Sex inequality in disease burden attributable to smoking, secondhand smoke and chewing tobacco: A systematic analysis for the Global Burden of Disease Study 2023

Summary

Background

Despite extraordinary progress in global tobacco control and prevention, smoking, secondhand smoke (SHS) and chewing tobacco continue to profoundly impact public health, affecting males and females differently. Existing studies have not systematically characterized sex disparities in disease burden attributable to these three tobacco risks within a unified framework, nor have they quantified the contributions of demographic and exposure-related drivers to these disparities or projected their future trajectories. We aimed to assess patterns of sex disparity in smoking-, SHS-, and chewing tobacco-attributable disease burden from 1990 to 2023, decompose the drivers of these disparities, and project sex-specific burden to 2050 under alternative scenarios.

Methods

We used data from the Global Burden of Disease Study (GBD) 2023, including sex-specific estimates of disability-adjusted life years (DALYs) and the age-standardized DALY rates (ASDR) attributable to smoking, SHS, and chewing tobacco. Sex disparities were examined across geographic, temporal, and age dimensions. Decomposition analysis was applied to quantify the contributions of population growth, population aging, change in risk exposure, and change in risk-deleted burden to sex-specific DALY changes between 1990 and 2023. Sex ratios were derived from ASDR and their 95% confidence intervals (CI) were estimated using the delta method. Future sex disparities were projected to 2050 under four counterfactual tobacco exposure scenarios.

Findings

Over the past three decades, the tobacco-attributable disease burden among males has consistently exceeded that among females. In 2023, the overall smoking-attributable ASDR remained substantially higher in males than in females (3146.9 [95% uncertainty interval 2591.8–3703.8] vs. 507.8 [361.0–677.1] per 100,000). The all-cause ASDR attributable to SHS was slightly higher in males than in females (540.7 [434.3–672.9] vs. 506.6 [402.8–612.5] per 100,000). For chewing tobacco, the attributable burden was also markedly greater in males (95.3 [61.3–144.1] vs. 47.8 [25.4–78.5] per 100,000). Decomposition analysis showed that population growth and aging were the predominant drivers of increasing absolute DALYs in both sexes, with consistently larger effects among males, while change in risk exposure contributed differently to sex-specific burden trajectories across the three tobacco risks. Among GBD regions, the sex ratio of smoking-attributable ASDR of all causes was notably lowest in Australasia and high-income North America. The sex ratio of SHS-attributable all-cause ASDR increased from below 1.0 in 1990 to above 1.0 in 2023 across several GBD regions like high-income Asia Pacific and Andean Latin America. Globally, the sex ratio of overall ASDR attributable to chewing tobacco has remained stable, from 1990 (ratio: 1.9 95% CI 1.0–4.0) to 2023 (ratio: 2.0 95% CI 1.0–4.0). Among age groups, smoking-attributable DALY rates increased with age, peaking in the 75+ group in 2023, and the burden among males always exceeded that among females in all age brackets. Compared to the sex ratios of DALY rates attributable to smoking risk in the same regions, the fluctuations in disease burden attributable to SHS and chewing tobacco between sexes across age groups were less pronounced in 2023. Projections suggested that sex disparities would persist to 2050, with alternative tobacco control scenarios yielding divergent patterns of future burden.

Interpretation

While smoking- and SHS-attributable ASDRs have declined and chewing tobacco-attributable ASDR has remained stable, sex disparities have varied by time, region, and life stage. These findings may inform the integration of sex-specific perspectives into tobacco control strategies, which could strengthen the implementation of the Framework Convention on Tobacco Control and help mitigate the future burden of tobacco use.

Funding

National Science and Technology Innovation 2030, Noncommunicable Chronic Diseases-National Science and Technology Major Project.

Research in context.

Evidence before this study

We searched PubMed, Scopus, Web of Science, and the Global Health Data Exchange (GHDx) from database inception to October 10, 2025, with no language restrictions, using combinations of the following terms: “smoking”, “secondhand smoke”, “chewing tobacco”, “sex disparity”, “gender disparity”, “disease burden”, “Global Burden of Disease”, “attributable burden”, and “tobacco-attributable”. Previous studies using prior rounds of the Global Burden of Disease Study (GBD) have provided sex-stratified estimates of the disease burden attributable to individual tobacco risk factors, and the broader literature has documented sex differences in tobacco-related health outcomes. However, most existing analyses have treated sex as a stratification variable rather than as a central analytical theme, and have rarely systematically examined the drivers of divergent burden trajectories in males and females or projected how sex disparities may evolve under alternative scenarios. Furthermore, the extent to which sex disparities differ across smoking, secondhand smoke (SHS), and chewing tobacco within a unified framework has not been assessed using the updated GBD 2023.

Added value of this study

Using the latest GBD 2023 data with newly updated risk–outcome pairs for tobacco risks, this study places sex inequality as the central analytical lens, comprehensively integrating smoking, SHS, and chewing tobacco within a unified comparative framework. To quantify the relative magnitude of sex disparities, we derived male-to-female sex ratios of age-standardized DALY rates (ASDR) and estimated their 95% confidence intervals using the delta method. Three specific advances are offered. First, by directly comparing the three tobacco risks, we reveal that they exhibit divergent trajectories of sex disparity: the male-to-female gap is narrowing for smoking, males currently bear a slightly higher SHS-attributable burden, and the gap remains persistently wide for chewing tobacco. Second, we employed decomposition analysis to quantify the contributions of population growth, population aging, change in risk exposure, and change in risk-deleted burden to sex-specific DALY changes between 1990 and 2023, highlighting how the relative influence of these drivers differed between males and females. Third, we projected future sex disparities to 2050 under a multi-scenario forecasting framework, including reference, moderate decline, aggressive decline, and worsening scenarios of tobacco exposure. Together, these analyses position sex inequality as a core investigative theme rather than a supplementary descriptive output.

Implications of all the available evidence

Tobacco-attributable mortality and DALYs remain substantial for both sexes. The distinct sex-disparity patterns across smoking, SHS, and chewing tobacco, together with the differential contributions of population growth, population aging, changes in risk exposure and risk-deleted burden by sex, suggest that uniform tobacco control strategies may be insufficient. These findings underscore the importance of sex-specific tobacco control policies that are tailored to the distinct risk profiles, disease patterns, and burden trajectories observed in males and females across different regions and life stages.

Introduction

Currently, over 8 million individuals worldwide die from tobacco use annually,¹ and the number will continue to rise in the future unless comprehensive policy measures are implemented.² Tobacco consumption harms not only users (through smoking and chewing tobacco) but also those exposed to secondhand smoke (SHS).³ Globally, annual healthcare expenditures attributable to tobacco use exceed one trillion dollars.² Importantly, the health impact of tobacco differs by sex,4, 5, 6 as it acts as a significant genetic modifier of health, contributing to disparities in disease susceptibility through genetic and developmental pathways.⁷ Therefore, identifying these sex-specific disparities in tobacco-related disease burden based on existing evidence is the vital first step toward reducing the global tobacco threat and advancing health equality and well-being.

The prevalence of tobacco exposure among men and women demonstrates considerable disparities. Globally, an estimated 32.7% of men are current smokers, compared to only 6.6% of women.⁴ According to a World Health Organization (WHO) report, the prevalence of SHS exposure among individuals aged 15 and older is 33% for men and 35% for women.⁸ Chewing tobacco also presents a noteworthy gap, with a global prevalence of 6.55% among men versus 2.87% among women.⁶ Beyond differences in prevalence, accumulating research has demonstrated the existence of sex differences in the health outcomes due to tobacco use.^4,9, 10, 11 Even with comparable patterns of use, men and women experience different disease burdens,¹² underscoring the need for tailored, sex-specific tobacco control strategies. Cardiovascular diseases (CVD) serve as a compelling illustration. A prior meta-analysis found that female smokers faced a 25% higher risk of coronary heart disease compared to their male counterparts.¹³ Similarly, a retrospective cohort study on smoking-attributable ST-elevation myocardial infarction (STEMI) risk revealed that smoking significantly elevated the risk of STEMI more in women than in men within the same age group, especially among those aged 50–64 years.¹⁴ Evidence on sex disparity in SHS exposure remains limited, though existing findings point to disadvantages for women, especially during adolescence and reproductive age.^15,16 A decade-long prospective cohort study elucidated the cardiovascular consequences of spousal SHS exposure among never-smoking women. After adjustment for major confounders, husbands’ current smoking status was associated with a 2.02-fold higher risk of ischemic heart disease, an 18% increased risk of stroke, and a 25% higher total CVD risk compared to non-smoking spouses.¹⁷ Current perspectives attribute the elevated health risk of both smoked and smokeless tobacco to the toxic mechanisms of nicotine, tar, and carbon monoxide during consumption. The complex pathophysiological risk mechanisms involve inflammatory responses, endothelial dysfunction, abnormalities in blood flow, thrombosis, and arrhythmias.¹⁸

While the Global Burden of Disease Study (GBD) tobacco collaborators have described the disease burden attributable to smoking, SHS, and chewing tobacco using previous rounds of GBD data,4, 5, 6^,19,20 often with some degree of sex stratification, a dedicated and comprehensive analysis that places sex differences as the core of investigation across all three tobacco risks has been lacking. Developing a clearer picture of these inequalities is essential to guide future restructuring of tobacco control regulations. In this study, we systematically assessed sex-specific differences in disease burden from three major tobacco risks, using data from the GBD 2023 with newly updated tobacco-related outcomes. It is also important to recognize that biological and social attributes of both sexes evolve across life stages and cultural contexts, shaping health outcomes. Thus, with the support of GBD 2023 data, we illustrated the temporal and age-related trends in sex disparities in tobacco-related disease burden, both globally and across regions and 204 countries and territories. We further employed decomposition analysis to quantify the contribution of demographic and exposure-related drivers to DALY changes among males and females, and projected future sex disparities to 2050, aiming to offer complementary insights into sex-specific burden patterns.

Methods

Data source and study population

This study utilized data from the GBD 2023 database (https://vizhub.healthdata.org/gbd-results/), produced by the Institute for Health Metrics and Evaluation (IHME). Detailed descriptions of GBD data sources, estimation methods, and statistical models have been published elsewhere.21, 22, 23 The GBD 2023 provides extensive estimates for 375 diseases and injuries and 88 risk factors across 204 countries and territories from 1990 to 2023. Risk factors are organized into four hierarchical levels, identified as Level 1 through Level 4, with smoking, SHS, and chewing tobacco categorized as Level 3 risk factors under the broader Level 2 risk of tobacco use.

For age groups, we used standard 5-year age groups provided by GBD 2023, along with the aggregated 15–49 and 75+ age groups. In terms of cross-regional comparisons, we presented the differences in smoking-, SHS- and chewing tobacco-attributable burden across seven GBD super-regions and 21 GBD regions.

The study was conducted and reported in accordance with the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER).²⁴

Definition

Smoking was defined as current daily or occasional use of any smoked tobacco product, and SHS pertained to any current exposure to secondhand smoke at home or workplace. Only non-daily smokers are deemed to be exposed to SHS. Chewing tobacco referred to current use of chewing tobacco products (such as betel quid with tobacco) of any frequency. The modeling process for smoking, SHS, and chewing tobacco exposure is detailed in the published GBD studies.²¹ The theoretical minimum-risk exposure level (TMREL) for the three tobacco risks is zero. The population attributable fraction (PAF) referred to the proportion of risk that would be decreased if the exposure was reduced to TMREL. Disability-adjusted life years (DALYs) serve as a comprehensive metric for assessing the overall burden of disease attributable to smoking, SHS, and chewing tobacco exposure. DALYs are the sum of the years of life lost (YLLs) due to premature mortality and the years lived with disability (YLDs) arising from health conditions related to tobacco exposure.

The Socio-demographic Index (SDI) is a composite indicator of background social and economic conditions that influence health outcomes in each location.^21,23 It is calculated as the geometric mean of indices of total fertility rate for those younger than 25 years, mean education for those aged 15 years and older, and lag-distributed income per capita, rescaled to a range of 0–100. SDI data for regions and countries and territories are publicly available online (https://ghdx.healthdata.org/record/gbd-2023-socio-demographic-index-sdi).

Sex serves as a genetic modifier of health, driving differences in disease susceptibility during genetic and developmental processes, whereas gender constitutes a framework of social and behavioral norms that influences cognitive and behavioral tendencies, thereby affecting exposure to health risks. Our study adhered to the sex binary stratification within the GBD framework, and utilized sex-specific death counts, DALYs, age-standardized DALY rates (ASDR) estimated by GBD to investigate the burden attributable to smoking, SHS, and chewing tobacco among males and females. The age-standardized rate (ASR) was calculated as $ASR=\frac{{\sum }_{i=1}^N{a}_i{w}_i}{{\sum }_{i=1}^N{w}_i}$, where a_i is the age-specific rate for the ith age group and w_i is the weight assigned to the ith age group based on the GBD world population age standard.

Statistical analysis

We extracted the latest dataset on the attributable burden of smoking, SHS, and chewing tobacco from the GBD results tool and analyzed sex differences in disease burden attributable to these three tobacco risk factors. All analyzed risk–outcome pairs are listed in Appendix pp 15–20. Firstly, to explore sex differences in health outcomes, we stratified the datasets from GBD 2023 by sex, allowing for nuanced insights into DALY patterns. National and regional disparities were examined by visualizing ASDRs for nine smoking- and SHS-attributable overlapping causes (ischemic heart disease [IHD], ischemic stroke, intracerebral hemorrhage, subarachnoid hemorrhage, chronic obstructive pulmonary disease [COPD], asthma, lower respiratory infections [LRI], tracheal, bronchus, and lung cancer [TBL], and diabetes mellitus type 2 [T2DM]) and eight chewing tobacco-attributable causes (lip and oral cavity cancer, nasopharynx cancer, other pharynx cancer, esophageal cancer, larynx cancer, ischemic stroke, intracerebral hemorrhage, and subarachnoid hemorrhage) between males and females across 204 countries and territories in 2023. For all estimates, we presented the 95% uncertainty intervals (95% UI) derived from the 2.5th and 97.5th percentiles of 250 draws.²²

Secondly, to disentangle the driving forces underlying the observed sex-specific changes in tobacco-attributable disease burden globally, we applied the decomposition method proposed by Das Gupta.²⁵ This approach constructs counterfactual scenarios in which the effect of one driver is evaluated while holding all other drivers constant. We quantified the contributions of four drivers (population growth, population aging, change in risk exposure, and change in risk-deleted burden) to the sex-specific changes in DALYs attributable to smoking, SHS, and chewing tobacco between 1990 and 2023. Risk-deleted burden refers to changes in DALYs attributable to all other factors apart from the specific risk factor, population growth, and population aging.^21,26 Population growth was determined by the overall increase in total population from 1990 to 2023. Population aging was measured by the change in the proportion of a specific age group relative to the total population over the same period. Change in risk exposure was represented by the change in the ratio PAF/(1−PAF). Risk-deleted burden was calculated as cause-specific DALYs × (1−PAF).²⁶

Thirdly, to quantify the relative sex disparities, we derived sex ratios using the ASDR. The point estimate for the ratio was obtained by dividing the male ASDR by the female ASDR. To evaluate the uncertainty of the ratio, we applied the delta method on the logarithmic scale to estimate its 95% confidence interval (CI).²⁷ This approach uses a first-order Taylor series expansion to approximate the variance of the ratio of two independent random variables. Specifically, for the ratio ${R=\hat{u}}_m/{\hat{\mu }}_f$, the variance was estimated as: $Var\left(R\right)\approx {\left(\frac{1}{{\hat{u}}_f}\right)}^2{\sigma }_m^2+{\left(-\frac{{\hat{\mu }}_m}{{\hat{\mu }}_f}\right)}^2{\sigma }_f^2$. Where ${\hat{\mu }}_m$ and ${\hat{u}}_f$ denote the point estimates of male and female ASDR, and ${\sigma }_m^2$ and ${\sigma }_f^2$ are their respective standard errors. Standard errors were derived from the 95% UIs reported by the GBD study, assuming approximate normality, such that the standard error was calculated as: $\sigma =\frac{\text{upper}\text{bound}-\text{lower}\text{bound}}{2\times 1.96}$. A 95% CI for R was then constructed as $R\pm 1.96\times \sqrt{Var\left(R\right)}$.

Using the sex ratio estimates, sex disparities in ASDRs were further quantified at the global level and across 21 GBD regions for 1990 and 2023. To explore the potential socioeconomic patterning of sex differences, we examined the correlation between sex-specific disease burden and the SDI across 21 geographic regions.

Fourthly, temporal patterns of sex differences were assessed through comparative analysis of sex-stratified burden estimates from 1990 to 2023. In addition, we projected the disease burden from 2024 to 2050 following the GBD forecasting framework's distinction between a reference scenario and alternative scenarios.²⁸ The former represents the most likely future based on historical trends, and the latter are constructed by replacing reference risk factor trajectories with hypothetical trajectories. In this study, the reference scenario burden projections were generated using the Bayesian age-period-cohort (BAPC) model, which integrated historical sex-stratified tobacco-attributable DALY rates and their posterior probability distributions, systematically disentangling age, period, and cohort effects through hierarchical parameter estimation.^29,30 Concurrently, to assess the potential impact of different tobacco control intensities, we constructed additional counterfactual scenarios based on sex-specific summary exposure values (SEVs) for smoking, SHS, and chewing tobacco. The SEV is a GBD metric representing risk-weighted exposure prevalence, ranging from 0 (no excess risk) to 100 (maximum risk), and serves as a standardized measure of population exposure to a given risk factor.²¹ For the reference scenario, historical SEV trends were extrapolated using an autoregressive integrated moving average (ARIMA) model. The moderate and aggressive decline scenarios assumed additional annual relative reductions in SEV of 3% and 5%, respectively, while the worsening scenario assumed that SEV remained constant at the 2023 level.

To translate SEV changes into attributable burden, we estimated cause- and sex-specific elasticities (k) using ordinary least-squares regression of log-transformed PAF on log-transformed SEV with historical data (1990–2023) as: ln(PAF) = α + k·ln(SEV). The coefficient $k$ quantifies the percentage change in PAF associated with a 1% change in SEV. For sex-specific diseases (e.g., breast cancer, prostate cancer), the elasticity estimation and scenario projection were performed only for the sex with available data. To avoid implausible results, whenever the estimated coefficient $k$ was ≤0, we conservatively set $k$ = 1, corresponding to a proportional change assumption. The attributable ASDR under each counterfactual scenario was then calculated as: $\mathit{ASD}{R}_{\mathit{scenario}}=\mathit{ASD}{R}_{\text{reference}}\times {\left(\frac{\mathit{SE}{V}_{\text{scenario}}}{\mathit{SE}{V}_{\text{reference}}}\right)}^k$. Uncertainty intervals for counterfactual ASDR were propagated by applying the same scenario-specific scaling factor ${\left(\frac{\mathit{SE}{V}_{\text{scenario}}}{\mathit{SE}{V}_{\text{reference}}}\right)}^k$ to the upper and lower bounds of the reference ASDR, preserving the relative uncertainty structure of the reference estimates. Additional methodological details of model selection, scenario construction and SEV translation are provided in the Appendix pp 21–23.

Finally, age-specific patterns were evaluated by comparing DALY rates across age groups in 2023, highlighting persistent sex disparities in tobacco-attributable burden from youth through older adulthood. The absence of overlap between the 95% UIs of the male and female ASDRs was considered indicative of a statistically significant difference. All statistical analyses were performed using R (version 4.4.1).

Ethics

The ethical approval and informed consent were not required for this study because the study involved only secondary analyses of publicly available, de-identified data from the GBD 2023.

Role of the funding source

The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Results

The overall burden among males and females

In 2023, there were an estimated 203.9 million (95% UI 171.4–236.6) tobacco-attributable DALYs and 7.4 million (6.4–8.5) deaths worldwide. Among three types of tobacco risk, smoking accounted for the largest burden, resulting in 161.7 million (131.8–192.5) DALYs and 5.8 million (4.8–7.0) deaths. SHS was responsible for 44.9 million (35.9–54.4) DALYs and 1.7 million (1.3–2.0) deaths. In contrast, chewing tobacco contributed 6.5 million (4.0–10.0) DALYs and 0.3 million (0.1–0.4) deaths (Table 1, Appendix pp 24–37). Over the past 30 years, tobacco-related deaths have increased among both sexes. However, there was a notable decline in DALYs attributable to SHS (Fig. 1).

Table 1.

DALYs, age-standardized DALY rates (per 100,000 population) in 2023, and percentage change in age-standardized DALY rates attributable to tobacco and three types of tobacco risk (smoking, secondhand smoke, and chewing tobacco) globally and regionally from 1990 to 2023.

	2023								1990–2023
	Tobacco		Smoking		Secondhand smoke		Chewing tobacco		Tobacco	Smoking	Secondhand smoke	Chewing tobacco
	DALYs (thousands)	ASDR (per 100,000)	DALYs (thousands)	ASDR (per 100,000)	DALYs (thousands)	ASDR (per 100,000)	DALYs (thousands)	ASDR (per 100,000)	Percentage change of ASDR (%)	Percentage change of ASDR (%)	Percentage change of ASDR (%)	Percentage change of ASDR (%)
Global	203921.4 (171360.6–236562.2)	2257.7 (1897.3–2612.0)	161709.6 (131789.1–192487.3)	1764.2 (1436.7–2105.0)	44878.8 (35856.3–54382.6)	522.2 (416.9–633.9)	6483.6 (3961.8–9929.2)	71.0 (43.4–108.5)	−48.3 (−53.6 to −42.8)	−47.7 (−54.4 to −40.7)	−53.1 (−60.9 to −44.8)	−2.0 (−44.8 to 71.3)
Sex
Males	155897.6 (131797.6–181238.4)	3607.8 (3055.2–4195.4)	137251.4 (113050.9–161438.9)	3146.9 (2591.8–3703.8)	22086.3 (17706.2–27323.6)	540.7 (434.3–672.9)	4183.0 (2672.8–6333.3)	95.3 (61.3–144.1)	−48.0 (−53.4 to −42.5)	−48.1 (−53.4 to −41.9)	−51.7 (−60.3 to −41.8)	−1.8 (−40.7 to 70.2)
Females	48023.7 (38961.1–55952.0)	1030.7 (836.7–1189.3)	24458.1 (17427.0–32615.9)	507.8 (361.0–677.1)	22792.4 (18018.8–27406.1)	506.6 (402.8–612.5)	2300.6 (1220.5–3791.2)	47.8 (25.4–78.5)	−51.0 (−57.4 to −44.0)	−50.2 (−62.0 to −35.4)	−54.3 (−61.3 to −46.9)	−4.4 (−50.5 to 71.9)
GBD super-regions
Central Europe, Eastern Europe, and Central Asia	18275.4 (15509.3–20779.6)	2894.3 (2453.6–3285.5)	15896.1 (13216.0–18807.0)	2465.5 (2039.5–2929.5)	3041.3 (2380.9–3778.2)	531.7 (416.2–662.2)	31.9 (10.6–71.0)	5.1 (1.7–11.3)	−40.0 (−46.5 to −33.8)	−36.7 (−44.9 to −29.0)	−52.6 (−60.7 to −42.7)	−27.2 (−73.8 to 120.4)
High-income	30392.6 (24404.8–36961.0)	1565.5 (1232.5–1895.8)	27900.4 (21714.2–34530.7)	1431.2 (1099.4–1767.6)	3400.5 (2681.2–4167.4)	182.7 (144.1–219.4)	116.0 (72.4–199.6)	5.7 (3.5–10.0)	−51.8 (−55.9 to −48.0)	−51.2 (−55.7 to −46.8)	−60.3 (−66.8 to −52.7)	−37.1 (−66.7 to 21.5)
Latin America and Caribbean	8579.5 (6782.7–10181.4)	1300.1 (1029.9–1543.2)	6790.2 (5119.7–8428.7)	1014.9 (762.6–1263.1)	2059.5 (1626.3–2517.9)	325.5 (256.9–398.3)	27.5 (9.0–61.1)	4.2 (1.4–9.3)	−58.0 (−63.4 to −52.3)	−58.0 (−64.4 to −50.8)	−60.8 (−66.8 to −53.6)	−41.7 (−79.3 to 84.3)
North Africa and Middle East	13109.9 (11122.9–15409.6)	2464.0 (2091.0–2900.1)	10078.2 (8044.1–12632.6)	1852.0 (1466.3–2322.3)	3612.8 (2900.0–4379.4)	719.3 (579.5–871.5)	60.9 (31.0–115.1)	12.5 (6.4–23.3)	−37.8 (−48.8 to −26.6)	−34.6 (−50.4 to −17.1)	−44.1 (−53.5 to −33.4)	−60.5 (−79.2 to −16.7)
South Asia	42468.6 (35557.0–50849.2)	2720.8 (2294.8–3251.6)	27109.3 (20807.2–34572.4)	1730.3 (1327.7–2196.9)	12495.2 (9954.6–15524.4)	814.1 (649.9–1009.1)	5434.6 (3421.8–8101.8)	342.6 (212.0–515.4)	−45.3 (−53.4 to −34.0)	−48.7 (−58.5 to −33.7)	−46.0 (−55.5 to −36.5)	−8.7 (−46.6 to 55.9)
Southeast Asia, East Asia, and Oceania	81802.5 (70288.0–95059.0)	2719.5 (2338.2–3146.2)	68496.9 (57061.6–81545.5)	2221.5 (1854.3–2649.5)	16367.7 (13036.1–19793.7)	597.2 (476.1–726.1)	669.0 (266.0–1361.9)	22.2 (8.8–44.9)	−51.7 (−57.9 to −45.1)	−49.1 (−57.1 to −40.2)	−59.4 (−66.0 to −52.7)	−42.9 (−76.5 to 54.3)
Sub-Saharan Africa	9292.8 (7359.1–11560.6)	1260.9 (984.0–1597.4)	5438.4 (3833.4–7769.5)	842.4 (595.9–1187.6)	3901.6 (2851.6–5128.1)	425.1 (322.7–543.0)	143.7 (60.0–298.1)	24.1 (9.6–50.1)	−41.0 (−56.2 to −21.9)	−41.7 (−62.0 to −14.1)	−40.8 (−54.4 to −25.7)	−17.7 (−67.8 to 109.0)
World bank regions
East Asia & Pacific	86669.4 (73872.3–100732.7)	2524.9 (2158.2–2933.1)	72992.6 (60738.3–86642.0)	2079.0 (1730.2–2468.7)	16926.1 (13464.9–20480.4)	537.3 (428.1–651.5)	641.6 (252.0–1322.0)	18.7 (7.4–38.3)	−51.2 (−57.1 to −45.0)	−48.2 (−55.6 to −39.8)	−60.2 (−66.6 to −53.4)	−41.1 (−76.1 to 60.7)
Europe & Central Asia	33007.0 (27497.9–38398.1)	2202.8 (1831.3–2565.3)	29362.7 (23876.9–35271.6)	1936.7 (1573.1–2338.4)	4843.4 (3919.0–5907.7)	345.9 (279.0–421.4)	46.4 (15.4–105.4)	3.0 (1.0–6.8)	−46.8 (−51.5 to −42.1)	−44.6 (−50.5 to −39.1)	−57.9 (−64.9 to −49.8)	−32.3 (−75.8 to 106.9)
Latin America & Caribbean	9931.4 (7849.2–11826.9)	1328.3 (1051.6–1581.1)	7915.1 (5940.5–9946.2)	1044.8 (781.9–1315.2)	2348.8 (1855.4–2867.5)	327.3 (258.7–400.3)	28.4 (9.3–63.0)	3.8 (1.2–8.5)	−57.3 (−63.0 to −51.1)	−57.1 (−64.3 to −49.5)	−60.4 (−66.5 to −53.0)	−40.8 (−79.0 to 87.2)
Middle East & North Africa	9874.5 (8295.5–11648.0)	2429.5 (2035.5–2870.8)	7551.0 (5980.0–9443.8)	1817.4 (1427.9–2301.3)	2750.2 (2202.4–3364.4)	714.0 (573.3–876.5)	47.8 (25.1–85.0)	13.4 (7.0–23.6)	−33.9 (−47.2 to −20.5)	−30.5 (−48.1 to −8.6)	−40.3 (−50.4 to −28.4)	−61.2 (−78.5 to −22.9)
North America	11220.0 (8964.6–13678.0)	1835.3 (1439.1–2245.6)	10408.7 (8175.6–12921.0)	1689.4 (1299.8–2110.8)	1083.6 (854.2–1317.0)	191.9 (152.0–230.5)	85.1 (52.1–135.4)	13.8 (8.4–22.2)	−49.2 (−54.5 to −43.3)	−48.8 (−54.7 to −42.2)	−60.0 (−65.7 to −52.8)	−40.4 (−64.7 to 1.9)
South Asia	43265.8 (36237.6–51776.2)	2693.5 (2270.9–3213.4)	27576.3 (21182.4–35124.3)	1711.9 (1314.3–2174.3)	12810.9 (10215.4–15890.0)	808.9 (643.6–1002.5)	5482.0 (3448.0–8191.9)	336.3 (208.1–507.0)	−45.2 (−53.2 to −34.0)	−48.5 (−58.4 to −33.7)	−45.9 (−55.3 to −36.3)	−8.6 (−46.5 to 56.0)
Sub-Saharan Africa	9762.6 (7737.3–12115.3)	1284.5 (1003.4–1618.9)	5747.1 (4046.6–8183.8)	855.2 (602.0–1200.2)	4073.9 (2997.1–5348.0)	437.3 (332.0–556.6)	151.2 (64.1–313.6)	24.3 (9.8–50.8)	−41.0 (−56.4 to −22.2)	−41.4 (−61.9 to −13.6)	−41.2 (−54.6 to −26.4)	−20.3 (−69.0 to 107.2)
World Bank income regions
World Bank Low Income	6753.7 (5248.8–8248.5)	1635.6 (1265.3–2037.9)	4050.6 (2796.2–5698.8)	1087.3 (751.4–1537.0)	2755.3 (2076.2–3598.9)	563.1 (426.9–725.4)	132.8 (69.9–252.9)	37.1 (19.8–69.2)	−34.3 (−52.0 to −13.3)	−33.0 (−58.1 to −1.8)	−36.4 (−50.7 to −20.1)	−32.5 (−66.2 to 44.4)
World Bank Lower Middle Income	62347.8 (52353.3–73013.8)	2574.3 (2170.1–3012.8)	41266.1 (32802.2–51735.8)	1710.0 (1351.3–2154.9)	18646.3 (15027.3–23114.2)	764.5 (610.5–944.2)	5816.4 (3609.9–8801.9)	243.2 (148.3–369.2)	−43.7 (−52.0 to −33.5)	−45.8 (−55.7 to −31.5)	−44.9 (−54.2 to −36.1)	−9.9 (−47.5 to 54.2)
World Bank Upper Middle Income	91582.1 (77567.3–107096.1)	2413.8 (2046.9–2810.2)	77069.0 (63682.4–91261.3)	1992.6 (1647.6–2365.9)	18195.3 (14476.1–22053.0)	516.5 (409.7–625.9)	395.7 (133.1–868.8)	10.4 (3.5–22.9)	−52.0 (−57.8 to −45.4)	−49.3 (−56.4 to −40.9)	−60.3 (−66.9 to −53.0)	−50.2 (−81.4 to 51.1)
World Bank High Income	43047.1 (35145.7–51246.4)	1796.7 (1460.1–2147.8)	39167.6 (31262.5–47732.6)	1629.4 (1294.5–1982.7)	5240.0 (4135.7–6472.3)	225.5 (179.0–273.6)	137.5 (82.3–247.9)	5.5 (3.2–10.0)	−49.9 (−53.8 to −45.9)	−48.8 (−53.2 to −44.3)	−59.8 (−67.2 to −52.0)	−35.9 (−68.1 to 34.6)

Data in parentheses are 95% uncertainty intervals. DALYs, disability-adjusted life years; ASDR, age-standardized DALY rate.

Fig. 1.

Temporal trends of death and DALY counts, and age-standardized rates attributable to tobacco use (A), smoking (B), secondhand smoke (C), and chewing tobacco (D) among males and females from 1990 to 2023, by year distribution. The bold lines are sex-specific estimates of age-standardized rates of mortality and DALYs, and the bars represent sex-specific estimates of deaths and DALYs, with error bars indicating their 95% uncertainty intervals. DALYs, disability-adjusted life years; ASMR, age-standardized mortality rate; ASDR, age-standardized DALY rate.

In 2023, the all-cause ASDR attributable to smoking for females (507.8 [361.0–677.1] per 100,000) was considerably lower than that for males (3146.9 [2591.8–3703.8] per 100,000), indicating a significant disparity (Table 1, Fig. 1B). For World Bank income regions, the ASDR in the Upper Middle Income region was 1992.6 (1647.6–2365.9) per 100,000, surpassing that of other regions. Among super-regions, Central Europe, Eastern Europe, and Central Asia regions exhibited a significantly higher ASDR of 2465.5 (2039.5–2929.5) per 100,000. This represented a stark contrast to the Sub-Saharan Africa region, which reported the ASDR of 842.4 (595.9–1187.6) per 100,000. Across 204 countries and territories, males generally exhibited higher smoking-attributable burden compared to females (Fig. 2A and B, Appendix pp 68–76). The most significant sex disparity in ASDR was observed in Azerbaijan, revealing a staggering difference of over 50 times, with males at 4749.5 (3783.8–5607.9) and females at 91.3 (36.9–204.2) per 100,000 population.

Fig. 2.

National age-standardized DALY rates (per 100,000 population) of overall causes among males and females attributable to smoking (A for males and B for females), secondhand smoke (C for males and D for females), and chewing tobacco (E for males and F for females) in 2023. DALYs, disability-adjusted life year; ASDR, age-standardized DALY rate.

For SHS, global attributable all-cause ASDRs were 540.7 (434.3–672.9) per 100,000 in males and 506.6 (402.8–612.5) per 100,000 in females in 2023 (Table 1). Regions with the highest DALY rates per 100,000 population were South Asia region (814.1 [649.9–1009.1]) across GBD super-regions and Lower Middle Income region (764.5 [610.5–944.2]) across word bank income regions. Across all countries and territories, SHS-attributable ASDR showed a relatively equal trend between sexes (Fig. 2C and D, Appendix pp 77–85). Nauru had the highest male and female ASDR values, at 3748.8 (2881.4–4669.1) and 3157.3 (2385.1–3924.0) per 100,000 population. Barbados presented a more balanced scenario in this regard, exhibiting a relatively equal burden, along with San Marino and Sweden (Fig. 2C and D).

The ASDR attributable to chewing tobacco for both males and females has remained stable, from 97.1 (50.3–161.6) for males and 50.0 (25.6–92.1) for females in 1990 to 95.3 (61.3–144.1) for males and 47.8 (25.4–78.5) for females per 100,000 population in 2023 (Table 1, Fig. 1D). South Asia had the highest ASDR among all super-regions, reaching 342.6 (212.0–515.4) per 100,000 population. The overall ASDR of Bangladesh attributable to chewing tobacco ranked highest among 204 countries and territories for both males (548.2 per 100,000) and females (771.8 per 100,000 population) (Fig. 2E and F). The cause-specific burden attributable to chewing tobacco also showed elevated levels in several South and Southeast Asian countries, including Bangladesh, Cambodia and the Lao People's Democratic Republic, for both males and females (Appendix pp 86–93).

Drivers of sex-specific DALY change

Decomposition of the absolute DALY change attributable to smoking, SHS, and chewing tobacco between 1990 and 2023 revealed distinct sex-specific patterns in the driving forces (Fig. 3, Appendix pp 94–95). Over the past three decades, the four drivers exhibited broadly similar directional effects on the change in smoking-attributable all-cause DALYs for both sexes: population growth and population aging contributed positively to the increase in burden, whereas changes in smoking exposure and smoking-deleted burden contributed to reductions in burden. However, the absolute magnitude of DALY change associated with each driver was consistently larger among males than among females. This pattern was also evident for most cause-specific DALYs attributable to smoking (Fig. 3A, Appendix p 94). Nonetheless, the net direction of DALY change differed by sex for certain diseases. For IHD and ischemic stroke, smoking-attributable DALYs increased by 6643.2 and 728.4 thousand among males but decreased by 703.2 and 266.2 thousand among females, whereas the opposite was observed for asthma. Of note, change in smoking-deleted burden contributed positively to DALY change for T2DM in both sexes; the reduction in DALYs driven by declining smoking exposure was insufficient to offset the positive contributions of the other three drivers (Fig. 3A).

Fig. 3.

Change in DALYs of the overlapping causes attributable to smoking (A), secondhand smoke (B), and causes attributable to chewing tobacco (C) due to population growth, population aging, change in risk exposure, and change in risk-deleted burden among males and females, globally from 1990 to 2023. The black dots represent the net DALY change attributable to the four drivers combined among males and females. DALYs, disability-adjusted life years.

Between 1990 and 2023, the reductions in SHS-attributable all-cause DALYs driven by changes in SHS exposure and SHS-deleted burden outweighed the increases driven by population growth and population aging, resulting in an overall net decrease of 4070.4 thousand DALYs for males and 4358.7 thousand DALYs for females (Fig. 3B). For IHD and the three stroke subtypes, the DALY reductions attributable to changes in SHS exposure and SHS-deleted burden were larger in females than in males. For SHS-attributable COPD, the negative contribution of change in SHS-deleted burden was substantially larger in males than in females, leading to a smaller net increase in DALYs among males (915.1 thousand) compared with females (1681.3 thousand) over the study period. For T2DM, consistent with the smoking-attributable decomposition, change in SHS-deleted burden again contributed positively to DALY change among both sexes (Fig. 3B).

Over the study period, the four drivers collectively exerted a positive influence on chewing tobacco-attributable all-cause and cause-specific DALYs in both males and females. The sex-specific differences in the magnitude of these contributions resulted in a larger net increase in DALYs among males than among females (Fig. 3C). Specifically, change in chewing tobacco exposure contributed to increasing DALYs for all assessed causes among males, and for all causes except subarachnoid hemorrhage among females. For other pharynx cancer, changes in chewing tobacco exposure and chewing tobacco-deleted burden, together with population growth and aging, all contributed to increasing DALYs in both sexes. Notably, change in chewing tobacco-deleted burden contributed positively to DALY change in both sexes for all cause-specific outcomes except subarachnoid hemorrhage (Fig. 3C).

The burden patterns by region

Globally, the sex ratio of smoking-attributable ASDRs increased from 1990 to 2023 (Fig. 4A). It is noteworthy that in various disease contexts, the male-to-female ratios in Australasia and high-income North America were often the smallest among the regions examined (Fig. 4A, Appendix pp 38–45). In Eastern Europe, the sex ratio of males to females decreased in 2023 compared to 1990 across all overlapping diseases. By contrast, in Southeast Asia, only the ratio of asthma decreased, while the male burden of other overlapping diseases further increased relatively (Fig. 4A, Appendix pp 38–45).

Fig. 4.

Change in the sex ratio of age-standardized DALY rates (per 100,000 population) for the overlapping diseases attributable to smoking (A), secondhand smoke (B), and chewing tobacco (C) between 1990 and 2023, globally and among GBD regions. The sex ratio is defined as the male-to-female ratio of cause-specific ASDR for a given risk and region. When the sex ratio is greater than 1 (i.e., male-to-female ratio >1), the point estimate of male ASDR exceeds that of females, shown in blue; conversely, when the sex ratio is less than 1 (i.e., female-to-male ratio >1), the point estimate of female ASDR exceeds that of males, shown in red. Lighter shades represent the ratios in 1990, while darker shades represent the ratios in 2023. DALYs, disability-adjusted life years; ASDR, age-standardized DALY rate.

The overall sex inequality in SHS-attributable ASDR underwent a shift between 1990 and 2023 across several GBD regions (high-income Asia Pacific, Andean Latin America, Caribbean, Central Latin America, Eastern Sub-Saharan Africa, and Western Sub-Saharan Africa), with the male-to-female ratio rising from below 1 to above 1, which may indicate a transition from a female-predominant to a male-predominant burden (Fig. 4B). For example, in high-income Asia Pacific, the male-to-female ratio of overall ASDR was 0.9 (95% CI 0.6–1.3) in 1990, and by 2023, this ratio increased to 1.7 (95% CI 1.2–2.5), indicating that males bore a relatively heavier SHS-related burden (Fig. 4B, Appendix pp 46–53). Globally, the male-to-female ratio for four CVDs and T2DM increased in 2023 compared to 1990. Conversely, the burden related to COPD, TBL, and asthma showed a decreasing sex ratio globally and across more than ten GBD regions (Fig. 4B, Appendix pp 46–53). Among 21 regions, high-income Asia Pacific region simultaneously observed an increase in the male-to-female ratio of ASDR to varying degrees across all overlapping diseases.

The sex ratio of overall ASDR attributable to chewing tobacco globally remained stable, from 1990 (ratio: 1.9 95% CI 0.8–4.6) to 2023 (ratio: 2.0 95% CI 1.0–4.0), and high-income North America showed the highest ratio among the 21 regions. This situation was observed in all specific diseases attributable to chewing tobacco (Fig. 4C, Appendix pp 54–60).

The disease burden experienced by different sexes may exhibit significant disparities in relation to regional economic development levels (Appendix pp 96–124). By SDI, smoking- and SHS-attributable ASDRs displayed two peaks, whereas chewing tobacco peaked only around an SDI of 0.4 (Appendix p 96 and p 106). The impact of smoking and SHS on COPD, asthma, and T2DM was most pronounced for both males and females at an SDI of 0.45 (Appendix pp 101, 102, 105, 111, 112 and 115). In contrast, the ASDR for TBL attributable to smoking and SHS reached its peak for both sexes at a higher SDI of approximately 0.75, with males experiencing significantly higher rates than females (Appendix p 104 and p 114). Eight diseases attributable to chewing tobacco exhibited a similar relationship between ASDR and SDI. At an SDI of 0.4, the ASDR for both sexes surpassed other levels, clearly linked to the persistent high burden of chewing tobacco in South Asia from 1990 to 2023 (Appendix pp 117–124).

The burden patterns by time

To characterize the temporal dimension of sex differences in the tobacco-attributable disease burden, sex-stratified ASDRs were examined over the period 1990–2023 and projected forward to 2050 (Fig. 5, Fig. 6, Appendix pp 125–163). In addition, four counterfactual scenarios were constructed based on alternative future trajectories of the SEV (Appendix pp 164–233), allowing exploration of how the burden might evolve under varying intensities of tobacco control.

Fig. 5.

Temporal patterns of global age-standardized DALY rates (per 100,000 population) of the overlapping diseases attributable to smoking (A), secondhand smoke (B), and causes attributable to chewing tobacco (C) among males and females, 1990–2023. DALYs, disability-adjusted life years; ASDR, age-standardized DALY rate.

Fig. 6.

Global projections of age-standardized DALY rates (per 100,000 population) and 95% CIs of the overlapping causes attributable to smoking (A), secondhand smoke (B), and causes attributable to chewing tobacco (C) among males and females, 2024–2050. To maintain consistency, the estimates for 1990–2023 are based on GBD 2023 observed values, while the values for 2024–2050 are projected estimates with 95% CIs. DALYs, disability-adjusted life years; ASDR, age-standardized DALY rate.

Globally, from 1990 to 2023, most smoking-attributable causes showed declining ASDRs, except T2DM, extensively drug-resistant tuberculosis and multidrug-resistant tuberculosis without extensive drug resistance, where burdens rose. This trend led to a gradual reduction in the sex disparity of ASDR for smoking-related diseases (Fig. 5A, Appendix p 125). For instance, the sex difference in ASDR for IHD decreased from 1111.7 per 100,000 in 1990 to 671.4 per 100,000 in 2023. However, a contrasting trend was observed in T2DM, where the ASDR for males increased robustly, while females experienced a slow decline. This divergence expanded the sex disparity in the disease burden of T2DM in 2023 (Males: 118.9 [87.4–154.5] per 100,000; Females: 19.9 [13.5–26.9] per 100,000) (Fig. 5A). Projections indicated that the overall smoking-attributable ASDR would continue to decline for both sexes through 2050, with the male rate (2802.4 [95% CI 2156.1–3448.7] per 100,000) remaining substantially higher than the female rate (391.3 [95% CI 304.5–478.2] per 100,000). This sex pattern was consistent across the majority of smoking-attributable causes. However, for ischemic stroke, asthma, and T2DM, the attributable ASDR was projected to increase in males while declining in females, thereby widening sex disparities (Fig. 6A). Under the moderate and aggressive decline scenarios, the smoking-attributable ASDR for these three diseases reversed to a declining trend in males. Under the worsening scenario, the attributable burden for multiple diseases increased in both sexes from 2024 to 2050, with TBL being a notable exception (Appendix p 166, p 170 and p 172).

Regarding SHS-attributable ASDR, from 2000 to 2023, the overall ASDR for males consistently exceeded that of females, with the largest disparity occurring in 2018 (males: 576.4 [460.3–708.6] per 100,000; females: 539.7 [421.4–659.7] per 100,000). Similarly, the ASDR related to IHD over the past thirty years also showed a pronounced burden among males (Fig. 5B). Moreover, the difference in ASDR for asthma between females and males increased annually. By 2023, the ASDR for females reached 20.9 (13.4–31.4) per 100,000 population, while for males it was 16.4 (10.3–23.7) per 100,000 population. For other conditions globally, such as ischemic stroke, intracerebral hemorrhage, and subarachnoid hemorrhage, the sex difference in ASDR has diminished to notably low levels globally (Fig. 5B). Similar to smoking risks, an upward trend in ASDR for T2DM attributable to SHS was observed among males while a slight decline was observed among females (Fig. 5B). Projections indicated a continued decline in SHS-attributable all-cause ASDR for both sexes, with rate of males (517.3 [95% CI 393.5–641.1] per 100,000) remaining slightly higher than that of females (500.9 [95% CI 367.0–634.9] per 100,000) in 2050. For TBL, the male attributable ASDR was projected to consistently exceed that of females, reaching 34.7 (95% CI 27.8–41.6) and 20.9 (95% CI 16.5–25.2) per 100,000 in 2050, respectively. The SHS-attributable ASDR for T2DM was projected to continue increasing in both sexes, with a higher burden among females (Fig. 6B). Under the aggressive decline scenario, the ASDRs for SHS-attributable IHD, ischemic stroke, intracerebral hemorrhage, subarachnoid hemorrhage, COPD, TBL, and T2DM declined in both sexes, but males became the more heavily burdened group by 2050 (Appendix pp 214–218 and 221–222).

From 1990 to 2023, the overall ASDR attributable to chewing tobacco remained stable, with males consistently experiencing higher rates than females (Fig. 5C). However, throughout the observation period, the chewing tobacco-attributable burden of esophageal cancer exhibited an upward trend among males, with ASDR increasing from 7.6 (3.7–13.1) in 1990 to 10.8 (6.5–17.9) per 100,000 population in 2023, and among females, rising from 3.3 (1.7–5.7) to 5.6 (2.8–10.0) per 100,000 population (Fig. 5C). In contrast, for intracerebral hemorrhage, a simultaneous decline in burden was observed for both sexes (Fig. 5C). Projections for chewing tobacco-attributable disease burden indicated that the sex-specific ASDRs for all causes combined, other pharynx cancer, larynx cancer, and ischemic stroke would remain stable, whereas concurrent declining trends were observed for nasopharynx cancer, intracerebral hemorrhage, and subarachnoid hemorrhage in both sexes, with male burden consistently higher than female burden (Fig. 6C). Under the moderate and aggressive decline scenarios, the ASDRs for all chewing tobacco-attributable diseases decreased rapidly, with male rates persistently exceeding female rates (Appendix pp 225–233).

The burden patterns by age

Age-specific patterns of sex disparities in tobacco-attributable disease burden were examined and shown in Fig. 7 and Appendix pp 234–270. Overall, smoking-attributable DALY rates increased with age, peaking in the 75+ group in 2023 and the burden among males always exceeded that among females in all age brackets (Fig. 7A, Appendix pp 61–62 and 234). Notably, with the exception of the Southeast Asia, East Asia, and Oceania, the burden of subarachnoid hemorrhage was particularly severe among males and females aged 55–59 years in other regions (Fig. 7A, Appendix pp 236, 243, 246, 251, 256 and 266). For sex-specific conditions, such as female cervical cancer, the age group of 55–59 years was identified as the stage of greatest burden globally (with DALY rate at 57.7 [32.7–96.7] per 100,000) (Appendix p 234). However, the global burden of male prostate cancer attributable to smoking escalated throughout the lifespan, with individuals over 75 years exhibiting the DALY rate of 75.1 (34.0–135.3) per 100,000 population.

Fig. 7.

Global trends of DALY rates (per 100,000 population) and 95% UIs of the overlapping diseases attributable to smoking tobacco (A), secondhand smoke (B), and causes attributable to chewing tobacco (C) among males and females across age groups in 2023. DALYs, disability-adjusted life years.

Compared to the sex ratios of DALY rates attributable to smoking risk in the same regions, the fluctuations in disease burden attributable to SHS (such as ischemic stroke, intracerebral hemorrhage, and COPD) between sexes across age groups were less pronounced in 2023 (Fig. 7B, Appendix pp 63–64). The sex ratio of DALY rates for TBL exhibited a general upward trend with age, peaking among the 70–74 age group (1.9 95% CI [1.3–2.8]) and showing the greatest sex disparity (males: 314.7 [231.7–404.0] per 100,000; females: 163.0 [124.4–214.6] per 100,000). Furthermore, with the exception of South Asia, Southeast Asia, East Asia, and Oceania, the DALY rate for TBL in males remained consistently higher than in females from age 50 onwards (Appendix pp 237, 242, 247, 252 and 267).

Among all outcomes attributable to chewing tobacco, only ischemic stroke exhibited a progressive elevation in DALY rates with advancing age, reaching peak values in the 75+ age group (males: 177.1 [88.6–329.9] per 100,000 population; females: 156.0 [70.3–285.2] per 100,000 population) (Fig. 7C). Notably, males exhibited consistently higher DALY rates for larynx cancer and other pharynx cancer than females across all age groups globally (Fig. 7C, Appendix pp 65–66), with this disparity being particularly pronounced in South Asia (Appendix p 258).

Discussion

This study provides a comprehensive sex-based analysis of disease burden attributable to smoking, SHS, and chewing tobacco using GBD 2023 data. The results revealed that, while the overall ASMR and ASDR for tobacco have declined significantly over the past 30 years, the absolute counts of deaths and DALYs remained a significant concern due to factors such as population growth and aging. Concurrently, the pattern of sex disparities in the global disease burden has diverged across the three tobacco risk factors: the ASDR differences between males and females reflected a gradual narrowing for smoking, a situation in which males currently appear to bear a marginally higher burden for SHS, and a persistently stable gap for chewing tobacco. These findings suggest that tobacco control policies may benefit from profound sex-specific adjustments to more effectively address the current health disparities.

Several previous rounds of GBD estimates have documented the global and regional prevalence and attributable burden of each tobacco risk factor with some degree of sex stratification.4, 5, 6^,19 For example, the GBD 2019 Tobacco Collaborators reported sex-stratified smoking prevalence and attributable DALYs, and noted that males accounted for over 80% (162 of 200 million) of the total smoking-attributable DALYs.⁴ However, relying on global prevalence distributions to characterize sex differences has left a gap in deeper pattern analysis and driver decomposition. Zheng et al.,⁵ examined the spatiotemporal trends in the global SHS-attributable burden and explored possible reasons for sex differences in summary exposure values, but did not extend to cause-specific sex disparities in attributable burden or their future trajectories. Building on these prior contributions, the present study placed sex inequality as the central analytical lens within a unified framework, integrating all three tobacco risks to reveal their distinct temporal, geographic, and age-related patterns of sex disparity. Furthermore, this study moved beyond descriptive reporting by employing decomposition analysis to quantify the contributions of population growth, population aging, change in risk exposure, and change in risk-deleted burden to sex-specific DALY changes between 1990 and 2023. This research further extended the analysis forward in time with a forecasting framework, projecting how sex disparities may evolve to 2050. This study focused on providing epidemiological and public-health insights rather than a mechanistic explanation of the underlying biological or social causes of sex disparities in tobacco-attributable burden. Together, these advances repositioned sex inequality from a supplementary finding to a core investigative theme, offering a more nuanced evidence base for designing targeted tobacco control policies.

Despite longstanding debates balancing public health priorities and tobacco-related economic interests,^31,32 substantial evidence has highlighted the mutually beneficial outcomes of a smoke-free future,33, 34, 35 prompting governments to advocate for the implementation of the WHO Framework Convention on Tobacco Control (FCTC),³⁶ which has significantly mitigated the adverse health effects of both active and passive smoking on the public. Our findings indicated that since 1990, the overall smoking-attributable ASDR has declined by 47.7% (from 3372.6 to 1762.4 per 100,000), while the rate for SHS has decreased by 53.1% (from 1112.8 to 521.8 per 100,000). However, the rate for chewing tobacco has remained stable (−2.0% [−44.8 to 71.3], from 72.5 to 71.0 per 100,000; non-significant). The timely regulation of tobacco, the foremost preventable risk factor for health, has yielded immeasurable benefits for global human health. Nevertheless, the absolute number of DALYs attributable to smoking and chewing tobacco continued to rise globally, whereas total SHS-attributable DALYs declined. Decomposition analysis revealed that, for smoking and SHS, population growth and aging acted as positive drivers, while changes in risk exposure and risk-deleted burden exerted countervailing effects. For chewing tobacco, risk exposure also contributed positively alongside demographic factors, with risk-deleted burden serving as the sole mitigating force. Comprehensive tobacco control policies should address not only reductions in tobacco exposure but also demographic factors and the broader health determinants such as healthcare accessibility and socioeconomic conditions.^5,26 Across all three risks, demographic contributions were consistently larger in males, contributing to wider sex disparities in absolute burden. For SHS, the net DALY decline reflected reductions in exposure and risk-deleted burden outweighing demographic growth, and this decline was greater in females. Despite these insights from decomposition, the disparities of disease burden between males and females appear to have been insufficiently addressed. Between 1990 and 2023, the global ASDR attributable to smoking and SHS has consistently followed a downward trajectory, with the decline among females consistently outpacing that of males, as supported by the latest research.²¹ It was speculated that this situation may be related to the sex composition of tobacco users.⁴

In 2003, the WHO urged the parties to the FCTC to undertake all feasible measures to mitigate life expectancy losses caused by tobacco use and to uphold their commitment to public health.³⁷ Chronologically, WHO issued four editions of global tobacco smoking/use report to monitor tobacco prevalence across nations and measure global trends of tobacco use, advocating for the goal of a 30% reduction in tobacco use by 2025 compared to 2010 levels.38, 39, 40, 41 According to estimates from the latest fourth edition, the gradual decline in the number of global tobacco users from 1.37 billion in 2000 to 1.27 billion in 2025 was predominantly driven by a rapid decrease in female tobacco users. It has been suggested that occupational stress and social role expectations may contribute to higher tobacco use among males.⁴² Correspondingly, it was not until 2015 that the number of male tobacco users ceased to increase. Both biological sex differences and gender-related social factors have been proposed to contribute to the disparities in disease burden associated with tobacco, with some evidence suggesting that females may experience a lower burden for certain conditions, although this advantage is not uniform across all diseases.^43,44 From a physiological perspective, it has been hypothesized that sex steroid-related pathways may contribute to cardiovascular protection in females,⁴⁵ however, this protective effect is not always observed in practice. One possible explanation for the increased risk among females is their lower sensitivity to nicotine compared to males, which may contribute to heightened vulnerability.⁴⁶ Furthermore, cognitive differences shaped by social constructs may be partly reflected in the differing perceptions of males and females towards tobacco advertising, potentially adding another layer of complexity to the issue. For example, menthol-flavored tobacco products tend to attract more interest from females compared to males.⁴⁷ The US Food and Drug Administration (FDA) is set to reconsider tobacco control policies and prevention initiatives, taking into account the influence of sex on tobacco consumption.⁴⁸ For nicotine restrictions, analyzing long-term exposure data by sex, while integrating findings from laboratory and cessation research, may help determine optimal nicotine reduction levels. To address sex-specific preferences, phasing out menthol cigarettes may require measures to discourage females from switching to other tobacco products. Additionally, effective public awareness campaigns and community education on sex sensitivity differences in tobacco use could help counteract the marketing strategies of tobacco industries.

The persistent prevalence of tobacco consumption has likely weakened the implementation efficacy of SHS control policies, particularly in settings with weak enforcement mechanisms and high public compliance gaps.^49,50 For example, in six European countries, strict enforcement of comprehensive smoke-free laws was associated with sharp declines in SHS exposure in Greece and Romania, while persistently high exposure was observed in Germany and Poland, where indoor smoking bans remained incomplete.⁵⁰ However, on a global scale, the risks of SHS exposure faced by males and females did not simply reflect the proportionality of tobacco users among each sex. Our study revealed that the all-cause ASDR attributable to SHS among males surpassed that of females in 2023 and was projected to remain higher under the reference scenario. This finding appeared to diverge from the currently reported percentages of SHS exposure and SEVs for males and females.^5,51,52 Traditionally, females have been considered the more vulnerable group in SHS research, given their heightened sensitivity to certain diseases compared with males—for example, their increased susceptibility to stroke¹¹ and the strong predisposition to lung cancer.^53,54 Indeed, in our study, the ASDR for three subtypes of stroke among females globally remained higher than that of males in 2023, although this pattern may also shift in the future under alternative scenarios (e.g., aggressive decline scenario in our study). Furthermore, extensive research has documented the broad and enduring harm caused by SHS to pregnant females and fetuses, including depression in pregnant females and adverse fetal outcomes in newborns, and increased incidence of T2DM in their adulthood.55, 56, 57, 58 These findings prompted tobacco control policies to increasingly emphasize the protection of females,^59,60 which may have contributed to enhanced awareness of and distancing from SHS exposure among them. However, the results of our study also suggested that males might bear a greater future burden attributable to SHS. This pattern suggests that a comprehensive approach to SHS control might consider the burden trajectory in both sexes, ensuring that the relatively higher burden in males receives adequate attention alongside existing female-oriented measures. The detrimental effects of SHS across sexes should be more widely communicated to the public. In high tobacco prevalence areas, such as restaurants and bars, smoke-free policies remain an important protective measure.⁶¹ Future legislation could address the current disease burden by continuing to fund sex-distinct research and emphasizing the importance of smoke-free home environments to better accommodate these sex disparities.

Compared to smoking and SHS, the burden attributable to chewing tobacco has received relatively less attention, yet its persistence over the past three decades is cause for concern. According to a 2019 global prevalence analysis, an estimated 273.9 million people worldwide used chewing tobacco products.⁶ Our findings indicated that in 2023, smokeless tobacco was responsible for 251.4 thousand deaths and 6.5 million DALYs, with the ASDR for both sexes remaining stable throughout the 34-year observation period. Although chewing tobacco has been traditionally perceived to offer benefits such as decongestion and mood relaxation,⁶² a growing body of evidence suggested that these potential advantages were substantially outweighed by its established links to various cancers and cardiovascular diseases.63, 64, 65 These findings highlighted the importance of integrating chewing tobacco into existing and future tobacco control frameworks, in order to mitigate the long-standing prevalence of a product often mistakenly regarded as a safer alternative to smoked tobacco.⁶⁶ In alignment with the GBD 2019 findings,⁶ our study identified South Asia as the region bearing the most severe burden of chewing tobacco use, particularly in India, among both male and female populations. This regional prevalence has been associated with the affordability of tobacco products, inadequate regulatory enforcement,⁶⁶ lower educational attainment and socioeconomic status.⁶⁷ Despite the implementation of ambitious policy measures in some countries, such as the ban on gutkha in India and the expansion of health warning label coverage on packaging,⁶⁸ their effectiveness has thus far remained limited.⁶⁹ In this context, strengthening legislative restrictions on smokeless tobacco and increasing taxation, along with evidence-based public health campaigns to address misconceptions regarding its safety, could help reduce the associated disease burden. Furthermore, targeted interventions for populations with limited educational access, particularly women in low-income regions, may contribute to lowering the social acceptability of chewing tobacco use.

Formulating sex-targeted tobacco control policies requires concerted efforts from governments worldwide. As our research findings suggested, the impact of tobacco-related risks on both sexes varies by region and age. First, tobacco control policies and media campaigns among nations have been recognized as influential factors, including increased tobacco taxes, implementation of smoke-free legislation, and health warnings on tobacco packaging and in public spaces.^70,71 These policies varied significantly across regions.^70,72, 73, 74 Moreover, the prevalence of tobacco use among females in high-income countries, categorized by the World Bank income group, surpassed that of any other region, whereas male tobacco use in this group was relatively low. Targeted marketing to female customers, through products like menthol cigarettes and slim cigarettes and couponing, may have contributed to this trend.⁷⁵ This may interact with rising awareness of sex equality and the accompanying elevated social status and social pressures.^76,77 Consistent with these observations, the sex ratio of smoking-attributable ASDR in Australasia and high-income North America tended to be the lowest. By contrast, for SHS, male ASDR consistently exceeded that of females in several high-income areas (Appendix pp 38–53). Differences in educational attainment, interpersonal factors, and sociocultural influences further shaped social disparities based on physiological development, which in turn may be reflected in attitudes towards smoking, SHS, and chewing tobacco.⁷⁸ Notably, societal constructs of masculinity and femininity have been exploited in tobacco industry marketing, which likely contributes to sex-specific patterns of tobacco use.^79,80 It has long been posited that tobacco control policy revisions should focus on sex differences and transcend the limitations of binary frameworks to address sex inequality in tobacco issues through considerations of gender roles, gender identities, and gender relations.^81,82 In this context, increased investment in sex-specific research may assist in developing more tailored clinical guidelines and medical and psychological support services grounded in sex-disaggregated data. Considering the sex differences in tobacco use prevalence and burden across various income levels, high-income areas may benefit from strategies that address the social drivers of tobacco use among females, including efforts to promote diverse representations of femininity. In contrast, in low- and middle-income regions, attention to the rising burden associated with increasing tobacco use among males and females, which often accompanies improvements in education and social development, could inform more targeted preventive actions.

Our research findings indicated that tobacco use exerted varying degrees of harm across different age groups. Overall, sex differences were evident early in the life course, with DALY rates increasing with age. However, the patterns for males and females were not entirely consistent. This discrepancy may partly reflect sex differences in tobacco use prevalence, which peaked at ages 45–54 years among males and 55–64 years among females.⁴¹ With aging and the stabilization of social roles, the health risks of tobacco use manifest in sex-specific ways. For instance, higher cumulative tobacco exposure in occupational settings, combined with workplace hazards such as polycyclic aromatic hydrocarbons, may accelerate lung function decline.⁸³ This partially explained the observed increase in COPD burden among both males and females aged over 40. Furthermore, although tobacco exposure has declined among adolescents and youth of both sexes, these groups remained differentially affected by novel smokeless tobacco products,⁸⁴ which may be associated with disparities in disease burdens during the early stages of life. The health risks of smokeless tobacco warrant prominent public communication and integration into comprehensive tobacco control measures. Moreover, regulatory frameworks could be refined by incorporating increasingly sophisticated sex-sensitive measures to address sex and age disparities and their interactions. Without such efforts, the inequalities in tobacco-related harm between males and females are likely to persist.

Although carefully conducted, our study should be considered in light of its limitations. First, as a secondary analysis of GBD 2023 estimates, the sex-stratified estimates of deaths, DALYs, ASDR and PAFs for each risk–outcome pair were obtained directly from the GBD results tool, thus our findings inevitably inherit the limitations of the GBD modeling framework. In regions with sparse or absent primary data, particularly for chewing tobacco and SHS exposure in low-income settings, GBD estimates rely heavily on covariates, spatial–temporal statistical models, and predictive extrapolations, which may not fully capture local epidemiological heterogeneity.⁶ Furthermore, self-reported data on smoking, SHS, and chewing tobacco may be subject to unmeasured residual confounding across sex, age, geographical region, and socioeconomic strata, introducing systematic reporting biases that current validation frameworks cannot fully quantify. For instance, low social acceptability of tobacco has obscured a significant portion of self-reported smoking prevalence rates among females in Asia.⁸⁵ Second, the newly derived analyses in our study carried their own assumptions and uncertainties. The confidence intervals for the sex ratios were constructed using the delta method on the logarithmic scale, which assumed that the male and female ASDR estimates were approximately normally distributed; this assumption might not hold in settings with sparse data or when the rates were very low, potentially yielding intervals that were too narrow. In addition, the decomposition analysis was generally subject to the uncertainty embedded in the age-, sex-, and cause-specific inputs. However, the PAFs and cause-specific DALY rates used in the decomposition in this study were single-point estimates and their full uncertainty distributions were not propagated through the Das Gupta decomposition. As a result, the reported contribution of each driver does not include uncertainty intervals, and the relative ordering of drivers should be interpreted as point estimates rather than statistically robust comparisons. Moreover, the BAPC and ARIMA models used for projections assumed that historical age-period-cohort effects and SEV autocorrelation structures would remain structurally stable over the forecast horizon and did not account for abrupt shifts in policy, behavior, or socioeconomic conditions. The use of SEV as the basis for counterfactual scenario construction relied on estimates from the GBD framework, which inherited uncertainties from the GBD modeling mentioned above. As a summary measure, SEV might not fully capture the relationship between exposure and attributable burden, potentially introducing uncertainty that propagates through to the scenario projections. The counterfactual SEV scenarios, while useful for policy exploration, imposed simplified annual reduction rates and assumed constant elasticities linking SEV to attributable burden, which might not capture the full complexity of real-world tobacco control dynamics. Third, our analysis consistently relied on a binary sex framework (male and female), thereby limiting a more nuanced understanding of the diverse differences within sex identities. Sexual and gender minorities often face disproportionately high health burdens related to smoking and SHS,86, 87, 88 yet remain underrepresented in tobacco research. More inclusive studies are needed to address their health needs. Smoking cessation researchers should also focus on these high-risk populations and enhance awareness of gender diversity, striving to create a more supportive and proactive smoking cessation environment. Fourth, while our analysis encompassed three manifestations of tobacco-related risk—smoking, SHS exposure, and chewing tobacco use—it did not comprehensively address the potential synergistic effects from simultaneous exposure to multiple risk modalities. This limitation may engender an underestimation of the aggregate tobacco-attributable disease burden. Finally, when comparing disease burden patterns across regions and age groups, simplistic consideration of differences between sexes appeared insufficient, as numerous intertwined factors—such as race, culture, social cognition, and socioeconomic status—underpin these disparities. More comprehensive research is needed to develop sex-differentiated health awareness and medical services tailored to specific sex populations.

To conclude, our research provided a sex-based epidemiological analysis of the disease burden associated with smoking, SHS, and chewing tobacco using GBD 2023, highlighting the evolving differences in disease burden between males and females and their drivers, over time, across geographical locations, levels of socioeconomic development, and life stages. These findings suggested that national tobacco control strategies may benefit from incorporating assessments of health burdens across whole life stages, addressing sex differences in current smoking prevalence. Targeted control strategies for disadvantaged groups remain essential. For instance, in some high-income countries, sustained public communication campaigns that challenge the masculine and feminine imagery historically associated with tobacco use may help accelerate progress in tobacco control. In low- and middle-income countries, enhancing health implementation systems is crucial, particularly by strengthening health surveillance for middle-aged and older populations to prevent potential consequences of rising tobacco trends. Understanding the interactive patterns of sex and gender across temporal, spatial, and life course contexts is essential for guiding research investments, healthcare initiatives, and public health strategies. To address tobacco-related health challenges, it is necessary to acknowledge the disparities between males and females in smoking and chewing tobacco initiation, SHS exposure, risk perceptions, cessation difficulties, and associated disease burdens. Systematic attention to these sex-specific dimensions would help to effectively mitigate the substantial health challenges posed by tobacco worldwide.

Contributors

YGW and XCD conceptualized and designed the study. Sheng Li, ZX, and Brook Morgan performed the methodology and analysis and drafted the manuscript. DL, LHZ, JL, BYG, and Shu Li performed the data collection. XCD and YGW accessed and verified the data. Sheng Li, ZX, DL, Brook Morgan, WLY, XLX, HYS, XCD, and YGW contributed to the revision of the manuscript and approved the final draft. All authors contributed to the intellectual content and critical revisions to the drafts of the paper and approved the final version. YGW and XCD accept full responsibility for the work and the conduct of the study, had access to the data, and controlled the decision to publish. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.

Data sharing statement

All data in the study are available at https://vizhub.healthdata.org/gbd-results/. All the R packages used are open-source for academic use. All of these data are de-identified and publicly available. All data generated in this study are included in the Supplementary Information. The code is available from the corresponding author upon reasonable request.

Editor note

The Lancet Group takes a neutral position with respect to territorial claims in published maps and institutional affiliations.

Declaration of interests

The authors declare no competing interests.