Global Burden of Lip and Oral Cavity Cancer Attributed to Different Tobacco Use From 1990 to 2021 and Projection to 2050 Authors: Zijing Yu, Min Zhang, Yan Zheng, Jiapeng Su, Shangzhi Han, Xue Chen, Yue Fang, Hongye Liu, Xiaoling Liu, Guoxing Wan Published: 2026 ABSTRACT AbstractIntroduction and AimsLip and oral cavity cancer (LOCC) remains a significant global health challenge, with tobacco use being its primary modifiable risk factor. However, a comprehensive understanding of the LOCC burden attributable to different forms of tobacco use is lacking. This study aims to systematically quantify the global, regional, and national burden of LOCC attributable to smoking and chewing tobacco from 1990 to 2021 and to project trends to 2050, providing evidence for targeted tobacco control policies.MethodsUsing data from the Global Burden of Disease (GBD) 2021 study, we extracted mortality and disability-adjusted life years (DALYs) for LOCC attributable to tobacco smoking and chewing tobacco. Age-standardized mortality rates (ASMR) and DALY rates (ASDR) were analysed by sex, region, and Socio-demographic Index (SDI). Cross-country inequality was assessed using the Slope Index and Concentration Index. A Bayesian age-period-cohort (BAPC) model was employed to forecast the burden to 2050.ResultsGlobally, between 1990 and 2021, absolute tobacco-attributable LOCC deaths nearly doubled, while the ASMR declined (from 1.07 to 0.94 per 100,000). This decline was driven entirely by smoking, whereas the burden from chewing tobacco increased slightly. Striking disparities emerged: the burden in females remained stable, driven primarily by chewing tobacco, while males saw a decline driven by reduced smoking. Regionally, South Asia bore the highest burden. Inequality analysis revealed narrowing disparities for smoking-attributable LOCC but persistent, unaddressed inequality for chewing tobacco. BAPC projections to 2050 forecast a continued decline in smoking-attributable burden but no improvement for chewing tobacco-attributable burden.ConclusionWhile global tobacco control efforts have successfully reduced the smoking-attributable LOCC burden, the persistent and, in some populations, increasing burden from chewing tobacco – particularly among females and in low-SDI regions – demands urgent attention. Projections indicate that without targeted intervention, chewing tobacco will remain an unmitigated cause of LOCC for decades.Clinical RelevanceTargeted tobacco control is urgently needed for the reduction of LOCC burden in specified populations.Key words: Lip and oral cavity cancer, Tobacco smoking, Chewing tobacco, Global burden of disease, Difference Introduction As a malignant neoplasm with a generally poor prognosis, lip and oral cavity cancer (LOCC) represents one of the most prevalent and devastating subtypes within head and neck cancers.1 According to the Global Cancer Statistics Report 2022, LOCC ranks as the sixteenth most commonly diagnosed cancer worldwide and the fifteenth leading cause of cancer-related mortality.2 Notably, in low- and middle-income countries (LMICs), the burden is disproportionately higher, where it stands as the fifth most common cancer and the eighth most frequent cause of cancer death.2 Despite dramatic increases in incidence over recent decades, survival outcomes remain dismal, approximately only 40% to 50% of patients survive beyond 5 years postdiagnosis. Notably, the prognosis of LOCC is strongly stage-dependent, with 5-year survival rates approaching 80% for stage I disease but declining sharply to approximately 20% for advanced-stage presentations (stages III/IV).3 LOCC also exhibits marked demographic disparities, and disease onset predominantly occurs in middle-aged and older populations, with a consistently higher incidence observed in males compared to females.4 Over the past three decades, the geographic distribution of LOCC burden has shifted considerably, with high-income countries achieving substantial reductions, whereas LMICs continue to experience persistently high disease burden.5 These inequities underscore an urgent need for targeted research and policy intervention. The aetiology of LOCC is multifactorial and well characterized, involving complex interactions between genetic susceptibility and environmental exposures, with substantial variation across geographic and demographic contexts.4,6 Among modifiable risk factors, deleterious oral habits – particularly tobacco use – are the most influential in LOCC pathogenesis.4,7 Epidemiological evidence indicates that nearly 85% of LOCC cases in industrialized nations are attributable to tobacco, and during the twentieth century, smoking accounted for an estimated 71% of LOCC mortality in developed countries and 37% in developing regions.8,9 Mechanistically, tobacco carcinogens – notably aromatic hydrocarbon benzopyrene and tobacco-specific nitrosamines – are absorbed directly through the oral mucosa, inducing deleterious mutations and activating oncogenic signalling pathways.6,10,11 Tobacco smoking elevates the risk of LOCC approximately fivefold, while sustained cessation for 3 to 5 years reduces risk by nearly 50%.12 Furthermore, younger patients (≤45 years) with a history of tobacco use exhibit significantly elevated risks of both recurrence and mortality compared to their nonexposed counterparts.13 Although global smoking prevalence has declined in recent decades, reductions have been modest in many LMICs, and due to population growth, the absolute number of smokers worldwide has increased.14,15 Concurrently, stricter smoking regulations have been associated with a notable rise in the consumption of smokeless tobacco, particularly chewing tobacco – an exposure that remains underaddressed in international public health research and tobacco control frameworks.16 The tobacco epidemic thus persists as one of the most consequential public health challenges globally.17 Despite the well-established association between tobacco and LOCC, the epidemiological patterns of LOCC burden attributable to distinct forms of tobacco use at the global level remain insufficiently characterized. Leveraging data from the Global Burden of Disease (GBD) 2021 study, the present investigation provides a comprehensive analysis of the global LOCC burden from 1990 to 2021. Specifically, we quantify temporal trends in mortality and disability-adjusted life years (DALYs) attributable to different tobacco exposures, emphasizing se and regional disparities, and project future trajectories through 2050. This study constitutes a pioneering effort to systematically evaluate the global impact of tobacco on LOCC, offering critical insights to inform evidence-based policymaking and strategic resource allocation for tobacco control and cancer prevention. Materials and methods Data sources This study utilized the most recent available data from the Global Burden of Disease (GBD) 2021 database, which provides comprehensive estimates of incidence, prevalence, mortality, and disability-adjusted life years (DALYs) for 371 diseases and injuries, along with associated estimates for 88 risk factors, across 204 countries and territories.14 Annual data on LOCC-related deaths, DALYs, age-standardized mortality rates (ASMR), and age-standardized DALY rates (ASDR) attributable to tobacco were extracted by year, age, region, and country for the period 1990 to 2021 using the Global Health Data Exchange (GHDx) query tool.17 The Socio-demographic Index (SDI), a composite indicator of socioeconomic development ranging from 0 (lowest) to 1 (highest), was employed to stratify countries and territories into five development quintiles and to categorize 21 geographic regions, as defined by the GBD framework.14 All data were publicly accessible and obtained from the GBD 2021 website (accessed August 19, 2025). The specific query parameters were as follows: GBD Estimate, ‘Risk factor’; Measure, ‘Deaths’ and ‘DALYs’; Metric, ‘Number’, ‘Percent’, and ‘Rate’; Risk, ‘Tobacco’, ‘Smoking’, ‘Chewing tobacco’, and ‘Secondhand smoke’; Cause, ‘Lip and oral cavity cancer’; Geographic regions, ‘Global’, 204 countries and territories, and 21 regions; Time period, ‘1990 to 2021’; Sex, ‘Both’, ‘Male’, and ‘Female’; Age, all standard GBD 5-year age groups from <5 to ≥95 years. As this study utilized deidentified, aggregated, and publicly available data, it was exempt from ethical review. In accordance with GBD estimation methodologies, all burden estimates are reported with their corresponding 95% uncertainty intervals (UIs). These UIs were calculated by the Institute of Health Metrics and Evaluation (IHME) using the Bayesian multilevel regression tool DisMod-MR 2.1, which integrates data from multiple sources to generate internally consistent estimates.4 Cross-country inequality analysis To evaluate socioeconomic disparities in the tobacco-attributable LOCC burden, we employed health inequality analysis methodologies established by the World Health Organization.18 Quantification of SDI-related inequality across countries was performed using the Slope Index of Inequality (SII) and the Concentration Index (CI), as previously described.19 The SII, a measure of absolute inequality, was derived from a regression model that assesses the relationship between country-level ASMR and ASDR for tobacco-attributable LOCC and the relative socioeconomic position of each country, defined by the midpoint of the cumulative class range of the population ranked by SDI.20 The CI, a measure of relative inequality, was calculated by fitting a Lorenz concentration curve to the cumulative distribution of the LOCC burden against the cumulative population ranked by SDI, with the index value derived from the area under this curve. The SII reflects absolute disparities across the entire SDI spectrum, with larger values indicating greater inequality. The CI ranges from -1 to 1; positive values indicate a concentration of the burden among populations with higher socioeconomic status, while negative values indicate a concentration among those with lower status, and values closer to zero suggest greater equity.21 Bayesian age-period-cohort analysis To forecast future trends, we applied a Bayesian age-period-cohort (BAPC) model. Building upon a generalized linear model (GLM) framework within a Bayesian context, this approach enables the dynamic integration analysis of age, period, and cohort effects. These effects are assumed to evolve smoothly over time and are modelled using a second-order random walk prior, leading to robust posterior predictive distributions.22 The BAPC model is particularly well-suited for epidemiological forecasting due to its capacity to handle age-structured population data and complex temporal effects. Using the ‘BAPC’ R package, we projected global mortality and DALYs trends for tobacco-attributable LOCC from 2022 to 2050. These projections are intended to inform future public health planning and preventive strategies. Statistical analysis The relationship between the LOCC burden, as measured by ASMR and ASDR, and national SDI was explored using locally weighted regression (LOESS) smoothing, which allowed for the visualization of nonlinear trends without imposing a predefined parametric form. For all stratified analyses involving tobacco types, data on smoking and chewing tobacco were included as covariates. Of note, estimates for secondhand smoke-attributable LOCC were not available in the GBD 2021 database for the specified cause and were therefore excluded from this analysis. All statistical analyses and data visualizations were performed using R software (version 4.1.1). Uncertainty intervals (UIs) for all point estimates were derived from the GBD study and are presented as per standard practice. Results Global mortality burden of tobacco-attributable LOCC (1990-2021) Globally, the number of deaths from LOCC attributable to tobacco nearly doubled from 1990 to 2021, increasing from 42,705 (95% UI: 34,703-50,564) to 81,947 (95% UI: 65,940-98,083) (Table 1). In contrast, the global age-standardized mortality rate (ASMR) declined over the same period, from 1.07 (95% UI: 0.87-1.26) to 0.94 (95% UI: 0.76-1.13) per 100,000 populations. Substantial sex-specific disparities were evident. Males consistently bore the majority of the mortality burden (32,971 deaths in 1990 vs 58,880 in 2021), yet the rate of increase in death counts was steeper among females. The ASMR for males decreased from 1.78 (95% UI: 1.37-2.14) to 1.44 (95% UI: 1.07-1.78) per 100,000, while the rate for females remained relatively stable between 2019 and 2021, at approximately 0.50 per 100,000. Regionally, the burden shifted markedly across SDI regions. While the absolute number of tobacco-attributable LOCC deaths increased in all SDI regions from 1990 to 2021, the increases were most pronounced in middle (139.09%), low-middle (128.67%), and low SDI (119.29%) regions (Table 1). Correspondingly, the ASMR declined substantially in high SDI (−67.39%), high-middle SDI (−37.5%), and middle SDI (−10.34%) regions, but no such downward trend was observed in low-middle and low SDI regions. Geographically, India recorded the highest number of tobacco-attributable LOCC deaths in both 1990 and 2021 (Figure 1A, B), whereas Pakistan exhibited the highest ASMR during the same period (Figure 1C, D). These data underscore the disproportionate concentration of the mortality burden in South Asia. Table 1Temporal trends in global burden of death due to lip and oral cavity cancer attributable to tobacco from 1990 to 2021.Table 1 dummy alt text19902021Dead cases,No. *102 (95% UI)ASMR per 105No. (95% UI)Dead cases,No. *102 (95% UI)ASMR per 105No. (95% UI)Global427.05 [347.03-505.64]1.07 [0.87-1.26]819.47 [659.4-980.83]0.94 [0.76-1.13]SexFemale97.34 [77.11-117.45]0.46 [0.37-0.56]230.67 [185.86-281.53]0.50 [0.40-0.61]Male329.71 [256.28-395.59]1.78 [1.37-2.14]588.8 [438.62-723.82]1.44 [1.07-1.78]SDIHigh83.18 [60.47-105.49]0.77 [0.56-0.98]92.63 [63.31-123.33]0.46 [0.32-0.61]High-middle77.72 [58.22-95.60]0.77 [0.57-0.95]111.34 [80.73-140.25]0.56 [0.41-0.70]Middle96.79 [78.50-114.50]0.96 [0.78-1.14]231.42 [182.63-282.58]0.87 [0.69-1.06]Low-middle139.34 [115.06-165.97]2.30 [1.9-2.72]318.63 [263.33-376.21]2.25 [1.86-2.65]Low29.60 [23.6-36.31]1.33 [1.07-1.62]64.91 [50.53-78.87]1.34 [1.06-1.62]ASMR, age-standardized mortality rate; SDI, sociodemographic index. Temporal trends in global burden of death due to lip and oral cavity cancer attributable to tobacco from 1990 to 2021. ASMR, age-standardized mortality rate; SDI, sociodemographic index. The number and age-standardized rates of deaths and DALYs of LOCC attributable to tobacco use in 204 countries and territories in 1990 and 2021. A, The number of deaths globally in 1990. B, The number of deaths globally in 2021. C, The age-standardized rates of deaths globally in 1990. D, The age-standardized rates of deaths globally in 2021. E, The number of DALYs globally in 1990. F, The number of DALYs globally in 2021. G, The age-standardized rates of DALYs globally in 1990. H, The age-standardized rates of DALYs globally in 2021. Global DALYs burden of tobacco-attributable LOCC (1990-2021) Consistent with mortality trends, the global number of DALYs due to tobacco-attributable LOCC rose from 1.26 million (95% UI: 1.03-1.49 million) in 1990 to 2.26 million (95% UI: 1.82-2.70 million) in 2021 (Table 2). The global age-standardized DALY rate (ASDR), however, declined from 30.09 (95% UI: 24.52-35.52) to 25.80 (95% UI: 20.76-30.84) per 100,000 over the study period. The DALY burden was also predominantly shouldered by males, although the rate of increase was greater in females. The ASDR for males decreased from 49.15 (95% UI: 38.24-58.90) to 39.42 (95% UI: 29.30-48.34), while the rate for females did not exhibit similar decrease (Table 2). The increase in absolute DALY counts was most pronounced in middle (120.87%), low-middle (115.97%), and low SDI (109.27%) regions. Conversely, ASDR declines were substantial in high and high-middle SDI regions (decreases of 42.87% and 30.23%, respectively) but only modest in lower SDI regions (ranging from approximately 5%-12% decrease) (Table 2). Mirroring the mortality findings, India ranked highest in DALY counts, and Pakistan in ASDR, further emphasizing the regional concentration of the burden in South Asia (Figure 1E-H). Table 2Temporal trends in global burden of disability-adjusted life-years (DALYs) due to lip and oral cavity cancer attributable to tobacco from 1990 to 2021.Table 2 dummy alt text19902021DALYs,No. *102 (95% UI)ASDRNo. (95% UI)DALYs,No. *102 (95% UI)ASDRNo. (95% UI)Global12609.41 [10296.60-14871.89]30.09 [24.52-35.52]22603.35 [18169.76-27023.77]25.80 [20.76-30.84]SexFemale2712.52 [2179.02-3292.73]12.57 [10.11-15.20]5938.38 [4766.82-7381.16]13.08 [10.49-16.27]Male9896.89 [7739.65-11873.32]49.15 [38.24-58.90]16664.96 [12392.83-20452.59]39.42 [29.30-48.34]SDIHigh2353.93 [1731.1-2954.64]22.51 [16.57-28.20]2357.15 [1633.92-3067.28]12.86 [8.99-16.62]High-middle2350.03 [1773.35-2866.04]22.59 [17.01-27.57]3106.42 [2255.96-3885.74]15.76 [11.45-19.69]Middle2835.90 [2308.11-3368.95]25.18 [20.45-29.81]6263.60 [4919.20-7746.85]22.24 [17.51-27.42]Low-middle4165.00 [3455.84-4976.66]61.61 [50.95-73.78]8995.03 [7366.87-10639.30]58.39 [48.09-69.09]Low891.72 [705.85-1103.26]35.39 [28.21-43.55]1866.13 [1442.16-2296.14]33.50 [26.11-40.95]ASDR, age-standardized DALYs rate; SDI, sociodemographic index. Temporal trends in global burden of disability-adjusted life-years (DALYs) due to lip and oral cavity cancer attributable to tobacco from 1990 to 2021. ASDR, age-standardized DALYs rate; SDI, sociodemographic index. Temporal trends in LOCC burden by sex and SDI region Globally, ASMR and ASDR for tobacco-attributable LOCC showed a sustained decline, particularly after 1995 (Figure 2). Sex-disaggregated analysis revealed that this downward trend was driven primarily by reductions in males, while rates in females showed no appreciable decline over the three-decade period (Figure 3A). By SDI region, marked decreases in both ASMR and ASDR were confined to high and high-middle SDI regions. In contrast, rates in middle, low-middle, and low SDI regions remained persistently elevated, showing no clear downward trajectory (Figure 3B). The burdens of LOCC attributable to different tobacco use among 204 countries and territories from 1990 to 2021. The burdens of LOCC attributable to tobacco use among 204 countries and territories from 1990 to 2021 according to different sexes and SDI regions. A, Sex difference. B, SDI regions difference. Burden attributable to smoking vs chewing tobacco, by sex and SDI region Due to data limitations in the GBD 2021 database for secondhand smoke, this analysis focused on smoking and chewing tobacco. Globally, while the burden attributable to smoking declined considerably, a slight but consistent upward trend was observed for chewing tobacco from 1990 to 2021 (Figure 2). Sex-specific patterns were strikingly divergent. In males, the ASMR and ASDR attributable to smoking decreased steadily, whereas the burden from chewing tobacco did not decline (Figure 4). In females, the smoking-attributable burden showed a modest decline, yet this was offset by a clear upward trend in the burden from chewing tobacco (Figure 4). Notably, chewing tobacco was the primary driver of the LOCC burden in females throughout the study period, whereas smoking remained the dominant risk factor in males (Figure 4). This dichotomy extended to regional patterns. In high, high-middle, and middle SDI regions, smoking was the predominant contributor to the LOCC burden, with rates showing a sustained decline (Figure 5). In contrast, in low-middle and low SDI regions, chewing tobacco was the principal risk factor, and its associated burden exhibited a slight but persistent increase from 1990 to 2021 (Figure 5). In these regions, the smoking-attributable burden, while smaller, showed a modest decline. The ASMR and ASDR of LOCC attributable to different tobacco use among 204 countries and territories from 1990 to 2021 stratified by sex. The ASMR and ASDR of LOCC attributable to different tobacco use among 204 countries and territories from 1990 to 2021 stratified by SDI levels. Age-stratified burden of LOCC (1990 vs 2021) The age-specific burden of tobacco-attributable LOCC increased with advancing age. In both 1990 and 2021, the ASMR began to rise noticeably from the 30 to 34 years age group. The peak ASMR shifted slightly from the 95+ years age group in 1990 to the 90 to 94 years group in 2021 (Figure 6). For DALYs, the burden peaked in the 60 to 64 years age group in 1990 and remained concentrated in the 60 to 69 years range in 2021 (Figure 6). The global ASMR and ASDR burden of LOCC attributable to tobacco use in 1990 and 2021 among different age groups. Cross-country inequality analysis Absolute inequality in the tobacco-attributable LOCC burden, as measured by the Slope Index of Inequality (SII), narrowed considerably between 1990 and 2021. The SII for ASMR decreased from 0.413 to 0.148, and for ASDR from 12.56 to 4.57, indicating a reduction in the burden gap between the highest and lowest SDI countries (Figure 7). Stratification by tobacco type revealed divergent trends. For smoking-attributable LOCC, both absolute (SII) and relative (Concentration Index) inequality decreased, with the SII for ASMR falling from 0.526 to 0.261 and the Concentration Index from 0.177 to 0.112 (Figure S1). In stark contrast, for chewing tobacco, the SII and Concentration Index for both ASMR and ASDR remained virtually unchanged from 1990 to 2021, indicating persistent and unaddressed inequality (Figure S2). Regression curves and concentration curves for health inequalities in LOCC attributable to tobacco use in 1990 and 2021: A, Regression curve for health inequalities in ASMR. B, Concentration curve for ASMR. C, Regression curve for health inequalities in ASDR. D, Concentration curve for ASDR. Correlation between SDI and LOCC burden by tobacco type The relationship between SDI and the age-standardized rates of tobacco-attributable LOCC followed distinct, nonlinear patterns. For total tobacco use, the ASMR increased gradually up to an SDI of approximately 0.4, after which it began to decline (Figure 8A). For smoking-attributable LOCC, a more complex pattern emerged with two inflection points (Figure 8B). The burden rose sharply below an SDI of 0.4, plateaued or slightly decreased between 0.4 and 0.6, increased modestly up to an SDI of ∼0.75, and then declined sharply in high-SDI settings. This indicates that the peak smoking-related burden is concentrated in low-middle and high-middle SDI regions. For chewing tobacco, the relationship mirrored that of total tobacco use, with the burden peaking at low SDI levels and declining thereafter (Figure 8C). The ASDR exhibited trends nearly identical to those of the ASMR (Figure 8D-F). The relationships between the SDI and the ASMR and ASDR burdens of LOCC related to different tobacco use among the 21 GBD regions. A-C, ASMR for tobacco use, tobacco smoking and chewing tobacco. D-F, ASDR for tobacco use, tobacco smoking and chewing tobacco. Projected burden of LOCC attributable to tobacco (2022-2050) The BAPC model projected a continued divergence in the burden attributable to different tobacco products. By 2050, the global ASMR for all tobacco-attributable LOCC is projected to decline by 18.29% relative to 2021 (Figure 9A). This decline is driven almost entirely by a steep 39.23% reduction in the smoking-attributable ASMR (Figure 9B). In stark contrast, no appreciable decline was projected for the ASMR attributable to chewing tobacco through 2050 (Figure 9C). Projections for the ASDR followed an identical pattern, with a 15.79% overall decline, a 43.27% decline for smoking, and a stable trajectory for chewing tobacco (Figure 9D-F). These projections underscore that without targeted intervention, chewing tobacco will remain a persistent and unmitigated cause of LOCC for decades to come. Predicted global trends in ASMR and ASDR for LOCC attributable to different tobacco use from 2022 to 2050. A-C, Predicted ASMR for tobacco use, tobacco smoking and chewing tobacco. D-F, Predicted ASDR for tobacco use, tobacco smoking and chewing tobacco. Discussion Lip and oral cavity cancer (LOCC) remains a significant global health challenge, with incidence and prevalence continuing to rise and considerable variation observed across geographic regions and demographic groups.4,5 Leveraging the most recent data from the GBD 2021 study, this investigation provides the most current and comprehensive assessment of the spatiotemporal trends in the global tobacco-attributable LOCC burden over the past three decades. Notably, this study is the first to profile this burden disaggregated by type of tobacco use – specifically smoking versus chewing tobacco – through analysis of age-standardized mortality and DALY rates (ASMR and ASDR), stratified by sex, age, and region. Our findings reveal that while the absolute numbers of deaths and DALYs from tobacco-attributable LOCC increased substantially between 1990 and 2021, the corresponding age-standardized rates declined globally, particularly for smoking tobacco and among males. In stark contrast, the burden attributable to chewing tobacco exhibited a slight but persistent upward trend, and among females, the tobacco-attributable burden remained virtually unchanged over the study period. It is important to recognize that shifts in absolute death and DALY counts are partly confounded by underlying demographic changes, including population growth and aging.4 Accordingly, our emphasis on age-standardized rates ensures comparability over time and across regions, providing a more accurate reflection of epidemiologic transitions in disease burden. According to the latest GBD studies, LOCC affected over 1.54 million people worldwide in 2021, and the burden exhibits notable disparities linked to geographic region, socioeconomic status, and healthcare access, compounded by risk factors like tobacco use.23 Consistent with previous GBD analyses, we observed a marked increase in the absolute number of tobacco-attributable LOCC deaths and DALYs from 1990 to 2021.24 However, this growth appears to be driven primarily by population aging and expansion, given the concurrent decline in ASMR and ASDR and the strong positive association between age and mortality. Our mapping of the tobacco-related LOCC landscape identified several countries – including Russia, Brazil, and South Africa – that achieved notable reductions in ASMR, while Kyrgyzstan, South Africa, and Australia exhibited significant declines in ASDR. These favorable trends may reflect improvements in LOCC management and healthcare access. Moreover, these declines are partially attributable to global reductions in smoking prevalence: between 1990 and 2020, smoking prevalence decreased by 27.2% among men and 37.9% among women worldwide.25 Nevertheless, the overall downward trajectory in ASMR and ASDR remained modest in most countries, underscoring the persistent and pervasive nature of this public health challenge. A particularly concerning finding is the concentration of high tobacco-attributable LOCC burden in South Asia, corroborating prior reports.24 Countries such as Pakistan and India consistently ranked among the highest globally for both ASMR and ASDR over the three-decade study period, with minimal evidence of decline. This enduring burden likely reflects deeply entrenched cultural practices surrounding tobacco use, coupled with insufficient implementation of effective tobacco control policies.14 These observations underscore an urgent need for enhanced surveillance and targeted public health interventions in this region. The global burden of tobacco-related disease arises from three primary exposures: smoked tobacco, smokeless tobacco (including chewing tobacco), and secondhand smoke. According to the GBD Tobacco Collaborators, the global age-standardized prevalence of smoking in 2019 was 28.5% among males and 5.96% among females, whereas for chewing tobacco, it was 6.55% among males and 2.87% among females.26,27 Secondhand smoke exposure also remains a significant public health issue, affecting approximately 40% of children, 33% of male nonsmokers, and 35% of female nonsmokers globally.28 While prevalence varies by product type, tobacco use remains the leading preventable cause of cancer worldwide, and cessation at any age confers substantial health benefits, including reduced cardiovascular risk, improved lung function, and decreased cancer incidence.29,30 Specifically, smoking increases LOCC risk fivefold and oral cavity cancer mortality 3.3-fold, whereas sustained cessation for 3 to 5 years reduces risk by approximately 50%.12,31 Comparable benefits have been observed with reduced use of chewing tobacco.32, 33, 34 Our temporal analysis by tobacco type reveals that the global LOCC burden attributable to smoking has consistently exceeded that attributable to chewing tobacco over the past 30 years, likely reflecting the higher global prevalence of smoking. However, a striking divergence emerged: while the smoking-attributable burden declined markedly, the chewing tobacco-attributable burden exhibited a slight but persistent upward trend, consistent with previous findings.16 This discrepancy may be explained by contrasting prevalence trends: between 1990 and 2019, the global age-standardized prevalence of smoking decreased significantly (annualized rate of change: −1.21%), whereas chewing tobacco prevalence showed a modest increase.26 These data suggest that while global tobacco control efforts have achieved measurable success in reducing smoking, preventive measures targeting chewing tobacco remain inadequate and underprioritized. Age-specific patterns further illuminate the burden dynamics. Across the study period, ASDR peaked earlier (60-64 years) than ASMR (>90 years), indicating that morbidity manifests at younger ages while mortality is deferred, potentially reflecting improvements in clinical management and survival despite persistent early-life risk exposures. Demographic factors – particularly population growth and aging – remain primary drivers of the absolute tobacco-attributable burden.16 Our BAPC projections to 2050 indicate a continued global decline in LOCC burden attributable to tobacco use overall, driven predominantly by reductions in smoking-related burden (projected decrease of 39.23% in ASMR). In stark contrast, the chewing tobacco-attributable burden is projected to remain essentially unchanged, with no appreciable decline. These forecasts carry critical public health implications: without intensified and targeted interventions – including taxation, comprehensive smoke-free policies, cessation support, and advertising bans – chewing tobacco will persist as an unmitigated cause of LOCC for decades.15 From a methodological standpoint, while BAPC projections provide plausible future trends, the width of uncertainty intervals varies across regions and tobacco types. Future research could enhance predictive accuracy by integrating advanced modelling strategies – such as multiplex network approaches – to capture complex interactions among multiple risk factors and demographic processes. Sex-specific disparities in tobacco-attributable LOCC burden followed sharply divergent patterns. From 1990 to 2021, males bore a substantially higher burden from smoking than from chewing tobacco, consistent with the higher global prevalence of smoking among men. Conversely, among females, the chewing tobacco-attributable burden predominated. Although global smoking prevalence among females exceeds chewing tobacco prevalence, the disparity is far less pronounced than in males.26,27 Critically, in seven of the twelve countries with the highest age-standardized chewing tobacco prevalence among females, chewing tobacco use exceeded smoking.26 Many of these countries – including Bangladesh, India, Bhutan, and Cambodia – also rank among the highest globally for LOCC burden, and a positive correlation between chewing tobacco prevalence and LOCC burden in females has been documented (r = 0.322, P < .001).35 Therefore, there is a possibility that the countries with the highest prevalence of chewing tobacco use have the heaviest burden globally, which accounts for the major contribution among females. Furthermore, the upward trend in chewing tobacco-attributable burden was more pronounced among females in this study, paralleling a higher annualized rate of increase in chewing tobacco prevalence among women (0.60%) compared to men (0.39%) over the past 30 years.26 Tobacco cultivation remains economically significant in many regions, particularly South Asia, where climate and soil conditions favour production. In several countries, chewing tobacco represents a culturally entrenched habit, especially among women.36 Between 1990 and 2019, countries with high chewing tobacco use exhibited virtually no decline in prevalence, even as smoking prevalence decreased substantially in the same settings,26 which was in collaboration with our findings. This stagnation likely reflects a combination of factors: less consistent application of the WHO Framework Convention on Tobacco Control (FCTC) provisions to smokeless tobacco, complex cultural norms, and targeted marketing.26 These findings underscore the imperative for sex-specific and product-specific tobacco control strategies that explicitly address smokeless tobacco use among both men and women. Beyond demographic and behavioural factors, cross-national disparities in LOCC burden are shaped by broader socioeconomic development and health system capacity.16 Our analyses confirm that SDI is a critical determinant of health inequalities for LOCC, with significantly higher tobacco-attributable burden concentrated in low-middle and low SDI regions. While high and high-middle SDI regions achieved substantial reductions in tobacco-related LOCC burden over the study period, rates in lower SDI settings remained virtually unchanged, aligning with previous study.24 This divergence likely reflects the stronger tobacco control policies, broader access to cancer screening, and improved treatment availability characteristic of higher-SDI countries.16 In contrast, low-resource settings face persistent exposure to multiple risk factors – including chewing tobacco, alcohol, and betel quid – compounded by limited preventive and diagnostic services, weak health policy implementation, and challenging living conditions.16 Importantly, our study reveals that the predominant tobacco type contributing to LOCC burden varies systematically with SDI. In high, high-middle, and middle SDI regions, smoking was the primary contributor; in low-middle and low SDI regions, chewing tobacco predominated. This pattern aligns with the higher prevalence of smokeless tobacco use in populations with lower income, educational attainment, and rural residence – characteristics concentrated in lower SDI settings.37 Studies have shown that patients in these groups often face barriers to timely diagnosis, standardized treatment, and comprehensive management, contributing to poorer survival outcomes and elevated burden.4 Although health inequalities related to smoking have narrowed over the study period, disparities in chewing tobacco-attributable burden have remained largely unchanged. The nonlinear relationships between SDI and LOCC burden by tobacco type further underscore the complexity of tobacco control across development levels and highlight the need for context-specific interventions. While the current trend for tobacco smoking is encouraging with effective policies in tobacco control, our findings reinforce the feasibility and urgency of targeted policy formulation: reallocating healthcare resources, strengthening health education, and developing tailored control strategies for low-middle SDI populations and for chewing tobacco users. This study advances the evidence base by providing policymakers, public health officials, and advocacy organizations with a comprehensive set of comparable estimates to inform strategies addressing this harmful exposure. Nevertheless, several limitations must be considered when interpreting our findings. First, although GBD 2021 encompasses 204 countries and territories, variability in data availability and quality across regions may affect analytic accuracy. In low SDI regions, limited healthcare infrastructure and data collection systems may lead to underestimation of tobacco-attributable LOCC mortality. Reliance on diverse data sources – including vital registration and verbal autopsy – may introduce diagnostic biases. Second, the GBD database includes only publicly accessible data, potentially omitting relevant unpublished sources. Third, as this is a secondary analysis of aggregated data, we lack individual-level covariates such as race, education, and occupation, which could aid in confounding control. The robustness of GBD estimates is inherently contingent upon the quality and scope of underlying input data. Fourth, due to data limitations, our analysis focused on smoking and chewing tobacco, precluding examination of secondhand smoke or emerging tobacco products. The increasing diversity of tobacco products in recent years suggests that our findings may not fully capture the contemporary landscape of tobacco-attributable LOCC burden. Fifth, our analysis focused only on smoking and chewing tobacco. In high‑burden regions like South Asia, areca nut and alcohol use are major LOCC risk factors.38 Their omission may underestimate the attributable burden, our findings therefore reflect tobacco‑attributable burden only, not the total preventable burden. Meanwhile, since this study utilizes aggregated data rather than individual-level data, and the aggregated GBD data do not allow interaction analyses (eg, tobacco with alcohol or areca nut), causal relationships cannot be inferred. Lastly, while the BAPC model provides valuable projections to 2050, long‑term forecasts inherently involve greater uncertainty due to potential changes in risk factors, healthcare, and demographics. Its assumptions regarding trend continuity may not fully capture nonlinear or irregular fluctuations as well. Nevertheless, the 2050 horizon captures the full impact of population aging and aligns with global planning benchmarks. We present 95% uncertainty intervals that widen over time, and the 2050 estimates should be interpreted as conditional on the continuation of historical trends. Conclusion In conclusion, this comprehensive analysis of the global burden of LOCC attributable to tobacco use from 1990 to 2021 reveals a complex and evolving epidemiologic landscape. While progress has been made in reducing the smoking-attributable burden – particularly in high-SDI settings and among males – the persistent and, in some contexts, increasing burden from chewing tobacco, especially among females and in low-SDI regions, demands urgent attention. Projections to 2050 underscore that without targeted intervention, chewing tobacco will remain an unmitigated cause of LOCC for decades. In line with this, our findings substantiate the necessity of accelerating the enforcement of comprehensive tobacco-free legislation, designing evidence-based health interventions, and optimizing the allocation of health resources according to region-specific risk profiles. Critically, tobacco control strategies must be expanded and differentiated to address both smoked and smokeless products, tailored to sex-specific and region-specific patterns, and designed to prioritize equity in access to prevention, early detection, and treatment. Furthermore, the development of robust population-based cancer registries, enhanced data collection systems, and investment in regionally tailored research are essential for formulating effective strategies to mitigate the growing global burden of LOCC. Compliance with ethics requirements As this study utilized deidentified, aggregated, and publicly available data from online source, it was exempt from ethical review. Author contributions ZJY: Conceptualization, Data curation, Formal analysis, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing. MZ: Formal analysis, Methodology, Validation, Visualization, Writing – review & editing. YZ: Formal analysis, Methodology, Validation, Visualization, Writing – review & editing. JS: Data curation, Methodology, Validation, Writing – review & editing. SZH: Funding acquisition, Data curation, Formal analysis, Writing – review & editing. XC: Validation, Writing – review & editing. YF: Validation, Writing – review & editing. HYL: Data curation, Writing – review & editing. XLL: Conceptualization, Investigation, Project administration, Resources, Supervision, Validation, Writing – review & editing. GXW: Conceptualization, Investigation, Project administration, Resources, Supervision, Validation, Writing – original draft, Writing – review & editing. Ethics statement Not applicable. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work the authors used Deepseek in order to improve the language and readability. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article. Funding This study was supported by a grant from S &T Program of Chengde (grant number: 202404B095), Hebei Province, China. Conflict of interest None disclosed. Acknowledgments Acknowledgements The authors wish to express their sincere gratitude to the Global Burden of Disease (GBD) study for access to its comprehensive data. We acknowledge the dedication and collaborative efforts of the GBD researchers and affiliated institutions, whose work provides essential insights into global health trends. Data availability All data in this study are publicly available through the GBD 2021 portal, which cloud be accessible at https://vizhub.healthdata.org/gbd-results/. Footnotes Supplementary material associated with this article can be found in the online version at doi:10.1016/j.identj.2026.109715. Contributor Information Xiaoling Liu, Email: liuxiaoling007@126.com. Guoxing Wan, Email: 15gxwan@stu.edu.cn. Supplementary materials Fig. S1 Regression curves and concentration curves for health inequalities in LOCC attributable to tobacco smoking in 1990 and 2021: (A) Regression curve for health inequalities in ASMR; (B) Concentration curve for ASMR; (C) Regression curve for health inequalities in ASDR; (D) Concentration curve for ASDR. Fig. S2 Regression curves and concentration curves for health inequalities in LOCC attributable to chewing tobacco in 1990 and 2021: (A) Regression curve for health inequalities in ASMR; (B) Concentration curve for ASMR; (C) Regression curve for health inequalities in ASDR; (D) Concentration curve for ASDR. References Associated Data Supplementary Materials Fig. S1 Regression curves and concentration curves for health inequalities in LOCC attributable to tobacco smoking in 1990 and 2021: (A) Regression curve for health inequalities in ASMR; (B) Concentration curve for ASMR; (C) Regression curve for health inequalities in ASDR; (D) Concentration curve for ASDR. Fig. S2 Regression curves and concentration curves for health inequalities in LOCC attributable to chewing tobacco in 1990 and 2021: (A) Regression curve for health inequalities in ASMR; (B) Concentration curve for ASMR; (C) Regression curve for health inequalities in ASDR; (D) Concentration curve for ASDR. Data Availability Statement All data in this study are publicly available through the GBD 2021 portal, which cloud be accessible at https://vizhub.healthdata.org/gbd-results/.