Criqui, Elizabeth Cromwell, M. Cross, J. Crump, A. Dadi, L. Dandona, R. Dandona, P. Dargan, A. Daryani, R. Gupta, J. Neves, T. Dasa, Gail Davey, A. Davis, D. Davițoiu, B. Courten, Fernando Hoz, Diego Leo, Jan‐Walter Neve, M. Degefa, L. Degenhardt, Selina Deiparine, R. Dellavalle, Gebre Demoz, Kebede Deribe, N. Dervenis, Don Jarlais, Getenet Dessie, Subhojit Dey, S. Dharmaratne, Mesfin Dinberu, M. Dirac, Shirin Djalalinia, L. Doan, K. Dokova, D. Doku, E. Dorsey, Kerrie Doyle, T. Driscoll, M. Dubey, E. Dubljanin, Eyasu Duken, B. Duncan, A. Durães, H. Ebrahimi, S. Ebrahimpour, M. Echko, D. Edvardsson, A. Effiong, Joshua Ehrlich, C. Bcheraoui, M. Zaki, Z. El-Khatib, H. Elkout, Iqbal Elyazar, A. Enayati, A. Endries, Benjamin Er, H. Erskine, B. Eshrati, S. Eskandarieh, Alireza Esteghamati, S. Esteghamati, H. Fakhim, Vahid Omrani, Mahbobeh Faramarzi, M. Fareed, F. Farhadi, Talha Farid, C. Farinha, Andrea Farioli, Andre Faro, M. Farvid, F. Farzadfar, Valery Feigin, Netsanet Fentahun, S. Fereshtehnejad, Eduarda Fernandes, Joao Fernandes, A. Ferrari, Garumma Feyissa, I. Filip, Florian Fischer, C. Fitzmaurice, N. Foigt, Kyle Foreman, Jack Fox, Tahvi Frank, Takeshi Fukumoto, N. Fullman, Thomas Fürst, João Furtado, Neal Futran, S. Gall, M. Ganji, F. Gankpé, A. García-Basteiro, William Gardner, A. Gebre, A. Gebremedhin, Teklu Gebremichael, Tilayie Gelano, J. Geleijnse, R. Gènova-Maleras, Y. Geramo, P. Gething, Kebede Gezae, K. Ghadiri, Khalil Falavarjani, Maryam Ghasemi-Kasman, M. Ghimire, Rakesh Ghosh, A. Ghoshal, Simona Giampaoli, P. Gill, Tiffany Gill, I. Ginawi, G. Giussani, E. Gnedovskaya, Ellen Goldberg, Srinivas Goli, H. Gómez-Dantés, P. Gona, S. Gopalani, Taren Gorman, A. Goulart, B. Goulart, A. Grada, Morgan Grams, Giuseppe Grosso, H. Gugnani, Yuming Guo, Prakash Gupta, Rahul Gupta, R. Gupta, Tanush Gupta, Bishal Gyawali, J. Haagsma, Vladimir Hachinski, N. Hafezi-Nejad, Hassan Bidgoli, Tekleberhan Hagos, G. Hailu, Arvin Haj-Mirzaian, Arya Haj-Mirzaian, R. Hamadeh, S. Hamidi, A. Handal, Graeme Hankey, Yuantao Hao, H. Harb, S. Harikrishnan, J. Haro, M. Hasan, H. Hassankhani, H. Hassen, Rasmus Havmoeller, Caitlin Hawley, Roderick Hay, S. Hay, A. Hedayatizadeh-Omran, B. Heibati, D. Hendrie, A. Henok, C. Herteliu, S. Heydarpour, D. Hibstu, H. Hoang, Hans Hoek, Howard Hoffman, M. Hole, Enayatollah Rad, P. Hoogar, H. Hosgood, Seyed Hosseini, M. Hosseinzadeh, M. Hostiuc, S. Hostiuc, Peter Hotez, D. Hoy, M. Hsairi, A. Htet, Guoqing Hu, John Huang, Chantal Huynh, K. Iburg, C. Ikeda, B. Ileanu, O Ilesanmi, Usman Iqbal, S. Irvani, C. Irvine, Sheikh Islam, F. Islami, K. Jacobsen, L. Jahangiry, N. Jahanmehr, Sudhir Jain, Mihajlo Jakovljevic, Mehdi Javanbakht, A. Jayatilleke, P. Jeemon, Ravi Jha, Vivekanand Jha, John Ji, C. Johnson, J. Jonas, J. Jozwiak, Suresh Jungari, Mikk Jürisson, Zubair Kabir, R. Kadel, Amaha Kahsay, Rizwan Kalani, T. Kanchan, Manoochehr Karami, B. Matin, André Karch, C. Karema, Narges Karimi, Seyed Karimi, A. Kasaeian, D. Kassa, G. Kassa, T. Kassa, N. Kassebaum, S. Katikireddi, Norito Kawakami, Ali Karyani, Masoud Keighobadi, P. Keiyoro, L. Kemmer, Grant Kemp, A. Kengne, Andre Keren, Yousef Khader, Behzad Khafaei, M. Khafaie, Alireza Khajavi, I. Khalil, E. Khan, Muhammad Khan, Muhammad Khan, Y. Khang, Mohammad Khazaei, Abdullah Khoja, A. Khosravi, Mohammad Khosravi, A. Kiadaliri, Daniel Kiirithio, Cho-il Kim, Daniel Kim, Pauline Kim, Young-Eun Kim, Yun Kim, R. Kimokoti, Y. Kinfu, A. Kisa, K. Kissimova-Skarbek, Mika Kivimäki, A. Knudsen, Jonathan Kocarnik, S. Kochhar, Yoshihiro Kokubo, T. Kolola, Jacek Kopec, S. Kosen, Georgios Kotsakis, P. Koul, Ai Koyanagi, M. Kravchenko, K. Krishan, Kristopher Krohn, B. Defo, B. Bicer, G. Kumar, Manasi Kumar, H. Kyu, Deepesh Lad, S. Lad, A. Lafranconi, R. Lalloo, T. Lallukka, Faris Lami, V. Lansingh, A. Latifi, Kathryn Lau, J. Lazarus, J. Leasher, J. Ledesma, P. Lee, J. Leigh, Janni Leung, Miriam Levi, S. Lewycka, Shanshan Li, Yichong Li, Yu Liao, Misgan Liben, Lee-Ling Lim, Stephen Lim, Shiwei Liu, R. Lodha, K. Looker, A. Lopez (2018)

    Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017

    Lancet (London, England), 392

  • D. Kurowicka, R. Cooke (2007)

    Sampling algorithms for generating joint uniform distributions using the vine-copula method

    Comput. Stat. Data Anal., 51

  • (2015)

    Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013

    BDJ, 219

  • A. Aplogan, V. Mangindula, G. Mwema, L. Okito, Herpesvirus Br (2015)

    From the Centers for Disease Control and Prevention

  • XMUMRF/2018­C2/ITCM/0001. K Krishan acknowledges support from UGC Centre of Advanced Study (CAS II) awarded to the Department of Anthropology, Panjab University, Chandigarh, India

  • A. Quinn, N. Bruce, E. Puzzolo, Katherine Dickinson, R. Sturke, Darby Jack, S. Mehta, A. Shankar, K. Sherr, J. Rosenthal (2018)

    An analysis of efforts to scale up clean household energy for cooking around the world.

    Energy for sustainable development : the journal of the International Energy Initiative, 46

  • C. Murray, Alan Lopez (1997)

    Global mortality, disability, and the contribution of risk factors: Global Burden of Disease Study

    The Lancet, 349

  • Kirk Smith, A. Woodward, D. Campbell-Lendrum, D. Chadee, Y. Honda, Qiyong Liu, J. Olwoch, B. Revich, R. Sauerborn, C. Aranda, H. Berry, C. Butler (2014)

    Human health: impacts, adaptation, and co-benefits

Publisher
Unpaywall
ISSN
0140-6736
DOI
10.1016/s0140-6736(20)30752-2
Publisher site
See Article on Publisher Site

Abstract

Global Health Metrics Global burden of 87 risk factors in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019 GBD 2019 Risk Factors Collaborators* Summary Background Rigorous analysis of levels and trends in exposure to leading risk factors and quantification of their effect on Lancet 2020; 396: 1223–49 human health are important to identify where public health is making progress and in which cases current efforts are *For the list of Collaborators see Viewpoint Lancet 2020; inadequate. The Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2019 provides a standardised and 396: 1135–59 comprehensive assessment of the magnitude of risk factor exposure, relative risk, and attributable burden of disease. Correspondence to: Prof Christopher J L Murray, Methods GBD 2019 estimated attributable mortality, years of life lost (YLLs), years of life lived with disability (YLDs), Institute for Health Metrics and and disability-adjusted life-years (DALYs) for 87 risk factors and combinations of risk factors, at the global level, Evaluation, University of Washington, Seattle, WA 98195, regionally, and for 204 countries and territories. GBD uses a hierarchical list of risk factors so that specific risk factors USA (eg, sodium intake), and related aggregates (eg, diet quality), are both evaluated. This method has six analytical steps. [email protected] (1) We included 560 risk–outcome pairs that met criteria for convincing or probable evidence on the basis of research studies. 12 risk–outcome pairs included in GBD 2017 no longer met inclusion criteria and 47 risk–outcome pairs for risks already included in GBD 2017 were added based on new evidence. (2) Relative risks were estimated as a function of exposure based on published systematic reviews, 81 systematic reviews done for GBD 2019, and meta-regression. (3) Levels of exposure in each age-sex-location-year included in the study were estimated based on all available data sources using spatiotemporal Gaussian process regression, DisMod-MR 2.1, a Bayesian meta-regression method, or alternative methods. (4) We determined, from published trials or cohort studies, the level of exposure associated with minimum risk, called the theoretical minimum risk exposure level. (5) Attributable deaths, YLLs, YLDs, and DALYs were computed by multiplying population attributable fractions (PAFs) by the relevant outcome quantity for each age- sex-location-year. (6) PAFs and attributable burden for combinations of risk factors were estimated taking into account mediation of different risk factors through other risk factors. Across all six analytical steps, 30 652 distinct data sources were used in the analysis. Uncertainty in each step of the analysis was propagated into the final estimates of attributable burden. Exposure levels for dichotomous, polytomous, and continuous risk factors were summarised with use of the summary exposure value to facilitate comparisons over time, across location, and across risks. Because the entire time series from 1990 to 2019 has been re-estimated with use of consistent data and methods, these results supersede previously published GBD estimates of attributable burden. Findings The largest declines in risk exposure from 2010 to 2019 were among a set of risks that are strongly linked to social and economic development, including household air pollution; unsafe water, sanitation, and handwashing; and child growth failure. Global declines also occurred for tobacco smoking and lead exposure. The largest increases in risk exposure were for ambient particulate matter pollution, drug use, high fasting plasma glucose, and high body- mass index. In 2019, the leading Level 2 risk factor globally for attributable deaths was high systolic blood pressure, which accounted for 10·8 million (95% uncertainty interval [UI] 9·51–12·1) deaths (19·2% [16·9–21·3] of all deaths in 2019), followed by tobacco (smoked, second-hand, and chewing), which accounted for 8·71 million (8·12–9·31) deaths (15·4% [14·6–16·2] of all deaths in 2019). The leading Level 2 risk factor for attributable DALYs globally in 2019 was child and maternal malnutrition, which largely affects health in the youngest age groups and accounted for 295 million (253–350) DALYs (11·6% [10·3–13·1] of all global DALYs that year). The risk factor burden varied considerably in 2019 between age groups and locations. Among children aged 0–9 years, the three leading detailed risk factors for attributable DALYs were all related to malnutrition. Iron deficiency was the leading risk factor for those aged 10–24 years, alcohol use for those aged 25–49 years, and high systolic blood pressure for those aged 50–74 years and 75 years and older. Interpretation Overall, the record for reducing exposure to harmful risks over the past three decades is poor. Success with reducing smoking and lead exposure through regulatory policy might point the way for a stronger role for public policy on other risks in addition to continued efforts to provide information on risk factor harm to the general public. Funding Bill & Melinda Gates Foundation. Copyright © 2020 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. www.thelancet.com Vol 396 October 17, 2020 1223 Global Health Metrics Research in context Evidence before this study the systematic reviews and meta-regressions, 47 new risk– The Global Burden of Diseases, Injuries, and Risk Factors Study outcome pairs have been included for risks that were previously (GBD) 2017 provided the most recent assessment of deaths, included. This includes outcomes linked to low birthweight and years of life lost because of premature mortality, years of life short gestational age as intermediate outcomes linked to lived with disability, and disability-adjusted life-years particulate matter with a diameter smaller than 2·5 μm (PM ), 2·5 attributable to metabolic, environmental and occupational, which has increased the burden attributable to PM . (8) New 2·5 and behavioural risk factors. GBD 2017 provided estimates from cohorts, trials, and case-control studies have been added for the 1990 to 2017 for 195 countries and territories. Many reports assessment of risk functions. (9) New sources have been added explore the burden of disease that can be attributed to a specific to the analysis of risk factor exposure by age, sex, and location. risk factor in a specific country or territory, region, or globally, (10) Corrections for non-reference method exposure but none attempts to assess an extensive list of risk factors in all measurements have been revised using network or related countries and regions. meta-regression. (11) For dietary risks, the theoretical minimum risk exposure level (TMREL) has been revised based on the new Added value of this study systematic reviews. (12) The distribution of alcohol use across GBD 2019 advances the technical quantification of attributable individuals has been revised to better capture the asymmetric burden in 12 ways. (1) In support of the agreement between GBD nature of the distribution. In addition to the technical and WHO, nine new countries have been added to the analysis: improvements in each step of the quantification of risk factor Cook Islands, Monaco, San Marino, Nauru, Niue, Palau, Saint Kitts exposure, relative risk, TMREL, and attributable burden, in this and Nevis, Tokelau, and Tuvalu. (2) Subnational assessments for study we have focused attention on the broad trends in risk Italy, Nigeria, Pakistan, the Philippines, and Poland have been exposure by computing summary exposure values for added to GBD 2019. (3) High and low non-optimal temperatures aggregations of risk factors. Isolating the long-term global and have been added as risk factors (54 new risk–outcome pairs). national trends in risk exposure reveals in which cases the world (4) For 81 risk–outcome pairs, new systematic reviews have been has been successful in reducing exposure to harmful risks. done as part of GBD 2019. (5) For 139 risk–outcome pairs, dose–response meta-regressions have been done to evaluate if Implications of all the available evidence the relationship between exposure and relative risk might not be Improved analysis of risk exposure and burden attributable to adequately captured by assuming a log-linear relationship risk factors at the national, regional, and global level can help to between risk and per unit increase in exposure. (6) On the basis of focus attention on risks for which exposure is increasing and in the systematic reviews and dose–response meta-regression, which locations. This quantification is an essential input into 12 risk–outcome pairs have been excluded from GBD 2019 public health prioritisation and evaluation of programme because they no longer met inclusion criteria. (7) On the basis of success. Introduction To help track risk exposures and the burden The mechanism for much of disease and injury preven­ attributable to these exposures, many studies are pub­ tion is through modifying environmental, occupational, lished each year on the burden of specific risks, often 4–7 behavioural, and metabolic risk factors. Other pathways, in specific countries or regions. To our knowledge, such as vaccination or addressing social determinants the only effort to examine attributable burden with of health, are crucially important, but a substantial standardised methods across a wide set of risk factors component of public health has targeted modifying the spanning all countries is the Global Burden of Diseases, 8–12 aforementioned risk factors. Whether the risk factor is Injuries, and Risk Factors Study (GBD). Many choices targeted through public policy such as taxation or go into the comparable quantification of the burden of regulation, through programmes such as water supply risk factors; GBD provides a rules­ based approach to improvement, or primary care advice and pharmacological evidence synthesis that follows the Guidelines on intervention, it is essential to track progress on risk Accurate and Transparent Health Estimates Reporting. exposure. Which risk factors are declining, stagnating, or Comparable quantification across risks over time even increasing gives insights into where current efforts and across populations facilitates identi fying relative are working or are insufficient. Understanding where the importance and how population health risks are promise of prevention is being realised might generate evolving over time. GBD also provides a framework lessons that can be applied to other risks in which progress to understand both the trends in risk exposure and is slow. Tracking the burden attributable to risk exposure, the trends in burden attributable to risks. Quantifying measured by deaths, years of life lost (YLLs), years lived and reporting both exposure and attributable burden is with disability (YLDs), or disability­ adjusted life­ years important because exposure might be increasing and (DALYs), can also help governments, donor agencies, attributable burden decreasing if other drivers of the international organisations, and civil society organisations underlying outcomes are declining at a fast enough 1–3 to identify new priorities. rate. 1224 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics In this study, we present new or updated results for Islands, Monaco, San Marino, Nauru, Niue, Palau, the quantification of 560 risk–outcome pairs including Saint Kitts and Nevis, Tokelau, and Tuvalu) were added, updated data for exposure, updated data for relative risks, such that the GBD location hierarchy now includes methods innovation in evaluating risk­ exposure functions, all WHO member states. These new locations were and the addition of two new risk factors—high and previously included in regional totals by assuming low non­ optimal temperatures. In addition to providing that age­ specific rates were equal to the regional rates. quantification of exposure and attributable burden in At the most detailed level, we generated estimates 204 locations over the period 1990–2019, we used summary for 990 locations. The GBD diseases and injuries exposure values (SEVs) for aggregates of risk factors to analytical framework generated estimates for every year understand where public health is making progress from 1990 to 2019. tackling the major environmental, occupational, beha­ vioural, and metabolic risk factors, and where it is not. GBD risk factor hierarchy Individual risk factors such as low birthweight or Methods ambient ozone pollution are evaluated in the GBD CRA. Overview In addition, there has been policy interest in groups of The GBD 2019 estimation of attributable burden fol­ risk factors such as household air pollution combined lowed the general framework established for comparative with ambient particulate matter. To accommodate these 14,15 risk assessment (CRA) used in GBD since 2002. Here, diverse interests, the GBD CRA has a risk factor hier­ we provide a general overview and details on major archy. Level 1 risk factors are behavioural, environmental innovations since GBD 2017. More detailed methods are and occupational, and metabolic; Level 2 risk factors available in appendix 1. CRA can be divided into six key include 20 risks or clusters of risks; Level 3 includes See Online for appendix 1 steps: inclusion of risk–outcome pairs in the analysis; 52 risk factors or clusters of risks; and Level 4 includes estimation of relative risk as a function of exposure; 69 spe cific risk factors. Counting all specific risk factors estimation of exposure levels and distributions; deter­ and aggregates computed in GBD 2019 yields 87 risks or mination of the counterfactual level of exposure, the level clusters of risks. For a full list of risk factors by level, see of exposure with minimum risk called the theoretical appendix 1 (section 5, table S2). minimum risk exposure level (TMREL); computation of population attributable fractions and attributable burden; Determining the inclusion of risk–outcome pairs in GBD and estimation of mediation of different risk factors Since GBD 2010, we have used the World Cancer Research through other risk factors such as high body­ mass index Fund criteria for convincing or probable evidence of risk– (BMI) and ischaemic heart disease, mediated through outcome pairs. For GBD 2019, we completely updated our elevated systolic blood pressure (SBP), elevated fasting systematic reviews for 81 risk–outcome pairs. Preferred plasma glucose (FPG), and elevated LDL cholesterol, to Reporting Items for Systematic Reviews and Meta­ compute the burden attributable to various combinations Analyses flowcharts on these reviews are available in of risk factors. appendix 1 (section 4). Convincing evidence requires more than one study type, at least two cohorts, no substantial Geographical units, age groups, and time periods unexplained heterogeneity across studies, good­ quality GBD 2019 estimated prevalence of exposure and attri­ studies to exclude the risk of confounding and selection butable deaths, YLLs, YLDs, and DALYs for 23 age bias, and biologically plausible dose–response gradients. groups; males, females, and both sexes combined; For GBD, for a newly proposed or evaluated risk–outcome and 204 countries and territories that were grouped pair, we additionally required that there was a significant into 21 regions and seven super­ regions. GBD 2019 association (p<0·05) after taking into account sources of includes subnational analyses for Italy, Nigeria, Pakistan, potential bias. To avoid risk–outcome pairs repetitively the Philippines, and Poland, and 16 countries previously entering and leaving the analysis with each cycle of GBD, estimated at subnational levels (Brazil, China, Ethiopia, the criteria for exclusion requires that with the available India, Indonesia, Iran, Japan, Kenya, Mexico, New studies the association has a p value greater than 0·1. Zealand, Norway, Russia, South Africa, Sweden, the UK, On the basis of these reviews and meta­ regressions, and the USA). All subnational analyses are at the 12 risk–outcome pairs included in GBD 2017 were excluded first level of administrative organi sation within each from GBD 2019: vitamin A deficiency and lower respiratory country except for New Zealand (by Māori ethnicity), infections; zinc deficiency and lower respiratory infections; Sweden (by Stockholm and non­ Stockholm), the UK diet low in fruits and four outcomes: lip and oral cavity (by local government authorities), and the Philippines cancer, nasopharynx cancer, other pharynx cancer, and (by province). In this publication, we present subnational larynx cancer; diet low in whole grains and two outcomes: estimates for Brazil, India, Indonesia, Japan, Kenya, intracerebral haemorrhage and subarachnoid haemor­ Mexico, Sweden, the UK, and the USA; given space rhage; intimate partner violence and maternal abortion constraints, these results are presented in appendix 2. and miscarriage; and high FPG and three outcomes: See Online for appendix 2 For this cycle, nine countries and territories (Cook chronic kidney disease due to hyper tension, chronic www.thelancet.com Vol 396 October 17, 2020 1225 Global Health Metrics kidney disease due to glomerulonephritis, and chronic sections 2.1, 4). For many risk factors, exposure data were kidney disease due to other and unspecified causes. modelled using either spatiotemporal Gaussian process 17,18 In addition, on the basis of multiple requests to begin regression or DisMod­ MR 2.1, which are Bayesian sta­ capturing important dimensions of climate change into tistical models developed over the past 12 years for GBD GBD, we evaluated the direct relationship between high analyses. For most risk factors, the distribution of exposure and low non­ optimal temperatures on all GBD disease and across individuals was estimated by modelling a measure injury outcomes. Rather than rely on a heterogeneous of dispersion, usually the SD, and fitting an ensemble of literature with a small number of studies examining parametric distributions to the predicted mean and SD. relationships with specific diseases and injuries, we ana­ Ensemble distributions for each risk were estimated lysed individual­ level cause of death data for all locations based on individual­ level data. Details for each risk factor with available information on daily temperature, location, modelling for mean, SD, and ensemble distribution are and International Classification of Diseases ­ coded cause of available in appendix 1 (section  4). Because of the strong death. These data totalled 58·9 million deaths covering dependency between birthweight and gestational age, eight countries. On the basis of this analysis, 27 GBD exposure for these risks was modelled as a joint distribution cause Level 3 outcomes met the inclusion criteria for each using the copula method. non­ optimal risk factor (appendix 1 section 2.2.1) and were In many cases, exposure data were available for the included in this analysis. Other climate­ related relation­ reference method of ascertainment and for alternative ships, such as between precipitation or humidity and methods, such as tobacco surveys reporting daily smoking health outcomes, have not yet been evaluated. versus total smoking; in these cases, we estimated the statistical relationship between the reference and alter­ Estimating relative risk as a function of exposure for native methods of ascertainment using network meta­ each risk–outcome pair regression and corrected the alternative data using this In GBD, we use published systematic reviews and for relationship. GBD 2019, we updated these where necessary to include any new studies that became available before Dec 31, 2019. Determining the TMREL We did meta­ analyses of relative risks from these studies For harmful risk factors with monotonically increasing as a function of exposure (appendix 1 sections 2.2.2, 4). For risk functions, the theoretical minimum risk level was GBD 2019, 81 new systematic reviews were done, including set to 0. For risk factors with J­ shaped or U­ shaped risk for 44 diet risk–outcome pairs. To allow for risk functions functions, such as for sodium and ischaemic heart that might not be log­ linear, we relaxed the meta­ regres­ disease or BMI and ischaemic heart disease, the TMREL sion assumptions to allow for monotonically increasing was determined as the low point of the risk function. or decreasing but potentially non­ linear functions for When the bottom of the risk function was flat or 147 risk–outcome pairs. Appendix 1 (section 2) provides the poorly determined, the TMREL uncertainty interval (UI) mathematical and computational details for how we captured the range over which risks are indistinguishable. implemented this approach for meta­ regression. 218 risk– For protective risks with monotonically declining risk outcome pairs were estimated assuming log­ linear functions with exposure, namely risk factors where relationships. For 126 risk–out come pairs, exposure was exposure lowers the risk of an outcome, the challenge is dichotomous or polytomous. For 37 risk–outcome pairs, selecting the level of exposure with the lowest level of the population attributable fractions were assumed by risk strongly supported by the available data. Projecting definition to be 100% (eg, 100% of diabetes is assumed to beyond the level of exposure supported by the available be, by definition, related to elevated FPG). For 32 risk– studies could exaggerate the attributable burden for a outcome pairs, other approaches were used that reflected risk factor. In these cases, for each risk–outcome pair, we the nature of the evidence that has been collected for determined the exposure level at the 85th percentile of those risks (appendix 1 section 4). For risks that affect exposure in the cohorts or trials used in the risk meta­ cardiovascular outcomes, we adjusted relative risks by age regression. We then generated the TMREL by weighting such that they follow the empirical pattern of attenuation each risk–outcome pair by the relative global magnitude seen in published studies for elevated SBP, FPG, and LDL of each outcome. Appendix 1 (section 2.4 and 4) provides cholesterol. details on the TMREL estimation for each risk. Estimation of the distribution of exposure for each risk Estimation of the population attributable fraction and by age-sex-location-year attributable burden For each risk factor, we systematically searched for For each risk factor j, we computed the population published studies, household surveys, censuses, admin­ attributable fraction (PAF) by age­ sex­ location­ year using istrative data, ground monitor data, or remote sensing the following general formula for a continuous risk: data that could inform estimates of risk exposure. To u ∫ RR (x)P (x)dx – RR (TMREL ) x=l joasg jasgt joasg jas estimate mean levels of exposure by age­ sex­ location­ year, PAF = joasgt specific methods varied across risk factors (appendix 1 ∫ RR (x)P (x)dx x = l joasg jasgt 1226 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics where PAF is the PAF for cause o, for age group a, distribution of exposure. We then averaged across joasgt sex s, location g, and year t; RR (x) is the relative risk as outcomes to compute the SEV for a given risk as joasg a function of exposure level x for risk factor j, for cause o controlled for confounding, age group a, sex s, and SEV = SEV r rc N(c) location g with the lowest level of observed exposure as l c and the highest as u; P (x) is the distribution of exposure jasgt at x for age group a, sex s, location g, and year t; and where N(c) is the total number of outcomes for a risk. TMREL is the TMREL for risk factor j, age group a, and The SEV is effectively excess risk ­ weighted prevalence, jas sex s. Where risk exposure is dichotomous or polytomous, which allows for comparisons across different types of this formula simplifies to the discrete form of the exposures. Maximum risk in the denominator of the SEV equation. is determined by the relative risk at the 99th percentile of Estimation of the PAF took into account the risk the global distribution of exposure. The SEV is on a function and the distribution of exposure across 0–100 scale where 100 means the entire population is at individuals in each age­ sex­ location­ year. By drawing maximum risk and 0 means everyone in the population 1000 samples from the risk function, 1000 distributions is at minimum risk. We computed age­ standardised of exposure for each age­ sex­ location­ year, and SEVs by age­ standardising age­ specific SEVs across the 1000 samples from the TMREL, we propagated all of age groups in which that risk factor was evaluated; this these sources of uncertainty into the PAF distributions. method is a change from GBD 2017 in which age­ PAFs were also applied at the draw level to the uncer­ standardisation included age groups in which the risk tainty distributions of each associated outcome for that was not evaluated. For example, the SEV for low age­ sex­ location­ year. birthweight is now age­ standardised across age groups 0–6 days to 7–27 days. Estimating the PAF and attributable burden for To estimate SEVs for groups of risk factors, we first combinations of risk factors estimated the value of RR without mediation through For the estimation of each specific risk factor, the risk 1 (RR ). 2/1 counterfactual distribution of exposure is the TMREL for that specific risk with no change in other risk factors. RR =MF (RR − 1) + 1 2/1 2/1 2 Thus, the sum of these risk­ specific estimates of attri­ butable burden can exceed 100% for some causes, such where RR is the relative risk of risk factor 2 and MF is 2 2/1 as cardiovascular diseases. It is also useful to assess the the mediation factor, or the proportion of the risk of risk PAF and attributable burden for combinations of factor 2 that is mediated through risk factor 1. We then risk factors, such as all diet components together or computed the PAF using the non­ mediated relative risk household air and ambient particulate matter pollution. (RR ) and computed the joint PAF as 1/2 To estimate the combined effects of risk factors, we should take into account how one risk factor might be PAF =1 −∏ (1 – PAF ). 1..j j mediated through another (eg, the effect of fruit intake j=1 might be partly mediated through fibre intake). We used the mediation matrix as developed in GBD 2017 to try We cannot simply multiply RR values used for the max to correct for overestimation of the PAF and the attri­ SEV of each component risk as this would exaggerate butable burden for combinations of risks if we were to the joint RR . We approximated the 99th percentile of max simply assume independence without any mediation. risk for the combination of risk factors by taking the Appendix 1 (section 5, table S6) provides the estimated geometric mean of the ratio between the individual mediation matrix. risk maximum risk and the individual risk global mean risk and multiplied that by the global mean joint risk. Summary exposure value Formally, As in previous rounds of GBD, we summarised expo­ sure distributions for dichotomous, polytomous, and RR N(r) max ∏RR ∏ global mean — continuous risk factors using the SEV. The SEV r r RR global mean compares the distribution of excess risk times exposure level to a population where everyone is at maximum risk. where N(r) is the total number of risks. ∫ P(x)RR(x)dx – 1 x=l Risk-deleted death rates SEV = rc RR – 1 We computed risk­ deleted death rates as the death max rates that would be observed if all risk factors were set For a given risk r and outcome c pair where RR is to their respective TMRELs. This was calculated as max the relative risk at the 99th percentile of the global the death rate in each age­ sex group multiplied by www.thelancet.com Vol 396 October 17, 2020 1227 Global Health Metrics 1 minus the all­ risk PAF for that age­ sex group in each (annual rate of change larger than –0·5%), substantial location. increases (annual rate of change greater than 0·5%) and the remainder of risks with either non­ significant rates Role of the funding source of change or significant rates of change between –0 ·5% The funders of the study had no role in study design, and 0·5%. The declining risks fall into two categories. data collection, data analysis, data interpretation, or First, a set of risks that are strongly linked to social writing of the report. The corresponding author had full and economic development, measured by the Socio­ access to all the data in the study and had final demographic Index (SDI): household air pollution; responsibility for the decision to submit for publication. unsafe water, sanitation, and handwashing; child growth failure; vitamin A deficiency; and zinc deficiency. The Results second set of declining risks includes tobacco smoking Global exposure to risks and lead, which historically have not been negatively The table shows the trends in risk exposure for each risk correlated with SDI. These risks could in fact increase as factor at the global level over two time intervals: the full countries and territories increase SDI, at least for a phase duration of the study, 1990–2019, and the past decade, in the development process. For a long list of risk factors, 2010–19. On the basis of this table, we can divide risks including some large risks, the annual rate of change into three groups based on the percentage change in was either statistically insignificant (p>0·05) or the the global SEV from 2010 to 2019: substantial declines annual rate of change was between –0·5% and 0·5% per SEV 1990 SEV 2010 SEV 2019 ARC 1990–2019 ARC 2010–19 All risk factors 23·09 (20·22 to 25·67) 21·21 (18·04 to 24·26) 21·22 (18·05 to 24·42) –0·29% (-0·46 to -0·15)* 0·00% (–0·18 to 0·20) Environmental and occupational risks 52·55 (48·66 to 55·92) 48·50 (44·44 to 52·15) 45·36 (41·16 to 49·19) –0·51% (–0·62 to –0·40)* –0·74% (–0·88 to –0·61)* Unsafe water, sanitation, and 55·40 (54·39 to 56·61) 49·70 (48·99 to 50·47) 47·13 (46·51 to 47·84) –0·56% (–0·61 to –0·51)* –0·59% (–0·67 to –0·52)* handwashing Unsafe water source 42·78 (41·06 to 44·39) 36·29 (34·57 to 37·92) 32·74 (30·82 to 34·41) –0·92% (–1·08 to –0·76)* –1·14% (–1·52 to –0·77)* Unsafe sanitation 56·28 (54·14 to 58·38) 38·21 (35·98 to 40·80) 28·93 (26·81 to 31·24) –2·29% (–2·52 to –2·07)* –3·09% (–3·68 to –2·47)* No access to handwashing facility 36·77 (36·54 to 37·03) 34·05 (33·80 to 34·32) 32·19 (31·92 to 32·48) –0·46% (–0·50 to –0·42)* –0·63% (–0·70 to –0·56)* Air pollution 45·11 (32·85 to 56·03) 38·36 (28·33 to 48·55) 34·68 (25·76 to 44·37) –0·91% (–1·21 to –0·60)* –1·12% (–1·48 to –0·81)* Particulate matter pollution 44·22 (31·97 to 55·06) 37·56 (27·57 to 47·75) 33·94 (25·11 to 43·56) –0·91% (–1·24 to –0·61)* –1·13% (–1·48 to –0·81)* Ambient particulate matter pollution 15·65 (10·62 to 21·58) 22·98 (18·28 to 27·62) 26·22 (21·57 to 30·50) 1·78% (0·95 to 2·71)* 1·46% (0·81 to 2·10)* Household air pollution from solid fuels 27·08 (16·20 to 38·13) 16·33 (9·59 to 24·52) 11·71 (6·64 to 18·27) –2·89% (–3·60 to –2·25)* –3·70% (–4·64 to –2·88)* Ambient ozone pollution 47·56 (22·76 to 60·54) 54·34 (29·48 to 65·36) 55·06 (32·21 to 67·16) 0·51% (0·27 to 1·24)* 0·15% (–0·10 to 1·08) Non-optimal temperature 29·57 (26·06 to 33·72) 30·21 (26·17 to 34·83) 29·53 (25·41 to 34·26) 0·00% (–0·13 to 0·11) –0·25% (–0·39 to –0·13)* High temperature 25·98 (22·07 to 30·21) 29·25 (24·92 to 33·82) 29·59 (25·16 to 34·26) 0·45% (0·29 to 0·59)* 0·13% (–0·01 to 0·26) Low temperature 33·21 (29·24 to 37·58) 33·47 (29·06 to 38·25) 32·92 (28·44 to 37·82) –0·03% (–0·13 to 0·06) –0·18% (–0·31 to –0·07)* Other environmental risks 50·81 (40·53 to 59·86) 45·11 (34·46 to 55·29) 39·67 (29·01 to 50·86) –0·85% (–1·18 to –0·55)* –1·43% (–1·95 to –0·93)* Residential radon 18·54 (12·37 to 25·82) 18·20 (12·23 to 25·41) 18·12 (12·17 to 25·43) –0·08% (–0·27 to 0·10) –0·05% (–0·25 to 0·14) Lead exposure 68·52 (53·18 to 80·97) 59·82 (43·52 to 74·40) 51·26 (35·09 to 67·32) –1·00% (–1·43 to –0·63)* –1·72% (–2·40 to –1·09)* Occupational risks 3·36 (2·99 to 3·90) 3·33 (2·97 to 3·89) 3·32 (2·96 to 3·87) –0·05% (–0·15 to 0·05) –0·05% (–0·22 to 0·13) Behavioural risks 16·80 (14·82 to 19·05) 15·38 (13·28 to 17·72) 15·09 (12·96 to 17·43) –0·37% (–0·50 to –0·25)* –0·21% (–0·36 to –0·07)* Child and maternal malnutrition 20·05 (19·06 to 21·19) 17·77 (16·61 to 19·07) 17·23 (15·98 to 18·55) –0·52% (–0·67 to –0·40)* –0·34% (–0·51 to –0·18)* Suboptimal breastfeeding 21·66 (20·28 to 22·96) 20·05 (18·26 to 21·34) 19·34 (17·42 to 20·68) –0·39% (–0·55 to –0·31)* –0·40% (–0·61 to –0·21)* Non-exclusive breastfeeding 21·34 (14·67 to 29·82) 19·40 (13·38 to 27·18) 18·39 (12·91 to 25·53) –0·51% (–0·61 to –0·40)* –0·59% (–0·83 to –0·31)* Discontinued breastfeeding 12·33 (12·04 to 12·65) 10·73 (10·50 to 10·99) 10·24 (9·96 to 10·54) –0·64% (–0·77 to –0·52)* –0·52% (–0·87 to –0·17)* Child growth failure 4·93 (4·41 to 5·57) 4·21 (3·70 to 4·78) 3·53 (3·01 to 4·10) –1·15% (–1·43 to –0·83)* –1·95% (–2·37 to –1·50)* Child underweight 13·32 (11·73 to 14·71) 10·51 (8·98 to 11·97) 8·13 (6·50 to 9·68) –1·70% (–2·05 to –1·45)* –2·86% (–3·54 to –2·37)* Child wasting 5·28 (4·50 to 5·98) 5·23 (4·41 to 5·97) 4·89 (4·08 to 5·61) –0·26% (–0·34 to –0·21)* –0·74% (–0·88 to –0·64)* Child stunting 24·07 (16·71 to 26·41) 19·65 (13·76 to 22·01) 16·24 (11·45 to 18·72) –1·36% (–1·63 to –1·17)* –2·11% (–2·68 to –1·74)* Low birthweight and short gestation 11·92 (10·66 to 13·44) 11·32 (10·15 to 12·67) 11·10 (9·99 to 12·42) –0·25% (–0·46 to –0·10)* –0·21% (–0·49 to 0·02) Short gestation 13·88 (12·81 to 15·20) 13·04 (12·19 to 13·96) 13·17 (12·30 to 14·13) –0·18% (–0·43 to –0·01)* 0·11% (–0·22 to 0·39) Low birthweight 11·03 (10·41 to 11·81) 10·11 (9·68 to 10·52) 9·69 (9·28 to 10·14) –0·45% (–0·69 to –0·28)* –0·47% (–0·76 to –0·21)* Iron deficiency 22·65 (21·51 to 23·98) 20·11 (18·78 to 21·59) 19·57 (18·11 to 21·12) –0·50% (–0·65 to –0·38)* –0·30% (–0·47 to –0·14)* Vitamin A deficiency 33·42 (30·78 to 36·10) 22·00 (19·70 to 24·45) 15·01 (13·55 to 16·86) –2·76% (–3·13 to –2·30)* –4·25% (–5·02 to –3·47)* Zinc deficiency 13·84 (5·91 to 24·06) 11·88 (4·96 to 21·34) 8·78 (2·89 to 17·60) –1·57% (–2·57 to –1·07)* –3·35% (–6·44 to –2·04)* (Table continues on next page) 1228 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics SEV 1990 SEV 2010 SEV 2019 ARC 1990–2019 ARC 2010–19 (Continued from previous page) Tobacco 30·54 (29·08 to 32·10) 25·32 (24·00 to 26·80) 24·03 (22·75 to 25·44) –0·83% (–0·89 to –0·77)* –0·58% (–0·69 to –0·47)* Smoking 14·85 (13·27 to 16·56) 12·41 (11·08 to 13·94) 11·14 (9·93 to 12·54) –0·99% (–1·04 to –0·94)* –1·20% (–1·29 to –1·11)* Chewing tobacco 4·58 (4·18 to 4·98) 4·95 (4·71 to 5·20) 5·11 (4·80 to 5·44) 0·37% (0·03 to 0·76)* 0·36% (–0·32 to 1·05) Secondhand smoke 43·20 (42·80 to 43·62) 37·76 (37·32 to 38·19) 37·51 (37·00 to 38·09) –0·49% (–0·54 to –0·43)* –0·07% (–0·20 to 0·06) Alcohol use 6·50 (4·62 to 8·84) 6·68 (4·81 to 9·02) 6·99 (4·98 to 9·41) 0·25% (0·00 to 0·56) 0·50% (0·05 to 0·95)* Drug use 0·18 (0·12 to 0·28) 0·18 (0·13 to 0·27) 0·19 (0·14 to 0·27) 0·28% (–0·19 to 0·69) 0·53% (0·06 to 0·97)* Dietary risks 51·31 (40·44 to 62·42) 48·28 (36·60 to 60·37) 47·10 (35·39 to 59·62) –0·30% (–0·50 to –0·15)* –0·28% (–0·50 to –0·10)* Diet low in fruits 66·70 (59·36 to 75·08) 59·09 (51·17 to 67·81) 56·86 (49·36 to 65·37) –0·55% (–0·71 to –0·42)* –0·43% (–0·58 to –0·29)* Diet low in vegetables 51·32 (38·33 to 65·78) 40·29 (29·88 to 52·52) 40·24 (29·59 to 52·46) –0·84% (–0·93 to –0·74)* –0·02% (–0·14 to 0·10) Diet low in legumes 69·46 (36·73 to 91·69) 61·20 (28·89 to 84·10) 59·67 (27·55 to 83·28) –0·52% (–1·08 to –0·32)* –0·28% (–0·67 to 0·00) Diet low in whole grains 79·92 (72·52 to 87·44) 79·57 (72·09 to 87·12) 78·81 (71·06 to 86·78) –0·05% (–0·07 to –0·03)* –0·11% (–0·17 to –0·06)* Diet low in nuts and seeds 57·76 (29·48 to 73·08) 50·13 (25·10 to 68·03) 47·47 (23·73 to 66·35) –0·68% (–0·92 to –0·29)* –0·61% (–0·91 to –0·26)* Diet low in milk 80·09 (68·47 to 89·10) 80·81 (70·31 to 89·37) 82·54 (71·88 to 91·12) 0·10% (0·05 to 0·18)* 0·23% (0·16 to 0·33)* Diet high in red meat 40·50 (33·75 to 47·06) 43·15 (36·95 to 49·10) 43·94 (38·03 to 49·58) 0·28% (0·15 to 0·47)* 0·20% (–0·04 to 0·50) Diet high in processed meat 30·95 (20·80 to 42·39) 30·56 (20·13 to 43·05) 29·81 (19·04 to 43·32) –0·13% (–0·39 to 0·12) –0·27% (–0·69 to 0·10) Diet high in sugar-sweetened beverages 29·97 (22·97 to 42·54) 29·35 (21·94 to 41·88) 30·36 (22·71 to 43·05) 0·04% (–0·43 to 0·37) 0·38% (–0·22 to 0·76) Diet low in fiber 36·87 (25·93 to 47·86) 31·43 (21·20 to 41·62) 27·62 (18·60 to 36·95) –1·00% (–1·23 to –0·81)* –1·43% (–1·78 to –1·11)* Diet low in calcium 52·64 (43·62 to 64·79) 48·63 (38·79 to 62·22) 46·02 (35·93 to 60·32) –0·46% (–0·68 to –0·23)* –0·61% (–0·89 to –0·31)* Diet low in seafood omega-3 fatty acids 96·35 (93·21 to 99·89) 93·13 (89·11 to 98·47) 93·52 (88·71 to 99·41) –0·10% (–0·18 to –0·01)* 0·05% (–0·07 to 0·15) Diet low in polyunsaturated fatty acids 69·53 (49·68 to 82·70) 62·66 (37·55 to 79·83) 61·86 (35·56 to 80·13) –0·40% (–1·08 to –0·08)* –0·14% (–0·50 to 0·14) Diet high in trans fatty acids 50·54 (43·82 to 63·48) 45·22 (38·20 to 58·98) 44·67 (37·57 to 58·75) –0·43% (–0·58 to –0·17)* –0·14% (–0·41 to 0·08) Diet high in sodium 48·42 (32·26 to 64·13) 46·04 (28·63 to 62·81) 44·97 (27·44 to 62·14) –0·25% (–0·59 to –0·09)* –0·26% (–0·60 to –0·07)* Intimate partner violence 22·48 (13·03 to 30·15) 22·17 (13·13 to 29·08) 22·98 (13·31 to 30·37) 0·07% (0·00 to 0·16) 0·40% (0·00 to 0·73) Childhood sexual abuse and bullying 7·55 (4·99 to 11·23) 8·46 (5·63 to 12·84) 9·10 (6·04 to 13·85) 0·65% (0·49 to 0·79)* 0·81% (0·64 to 0·96)* Childhood sexual abuse 8·68 (6·85 to 10·90) 8·65 (6·89 to 10·78) 9·36 (7·40 to 11·79) 0·26% (0·18 to 0·33)* 0·87% (0·60 to 1·15)* Bullying victimisation 5·51 (2·34 to 11·04) 6·83 (3·04 to 13·36) 7·31 (3·25 to 14·34) 0·98% (0·82 to 1·28)* 0·76% (0·54 to 0·90)* Unsafe sex ·· ·· ·· ·· ·· Low physical activity 3·34 (1·79 to 6·00) 3·43 (1·90 to 6·08) 3·54 (1·95 to 6·26) 0·20% (0·06 to 0·41)* 0·37% (–0·13 to 0·87) Metabolic risks 14·90 (12·02 to 18·55) 19·40 (16·12 to 23·38) 22·14 (18·63 to 26·36) 1·37% (1·17 to 1·56)* 1·46% (1·26 to 1·69)* High fasting plasma glucose 7·88 (6·96 to 8·85) 10·41 (9·43 to 11·42) 11·72 (10·56 to 12·94) 1·37% (1·27 to 1·46)* 1·32% (1·01 to 1·64)* High LDL cholesterol 35·68 (32·92 to 38·73) 32·67 (29·73 to 35·84) 32·44 (29·49 to 35·57) –0·33% (–0·38 to –0·28)* –0·08% (–0·12 to –0·05)* High systolic blood pressure 27·12 (25·51 to 28·87) 26·50 (24·51 to 28·46) 27·74 (25·70 to 29·72) 0·08% (–0·12 to 0·28) 0·51% (0·04 to 1·00)* High body-mass index 11·09 (7·96 to 15·23) 16·46 (12·79 to 21·04) 19·45 (15·57 to 24·39) 1·94% (1·56 to 2·35)* 1·86% (1·55 to 2·19)* Low bone mineral density 17·06 (12·11 to 23·39) 16·42 (11·66 to 22·72) 16·26 (11·41 to 22·60) –0·16% (–0·25 to –0·10)* –0·10% (–0·34 to 0·09) Kidney dysfunction 20·56 (14·29 to 27·97) 22·35 (15·82 to 29·79) 22·74 (16·24 to 30·25) 0·35% (0·26 to 0·47)* 0·19% (0·13 to 0·28)* Data in parentheses are 95% uncertainty intervals. SEVs are measured on a 0 to 100 scale, in which 100 is when the entire population is exposed to maximum risk and 0 is when the entire population is at minimum risk. SEVs are shown for all levels of the risk factor hierarchy. ARC=annualised rate of change. SEVs=summary exposure values. *Statistically significant increase or decrease. Table: Global age-standardised SEVs for both sexes combined in 1990, 2010, and 2019, and annualised rate of change between 1990 and 2019 and 2010 and 2019 year: ambient ozone pollution, high temperature, low pollution, alcohol use, drug use, childhood sexual temperature, residential radon, occupational risks, sub­ abuse, bullying victimisation, high FPG, high SBP, and optimal breastfeeding, short gestation, low birthweight, high BMI. Many of the increasing risks are metabolic iron deficiency, chewing tobacco, dietary risks as a group, risk factors; in fact, taken together, the exposure to intimate partner violence, low physical activity, high metabolic risks increased 1·37% per year (95% UI LDL cholesterol, low bone mineral density, and kidney 1·17–1·56) from 1990 to 2019 and 1·46% per year dysfunction. Many of these stagnating risks have been or (1·26–1·69) from 2010 to 2019. Figure 1A, which shows are targets of concerted public health efforts spanning the trends in the age­ standardised SEV for each risk public policy, targeted programmes, and primary care factor compared with the fraction of global DALYs intervention. attributable to each risk factor, further emphasises Concerning for both current and future health are these patterns. In 2019, there were three risks that the exposures that are increasing at more than 0·5% per accounted for more than 1% of DALYs and were year. This list includes ambient particulate matter increasing in age­ standardised SEVs by more than www.thelancet.com Vol 396 October 17, 2020 1229 Global Health Metrics Behavioural Environmental or occupational Metabolic Global High SDI 2·0 3·0 High body-mass index Drug use Ambient particulate matter pollution High fasting plasma glucose 1·0 2·0 Alcohol use High systolic blood pressure High fasting plasma glucose Drug use Kidney dysfunction Secondhand Short gestation Low temperature smoke High systolic blood pressure Dietary risks Occupational risks High LDL cholesterol Iron deficiency 1·0 High body-mass index Low birthweight Kidney dysfunction No access to handwashing facility Low bone –1·0 Smoking mineral density Alcohol use Unsafe water source Dietary risks Occupational risks Child growth failure High LDL cholesterol Low temperature –2·0 Ambient particulate matter pollution –1·0 –3·0 Unsafe sanitation Household air pollution from solid fuels Smoking –4·0 –2·0 0 2·5 5·0 7·5 10·0 0 5·0 10·0 15·0 High–middle SDI Middle SDI 2·0 2·5 High body-mass index Alcohol use High body-mass index High fasting plasma glucose Kidney Drug use Short gestation Ambient particulate matter pollution High systolic blood pressure dysfunction Secondhand smoke High LDL cholesterol Dietary risks Low birthweight Low temperature –1·0 Occupational risks Smoking Child growth failure Alcohol use Lead exposure –2·5 Short gestation High fasting plasma glucose Low birthweight Kidney dysfunction Dietary risks Occupational risks High LDL cholesterol Drug use High systolic blood pressure Secondhand smoke Ambient particulate matter pollution –5·0 Low temperature Smoking Household air pollution from solid fuels –1·0 –7·5 0 5·0 10·0 15·0 0 5·0 10·0 15·0 Low–middle SDI Low SDI 5·0 5·0 Ambient particulate matter pollution Ambient particulate matter pollution High body-mass index High body-mass index 2·5 Alcohol use High fasting plasma glucose 2·5 High fasting plasma glucose High systolic blood pressure Kidney Secondhand smoke dysfunction High LDL cholesterol Alcohol use Kidney Short gestation High systolic blood pressure Dietary risks dysfunction Iron deficiency Occupational risks Short gestation Occupational risks Smoking Dietary risks Low birthweight No access to Unsafe water source High LDL cholesterol Lead No access to handwashing facility Non-exclusive breastfeeding handwashing exposure Iron deficiency facility Child growth failure Low birthweight Unsafe water source –2·5 Smoking Unsafe sanitation Child growth failure Unsafe sanitation –2·5 Household air pollution from solid fuels –5·0 Household air pollution from solid fuels –7·5 –5·0 0 2·5 5·0 7·5 10·0 0 5·0 10·0 15·0 DALYs attributable to each risk, 2019 (%) DALYs attributable to each risk, 2019 (%) (Figure 1 continues on next page) 1230 www.thelancet.com Vol 396 October 17, 2020 ARC in age-standardised SEVs, 2010–19 (%) ARC in age-standardised SEVs, 2010–19 (%) ARC in age-standardised SEVs, 2010–19 (%) Global Health Metrics 5·0 Global Low SDI Low–middle SDI Middle SDI High–middle SDI 4·0 High SDI 3·0 2·0 1·0 –1·0 –2·0 High systolic Smoking High fasting Low birthweight High body-mass Short gestation Ambient particulate High LDL Alcohol use blood pressure plasma glucose index matter pollution cholesterol Figure 1: ARC in age-standardised SEVs, globally and by SDI quintile, 2010–19 (A) Level 4 risks and occupational risks, dietary risks, and child growth failure, compared with percentage of DALYs attributable to each risk. (B) Top nine Level 4 risks by attributable DALYs. Only risk factors causing more than 1% of DALYs are shown in panel A. SEVs are measured on a 0–100 scale in which 100 is when the entire population is exposed to maximum risk and 0 is when the entire population is at minimum risk. ARC=annualised rate of change. DALYs=disability-adjusted life-years. SDI=Socio-demographic Index. SEVs=summary exposure values. 1% per year, dominating the figure: high FPG, high in this study leads to large percentage reductions in BMI, and ambient particulate matter pollution. Reductions mortality in those younger than 5 years and in the middle in risks that currently still have large attributable burden and older age groups. Risk reduction can have a slightly are almost exclusively those inversely associated with larger effect on male mortality than female mortality; in rising SDI, except smoking. It might be assumed that other words, some of the difference between male and effective efforts to reduce risk exposure have been female life expectancy can be explained by risk exposures. concentrated on the world’s largest risk factors, but we The percentage of age­ specific mortality explained by all see no discernible pattern between trends in exposure risk factors combined in 1990 is very similar to the and attributable burden. The global trends shown in share shown in figure 2A (appendix 2 table S3). Figure 2B the table and figure 1 give a high­ level view of how well shows the annualised rate of decline in risk­ deleted the world is managing exposure to an extensive list age­ specific mortality from 1990 to 2019. Risk ­ deleted of harmful risks, but regional and country trends can mortality rates declined from 1990 to 2019 in all age be markedly variable. Figure 1B shows trends for groups other than in those aged 95 years and older, the largest risks in terms of global attributable age­ declining between 1·0% and 3·3% per year for all the age standardised DALY rates for countries grouped into groups younger than 75 years, and at lower rates for quintiles of SDI in 2019. There is considerable variation those aged 75 years and older. The substantial declines in across quintiles in trends in exposure. Notably, ambient risk­ deleted mortality rates are likely to be related to particulate matter pollution exposure is increasing in reductions in risks not included in our assessment, the low SDI up to middle SDI quintiles but decreasing reductions in case­ fatality rates, or other factors. The in the high SDI quintile. High FPG and high BMI are observed rates of decline for all­ cause mortality for ages increasing in all quintiles, as is alcohol use. Smoking is younger than 10 years and older than 65 years have been declining in all SDI quintiles. Regional and national faster than the risk­ deleted rates, suggesting reduction trends in SEVs are available in appendix 2 (table S1). of risks included in our analysis has played a role in Figure 2A provides an alternative way to consider progress in these age groups, particularly in those the link between risk exposures and overall trends in younger than 5 years. Notably, risk­ deleted death rates mortality. Removing the effect of all risk factors included have declined faster than observed rates, particularly for www.thelancet.com Vol 396 October 17, 2020 1231 ARC in age-standardised SEVs, 2010–19 (%) Global Health Metrics five risks for attri butable deaths for females were high SBP (5·25 million [95% UI 4·49–6·00] deaths, or 20·3% [17·5–22·9] of all female deaths in 2019), dietary risks (3·48 million [2·78–4·37] deaths, or 13·5% [10·8–16·7] of –10 all female deaths in 2019), high FPG (3·09 million [2·40–3·98] deaths, or 11·9% [9·4–15·3] of all female –20 deaths in 2019), air pollution (2·92 million [2·53–3·33] deaths or 11·3% [10·0–12·6] of all female deaths in 2019), –30 and high BMI (2·54 million [1·68–3·56] deaths or 9·8% [6·5–13·7] of all female deaths in 2019). For males, the top five risks differed slightly. In 2019, the leading Level 2 risk –40 factor for attributable deaths globally in males was tobacco (smoked, second­ hand, and chewing), which accounted –50 for 6·56 million (95% UI 6·02–7·10) deaths (21·4% [20·5–22·3] of all male deaths in 2019), followed by high –60 SBP, which accounted for 5·60 million (4·90–6·29) deaths (18·2% [16·2–20·1] of all male deaths in 2019). The third –70 largest Level 2 risk factor for attributable deaths among Females males in 2019 was dietary risks (4·47 million [3·65–5·45] Males deaths, or 14·6% [12·0–17·6] of all male deaths in 2019) –80 followed by air pollution (ambient particulate matter and ambient ozone pollution, accounting for 3·75 million 0·5 [3·31–4·24] deaths (12·2% [11·0–13·4] of all male deaths in 2019), and then high FPG (3·14 million [2·70–4·34] deaths, or 11·1% [8·9–14·1] of all male deaths in 2019). Outside of the top five, there were large differences between attributable deaths in males and females due to –0·5 alcohol use, which accounted for 2·07 million (1·79–2·37) deaths in males and 0·374 million (0·298–0·461) deaths in –1·0 females in 2019. Newly included in GBD 2019, non­ optimal temperature accounted for 1·01 million (0·880–1·15) –1·5 deaths in males and 0·946 million (0·812–1·09) deaths in females. For both sexes combined, the leading Level 2 risk –2·0 factor for deaths was high SBP, accounting for 10·8 million (9·51–12·1) deaths in 2019 (19·2% [16·9–21·3] of all deaths that year), followed by tobacco, with 8·71 million –2·5 (8·12–9·31) deaths (15·4% [14·6–16·2] of all deaths that year). –3·0 When viewed in terms of DALYs (figure 3C, D), the ranking of Level 2 risk factors reflects the differential ages –3·5 of death and the contribution of non­ fatal disease burden. Most notably, child and maternal malnutrition –4·0 (including low birthweight, short gestation, child growth failure, non­ optimal breastfeeding, and low intake of micronutrients), which has predominant health effects among the young, was the second leading Level 2 risk Age (neonatal stage or years) factor for males and leading risk factor for females in Figure 2: Change in global mortality rates after risk deletion, by age group and sex 2019, accounting for 11·5% (95% UI 10·1–13·1) of DALYs (A) Percentage change in age-specific mortality rates in 2019 after removing the effect of all risk factors in this for males and 11·7% (10·5–13·2) of DALYs for females. study. (B) ARC in risk-deleted mortality rates from 1990 to 2019. ARC=annualised rate of change. Tobacco was ranked first for males and seventh for women aged between 25 and 59 years, implying that risk females in terms of attributable DALYs. For both sexes exposure has increased in those age groups. combined, the leading Level 2 risk factor globally for attributable DALYs was child and maternal malnutri­ Global attributable burden tion, at 295 million (95% UI 253–350) DALYs in 2019 Figure 3A and 3B show global attributable deaths for (11·6% [10·3–13·1] of all DALYs that year). females and males in 2019 for the 20 risk factors at Level 2 Figure 4 shows the ranking of Level 2 risk factors by of the risk factor hierarchy (appendix 2 table S3) The top attributable DALYs, both for SDI quintiles and the 21 GBD 1232 www.thelancet.com Vol 396 October 17, 2020 Early neonatal Late neonatal Post neonatal 1–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 80–84 85–89 90–94 ≥95 ARC in risk-deleted mortality rates, 1990–2019 (%) Change in mortality rate due to risk deletion, 2019 (%) Global Health Metrics regions. Risk factors are shaded by the trend in the handwashing. In the low­ middle SDI quintile, malnutri­ attributable DALY rates over the past decade. In the tion and air pollution were still the leading risk factors for low SDI quintile, the most important risk factors were attributable DALYs, but high SBP rose to third. In the malnutrition; air pollution; and water, sanitation, and middle to high SDI quintiles, the leading risks transitioned A Global attributable deaths from Level 2 risk factors for females in 2019 High systolic blood pressure Dietary risks High fasting plasma glucose Air pollution High body-mass index Tobacco High LDL cholesterol Kidney dysfunction Cardiovascular diseases Child and maternal malnutrition Chronic respiratory diseases Non-optimal temperature Diabetes and kidney diseases Digestive diseases Unsafe water, sanitation, and handwashing Enteric infections Unsafe sex HIV/AIDS and sexually transmitted infections Maternal and neonatal disorders Low physical activity Musculoskeletal disorders Alcohol use Neoplasms Neurological disorders Other environmental risks Nutritional deficiencies Occupational risks Other infectious diseases Other non-communicable diseases Low bone mineral density Respiratory infections and tuberculosis Self-harm and interpersonal violence Drug use Substance use disorders Intimate partner violence Transport injuries Unintentional injuries Childhood sexual abuse and bullying B Global attributable deaths from Level 2 risk factors for males in 2019 Tobacco High systolic blood pressure Dietary risks Air pollution High fasting plasma glucose High body-mass index High LDL cholesterol Alcohol use Kidney dysfunction Child and maternal malnutrition Non-optimal temperature Occupational risks Unsafe water, sanitation, and handwashing Other environmental risks Low physical activity Drug use Unsafe sex Low bone mineral density Childhood sexual abuse and bullying Intimate partner violence 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 Number of deaths (in 1000s) (Figure 3 continues on next page) www.thelancet.com Vol 396 October 17, 2020 1233 Global Health Metrics C Global attributable DALYs from Level 2 risk factors for females in 2019 Child and maternal malnutrition High systolic blood pressure Air pollution High fasting plasma glucose High body-mass index Dietary risks Tobacco Unsafe water, sanitation, and handwashing Cardiovascular diseases Chronic respiratory diseases High LDL cholesterol Diabetes and kidney diseases Kidney dysfunction Digestive diseases Enteric infections Unsafe sex HIV/AIDS and sexually transmitted infections Occupational risks Maternal and neonatal disorders Musculoskeletal disorders Non-optimal temperature Neoplasms Alcohol use Neurological disorders Nutritional deficiencies Drug use Other infectious diseases Low bone mineral density Other non-communicable diseases Respiratory infections and tuberculosis Other environmental risks Self-harm and interpersonal violence Intimate partner violence Sense organ diseases Substance use disorders Low physical activity Transport injuries Unintentional injuries Childhood sexual abuse and bullying D Global attributable DALYs from Level 2 risk factors for males in 2019 Tobacco Child and maternal malnutrition High systolic blood pressure Air pollution Dietary risks High fasting plasma glucose High body-mass index Alcohol use High LDL cholesterol Occupational risks Unsafe water, sanitation, and handwashing Kidney dysfunction Non-optimal temperature Drug use Unsafe sex Other environmental risks Low bone mineral density Low physical activity Childhood sexual abuse and bullying 0 2 4 6 8 10 12 14 DALYs (%) Figure 3: Global number of deaths and percentage of DALYs attributable to Level 2 risk factors, by cause and sex, 2019 DALYs=disability-adjusted life-years. to tobacco, high SBP, dietary risks, high BMI, and high dysfunction, dietary risks, and high LDL cholesterol. In FPG. The risk transition is evident across quintiles. Select seven regions, child and maternal malnutrition is the regional patterns stand out. In the Caribbean and central leading risk factor, and in another seven regions, tobacco is Latin America, large increases were seen in attributable the leading risk factor. In the remainder of regions, the burden for high FPG, high BMI, high SBP, kidney leading risk factor is high SBP (four regions), high FPG 1234 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics Annual rate of change in all-age DALYs from 2010 to 2019 –9·7% to <–3·3% –3·3% to <–1·4% –1·4% to <–0·7% –0·7% to <–0·3% –0·3% to <0·0% 0·0% to <0·4% 0·4% to <0·8% 0·8% to <1·1% 1·1% to <1·6% 1·6% to <5·0% 12 345678 910 High fasting High body- High systolic Tobacco Low SDI Malnutrition Air pollution WaSH Dietary risks Unsafe sex Alcohol use blood pressure plasma glucose mass index High fasting High body- High LDL High systolic Dietary risks WaSH Low-middle SDI Malnutrition Air pollution Tobacco Alcohol use plasma glucose mass index cholesterol blood pressure High systolic High body- High LDL High fasting Kidney Middle SDI Tobacco Dietary risks Air pollution Malnutrition Alcohol use blood pressure mass index cholesterol plasma glucose dysfunction High fasting Kidney High systolic High LDL Occupational High body- High-middle SDI Tobacco Dietary risks Air pollution Alcohol use plasma glucose dysfunction risks blood pressure cholesterol mass index High body- High systolic High LDL High fasting Kidney Occupational Dietary risks Alcohol use Drug use High SDI Tobacco cholesterol mass index blood pressure plasma glucose dysfunction risks High systolic High body- High fasting Kidney High LDL Air pollution Andean Latin America Malnutrition Dietary risks Alcohol use Tobacco blood pressure cholesterol mass index plasma glucose dysfunction High body- Kidney High systolic High fasting Occuptional High LDL Alcohol use Australasia Tobacco Dietary risks Drug use mass index blood pressure risks dysfunction plasma glucose cholesterol High fasting High systolic High body- High LDL Kidney Tobacco Dietary risks Air pollution Alcohol use Caribbean Malnutrition plasma glucose blood pressure mass index cholesterol dysfunction Kidney High systolic High body- High fasting High LDL Malnutrition Central Asia Dietary risks Tobacco Air pollution Alcohol use cholesterol dysfunction blood pressure mass index plasma glucose High systolic High body- High fasting High LDL Kidney Occupational Alcohol use Air pollution Central Europe Tobacco Dietary risks blood pressure mass index plasma glucose cholesterol dysfunction risks High fasting High body- Kidney High LDL High systolic Dietary risks Alcohol use Malnutrition Tobacco Central Latin America Air pollution plasma glucose mass index blood pressure dysfunction cholesterol Occupational High systolic High fasting High body- Central sub-Saharan Africa Malnutrition WaSH Air pollution Unsafe sex Alcohol use Dietary risks risks blood pressure plasma glucose mass index High systolic High fasting High body- High LDL Occupational Kidney Dietary risks Alcohol use East Asia Tobacco Air pollution blood pressure plasma glucose mass index cholesterol risks dysfunction High fasting High systolic High body- High LDL Kidney Tobacco Dietary risks Alcohol use Drug use Eastern Europe Air pollution blood pressure plasma glucose mass index cholesterol dysfunction High systolic High fasting High body- Alcohol use Eastern sub-Saharan Africa Malnutrition Air pollution WaSH Unsafe sex Dietary risks Tobacco blood pressure plasma glucose mass index High systolic High fasting High body- Kidney Occupational High LDL Alcohol use High-income Asia Pacific Tobacco Dietary risks Air pollution plasma glucose cholesterol blood pressure mass index dysfunction risks High fasting Kidney High body- High systolic High LDL Occupational Drug use Dietary risks Alcohol use High-income North America Tobacco plasma glucose dysfunction mass index blood pressure cholesterol risks High body- High fasting Kidney Occupational High systolic High LDL Air pollution North Africa and Middle East Malnutrition Tobacco Dietary risks blood pressure mass index plasma glucose dysfunction risks cholesterol High fasting High body- High systolic High LDL Dietary risks WaSH Oceania Malnutrition Air pollution Tobacco Unsafe sex plasma glucose mass index blood pressure cholesterol Kidney High systolic High fasting High body- High LDL Malnutrition Air pollution Tobacco Dietary risks WaSH South Asia plasma glucose dysfunction blood pressure mass index cholesterol High systolic High fasting High body- Kidney High LDL Southeast Asia Tobacco Dietary risks Air pollution Malnutrition Alcohol use blood pressure plasma glucose mass index dysfunction cholesterol High systolic High fasting Kidney High LDL High body- Occupational Dietary risks Alcohol use Southern Latin America Tobacco Malnutrition blood pressure plasma glucose dysfunction cholesterol risks mass index High systolic High body- High fasting Air pollution Southern sub-Saharan Africa Unsafe sex Malnutrition Tobacco Alcohol use WaSH Dietary risks blood pressure mass index plasma glucose High fasting High body- High systolic High LDL Kidney Alcohol use Tropical Latin America Tobacco Dietary risks Malnutrition Air pollution mass index plasma glucose blood pressure cholesterol dysfunction High systolic High fasting High body- High LDL Occupational Kidney Dietary risks Alcohol use Western Europe Tobacco Air pollution blood pressure plasma glucose mass index cholesterol risks dysfunction High systolic High body- High fasting Western sub-Saharan Africa Malnutrition WaSH Air pollution Unsafe sex Dietary risks Alcohol use Tobacco blood pressure mass index plasma glucose Figure 4: Leading ten Level 2 risk factors for attributable DALYs by GBD region and SDI quintile, 2019 For each region and SDI quintile, Level 2 risks are ranked by attributable DALYs from left (first) to right (tenth). Risks are coloured by their annualised rate of change in all-age DALY rates from 2010 to 2019. GBD=Global Burden of Diseases, Injuries, and Risk Factors Study. Malnutrition=child and maternal malnutrition. SDI=Socio-demographic Index. WaSH=water, sanitation, and handwashing. (one region), high BMI (one region), and unsafe sex (one Attributable burden by age group region). The largest rates of increase in attributable DALYs The pattern of risk­ factor­ attributable burden varied have been seen for high FPG in ten of 21 regions and for considerably by age and over time, as shown in figure 5, high BMI in ten of 21 regions. which includes arrows plots for all age groups combined www.thelancet.com Vol 396 October 17, 2020 1235 Global Health Metrics and for five broad age groups (0–9, 10–24, 25–49, 50–74, led to marked changes in risk rankings from 1990 to 2019. and 75 years and older). These figures show specific risk In 1990, the leading risk factors were child wasting, factors at Level 4 of the risk factor hierarchy. Figure 5A low birthweight, short gestation, and household air shows how risk exposure trends, underlying changes in pollution, all of which have dropped substantially in disease rates, and rising mean age of populations have magnitude in terms of percentage of attributable DALYs A All ages Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 Child wasting 11·4 (9·5 to 13·6) 1 High systolic blood pressure 9·3 (8·2 to 10·5) 53·1 (43·0 to 62·7) –27·0 (–31·7 to –22·6) 2 Low birthweight 10·6 (9·9 to 11·4)7 2 Smoking ·9 (7·2 to 8·6) 24·3 (15·9 to 33·9) –39·0 (–43·1 to –34·4) 3 Short gestation 8·7 (8·1 to 9·5) 3 High fasting plasma glucose 6·8 (5·8 to 8·0) 122·9 (110·0 to 135·7) 7·4 (1·5 to 13·8) 4 Household air pollution 8·0 (6·2 to 10·0) 4 Low birthweight 6·3 (5·5 to 7·3) –41·4 (–49·7 to –31·0) –41·3 (–49·6 to –30·8) 5 Smoking 6·2 (5·8 to 6·6) 5 High body-mass index 6·3 (4·2 to 8·6) 138·2 (106·1 to 186·9) 18·0 (2·2 to 42·3) 6 Unsafe water 6·2 (4·7 to 7·6) 6 Short gestation 5·5 (4·7 to 6·3) –38·9 (–47·3 to –28·0) –38·9 (–47·4 to –27·9) 0·3 (–21·2 to 30·7) 7 High systolic blood pressure 5·9 (5·3 to 6·5) 7 Ambient particulate matter 4·7 (3·8 to 5·5) 67·7 (27·9 to 126·1) 8 Child underweight 4·9 (3·9 to 6·3) 8 High LDL cholesterol 3·9 (3·2 to 4·7) 41·5 (31·1 to 50·4) –32·2 (–36·7 to –27·8) 9 Unsafe sanitation 4·6 (3·7 to 5·6) 9 Alcohol use 3·7 (3·3 to 4·1) 37·1 (27·3 to 47·9) –23·7 (–29·2 to –17·7) 10 Handwashing 3·2 (2·3 to 4·0) 10 Household air pollution 3·6 (2·7 to 4·6) –56·1 (–64·7 to –46·0) –68·2 (–74·0 to –61·6) 11 High fasting plasma glucose 3·0 (2·5 to 3·5) 11 Child wasting 3·3 (2·6 to 4·1) –71·7 (–77·4 to –65·2) –72·9 (–78·4 to –66·6) 13 Ambient particulate matter 2·7 (1·8 to 3·8) 13 Unsafe water 2·6 (1·9 to 3·3) –59·3 (–68·1 to –46·7) –65·9 (–73·0 to –55·4) 14 High LDL cholesterol 2·7 (2·2 to 3·2) 17 Unsafe sanitation 1·6 (1·3 to 2·1) 65·5 (–72·9 to –54·8) –71·0 (–77·0 to –61·8) 15 Alcohol use 2·6 (2·3 to 2·9) 19 Handwashing 1·3 (0·9 to 1·8) –58·7 (–65·9 to –49·8) –64·2 (–70·5 to –56·3) 16 High body-mass index 2·6 (1·5 to 4·0) 22 Child underweight 1·1 (0·9 to 1·4) –77·8 (–82·7 to –71·7) –79·5 (–84·0 to –73·8) B 0–9 years Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 Child wasting 24·7 (20·7 to 28·9) 1 Low birthweight 28·9 (27·3 to 30·4) –43·3 (–51·8 to –33·0) –42·6 (–51·2 to –32·2) 2 Low birthweight 23·1 (22·1 to 24·1) 2 Short gestation 24·7 (23·3 to 26·1) –41·2 (–49·6 to –30·2) –40·4 (–49·0 to –29·3) 3 Short gestation 19·0 (18·1 to 19·9) 3 Child wasting 14·8 (12·3 to 17·3) –72·9 (–78·4 to –66·3) –73·6 (–79·1 to –67·3) 4 Household air pollution 11·2 (8·7 to 14·2) 4 Household air pollution 7·7 (6·0 to 9·5) –68·8 (–75·2 to –60·6) –68·9 (–75·4 to –60·9) 5 Unsafe water 11·0 (8·5 to 13·3) 5 Unsafe water 7·7 (5·9 to 9·4) –68·3 (–75·8 to –57·4) –68·9 (–76·4 to –58·6) 6 Child underweight 10·4 (8·2 to 13·3) 6 Unsafe sanitation 5·1 (4·3 to 6·0) –72·0 (–78·7 to –62·0) –72·5 (–79·3 to –63·0) 7 Unsafe sanitation 8·2 (6·8 to 9·7) 7 Handwashing 4·5 (3·2 to 5·8) –66·0 (–72·9 to –57·0) –66·7 (–73·6 to –58·0) 8 Child stunting 6·2 (3·2 to 10·5) 8 Child underweight 4·4 (3·6 to 5·4) –80·8 (–85·2 to –75·3) –81·4 (–85·7 to –76·1) 9 Handwashing 6·0 (4·3 to 7·6) 9 Ambient particulate matter 4·0 (2·8 to 5·2) –23·3 (–45·9 to 11·5) –20·5 (–46·3 to 10·8) 10 Non-exclusive breastfeeding 3·8 (2·8 to 4·9) 10 Child stunting 2·7 (1·3 to 4·8) –80·3 (–85·8 to –74·5) –81·1 (–86·4 to –75·5) 11 Ambient particulate matter 2·3 (1·3 to 3·9) 11 Non-exclusive breastfeeding 2·4 (1·8 to 3·0) –72·1 (–77·8 to –65·3) –72·1 (–77·8 to –65·3) C 10–24 years Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 Occupational injury 3·2 (2·8 to 3·7) 1 Iron deficiency 3·0 (2·3 to 3·8) –0·9 (–11·4 to 9·5) –17·6 (–26·4 to –8·8) 2 Iron deficiency 2·8 (2·1 to 3·6) 2 Alcohol use 2·6 (2·1 to 3·1)– –6·3 (–12·9 to 0·3) 22·6 (–28·0 to –17·2) 3 Unsafe water 2·7 (1·7 to 4·2) 3 Unsafe sex 2·1 (1·5 to 2·9) 108·3 (78·5 to 140·5) 73·4 (47·4 to 98·5) 4 Alcohol use 2·6 (2·1 to 3·0) 4 Unsafe water 2·0 (1·3 to 3·0) –29·8 (–43·6 to –4·4) –40·5 (–53·1 to –20·4) 5 Unsafe sanitation 2·0 (1·3 to 3·1) 5 Occupational injury 1·8 (1·6 to 2·1) –47·7 (–54·0 to –40·8) –56·6 (–61·9 to –51·0) 6 Drug use 1·4 (1·1 to 1·7) 6 Drug use 1·8 (1·4 to 2·3) 20·4 (13·7 to 27·1) –0·6 (–6·2 to 4·9) 7 Handwashing 1·0 (0·7 to 1·5) 7 Short gestation 1·3 (1·0 to 1·6) 84·6 (68·2 to 99·4) 54·1 (40·0 to 66·0) 8 Unsafe sex 1·0 (0·7 to 1·4) 8 Low birthweight 1·3 (1·0 to 1·6) 84·6 (68·2 to 99·4) 54·1 (40·0 to 66·0) 9 Kidney dysfunction 0·9 (0·8 to 1·0) 9 Unsafe sanitation 1·2 (0·9 to 1·8) –42·4 (–53·9 to –21·6) –51·1 (–61·7 to –34·7) 10 Bullying 0·7 (0·2 to 1·4) 10 Bullying 1·1 (0·4 to 2·2) 50·7 (41·3 to 69·4) 26·9 (17·5 to 41·3) 11 Short gestation 0·6 (0·5 to 0·8) 11 Kidney dysfunction 1·1 (0·9 to 1·3) 19·0 (9·1 to 28·6) –1·2 (–9·4 to 6·7) 12 Low birthweight 0·6 (0·5 to 0·8) 12 Handwashing 0·8 (0·6 to 1·1) –28·8 (–41·4 to –8·1) –40·0 (-51·2 to –23·6) Environmental and occupational risks Behavioural risks Metabolic risks (Figure 5 continues on next page) 1236 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics D 25–49 years Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 Alcohol use 6·7 (5·9 to 7·5) 1 Alcohol use 6·3 (5·5 to 7·3) 26·7 (18·0 to 35·7) –23·5 (–28·8 to –18·0) 2 Smoking 6·6 (5·9 to 7·2) 2 High systolic blood pressure 6·0 (4·9 to 7·1) –15·1 (–23·0 to –7·4) 48·4 (34·4 to 61·8) 5·9 (4·2 to 7·8) 3 High systolic blood pressure 5·4 (4·4 to 6·4) 3 High body-mass index 136·1 (95·0 to 203·5) 40·5 (12·1 to 73·9) 4 Occupational injury 3·9 (3·5 to 4·3) 4 Smoking 5·0 (4·5 to 5·6) 1·9 (–5·7 to 9·7) –42·8 (–47·1 to –38·4) 5 High LDL cholesterol 3·5 (3·0 to 4·1) 5 Unsafe sex 4·9 (4·1 to 6·0) 45·1 (26·9 to 67·2) 131·3 (102·8 to 171·1) 4·0 (3·4 to 4·6) 6 Household air pollution6 3·4 (2·6 to 4·3) High fasting plasma glucose 90·9 (76·6 to 104·3) 10·0 (1·9 to 17·8) 7 High body-mass index7 3·3 (1·9 to 5·2) High LDL cholesterol 3·8 (3·1 to 4·5) 41·4 (28·4 to 54·5) –18·9 (–26·2 to –11·6) 8 Unsafe sex 2·8 (2·1 to 3·7) 8 Drug use 2·9 (2·5 to 3·3) 22·9 (16·6 to 30·4) 94·4 (84·5 to 106·7) 2·9 (2·4 to 3·5) 9 High fasting plasma glucose9 2·8 (2·4 to 3·2) Ambient particulate matter 120·4 (76·2 to 180·7) 29·4 (1·8 to 62·5) 10 Drug use 2·0 (1·7 to 2·3) 10 Kidney dysfunction 2·4 (2·0 to 2·7) 62·4 (49·9 to 75·0) –4·2 (–11·8 to 3·1) 11 Kidney dysfunction 1·9 (1·7 to 2·2) 11 Occupational injury 2·3 (2·1 to 2·6) –21·3 (–30·0 to –11·8) –50·4 (–55·9 to –44·5) 1·7 (1·2 to 2·3) –34·2 (–47·2 to –21·0) –61·9 (–69·4 to –54·2) 12 Ambient particulate matter 1·8 (1·2 to 2·3) 12 Household air pollution E 50–74 years Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 Smoking 19·4 (18·2 to 20·6) 1 High systolic blood pressure 16·1 (14·2 to 18·0) 47·7 (36·9 to 58·0) –28·3 (–33·6 to –23·3) 2 High systolic blood pressure 16·8 (14·9 to 18·7) 2 Smoking 15·5 (14·1 to 16·7) 22·6 (13·9 to 32·6) –40·3 (–44·6 to –35·5) 3 Household air pollution 8·5 (6·3 to 10·7) 3 High fasting plasma glucose 12·2 (10·4 to 14·4) 127·2 (113·4 to 141·5) 10·2 (3·4 to 17·0) 4 High fasting plasma glucose 8·3 (7·0 to 9·8) 4 High body-mass index 11·8 (7·9 to 16·0) 138·4 (106·5 to 186·2) 19·1 (0·7 to 39·5) 5 High body-mass index 7·6 (4·3 to 11·6) 5 Ambient particulate matter 6·8 (5·7 to 8·0) 122·5 (78·2 to 185·1) 9·8 (–13·6 to 38·3) 6 High LDL cholesterol 7·0 (5·6 to 8·5) 6 High LDL cholesterol 6·2 (4·9 to 7·7) 37·8 (27·8 to 47·5) –32·6 (–37·5 to –27·8) 7 Alcohol use 5·1 (4·5 to 5·7) 7 Alcohol use 5·0 (4·4 to 5·7) 51·2 (37·6 to 65·1) –25·8 (–32·6 to –19·1) 8 Ambient particulate matter 4·7 (3·3 to 6·3) 8 Kidney dysfunction 4·7 (4·0 to 5·3) 92·8 (80·4 to 105·3) –6·4 (–12·6 to –0·5) –36·7 (–50·4 to –21·6) –69·3 (–76·0 to –62·0) 9 High sodium 4·0 (1·4 to 8·0) 9 Household air pollution 3·5 (2·4 to 4·8) 10 Kidney dysfunction 3·7 (3·2 to 4·2) 10 High sodium 3·4 (1·1 to 7·1) 31·9 (–1·6 to 51·0) –37·1 (–52·1 to –26·5) F ≥75 years Leading risks 1990 Percentage of DALYs Leading risks 2019 Percentage of DALYs Percentage change in Percentage change in 1990 2019 number of DALYs, age-standardised DALY rate, 1990–2019 1990–2019 1 High systolic blood pressure 22·0 (18·6 to 25·3) 1 High systolic blood pressure 19·5 (16·3 to 22·7) 69·6 (58·6 to 80·5) –30·0 (–34·3 to –25·7) 2 Smoking 14·8 (13·9 to 15·7) 2 High fasting plasma glucose 13·5 (10·2 to 18·0) 144·5 (130·1 to 158·7) 1·8 (–4·8 to 7·9) 3 High fasting plasma glucose 10·5 (7·8 to 14·4) 3 Smoking 12·3 (11·4 to 13·0) 58·2 (48·9 to 69·1) –31·9 (–35·8 to –27·3) 4 High LDL cholesterol 9·2 (6·0 to 13·2) 4 High body-mass index 7·3 (4·3 to 11·1) 145·1 (123·1 to 180·2) 4·7 (–6·0 to 17·9) 5 Household air pollution 7·8 (5·7 to 10·2) 5 High LDL cholesterol 7·2 (4·5 to 10·6) 50·0 (39·2 to 58·7) –40·2 (–43·5 to –37·1) 6 High body-mass index 5·7 (3·0 to 9·2) 6 Ambient particulate matter 6·7 (5·6 to 7·8) 143·7 (94·6 to 211·9) 4·1 (–18·2 to 31·0) 7 Ambient particulate matter 5·2 (3·7 to 6·8) 7 Kidney dysfunction 5·9 (4·9 to 6·9) 121·7 (108·6 to 134·1) –8·6 (–14·2 to –3·6) 8 Kidney dysfunction 5·1 (4·1 to 6·1) 8 Low temperature 3·4 (2·9 to 3·9) 42·2 (32·5 to 53·1) –41·8 (–45·7 to –37·6) 9 Low temperature 4·6 (3·9 to 5·3) 9 Household air pollution 3·1 (2·1 to 4·3) –24·5 (–41·1 to –4·8) –67·7 (–74·9 to –59·4) 10 Low whole grains 3·5 (1·8 to 4·4) 10 Low whole grains 3·0 (1·6 to 3·9) 66·2 (57·6 to 74·6) –32·3 (–35·7 to –28·8) Environmental and occupational risks Behavioural risks Metabolic risks Figure 5: Leading ten Level 4 risks by attributable DALYs, 1990–2019 For all ages (A), 0–9 years (B), 10–24 years (C), 25–49 years (D), 50–74 years (E), and 75 years and older (F). DALYs=disability-adjusted life-years. and rank by 2019. The leading risks in 2019 were high The largest declines among the leading ten risks were SBP, smoking, high FPG, low birthweight, and high for child growth failure (child underweight, stunting, BMI. Other notable shifts include the large increase in and wasting); water, sanitation, and handwashing; percentage of attributable DALYs and rank for ambient and house hold air pollution. Large but more moderate particulate matter pollution, high LDL cholesterol, and declines in attributable burden occurred for short alcohol use. Among the youngest age group (0–9 years), gestation and low birthweight, with the smallest reduc­ shown in figure 5B, the leading Level 4 risk factors tion observed for ambient particulate matter pollution. were composed exclusively of malnutrition and environ­ Among adolescents and young adults (aged 10–24 years; mental risk factors. Over the 1990–2019 period, there figure 5C), the pattern of risk factor burden was notably were substantial reductions in the burden attributable to different from the 0–9 years age group, with iron these risk factors in both absolute numbers and rates. deficiency, alcohol use, and unsafe sex ranking first to www.thelancet.com Vol 396 October 17, 2020 1237 Global Health Metrics A Child and maternal malnutrition DALYs attributable 0% to <2% 10% to <12·5% 2% to <4% 12·5% to <15% 4% to <6% 15% to <17·5% 6% to <8% 17·5% to <20% 8% to <10% ≥20% Caribbean and central America Persian Gulf Balkan Peninsula Southeast Asia West Africa Eastern Mediterranean Northern Europe B High systolic blood pressure DALYs attributable 0% to <2% 10% to <12·5% 2% to <4% 12·5% to <15% 4% to <6% 15% to <17·5% 6% to <8% 17·5% to <20% 8% to <10% ≥20% Caribbean and central America Persian Gulf Balkan Peninsula Southeast Asia West Africa Eastern Mediterranean Northern Europe (Figure 6 continues on next page) 1238 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics C Tobacco DALYs attributable 0% to <2% 10% to <12·5% 2% to <4% 12·5% to <15% 4% to <6% 15% to <17·5% 6% to <8% 17·5% to <20% 8% to <10% ≥20% Eastern Caribbean and central America Persian Gulf Balkan Peninsula Southeast Asia West Africa Mediterranean Northern Europe D Air pollution DALYs attributable 0% to <2% 10% to <12·5% 2% to <4% 12·5% to <15% 4% to <6% 15% to <17·5% 6% to <8% 17·5% to <20% 8% to <10% ≥20% Eastern Caribbean and central America Persian Gulf Balkan Peninsula Southeast Asia West Africa Mediterranean Northern Europe (Figure 6 continues on next page) www.thelancet.com Vol 396 October 17, 2020 1239 Global Health Metrics E Dietary risks DALYs attributable 0% to <2% 10% to <12·5% 2% to <4% 12·5% to <15% 4% to <6% 15% to <17·5% 6% to <8% 17·5% to <20% 8% to <10% ≥20% Eastern Caribbean and central America Persian Gulf Balkan Peninsula Southeast Asia West Africa Mediterranean Northern Europe Figure 6: Percentage of all DALYs attributable to the five leading Level 2 risk factors, 2019 DALYs attributable to child and maternal malnutrition (A), high systolic blood pressure (B), tobacco (C), air pollution (D), and dietary risks (E). DALYs=disability-adjusted life-years. third for attributable DALYs in this age group in 2019. National findings There were declines in unsafe sex DALYs in the second The leading risk factors for attributable DALYs had half of the study period, but due to rapid increases highly varied geographical patterns, as shown in figure 6, from 1990 to 2004, there was still a 108·3% (95% UI which presents maps of the percentage of burden 78·5–140·5) increase in unsafe sex DALYs from 1990 to attributable to the top five Level 2 risk factors globally 2019. The long­ term consequences of short gestation and in 2019. The highest proportions (greater than 20%) of low birthweight also increased in importance. burden attributable to the leading Level 2 risk factor In the 25–49 years age group (figure 5D), alcohol use was in 2019, child and maternal malnutrition, were seen the leading Level 4 risk factor for attributable burden, in most of western, central, and eastern sub­ Saharan followed by high SBP and then high BMI, smoking, unsafe African regions (figure 6A). In addition, rates greater sex, and high FPG. The number of DALYs increased for all than 20% were seen in Afghanistan, Pakistan, states in the top ten risks, but age­ standardised attributable DALY northern India, Yemen, and Papua New Guinea. Rates rates increased only for high BMI, unsafe sex, high FPG, between 10% and 20% were seen in a diverse set of drug use, and ambient particulate matter pollution. In the central American countries, states in Brazil, Tajikistan, two oldest age groups, the set of leading risks are quite Uzbekistan, Myanmar, regions of the Philippines, and similar to one another, dominated by high SBP at the top, some Indonesian provinces. and followed by other metabolic risk factors including Figure 6B shows the burden attributable to the second high FPG, high BMI, high LDL cholesterol, and kidney leading Level 2 risk factor in 2019, high SBP. Locations dysfunction. Smoking also contributed substantially to with more than 20% of DALYs attributable to high SBP the risk attributable burden in these age groups, ranked included Georgia and most of central and eastern second in ages 50–74 years (figure 5E) and third in ages Europe. Most countries in north Africa and the Middle 75 years and older (figure 5F). In the oldest age group, East had between 10% and 20% of DALYs attributable to low temperature was also one of the top ten risks, high SBP as did states in southern India and many parts although age­standardised attributable DALY rates of southeast Asia. The only countries with less than declined from 1990 to 2019. Sex­ specific rankings by age 2% of all­ age DALYs attributable to high SBP were in group are available in appendix 2 (figures S4, S5). western sub­ Saharan Africa. 1240 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics The third leading Level 2 risk factor, tobacco, is shown DALYs were attributed to present and past exposure for in figure 6C. Locations with more than 20% of DALYs the 87 environmental, occupational, behavioural, and attributable in 2019 include countries in the Balkan metabolic risk factors and combinations of risk factors Peninsula and two provinces in China—Liaoning and included in this analysis. Overall, combined global expo­ Heilongjiang. Most countries in Europe had between sure to the risks included in this study has remained 10% and 20% of DALYs attributable to smoking; Canada, remarkably constant over the past 30 years. Risk­ deleted most states in the USA, Russia, the rest of China, and mortality rates over the same period have declined, many parts of southeast Asia were also in this category. ranging from a 3·3% decline per year in females aged Attributable burden remains less than 6% in most of 1–4 years to a 0·3% decline per year in males aged Mexico, central America, and Andean Latin America. 90–94 years. Despite this overall pattern, reductions in The burden attributable to tobacco is less than 2% in key risks highly correlated with SDI—unsafe water, much of western and eastern sub­ Saharan Africa. sanitation, and handwashing; household air pollution; Figure 6D shows the burden attributable to air pollution child growth failure; and vitamin A and zinc defi­ (ambient particulate matter, household air pollution, and ciencies—have contributed to reductions in global child ambient ozone pollution). No location had more than death rates. Among the most detailed major non­ commu­ 20% of DALYs attributable to air pollution. But a wide nicable disease risks, only tobacco smoking has declined range of countries in western and eastern sub­ Saharan steadily. At the global level in 2019, there were three risk Africa had attributable burden percentages between 10% factors that accounted for more than 1% of DALYs and and 15%. Similarly, nearly all locations in south Asia, were increasing in exposure by more than 1% per year: many parts of southeast Asia, and most provinces in high BMI, ambient particulate matter pollution, and China also had the same levels of attributable burden. The high FPG. There is large scope for public regulatory spatial patterns of the constituent risks included in air policy, community programmes, and primary care inter­ pollution—particularly ambient particulate matter and ventions on risks to have a greater effect on prevention. household air pollution—were quite different (see GBD These broad global patterns mask considerable hetero­ For GBD Compare see https://vizhub.healthdata.org/ Compare for data), with ambient particulate matter geneity in risk levels and trends at the country level, gbd-compare/ pollution playing a much greater role in Asia than in reinforcing the need for country assessments and Africa. country­ specific prevention planning. The fifth most important Level 2 risk factor was the dietary risks that are based on the joint effects of 15 diet Important risk factor trends quality components (figure 6E). In Bulgaria, dietary risks Some risk exposures are highly correlated with social accounted for more than 20% of attributable DALYs. and economic development, as measured by SDI. However, diet accounts for more than 10% of DALYs in As countries and territories increase SDI through many locations in central and eastern Europe, central higher levels of education, particularly among women; Asia, and most of China. The lowest shares of DALYs increased GDP per capita; and improved access to attributed to dietary risks are in sub­ Saharan Africa, modern contraceptives, we should expect progress on particularly countries in the Sahel. these risks. The incremental effect of campaigns, policies, and programmes on top of this social and Risk-specific trends economic development process is yet to be established Two­ page risk­ specific summaries provide detailed in this analysis. Bending the development curve is results on attri butable deaths, YLLs, YLDs, and DALYs possible, as evidenced by the abrupt accelerations in the for a selection of the 87 risk factors in the GBD risk decline in some risk factors, such as the recent decline hierarchy. These summaries include 2019 counts, age­ in unsafe sanitation in India. Even for risks that standardised rates, and rankings for attributable burden; are historically highly correlated with SDI, intervention the com position of attributable burden for leading can accelerate progress. The range of policy initiatives to causes; patterns of attributable burden over time and accelerate the transition to cleaner cooking fuels is 21,22 age; and age­ standardised SEVs by location and SDI. another example of this effort. Analysis of exemplars, They were written to increase the accessibility to and countries with lower SEVs for these risks for their level transparency of GBD estimates for each risk factor. of SDI or faster progress than expected for the change in Summaries for select risk factors are highlighted in print SDI, could yield further insights. For all two-page summaries see (pp S216–319); summaries for all risk factors can be Two risk factors that have not been highly correlated https://www.thelancet.com/ found online. with SDI in the past have also seen declines in exposure gbd/summaries at nearly 1% per year over the study period: tobacco Discussion smoking and lead exposure. Progress on exposure to Main findings these risks stands out compared with the increases in Our analysis of risk­attributable burden using exposure to many metabolic risks and no substantial 30 652 sources for exposure, relative risk, and the TMREL change for others such as diet quality. In both of these showed that in 2019, 47·8% (95% UI 45·3–50·1) of global cases, government action through taxation and regulatory www.thelancet.com Vol 396 October 17, 2020 1241 Global Health Metrics policy for tobacco smoking, including advertising bans intake above energy requirements. Some studies sug­ 24,25 and clean air legislation, and regulation of lead content, gest that certain diet components are more likely to have had a major effect. Tobacco interventions highlight contribute to increased BMI than others; the mechanism how regulatory policy can lead to behaviour change. of these effects can be complex and include effects on International efforts for tobacco control have also been appetite, absorption, and displacement of other foods. It bolstered by the Framework Convention on Tobacco is currently hard to understand the role of physical Control. Despite the more than 1% per year decline in inactivity, excess caloric intake, and diet quality in driving age­ standardised tobacco smoking exposure between the increase in BMI. The large combined burden of 2010 and 2019, tobacco remains the third leading risk diet quality, physical inactivity, and high BMI (11·9% factor for attributable DALYs among Level 2 risks. For [95% UI 9·6–14·5] of all DALYs in 2019) indicates just the three major and rapidly increasing risks, the role of how profoundly important the nexus of diet and physical taxation and regulatory policy should be examined. For activity can be to current and future health. The setting ambient particulate matter pollution, regulation can for understanding the potential of changes in overall diet 27,28 clearly have a direct impact. For the nexus of high FPG is to use future health scenarios to trace how public and high BMI, regulatory strategies are less clear. We are policies such as subsidies, taxes, information campaigns, failing to deal with these risks, and concerted research and improving accessibility can affect health in each and policy efforts are needed to reverse the trends. country. In this study, no country or territory has had a The marked rise of metabolic risks as a group, in significant decline in the proportion of the population particular high FPG and high BMI, and their large with high BMI between 1990 to 2019 or in the past contribution to attributable burden is perhaps most decade. The complete failure to reduce BMI at the disturbing. During this period of rising metabolic risk national level implies that efforts to modify the nexus of exposure, global cardiovascular disease age­ standardised physical inactivity, diet quality, and excess energy intake mortality has been declining as documented in GBD 2019 might be very challenging. Tackling this diet quality and for diseases and injuries. The seeming paradox could, to excess energy intake will not only be important for a large extent, be explained by the effect of access to care, human health but has important ramifications for social determinants of health, cohort effects, and other environmental sustainability. behavioural, occupational, and environmental risks not The two types of exposure to particulate matter with a quantified here. Rising metabolic risks might at some diameter of less than 2·5 μm (PM ) have profoundly 2·5 point overwhelm these other drivers and eventually lead to different relationships with socio ­ demographic develop­ rising cardiovascular mortality in the future. This situation ment: household air pollution is strongly related to SDI might have arrived in some high­ income countries in and tends to decrease steadily with socio­ demographic which age­ standardised cardio vascular disease mortality development. By contrast, ambient particulate matter has plateaued or increased since 2017. If year­ on­ year pollution tends to increase with industrialisation and declines in cardiovascular disease mortality come to an then decline with air­ quality management at higher levels 37,38 end, the effect on mortality and longevity at the global level of SDI. The global increases in ambient particulate could be massive. While high BMI and high FPG have matter pollution exposure are being driven by the middle steadily increased, high LDL cholesterol has remained SDI quintiles, as seen in figure 1B. Studies have shown constant over the past decade despite the expected that for ambient particulate matter pollution (ambient correlation with BMI; this finding warrants further PM ), the main sources of exposure are residential 2·5 investigation and could be related to changes in diet energy use, industry, and power generation. The con­ quality, pharmacological intervention, or other factors. centration of PM burden in south Asia highlights how 2·5 Although not increasing at the rate of high BMI or high the absence of national policy actions can have a major FPG, high SBP has become the leading risk factor for effect. Among the large risk factors in which exposure is disease burden at the global level, among the most increasing, ambient particulate matter pollution stands detailed risks in this analysis. A range of strategies out because exposure is declining in countries with a including primary care management and reductions in higher SDI. Like tobacco and lead, regulation can have a sodium intake are known to be potentially effective in profound effect on exposure to and health effects of 32,33 reducing the burden of this critical risk factor. ambient particulate matter pollution and does not require 40,41 The rise of high BMI and its probable role in increasing individual action. There is a clear role for global high FPG needs further examination. Increased BMI can organisations to encourage regulatory change in middle be traced to the combination of physical inactivity, excess SDI countries with large and increasing exposure to caloric intake, and diet quality. At the global level, we ambient particulate matter pollution. This agenda is all find that high BMI is rising considerably faster than low the more urgent because of the direct linkage to global physical activity and poor diet quality. Diet quality on its climate change. own is the fifth leading Level 2 risk factor for attributable Because of profound global interest in the potential DALYs. The effect of diet on human health goes beyond health effects of climate change, we have included high diet quality and should include the contribution of diet and low non­ optimal temperatures in GBD 2019. Climate 1242 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics change will have impacts on human health through between avoidable and attributable burden can vary many mechanisms: direct effects of temperature rise, across countries; in high mortality settings, competing humidity changes, sea­ level rise, extreme weather events, risks might mean that avoidable burden will be sys­ and reduced agricultural yields and increased rural tematically smaller than attributable burden. 42,43 poverty. We have so far included only one of these For GBD 2019 and all previous GBD CRA efforts, our pathways in the GBD analysis, namely the direct effects inclusion criteria for a risk–outcome pair were based on of ambient temperature on different disease outcomes. the World Cancer Research Fund criteria for convincing Our analysis showed that the TMREL varies as a function or probable evidence. We also required that published of mean annual temperature. Locations where mean studies, when meta­ analysed together, yielded a sig­ temperature is higher tend to have higher optimal nificant (p<0·05) relative risk for any risk–outcome pair temperatures, probably through physical and social meeting these criteria. To avoid risk–outcome pairs on adaptation. In 2019, the burden (as measured by the cusp of statistical significance coming in and out of percentage of total DALYs) attributable to low temperature GBD with different cycles, we introduced a threshold of was 2·2 times greater than the burden attributable to p>0·1 to exclude a risk–outcome pair that has previously high temperature. This balance does not, however, hold been included in GBD. Among the included risk– true when looking at specific locations or regions. While outcome pairs, the consistency of the evidence and risk for high SDI countries, the cold­ related burden is of bias varies considerably. The evidence linking smoking 15·4 times greater than the heat­ related burden, this to lung cancer is clearly far stronger than the evidence on relationship is switched for other regions, such as south omega­ 3 and ischaemic heart disease. The UI of the Asia where we observed a 1·7 times greater heat­ related mean effect does not fully capture this difference in the burden and sub­ Saharan Africa where we observed consistency of evidence or the risk of bias. A more robust a  3·6 times greater heat­ related burden. Rising tem­ measure needs to take into account various risks of bias perature will probably have a substantial effect in and the unexplained variation in effect after taking into locations with less capacity to adapt to increased account these risks of bias as well as the magnitude of temperature, potentially exacerbating health inequalities the effect size. The relative risk of lung cancer from across countries. The social capacity to adapt is also smoking is very high across levels of exposure, with a probably tied to economic development: for example, air relative risk of 3·4 at ten pack­ years of smoking and a conditioners in the USA have mitigated the impact of relative risk of 6·5 at 20 pack­ years of smoking, which heat waves over the past 50 years. In terms of trends, makes it far more likely to be causal than an exposure there was a marked increase in exposure to high with a relative risk of 1·1. If we included the unexplained temperature from 1990 to 2010 and then a slight decline heterogeneity across studies after adjusting for risk of from 2010 to 2019; there are major annual fluctuations in bias in the UI of the relative risks, several risk–outcome temperature exposure on top of long­ term warming pairs might not meet inclusion criteria. We are working trends, and 2010 stood out as a year with high tempe­ on developing an evidence scoring system that quantifies ratures in many regions. Our analysis does not provide a consistency and risk of bias for GBD and would allow basis for understanding the full effects of future climate readers to understand that not all risk–outcome pairs change, which will operate through many different have the same evidence base. It would, however, be pathways in addition to the direct effects of temperature. misleading and potentially harmful to argue that we In the GBD CRA work to date, we have estimated the should only examine the GBD risk–outcome pairs burden attributable to past exposure in a given year. The with the highest grade of evidence. The precautionary CRA framework also laid out the important utility for principle for public policy implies that governments have policy making of estimating how changes in current and a duty to act on risk factors that are probably or potentially future exposure can change future levels of health; this harmful and not only those that have overwhelming concept is called avoidable burden. Most CRA work to evidence. Relying only on effects on the basis of the date has focused on estimating attributable burden, highest degree of evidence will very seriously delay public even though avoidable burden is arguably more relevant recognition and proactive policies, which in turn would to policy prioritisation. The dominance of work on attri­ result in perpetuating preventable burden. We hope to butable burden is founded on two premises: it is very inform both individuals and public policy makers with difficult to estimate avoidable burden as this estimate quantification of burden and strength of evidence so they requires a comprehensive future health scenarios frame­ are empowered to make sense of the available data. work; and attributable burden is likely to be highly Analysing relative risks to inform choices by individuals correlated with avoidable burden. With the availability of and choices by population health decision makers might 45,46 a GBD­ informed future health scenarios platform, the legitimately have different perspectives. Some guidelines 48,49 possibility of estimating avoidable burden is much more on systematic reviews recommend reporting absolute tractable. Future work on avoidable burden for each GBD risk levels related to exposure; this method is perhaps risk factor might allow us to examine the true relationship appropriate for informing individual choice. For public between the two approaches to CRA. The relationships policy, attributable burden might be more relevant. If a www.thelancet.com Vol 396 October 17, 2020 1243 Global Health Metrics risk factor is related to 1000 deaths, from a public policy cardiovascular diseases, particularly stroke, due to newly perspective, the concentration of the risk in a smaller or added data and changes in fitting the exposure ­ response larger group of individuals might matter less than it does curves. Given the very large burden of low birthweight to individuals. To provide a synthesis of the evidence for and short gestation on neonatal mortality, the inclusion different users, we included estimation of the all ­ cause of these intermediates has been an important change in mortality relative risk associated with exposure levels of our assessment. The burden of stroke and ischaemic each risk factor using the global distribution of burden heart disease attributable to kidney dysfunction increased across outcomes. However, risks that cause relatively between GBD 2017 and GBD 2019 after updating the modest increases for individuals but are highly prevalent, relative risks with new data from 44 cohorts. For instance, such as air pollution, are, nevertheless, legitimate targets the comparable estimate of the proportion of cardio­ for public policy. vascular DALYs due to kidney dysfunction in 2010 increased from 6·8% (95% UI 6·0–7·6) to 8·5% Substantial changes compared with GBD 2017 (6·8–10·3). Compared with GBD 2017, our GBD 2019 estimates of the burden (as measured by percentage of total DALYs) Limitations attributable to diet quality in 2017 were 29·7% lower. In GBD 2019, we undertook a reassessment of dose– These reductions stem from three major sources: response relationships and relaxed previous assumptions changes in the crosswalks between alternative and that the risk curve is log­ linear. This reassessment was reference methods for estimating diet intake, new limited, however, to dietary risks, kidney dysfunction, and systematic reviews and meta­ regressions, and more air pollution. Future reassess ments of other continuous risk empirical standardised methods for selecting the TMREL factors that currently assume a log­ linear relationship could for protective factors. Although there were changes in materially change risk factor rankings in the future and the overall burden of diet, there were larger changes could also lead to exclusion of other risk–outcome pairs. in the diet components themselves, particularly the Assessment of the joint effects of risk factors depends substantial increase in the attributable burden from on two critical factors: the correlation of risk exposure red meat and the decline in the burden attributable to and the estimation of the joint effects of groups of risks low vegetable intake. The sources of the changes were together. For exposure, we assumed that for each age­ sex­ the same as for diet quality overall. One of the most location­ year, the estimates of the prevalence of exposure important insights from this enriched analysis is that for were independent. Previous simulation analyses under­ many harmful and protective factors, the relative risk taken for GBD 2010 with use of US data from the functions tend to flatten out at higher exposure levels; National Health and Nutrition Examination Survey the previous practice of imposing a log­ linear functional suggested this assumption did not materially bias our form on the risk equation—widely used in the scientific findings. To assess the joint effects of risk factors, we literature—might have led to overestimation. For protec­ assumed in general that relative risks are multiplicative. tive diet components (whole grains, fruit, fibre, nuts This simple assumption has been modified to take into and seeds, omega­ 3, polyunsaturated fatty acids, veg­ account known pathways in which one risk factor, such etables, milk, and calcium), we set the TMREL to the as fruit consumption, is mediated through another risk 85th percentile of levels of exposure included in the factor such as fibre intake. To avoid over ­ estimation of published cohort studies or randomised controlled trials. the joint effects, we computed the non ­ mediated relative With further study of individuals with higher levels of risks and then assumed that non­ mediated relative intake, it is possible that the level of intake associated risks are multiplicative. This approach does not capture with the lowest risk is in fact higher than the TMREL potential synergy between relative risks in which some set for protective diet components in GBD 2019. 12 diet combinations might be super­ multiplicative. For some risk–outcome pairs from GBD 2017 were excluded areas such as diet, the joint estimation is very important from GBD 2019 because our re­ analysis with updated for public policy. Further, more detailed work is needed data suggested that the effects were no longer signifi­ to strengthen the evidence base for understanding cant. Some risk–outcome pairs, such as omega­ 3 and mediation. In particular, mediation implies necessarily ischaemic heart disease, which remained in the analysis that exposure between mediated risks is correlated. as the result of the new meta­ regression of 21 trials and Factoring in that implied correlation into risk exposure 27 cohort studies, met inclusion criteria but future estimation could strengthen estimates in the future. studies could shift the balance of the evidence to be The main limitation of our estimates of risk­ excluded. attributable burden is the availability and quality of Particulate matter pollution burden in 2017 was 44·6% primary data that underpin the analysis. Data for risk higher in GBD 2019 than in GBD 2017. The increase was relationships of several risk factors, such as ambient due to the inclusion of low birthweight and short ozone pollution, residential radon, occupational risks, gestation as risk factors that are themselves affected by child hood sexual abuse, intimate partner violence, PM , as well as increases in the relative risk curve for bullying victimisation, and child growth failure are 2·5 1244 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics sparse. For exposure measurement, patterns of data Conclusion availability are non­ uniform across geography and over Using the most up­ to­ date assessment of the data for time and, where available, might be based on less exposure and relative risk, we found that global exposure reliable modes of data collection such as self­ report. In to harmful environmental risks has been declining, with GBD 2019, we implemented more explicit corrections the notable exception of ambient particulate matter for bias associated with non­ reference methods of pollution. Environmental risk reduction is making an exposure measurement that improved the estimation of important contribution to reductions in child mortality. In risk exposure. Furthermore, these assessments can be aggregate, there has been no real progress reducing used to guide future data collection efforts by identifying exposure to behavioural risks, while metabolic risks are, those populations with not only sparse but low­ quality on average, increasing every year. As a world, we are data based on the collection mode. failing to change some behaviours, particularly those Our analysis, particularly the overall assessment of related to diet quality, caloric intake, and physical activity. burden attributable to all risks combined and risk­ Progress on reducing harm from one crucial behaviour, deleted mortality, is limited by several potentially tobacco smoking, shows the power of taxation and important risk factors not included in this analysis. The regulation. The promise of prevention through risk most important set is likely to be social determinants of modification is not being realised in adult populations health such as educational attainment, poverty, or social around the world. Urgent attention on more successful exclusion. We are currently doing systematic reviews on strategies to reduce risks is needed. educational attainment, which will be the first social Contributors Please see appendix 1 for more detailed information about individual determinant to be incorporated into future rounds of author contributions to the research, divided into the following the GBD CRA. There is also a wide range of other categories: managing the estimation process; writing the first draft of the risk factors not yet included such as nitrous oxide, heavy manuscript; providing data or critical feedback on data sources; metals, environ mental noise, sleep, stress, UV radiation, developing methods or computational machinery; applying analytical methods to produce estimates; providing critical feedback on methods or among others. Future rounds of GBD might evaluate results; drafting the work or revising it critically for important intellectual whether these risk factors meet inclusion criteria. content; extracting, cleaning, or cataloguing data; designing or coding To date, GBD has not included Mendelian ran­ figures and tables; and managing the overall research enterprise. domisation studies in meta­ regression. These studies Declaration of interests could provide new insights on the causal connections C A T Antonio reports personal fees from Johnson & Johnson between risks and outcomes. Not all Mendelian (Philippines), outside of the submitted work. E Beghi reports grants from 51–53 Italian Ministry of Health and SOBI, and personal fees from Arvelle randomisation studies are appropriate for inclusion. Therapeutics, outside of the submitted work. Y Béjot reports personal Future rounds of GBD will give careful consideration to fees from AstraZeneca, Bristol­ Myers Squibb, Pfizer, Medtronic, Merck including these studies for some risk–outcome pairs. Sharpe & Dohme, and Amgen; grants and personal fees from Boehringer For harmful risks with monotonically increasing risk Ingelheim; personal fees and non­ financial support from Servier; and non­ financial support from Biogen, outside of the submitted work. functions, we have generally assumed that the TMREL M L Bell reports grants from US Environmental Protection Agency, is 0. For protective risks such as fruit or whole grain National Institutes of Health (NIH), and from Wellcome Trust intake, selecting the level of exposure that is minimum Foundation, during the conduct of the study; and Honorarium or travel risk is more challenging. Extrapolating the risk function reimbursement from the NIH (for review of grant proposals), American Journal of Public Health (participation as editor), Global Research beyond where the available cohort studies or trials Laboratory and Seoul National University, Royal Society, Ohio University, support the protective effect could easily lead to both Atmospheric Chemistry Gordon Research Conference, Johns Hopkins exaggerated estimates of attributable burden and implau­ Bloomberg School of Public Health, Arizona State University, Ministry of sible recom mendations on consumption. To avoid this the Environment Japan, Hong Kong Polytechnic University, University of Illinois–Champaign, and the University of Tennessee–Knoxville, outside exaggeration, we set the TMREL for protective risks to be of the submitted work. H Christensen reports personal fees from Bristol equal to the 85th percentile of exposure in the available Myers Squibb, Bayer, and Boehringer Ingelheim, outside of the cohorts and trials. The 85th percentile is arbitrary, but submitted work. S­ C Chung reports grants from GlaxoSmithKline, sensitivity analysis did not suggest major changes if we outside of the submitted work. L Degenhardt reports grants from Indivior and Seqirus, outside of the submitted work. S M S Islam reports selected the 90th or 80th percentiles. grants from National Heart Foundation of Australia and Deakin Lastly, in most cases, we assume that relative risks as University, during the conduct of the study. S L James reports grants a function of exposure are universal and apply in from Sanofi Pasteur and employment from Genentech, outside of the all locations and time periods. Exceptions include submitted work. V Jha reports grants from Baxter Healthcare, GlaxoSmithKline, Zydus Cadilla, and Biocon and personal fees from temperature, in which the risk functions clearly depend NephroPlus, outside of the submitted work. J J Jozwiak reports personal on the annual mean temperature, and the relative risks fees from Amgen, ALAB Laboratoria, Teva, Synexus, and Boehringer for high BMI for breast cancer that differ in Asian and Ingelheim, outside of the submitted work. S V Katikireddi reports grants from NRS Senior Clinical Fellowship, UK Medical Research Council, and non­ Asian populations. Our rules require that there is the Scottish Government Chief Scientist Office, during the conduct of evidence of significant differences in the relative risk for the study. M Kivimäki reports grants from the Medical Research Council, different subgroups; to date, few cases have met this UK (MR/R024227/1), during the conduct of the study. S Lorkowski standard. As evidence accumulates, more location­ reports personal fees from Akcea Therapeutics, Amedes, Amgen, Berlin­ Chemie, Boehringer Ingelheim Pharma, Daiichi Sankyo, Merck Sharp & specific or subgroup relative risks might be identified. www.thelancet.com Vol 396 October 17, 2020 1245 Global Health Metrics Dohme, Novo Nordisk, Sanofi ­ Aventis, Synlab, Unilever, and Upfield, The authors are grateful to the Ministry of Health, the survey copyright and non­ financial support from Preventicus, outside of the submitted owner, for allowing them to have the database. All results of the study work. R V Martin reports grants from the Natural Science and are those of the authors and in no way committed to the Ministry. Engineering Research Council of Canada, during the conduct of the This research used data from China Family Panel Studies, funded by study. T R Miller reports having a contract from the AB InBev 985 Program of Peking University and carried out by the Institute of Foundation, outside of the submitted work. J F Mosser reports grants Social Science Survey of Peking University. The Costa Rican Longevity from the Bill & Melinda Gates Foundation, during the conduct of the and Healthy Aging Study project is a longitudinal study by the study. S Nomura reports grants from the Ministry of Education, Culture, University of Costa Rica’s Centro Centroamericano de Población and Sports, Science, and Technology. S B Patten reports funding from the Instituto de Investigaciones en Salud, in collaboration with the Cuthbertson & Fischer Chair in Pediatric Mental Health at the University University of California at Berkeley. The original pre­ 1945 cohort was of Calgary, during the conduct of the study. C D Pond reports personal funded by the Wellcome Trust (grant 072406), and the 1945­ 1955 fees from Nutricia, outside of the submitted work; and grants from the Retirement Cohort was funded by the US National Institute on Aging National Medical Research council in relation to dementia, and travel (grant R01AG031716). The Principal Investigators are Luis Rosero­ Bixby grants and remuneration related to education of primary care and William H Dow, and co­ Principal Investigators Xinia Fernández and professionals in relation to dementia. M J Postma reports grants from Gilbert Brenes. This paper uses data from Eurostat. The responsibility BioMerieux, WHO, European Union, FIND, Antilope, DIKTI, LPDP, for all conclusions drawn from the data lies entirely with the authors. Bayer, and Budi; personal fees from Quintiles, Novartis, and Pharmerit; The Health Behaviour in School­ Aged Children study is an international grants and personal fees from IQVIA, Bristol­ Myers Squibb, Astra study carried out in collaboration with WHO/EURO. The International Zeneca, Seqirus, Sanofi, Merck Sharpe & Dohme, GlaxoSmithKline, Coordinator of the 1997–98, 2001–02, 2005–06, and 2009–10 surveys was Pfizer, Boehringer Ingelheim, and Novavax; stocks from Ingress Health, Candace Currie and the Data Bank Manager for the 1997–98 survey was and PAG; and is acting as adviser to Asc Academics, all outside of the Bente Wold, whereas for the following survey Oddrun Samdal was the submitted work. I Rakovac reports grants from WHO, during the Databank Manager. A list of principal investigators in each country can conduct of the study. A E Schutte reports personal fees from Omron, be found at http://www.hbsc.org. This paper uses data from the WHO Servier, Takeda, Novartis, and Abbott, outside of the submitted work. Study on global AGEing and adult health. Researchers interested in J A Singh reports personal fees from Crealta/Horizon, Medisys, Fidia, using Irish Longitudinal Study on Ageing data can access the data for UBM LLC, Trio Health, Medscape, WebMD, Clinical Care Options, free from the following sites: Irish Social Science Data Archive at Clearview Healthcare Partners, Putnam Associates, Spherix, Practice University College Dublin (http://www.ucd.ie/issda/data/tilda/); Point Communications, the NIH, and the American College of Interuniversity Consortium for Political and Social Research at the Rheumatology; personal fees from Simply Speaking; stock options in University of Michigan (http://www.icpsr.umich.edu/icpsrweb/ICPSR/ Amarin Pharmaceuticals and Viking Pharmaceuticals; membership in studies/34315). Data used in this paper come from the 2009–10 Ghana the steering committee of OMERACT (an international organisation that Socioeconomic Panel Study Survey, which is a nationally representative develops measures for clinical trials and receives arm’s length funding survey of more than 5000 households in Ghana. The survey is a joint from 12 pharmaceutical companies), the FDA Arthritis Advisory effort undertaken by the Institute of Statistical, Social and Economic Committee, and the Veterans Affairs Rheumatology Field Advisory Research (ISSER) at the University of Ghana, and the Economic Growth Committee; and non­ financial support from UAB Cochrane Centre (EGC) at Yale University. It was funded by EGC. At the same Musculoskeletal Group Satellite Center on Network Meta­ analysis, time, ISSER and the EGC are not responsible for the estimations outside of the submitted work. S T S Skou reports personal fees from reported by the analysts. The Palestinian Central Bureau of Statistics Journal of Orthopaedic & Sports Physical Therapy and Munksgaard and granted the researchers access to relevant data in accordance with grants from The Lundbeck Foundation, outside of the submitted work; license number SLN2014­ 3­ 170, after subjecting data to processing and being co­ founder of GLA:D. GLA:D is a non­ profit initiative hosted at aiming to preserve the confidentiality of individual data in accordance University of Southern Denmark aimed at implementing clinical with the General Statistics Law, 2000. The researchers are solely guidelines for osteoarthritis in clinical practice. J D Stanaway reports responsible for the conclusions and inferences drawn upon available grants from the Bill & Melinda Gates Foundation, during the conduct of data. The authors thank the Russia Longitudinal Monitoring Survey, the study. D J Stein reports personal fees from Lundbeck, and Sun, RLMS­ HSE, conducted by the National Research University Higher outside of the submitted work. F Topouzis reports grants from Pfizer, School of Economics and ZAO Demoscope together with Carolina Thea, Novartis, Rheon, Omikron, Pharmaten, Bayer, and Population Center, University of North Carolina at Chapel Hill, and the Bausch & Lomb; and personal fees from Novartis, and Omikron, outside Institute of Sociology, RAS for making data available. This paper uses of the submitted work. R Uddin reports travel and accommodation data from the Armenia 2016, Bangladesh 2009–10, Belarus 2016–17, reimbursement from Deakin University Institute for Physical Activity Benin 2015, Bhutan 2014, Iraq 2015, Kuwait 2006 and 2014, Libya 2009, and Nutrition, outside of the submitted work. All other authors declare Malawi 2009, Moldova 2013, and Sudan 2016 STEP Surveys, all no competing interests. implemented with the support of WHO. This paper uses data from Survey of Health, Ageing and Retirement in Europe (SHARE) Waves 1 For more on SHARE see Data sharing (DOI:10.6103/SHARE.w1.700), 2 (10.6103/SHARE.w2.700), 3 (10.6103/ http://www.share-project.org To download the data used in these analyses, please visit the Global SHARE.w3.700), 4 (10.6103/SHARE.w4.700), 5 (10.6103/SHARE.w5.700), For the Global Health Data Health Data Exchange GBD 2019 website. 6 (10.6103/SHARE.w6.700), and 7 (10.6103/SHARE.w7.700); see Börsch­ Exchange GBD 2019 website Acknowledgments Supan and colleagues (2013) for methodological details. The SHARE see http://ghdx.healthdata.org/ Research reported in this publication was supported by the Bill & data collection has been funded by the European Commission through gbd−2019 Melinda Gates Foundation; Bloomberg Philanthropies; the University of FP5 (QLK6­ CT­ 2001­ 00360), FP6 (SHARE­ I3: RII­ CT­ 2006­ 062193, Melbourne; Queensland Department of Health, Australia; the National COMPARE: CIT5­ CT­ 2005­ 028857, SHARELIFE: CIT4­ CT­ 2006­ 028812), Health and Medical Research Council, Australia; Public Health England; FP7 (SHARE­ PREP: GA number 211909, SHARE­ LEAP: GA number the Norwegian Institute of Public Health; St Jude Children’s Research 227822, SHARE M4: GA number 261982), and Horizon 2020 Hospital; the Cardiovascular Medical Research and Education Fund; (SHARE­ DEV3: GA number 676536, SERISS: GA number 654221) and the National Institute on Ageing of the National Institutes of Health by DG Employment, Social Affairs & Inclusion. Additional funding from (NIH; award P30AG047845); and the National Institute of Mental Health the German Ministry of Education and Research, the Max Planck Society of the NIH (award R01MH110163). The content is solely the for the Advancement of Science, the US National Institute on Aging responsibility of the authors and does not necessary represent the official (U01_AG09740­ 13S2, P01_AG005842, P01_AG08291, P30_AG12815, views of the funders. Data for this research was provided by MEASURE R21_AG025169, Y1­ AG­ 4553­ 01, IAG_BSR06­ 11, OGHA_04­ 064, Evaluation, funded by the United States Agency for International HHSN271201300071C) and from various national funding sources is Development (USAID). Views expressed do not necessarily reflect those gratefully acknowledged (www.share­ project.org). The United States of USAID, the US Government, or MEASURE Evaluation. This research Aging, Demographics, and Memory Study is a supplement to the Health used data from the Chile National Health Survey 2003 and 2009–10. and Retirement Study, which is sponsored by the National Institute of 1246 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics Aging (grant number NIA U01AG009740). It was conducted jointly by Research, Kalaghatgi. B­ F Hwang acknowledges support from China Duke University and the University of Michigan. This paper uses data Medical University (107­ Z­ 04), Taichung, Taiwan. N Ikeda acknowledges from Add Health, a programme project designed by J Richard Udry, support from The Japan Society for the Promotion of Science (grant Peter S Bearman, and Kathleen Mullan Harris, and funded by a grant number 15K08762 and 18H03063). S M S Islam acknowledges support P01­ HD31921 from the Eunice Kennedy Shriver National Institute of from the National Heart Foundation of Australia and Deakin University. Child Health and Human Development, with cooperative funding from M Jakovljevic acknowledges grant OI175014 of the Ministry of Education 17 other agencies. Special acknowledgment is due to Ronald R Rindfuss Science and Technological Development of the Republic of Serbia for and Barbara Entwisle for assistance in the original design. Information supporting the Serbian portion of this GBD study. P Jeemon supported on how to obtain the Add Health data files is available on the Add Health by a Clinical and Public Health intermediate fellowship (grant number For the Add Health website see website. No direct support was received from grant P01­ HD31921 for this IA/CPHI/14/1/501497) from the Wellcome Trust­ Department of http://www.cpc.unc.edu/ analysis. This paper uses data from the NIH. The content is solely the Biotechnology, India Alliance (2015–20). O John acknowledges receiving addhealth responsibility of the authors and does not necessarily represent the the UIPA scholarship from UNSW, Sydney. S V Katikireddi official views of the NIH. This analysis uses data or information from acknowledges funding from a NRS Senior Clinical Fellowship the Longitudinal Aging Study in India (LASI) Pilot microdata and (SCAF/15/02), the Medical Research Council (MC_UU_12017/13 & documentation. The development and release of the LASI Pilot Study MC_UU_12017/15), and the Scottish Government Chief Scientist Office was funded by the National Institute on Aging, a division of the NIH (SPHSU13 & SPHSU15). C Kieling acknowledges employment with the (R21AG032572, R03AG043052, and R01 AG030153). L G Abreu Conselho Nacional de Desenvolvimento Científico e Tecnológico acknowledges support from Coordenação de Aperfeiçoamento de (a Brazilian public funding agency) and being a UK Academy of Medical Pessoal de Nível Superior, Brazil (finance Code 001) and Conselho Sciences Newton Advanced Fellow. Y J Kim acknowledges support from Nacional de Desenvolvimento Científico e Tecnológico. L Abu­ Raddad the Research Management Centre, Xiamen University Malaysia, acknowledges the support of Qatar National Research Fund XMUMRF/2018­ C2/ITCM/0001. K Krishan acknowledges support from (NPRP 9­ 040­ 3­ 008). O Adetokunboh acknowledges the South African UGC Centre of Advanced Study (CAS II) awarded to the Department of Department of Science and Innovation, and National Research Anthropology, Panjab University, Chandigarh, India. B Lacey Foundation. A Agrawal acknowledges support from the Wellcome Trust acknowledges support from the NIHR Oxford Biomedical Research DBT India Alliance Senior Fellowship. R Akinyemi acknowledges Centre and the BHF Centre of Research Excellence, Oxford. T Lallukka support by Grant U01HG010273 from the NIH as part of the H3Africa acknowledges support from the Academy of Finland (grant Consortium and support by the FLAIR fellowship funded by the number 319200). I Landires is member of the Sistema Nacional de UK Royal Society and the African Academy of Sciences. S Aljunid Investigación, which is supported by the Secretaría Nacional de Ciencia, acknowledges the Department of Health Policy and Management, Tecnología e Innovación, Panamá. S Langan acknowledges support from Faculty of Public Health, Kuwait University and International Centre for a Wellcome Trust Senior Clinical Fellowship (205039/Z/16/Z). J Lazarus Casemix and Clinical Coding, Faculty of Medicine, and the National acknowledges support from a Spanish Ministry of Science, Innovation University of Malaysia for the approval and support to participate in this and Universities Miguel Servet grant (Instituto de Salud Carlos III/ESF, research project. T Bärnighausen acknowledges support from the EU [CP18/00074]). S Lorkowski acknowledges support from the German Alexander von Humboldt Foundation through the Alexander von Federal Ministry of Education and Research (nutriCARD, grant Humboldt Professor award, funded by the German Federal Ministry of agreement number 01EA1808A). A M Mantilla­ Herrera is affiliated with Education and Research. A Badawi acknowledges support from the the Queensland Centre for Mental Health Research which receives core Public Health Agency of Canada. G Britton acknowledges support from funding from the Department of Health, Queensland Government. the Sistema Nacional de Investigación of SENACYT, Panamá. J J Carrero R Martin is grateful to Washington University for additional institutional acknowledges support from the Swedish Research Council (2019­ 01059). support. J McGrath acknowledges support from the Danish National F Caravlho acknowledges support from UID/MULTI/04378/2019 and Research Foundation (Niels Bohr Professorship), and the Queensland UID/QUI/50006/2019 with funding from FCT/MCTES through national Health Department (via West Moreton HHS). W Mendoza acknowledges funds. A Cohen acknowledges support from Health Effects Institute, Population and Development department at the United Nations Boston. V M Costa acknowledges support from Grant Population Fund Country Office in Peru, which does not necessarily SFRH/BHD/110001/2015, received by Portuguese national funds endorse this study. U Mueller acknowledges support from the German through Fundação para a Ciência e Tecnologia, IP, under the Norma National Cohort Study BMBF grant number 01ER1801D. S Nomura Transitória DL57/2016/CP1334/CT0006. M DeLang acknowledges the acknowledges support from the Ministry of Education, Culture, Sports, NASA grant number NNX16AQ30G and NIOSH grant number Science, and Technology of Japan (18K10082). A Ortiz acknowledges T42­ OH008673 for supporting the work to estimate global ozone. support from ISCIII PI19/00815, DTS18/00032, ISCIII­ RETIC H Erskine acknowledges support from the Australian National Health REDinREN RD016/0009 Fondos FEDER, FRIAT, Comunidad de Madrid and Medical Research Council Early Career Fellowship (APP1137969) B2017/BMD­ 3686 CIFRA2­ CM; these funding sources had no role in the and is employed by the Queensland Centre for Mental Health Research, writing of the manuscript or the decision to submit it for publication. which receives core funding from Queensland Health. E Fernandes S Patten acknowledges support from the Cuthbertson and Fischer Chair acknowledges support from UID/MULTI/04378/2019 and in Pediatric Mental Health at the University of Calgary. G Patton UID/QUI/50006/2019 with funding from FCT/MCTES through national acknowledges support from a National Health & Medical Research funds. A Ferrari acknowledges support from a National Health and Council Fellowship. M R Phillips acknowledges support in part by the Medical Research Council Early Career Fellowship Grant (APP1121516) National Natural Science Foundation of China (number 81371502 and and the Queensland Centre for Mental Health Research, which receives 81761128031). A Raggi, D Sattin, and S Schiavolin acknowledge support funding from the Queensland Department of Health. M Ferreira by a grant from the Italian Ministry of Health (Ricerca Corrente, acknowledges support from a National Health and Medical Council of Fondazione Istituto Neurologico C Besta, Linea 4 ­ Outcome Research: Australia fellowship. M Freitas acknowledges financial support from the dagli Indicatori alle Raccomandazioni Cliniche). D Ribeiro acknowledges EU (FEDER funds through COMPETE POCI­ 01­ 0145­ FEDER­ 029248), the financial support from the EU (FEDER funds through the and National Funds (FCT, Fundação para a Ciência e Tecnologia) Operational Competitiveness Program [COMPETE]; POCI­ 01­ 0145­ through project PTDC/NAN­ MAT/29248/2017. M Ausloos, C Herteliu, FEDER­ 029253). D C Ribeiro acknowledges support from The and A Pana acknowledge partial support by a grant of the Romanian Sir Charles Hercus Health Research Fellowship (number 18/111) Health National Authority for Scientific Research and Innovation, Research Council of New Zealand. P Sachdev acknowledges support CNDS­ UEFISCDI, project number PN­ III­ P4­ ID­ PCCF­ 2016­ 0084. from the National Health and Medical Research Council of Australia C Herteliu acknowledges partial support by a grant co­ funded by Program Grant. A M Samy acknowledges support from a fellowship European Fund for Regional Development through Operational from the Egyptian Fulbright Mission Program. D Santomauro Program for Competitiveness, Project ID P_40_382. P Hoogar acknowledges affiliation with the Queensland Centre for Mental Health acknowledges the Bio Cultural Studies, Manipal Academy of Higher Research, which receives core funding from the Department of Health. Education, and Manipal and Centre for Holistic Development and M Santric­ Milicevic acknowledges support from the Ministry of www.thelancet.com Vol 396 October 17, 2020 1247 Global Health Metrics Education, Science and Technological Development of the Republic of 11 Gakidou E, Afshin A, Abajobir AA, et al. Global, regional, and national comparative risk assessment of 84 behavioural, Serbia (contract number 175087). R Sarmiento acknowledges support environmental and occupational, and metabolic risks or clusters of from the Applied and Environmental Sciences University in Bogota, risks, 1990–2016: a systematic analysis for the Global Burden of Colombia and Carlos III Institute of Health Madrid Spain. A Schutte Disease Study 2016. Lancet 2017; 390: 1345–422. acknowledges support from the South African National Research 12 Stanaway JD, Afshin A, Gakidou E, et al. Global, regional, and national Foundation SARChI Chair initiative (GUN 86895) and the South African comparative risk assessment of 84 behavioural, environmental and Medical Research Council. M Serre acknowledges the NASA grant occupational, and metabolic risks or clusters of risks for 195 countries number NNX16AQ30G for supporting the work to estimate global and territories, 1990–2017: a systematic analysis for the Global Burden ozone. S T Skou acknowledges support from a grant from Region of Disease Study 2017. Lancet 2018; 392: 1923–94. Zealand (Exercise First) and a grant from the European Research 13 Stevens GA, Alkema L, Black RE, et al. Guidelines for accurate and Council under the EU’s Horizon 2020 research and innovation transparent health estimates reporting: the GATHER statement. programme (grant agreement number 801790). J B Soriano Lancet 2016; 388: e19–23. acknowledges support by Centro de Investigación en Red de 14 Murray CJ, Lopez AD. Global mortality, disability, and the Enfermedades Respiratorias, Instituto de Salud Carlos III, Madrid, contribution of risk factors: Global Burden of Disease Study. Lancet Spain. R Tabarés­ Seisdedos acknowledges support in part by the 1997; 349: 1436–42. national grant PI17/00719 from ISCIII­ FEDER. N Taveira acknowledges 15 Murray CJ, Ezzati M, Lopez AD, Rodgers A, Vander Hoorn S. partial support from the European & Developing Countries Clinical Comparative quantification of health risks conceptual framework Trials Partnership, EU (LIFE project, reference RIA2016MC­ 1615). and methodological issues. Popul Health Metr 2003; 1: 1. M Tonelli acknowledges support from the David Freeze chair in health 16 World Cancer Research Fund, American Institute for Cancer services research, University of Calgary. S Tyrovolas acknowledges Research. Food, nutrition, and physical activity, and the prevention support from the Foundation for Education and European Culture, of cancer: a global perspective. Washington DC: AICR, 2007. the Sara Borrell postdoctoral program (reference number CD15/00019 17 James SL, Abate D, Abate KH, et al. Global, regional, and national from the Instituto de Salud Carlos III), and the Fondos Europeo de incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, Desarrollo Regional. M Wei acknowledges support from the National 1990–2017: a systematic analysis for the Global Burden of Disease Institute on Aging at the NIH (K23AG056638). J J West acknowledges Study 2017. Lancet 2018; 392: 1789–858. support from NASA grant number NNX16AQ30G and NIOSH number 18 Griswold MG, Fullman N, Hawley C, et al. Alcohol use and burden T42­ OH008673. for 195 countries and territories, 1990–2016: a systematic analysis Editorial note: the Lancet Group takes a neutral position with respect to for the Global Burden of Disease Study 2016. Lancet 2018; 392: 1015–35. territorial claims in published maps and institutional affiliations. 19 Kurowicka D, Cooke R. Sampling algorithms for generating joint References uniform distributions using the vine­ copula method. 1 WHO. Global action plan for the prevention and control of Comput Stat Data Anal 2007; 51: 2889–906. noncommunicable diseases. 2013–2020. Geneva: World Health 20 Curtis V. Explaining the outcomes of the ‘Clean India’ campaign: Organization, 2013. institutional behaviour and sanitation transformation in India. 2 WHO. UN Task Force on NCDs: The government of Ethiopia is BMJ Glob Health 2019; 4: e001892. working jointly with the UN system to strengthen the national 21 Quinn A, Bruce N, Puzzolo E, et al. An analysis of efforts to scale NCD response. 2017. http://www.who.int/ncds/un­ task­ force/joint­ up clean household energy for cooking around the world. programming­ missions/ethiopia­ november­ 2017/en/ (accessed Energy Sustain Dev 2018; 46: 1–10. Feb 25, 2020). 22 Dabadge A, Sreenivas A, Josey A. What has the pradhan mantri 3 Centers for Disease Control and Prevention. The National ujjwala yojana achieved so far? Econ Polit Wkly 2018; 53: 69–75. Environmental Public Health Tracking Program. 2019. https://www. 23 Yach D, McKee M, Lopez AD, Novotny T. Improving diet and cdc.gov/nceh/information/tracking_network.htm (accessed physical activity: 12 lessons from controlling tobacco smoking. Feb 25, 2020). BMJ 2005; 330: 898–900. 4 WHO. Global health risks: mortality and burden of disease 24 Davis JM, Elias RW, Grant LD. Current issues in human lead attributable to selected major risks. Geneva: World Health exposure and regulation of lead. Neurotoxicology 1993; 14: 15–27. Organization, 2009. 25 Dignam T, Kaufmann RB, LeStourgeon L, Brown MJ. Control of 5 Nowbar AN, Gitto M, Howard JP, Francis DP, Al­ Lamee R. Mortality lead sources in the United States, 1970–2017: public health progress from ischemic heart disease. Circ Cardiovasc Qual Outcomes 2019; and current challenges to eliminating lead expsoure. 12: e005375. J Public Health Manag Pract 2019; 25: S13–22. 6 Zhou B, Bentham J, Di Cesare M, et al. Worldwide trends in blood 26 Chung­ Hall J, Craig L, Gravely S, Sansone N, Fong GT. Impact of pressure from 1975 to 2015: a pooled analysis of 1479 population­ the WHO FCTC over the first decade: a global evidence review based measurement studies with 19·1 million participants. Lancet prepared for the Impact Assessment Expert Group. Tob Control 2017; 389: 37–55. 2019; 28 (suppl 2): 119–28. 7 NCD Risk Factor Collaboration (NCD­ RisC). Trends in adult body­ 27 United States Environmental Protection Agency. National air mass index in 200 countries from 1975 to 2014: a pooled analysis of quality status and trends through 2007. Research Triangle Park, 1698 population­ based measurement studies with 19·2 million North Carolina: U.S. Environmental Protection Agency. Office of participants. Lancet 2016; 387: 1377–96. Air Quality Planning and Standards, 2008. https://www.epa.gov/ 8 Lim SS, Vos T, Flaxman AD, et al. A comparative risk assessment sites/production/files/2017 ­ 11/documents/trends_brochure_2007. of burden of disease and injury attributable to 67 risk factors and pdf (accessed Feb 25, 2020). risk factor clusters in 21 regions, 1990–2010: a systematic analysis 28 Zalakeviciute R, Rybarczyk Y, López­ Villada J, Diaz Suarez MV. for the Global Burden of Disease Study 2010. Lancet 2012; Quantifying decade­ long effects of fuel and traffic regulations on 380: 2224–60. urban ambient PM pollution in a mid­ size South American city. 2·5 9 Forouzanfar MH, Alexander L, Anderson HR, et al. Atmos Pollut Res 2018; 9: 66–75. Global, regional, and national comparative risk assessment of 29 GBD 2019 Diseases and Injuries Collaborators. Global burden of 79 behavioural, environmental and occupational, and metabolic 369 diseases and injuries in 204 countries and territories, risks or clusters of risks in 188 countries, 1990–2013: a systematic 1990–2019: a systematic analysis for the Global Burden of Disease analysis for the Global Burden of Disease Study 2013. Lancet 2015; Study 2019. Lancet 2020; 396: 1204–22. 386: 2287–323. 30 Fullman N, Yearwood J, Abay SM, et al. Measuring performance on 10 Forouzanfar MH, Afshin A, Alexander LT, et al. Global, regional, the Healthcare Access and Quality Index for 195 countries and and national comparative risk assessment of 79 behavioural, territories and selected subnational locations: a systematic analysis environmental and occupational, and metabolic risks or clusters of from the Global Burden of Disease Study 2016. Lancet 2018; risks, 1990–2015: a systematic analysis for the Global Burden of 391: 2236–71. Disease Study 2015. Lancet 2016; 388: 1659–724. 1248 www.thelancet.com Vol 396 October 17, 2020 Global Health Metrics 31 Lopez AD, Adair T. Is the long­ term decline in cardiovascular­ 43 IPCC. Global warming of 1·5°C. An IPCC special report on the disease mortality in high­ income countries over? Evidence from impacts of global warming of 1·5°C above pre­ industrial levels national vital statistics. Int J Epidemiol 2019; 48: 1815–23. and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of 32 Saiz LC, Gorricho J, Garjón J, Celaya MC, Erviti J, Leache L. Blood climate change, sustainable development, and efforts to eradicate pressure targets for the treatment of people with hypertension and poverty. Geneva, Switzerland: World Meteorological cardiovascular disease. Cochrane Database Syst Rev 2018; 7: CD010315. Organization, 2018. 33 He FJ, Li J, MacGregor GA. Effect of longer ­ term modest salt 44 Barreca A, Clay K, Deschenes O, Greenstone M, Shapiro JS. Adapting reduction on blood pressure. Cochrane Database Syst Rev 2013; to climate change: the remarkable decline in the US temperature­ 4: CD004937. mortality relationship over the twentieth century. J Polit Econ 2016; 34 The Lancet. The global syndemic of obesity, undernutrition, 124: 105–59. and climate change: The Lancet Commission report. Lancet 2019; 45 Foreman KJ, Marquez N, Dolgert A, et al. Forecasting life 393: 791–846. expectancy, years of life lost, and all­ cause and cause­ specific 35 Mozaffarian D, Hao T, Rimm EB, Willett WC, Hu FB. Changes in mortality for 250 causes of death: reference and alternative diet and lifestyle and long­ term weight gain in women and men. scenarios for 2016–40 for 195 countries and territories. Lancet 2018; N Engl J Med 2011; 364: 2392–404. 392: 2052–90. 36 Willett W, Rockström J, Loken B, et al. Food in the Anthropocene: 46 Vollset S, Goren E, Yuan C­ W, et al. Fertility, mortality, migration, the EAT–Lancet Commission on healthy diets from sustainable and population scenarios for 195 countries and territories from 2017 food systems. Lancet 2019; 393: 447–92. to 2100: a forecasting analysis for the Global Burden of Disease 37 Adesina JA, Piketh SJ, Qhekwana M, Burger R, Language B, Study. Lancet 2020; 396: 1285–306. Mkhatshwa G. Contrasting indoor and ambient particulate matter 47 Saltelli A, Funtowicz S. The Precautionary Principle: implications concentrations and thermal comfort in coal and non­ coal burning for risk management strategies. Int J Occup Med Environ Health households at South Africa Highveld. Sci Total Environ 2020; 2004; 17: 47–57. 699: 134403. 48 Alonso­ Coello P, Carrasco­ Labra A, Brignardello­ Petersen R, et al. 38 HEI International Scientific Oversight Committee. Section 11. A methodological survey of the analysis, reporting and Development, air pollution expsoure, and population health. In: interpretation of Absolute Risk ReductiOn in systematic revieWs Outdoor air pollution and health in developing countries of Asia: (ARROW): a study protocol. Syst Rev 2013; 2: 113. a comprehensive review. Special Report 18. Boston, MA: Health 49 Nuovo J, Melnikow J, Chang D. Reporting number needed to treat Effects Institute, 2010. https://www.healtheffects.org/system/files/ and absolute risk reduction in randomized controlled trials. JAMA SR18AsianLitReview.pdf (accessed Nov 27, 2019). 2002; 287: 2813–14. 39 Weagle CL, Snider G, Li C, et al. Global sources of fine particulate 50 Millwood IY, Walters RG, Mei XW, et al. Conventional and genetic matter: interpretation of PM chemical composition observed by 2·5 evidence on alcohol and vascular disease aetiology: a prospective SPARTAN using a global chemical transport model. study of 500 000 men and women in China. Lancet 2019; Environ Sci Technol 2018; 52: 11670–81. 393: 1831–42. 40 Correia AW, Pope CA, Dockery DW, Wang Y, Ezzati M, Dominici F. 51 Adam D. The gene­ based hack that is revolutionizing epidemiology. Effect of air pollution control on life expectancy in the United States: Nature 2019; 576: 196–99. an analysis of 545 U.S. counties for the period from 2000 to 2007. Epidemiology 2013; 24: 23–31. 52 Burgess S, Davey Smith G. How humans can contribute to Mendelian randomization analyses. Int J Epidemiol 2019; 48: 661–64. 41 Dedoussi IC, Eastham SD, Monier E, Barrett SRH. Premature mortality related to United States cross­ state air pollution. Nature 53 Burgess S, Timpson NJ, Ebrahim S, Davey Smith G. Mendelian 2020; 578: 261–65. randomization: where are we now and where are we going? Int J Epidemiol 2015; 44: 379–88. 42 Smith, KR, Woodward A, Campbell­ Lendrum D, et al. Human health: impacts, adaptation, and co­ benefits. In: Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Field CB, Barros VR, Dokken DJ, et al (eds)]. Cambridge, UK and New York, NY: Cambridge University Press. 2014. www.thelancet.com Vol 396 October 17, 2020 1249

Journal

The Lancet – Unpaywall

Published: Oct 1, 2020

There are no references for this article.