Different change patterns and sex difference for stroke and its subtypes mortality attributable to tobacco in China from 1990 to 2021

Abstract

Background

Tobacco use is one of the major risk factors of stroke. However, the dynamic changes in stroke attributable to tobacco over the past three decades in China remain unclear, as does the potential differential effect of tobacco on different stroke subtypes.

Methods

Data was derived from the Global Burden of Disease 2021 study. Tobacco use was defined as current or past usage of any smoked tobacco product and the average daily exposure to air particulate matter from second-hand smoking. Joinpoint regression and age-period-cohort analysis were adopted to analyze the trends in stroke mortality due to tobacco.

Result

From 1990 to 2021, the age-standardized mortality rates (ASMR) for stroke attributable to tobacco exhibited a declining trend. Ischemic stroke (IS) exhibited the smallest decline, while subarachnoid hemorrhage (SAH) showed the largest decline. The curve of the net drifts showed a downward trend for all stroke subtypes irrespective of sex. Among stroke subtypes, mortality reductions were most pronounced in SAH, followed by intracerebral hemorrhage (ICH), and IS. The mortality of IS and ICH increased rapidly with age, while the mortality rate for SAH showed fluctuations with age. Both period and cohort rate ratios for stroke mortality attributable to tobacco have declined, with the most significant reduction observed in SAH and the least in IS. Stroke mortality attributable to tobacco in men was chiefly associated with active smoking, whereas in women, second-hand smoking was the predominant factor, exhibiting a gradual upward trend.

Conclusion

In the past three decades, the risk ratio and ASMR of stroke attributable to tobacco showed the declining trend, and different stroke subtypes exhibited different change patterns in China. There was a gender disparity in the effects of active smoking and second-hand smoke exposure on men and women. Policies aimed at preventing tobacco-related stroke in the future must take the varying effects across different stroke subtypes and genders into account.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-24336-1.

Introduction

Stroke is estimated to be the second-leading cause of death and the third-leading cause of disability and death globally, with China confronting the greatest burden of stroke among all nations [1, 2]. Stroke survivors commonly experience difficulty in managing daily activities because of physical and cognitive deficits, including limb weakness, speech problems, fatigue, and cognitive impairments, bringing a heavy burden to family and society [3]. With the aging of populations, the incidence of stroke and its long-term effects, along with the associated healthcare costs, are anticipated to increase significantly [4]. Therefore, preventing stroke by targeting modifiable risk factors is a critical focus for public health strategies.

Tobacco smoking is one of the most important modifiable risk factors of stroke, contributing significantly to the global burden of stroke [1]. Tobacco contains more than 7,000 chemical compounds, including tar, nicotine, and carbon monoxide, which impair vascular endothelial function, increase blood viscosity, and reduce oxygen transport, thus raising the risk of stroke [5, 6]. Second-hand smoking exposure could also adversely influence the onset and prognosis of stroke. Research indicates that the risk of stroke-related death increases by 10% in individuals living with a smoking spouse, and nonsmoking women living with smoking husbands experience a higher stroke prevalence, with the likelihood increasing as the intensity and duration of the husband's smoking habits rise [7, 8]. Smoking cessation offers significant and rapid benefits in reducing stroke risk, especially in light smokers. For stroke survivors, quitting smoking is linked to a decrease in recurrent vascular events and mortality [9, 10]. As the world's largest tobacco consumer, China bears a significant burden from tobacco-related diseases, with stroke being especially prevalent [11]. Therefore, investigating the dynamic changes in stroke associated with tobacco use, and implementing targeted public health policies to control tobacco consumption, could serve as an effective and cost-efficient strategy to reduce the stroke burden in China.

While previous studies have simply reported stroke mortality, disability-adjusted life years, and tobacco-related stroke trends in China, no research has specifically examined stroke mortality attributable to tobacco across different time periods, age groups, or cohorts. Furthermore, there is a lack of studies that separately analyze the temporal trends and disparities in stroke subtypes, including ischemic stroke (IS), intracerebral hemorrhage (ICH), and subarachnoid hemorrhage (SAH) in China [12, 13]. Moreover, the differential impact of active and second-hand smoking on stroke between men and women has not been adequately addressed. In this context, we employed joinpoint regression and an age-period-cohort model to estimate the mortality from stroke and its subtypes attributable to tobacco over the past three decades in China. Additionally, we examined gender differences in the effects of active and second-hand smoking on stroke risk. The findings could inform the development of targeted public health policies to address the specific needs of different population groups.

Methods and materials

Data source

The data on the attributable burden of stroke, both in China and globally, was derived from the Global Burden of Disease (GBD) 2021. This extensive study, conducted by the Institute for Health Metrics and Evaluation (IHME) in partnership with global collaborators, provides a thorough assessment of age- and sex-specific, as well as cause-specific, incidence and mortality rates for 369 diseases and injuries across 204 countries and territories, spanning from 1990 to 2021 [14]. The anonymized data are freely available to the public on the IHME website and can be accessed online (http://ghdx.healthdata.org/gbd-results-tool).

Stroke mortality data for the Chinese population were sourced from the Cause of Death Reporting System of the Chinese Center for Disease Control and Prevention and the Disease Surveillance Points, which offer nationally representative statistics [15]. Subsequently, the data were processed using the Cause of Death Ensemble modeling framework in GBD 2021 to accurately account for all pathological types [1]. Stroke was diagnosed and defined according to the WHO clinical criteria and the International Statistical Classification of Diseases. The stroke subtypes, including IS, ICH, and SAH, were classified using the tenth revision of the International Classification of Diseases and Injuries. In the GBD 2021 study, tobacco use encompassed multiple modalities, including current or past usage of any smoked tobacco product, current use of any chewing tobacco product, and the average daily exposure to air particulate matter from second-hand smoking [16]. Specifically, smoking exposure is defined as current or past use of any tobacco product, excluding electronic cigarettes or vaporizers, and second-hand smoke exposure was defined as inhalation of ambient tobacco smoke in residential, occupational, and public environments [17–19].

The informed consent was reviewed and approved by the University of Washington Institutional Review Board. To ensure consistency across rates of incidence, prevalence, remission, and cause of death for the condition under study, the primary estimation method employed was DisMod-MR 2.1, a Bayesian meta-regression tool. This approach enabled the comprehensive annual estimation of global, regional, and national incidence, prevalence, mortality, as well as causes of death and associated risk factors across 204 nations and territories from 1990 to 2021 [20]. The strength of the GBD approach lies in its application of standardized methodologies to critically assess available data for each condition. This ensures comparability and systematic evaluation, facilitates the estimation of results from countries with incomplete data, and allows for the reporting of disease burden using consistent, standardized metrics [21].

Statistical analysis

The Joinpoint regression model, a series of Linear statistical models, was employed to analyze trends in tobacco-attributable stroke and its subtypes from 1990 to 2021. Through this analysis, long-term trend lines were segmented into distinct phases, each characterized by continuous linearity, based on the model's fitting process [22]. Temporal trends in the age-standardized mortality rate (ASMR) for stroke and its subtypes from 1990 to 2021 were modeled using one to five joinpoints to identify the optimal fit. The average annual percentage change (AAPC), as well as the annual percentage change (APC) for each segment, along with their corresponding 95% confidence intervals (CIs), were calculated to assess the direction and Magnitude of the trends. Joinpoint analysis was performed using Joint Command Line Version 4.9.1.0, made available by the Surveillance Research Program of the U.S. National Cancer Institute.

Age-period-cohort models are commonly employed in both sociology and epidemiology to examine disease trends in relation to age, population-level demographic shifts (period), and the effects of early-life or generational exposures (cohort) [23]. The age effect refers to the biological and social processes associated with aging. The period effect reflects changes in stroke mortality and its subtypes due to tobacco over time, affecting all age groups simultaneously. The cohort effect refers to changes resulting from distinct risk factors and environmental exposures experienced by a specific group (cohort) over time [24]. In this study, the data were stratified into successive 5-year age groups, consecutive 5-year periods, and corresponding 5-year birth cohorts. The incidence and mortality rates for stroke and its subtypes attributable to tobacco were recorded for age groups (from 25–29 to 90–94 years), periods (from 1992 to 2021), and birth cohorts (from 1892–1897 to 1992–1997). Since tobacco-attributable stroke mortality in individuals under 25 years was rare, and all individuals aged 95 years and older were grouped together in the GBD 2021 database, these groups were excluded from the analysis. Data from 1990 to 1992 were excluded as they did not constitute a complete 5-year period. The reference groups were defined as the 55–59 years old age group, the 2002–2006 period, and the 1947 birth cohort. The age-period-cohort web tool (Biostatistics Branch, National Cancer Institute, Bethesda, MD; https://analysistools.nci.nih.gov/apc/) was used for the age-period-cohort analysis. All statistical tests were two-sided, and p < 0.05 was considered significant.

Result

We analyzed the trend in ASMR for stroke and its subtypes attributable to tobacco in China and globally from 1990 to 2021, as shown in Fig. 1A. Among the stroke subtypes attributed to tobacco, the ASMR was highest for ICH, followed by IS and SAH, and they all exhibited a downward trajectory from 1990 to 2021 both in China and globally. The ASMR for stroke and its subtypes in China were notably higher than those globally. Between 1990 and 2021, global ASMRs per 100,000 attributed to tobacco decreased from 23.6 to 12.7 for stroke, from 9.5 to 5.4 for IS, from 12.0 to 6.6 for ICH, and from 2.8 to 0.7 for SAH. In China, the ASMRs per 100,000 attributed to tobacco dropped from 50.0 to 25.6 for stroke, from 14.3 to 10.9 for IS, from 29.8 to 13.7 for ICH, and from 5.8 to 0.9 for SAH.

Fig. 1.

Trends in age-standardized mortality rates (ASMRs) of stroke and its subtypes attributable to tobacco1990 to 2021. A Trends of ASMRs for both sexes in China and globally. B Trends of ASMRs for men and women in China. Abbreviations: ICH, intracerebral hemorrhage; IS, ischemic stroke; SAH, subarachnoid hemorrhage

From a gender perspective, in China, the ASMR for stroke and its subtypes attributable to tobacco was significantly higher in males than in females (Fig. 1B). Tobacco use in the GBD 2021 study was primarily categorized into active smoking and second-hand smoking. Figure 2 illustrates the proportion of stroke mortality attributable to these two factors in both men and women from 1990 to 2021. In men, stroke mortality attributable to tobacco was predominantly driven by active smoking, whereas in women, it was mainly influenced by second-hand smoking, which showed a gradual upward trend (From 59.3% in 1990 to 67.1% in 2021) (Fig. 2).

Fig. 2.

Proportion of stroke attributable to smoking and second-hand smoking by gender from 1990 to 2019 in China

Joinpoint regression analysis was conducted to examine the dynamic changes in sex-specific ASMR for stroke and its subtypes in China and globally (Table 1 and Table S1). In China, IS exhibited the smallest decline (AAPC = −0.91, 95%CI: −1.23 to −0.59), while SAH showed the largest decline (AAPC = −5.77, 95% CI: −6.09 to −5.44). The joinpoint analysis indicated that the most significant declines for stroke, IS, ICH, and SAH occurred between 2004 and 2007. Both male and female ASMRs for stroke and its subtypes exhibited a downward trend, with women showing more pronounced declines than men.

Table 1.

Trends of age-standardized mortality rates of stroke and its subtypes attributable to tobacco from 1990 to 2019 in China using joinpoint regression

	Both sexes				Male				Female
	Segment	Period	APC (95%CI)	p	Segment	Period	APC (95%CI)	p	Segment	Period	APC (95%CI)	p
Stroke	1	1990 ~ 1999	−1.72 (−1.90 to −1.55)	< 0.001	1	1990–1999	−1.63 (−1.77 to −1.49)	< 0.001	1	1990–1994	−1.06 (−1.71 to −0.41)	< 0.001
	2	1999–2004	−0.06 (−0.67 to 0.56)	0.851	2	1999–2004	1.23 (0.74 to 1.73)	< 0.001	2	1994–2004	−2.32 (−2.50 to −2.13)	< 0.001
	3	2004–2007	−5.14 (−6.99 to −3.26)	< 0.001	3	2004–2007	−4.42 (−5.89 to −2.91)	< 0.001	3	2004–2007	−6.86 (−8.80 to −4.87)	< 0.001
	4	2007–2010	−0.73 (−2.68 to 1.25)	0.440	4	2007–2010	−0.29 (−1.83 to 1.27)	0.696	4	2007–2010	−2.85 (−4.87 to −0.79)	0.010
	5	2010–2018	−3.41 (−3.66 to −3.16)	< 0.001	5	2010–2018	−3.21 (−3.41 to −3.02)	< 0.001	5	2010–2015	−5.43 (−6.05 to −4.82)	< 0.001
	6	2018–2021	−1.98 (−2.87 to −1.08)	< 0.001	6	2018–2021	−1.97 (−2.66 to −1.27)	< 0.001	6	2015–2021	−2.70 (−3.02 to −2.38)	< 0.001
	AAPC	1990–2021	−2.16 (−2.44 to −1.88)	< 0.001	AAPC	1990–2021	−1.76 (−1.99 to −1.54)	< 0.001	AAPC	1990–2021	−3.24 (−3.53 to −2.94)	< 0.001
IS	1	1990–1999	−0.71 (−0.91 to −0.50)	< 0.001	1	1990–1999	−0.78 (−1.01 to −0.56)	< 0.001	1	1990–2004	−0.42 (−0.57 to −0.27)	< 0.001
	2	1999–2004	1.96 (1.24 to 2.69)	< 0.001	2	1999–2004	3.40 (2.57 to 4.23)	< 0.001	2	2004–2007	−5.78 (−8.79 to −2.68)	< 0.001
	3	2004–2007	−4.30 (−6.48 to −2.08)	< 0.001	3	2004–2007	−3.53 (−5.97 to −1.03)	0.008	3	2007–2010	−1.25 (−4.36 to 1.95)	0.419
	4	2007–2010	1.52 (−0.81 to 3.89)	0.188	4	2007–2010	2.07 (−0.51 to 4.73)	0.110	4	2010–2015	−3.99 (−4.93 to −3.03)	< 0.001
	5	2010–2021	−2.07 (−2.22 to −1.93)	< 0.001	5	2010–2021	−1.91 (−2.07 to −1.75)	< 0.001	5	2015–2021	−2.13 (−2.62 to −1.64)	< 0.001
	AAPC	1990–2021	−0.91 (−1.23 to −0.59)	< 0.001	AAPC	1990–2021	−0.52 (−0.88 to −0.16)	0.004	AAPC	1990–2021	−1.94 (−2.38 to −1.50)	< 0.001
ICH	1	1990–1999	−1.73 (−1.87 to −1.58)	< 0.001	1	1990–1999	−1.61 (−1.76 to −1.46)	< 0.001	1	1990–2000	−2.08 (−2.29 to −1.87)	< 0.001
	2	1999–2004	0.64 (0.11 to 1.18)	0.021	2	1999–2004	1.77 (1.22 to 2.32)	< 0.001	2	2000–2004	−1.28 (−2.64 to 0.09)	0.066
	3	2004–2007	−5.54 (−7.12 to −3.92)	< 0.001	3	2004–2007	−4.78 (−6.41 to −3.11)	< 0.001	3	2004–2007	−7.71 (−10.25 to −5.09)	< 0.001
	4	2007–2010	−1.70 (−3.37 to −0.01)	0.0491	4	2007–2010	−1.61 (−3.30 to 0.12)	0.065	4	2007–2010	−3.83 (−6.47 to −1.12)	0.009
	5	2010–2015	−4.92 (−5.43 to −4.41)	< 0.001	5	2010–2017	−4.30 (−4.58 to −4.03)	< 0.001	5	2010–2015	−6.69 (−7.49 to −5.87)	< 0.001
	6	2015–2021	−3.05 (−3.31 to 2.78)	< 0.001	6	2017–2021	−2.44 (−2.93 to −1.94)	< 0.001	6	2015–2021	−3.22 (−3.65 to −2.79)	< 0.001
	AAPC	1990–2021	−2.50 (−2.74 to −2.26)	< 0.001	AAPC	1990–2021	−2.11 (−2.35 to −1.86)	< 0.001	AAPC	1990–2021	−3.68 (−4.08 to −3.27)	< 0.001
SAH	1	1990–1995	−1.91 (−2.28 to −1.54)	< 0.001	1	1990–1995	−2.06 (−2.46 to −1.65)	< 0.001	1	1990–1994	−0.92–1.79 to −0.05	0.040
	2	1995–2000	−7.71 (−8.22 to −7.19)	< 0.001	2	1995–2000	−6.98 (−7.58 to −6.37)	< 0.001	2	1994–1997	−6.80 (−9.40 to −4.13)	< 0.001
	3	2000–2003	−15.29 (−17.11 to −13.43)	< 0.001	3	2000–2006	−13.22 (−13.86 to −12.57)	< 0.001	3	1997–2000	−10.61 (−13.43 to −7.71)	< 0.001
	4	2003–2006	−13.06 (−15.44 to −10.62)	< 0.001	4	2006–2015	−1.98 (−2.39 to −1.57)	< 0.001	4	2000–2004	−17.66 (−19.40 to −15.88)	< 0.001
	5	2006–2016	−2.40 (−2.64 to −2.16)	< 0.001	5	2015–2021	−3.84 (−4.37 to −3.31)	< 0.001	5	2004–2007	−11.74 (−15.80 to −7.49)	< 0.001
	6	2016–2021	−3.59 (−4.12 to −3.07)	< 0.001	6	…	…	…	6	2007–2021	−2.92 (−3.12 to −2.72)	< 0.001
	AAPC	1990–2021	−5.77 (−6.09 to −5.44)	< 0.001	AAPC	1990–2021	−5.43 (−5.65 to −5.21)	< 0.001	AAPC	1990–2021	−6.71 (−7.30 to −6.11)	< 0.001

Abbreviations: AAPC average annual percentage change, APC annual percentage change, ICH intracerebral hemorrhage, IS ischemic stroke, SAH subarachnoid hemorrhage

We further employed age-period-cohort analysis to calculate local drift values and estimate the AAPC in stroke mortality attributable to tobacco across age groups in China, as shown in Fig. 3. The local drift curve for stroke and its subtypes exhibited a U-shaped pattern across age groups for both sexes, with the most pronounced declines observed among middle-aged and older individuals. Almost all local drift values across age groups were less than 0, irrespective of sex, indicating a gradual reduction in tobacco-related stroke mortality across all age groups. Notably, nearly all local drift values for women were lower than those for men. Among stroke subtypes, mortality reductions were most pronounced in SAH, followed by ICH and IS.

Fig. 3.

Local drift for stroke mortality attributable to tobacco for both sex (A), men (B), and women (C) in China. Local drift values represent the annual percentage change in the respective age group. The local drift values < 0, which revealed a decreasing trend of high SBP-attributable stroke and subtype mortality across the study period. Error bars represent the 95% CIs for the local drift values. Abbreviations: ICH, intracerebral hemorrhage; IS, ischemic stroke; SAH, subarachnoid hemorrhage

The longitudinal age curves of tobacco-attributable stroke mortality rates, stratified by subtypes and gender, were presented in Fig. 4. The analysis revealed a pronounced increase in stroke mortality rates attributable to tobacco with advancing age. Among the subtypes, the steepest age-related increase was observed in IS, with mortality rates rising from 0.07 per 100,000 in the 25–29 age group to 336.35 per 100,000 in the 90–94 age group. For ICH, mortality increased from 0.61 to 192.42 per 100,000 over the same age range. IS mortality exceeded ICH mortality starting from the 75–80 age group. In contrast, SAH mortality remained consistently the lowest across all age groups for both genders. The mortality rate for SAH showed fluctuations with age, peaking in the 50–60 age group.

Fig. 4.

Longitudinal age curves of tobacco-attributable stroke mortality and subtypes for both sex (A), men (B), and women (C) in China. Fitted longitudinal age-specific rates of stroke mortality rates (per 100,000 person-years) and the corresponding 95% confidence intervals. Abbreviations: ICH, intracerebral hemorrhage; IS, ischemic stroke; SAH, subarachnoid hemorrhage

The estimated period and cohort effects on tobacco-attributable stroke mortality and its subtypes by gender in China were shown in Figs. 5 and 6. The results indicate a general decline in the risk of tobacco-attributable stroke mortality from 1992 to 2021, with the most pronounced decreases in SAH and the least in IS (from 2.64 to 0.53 for SAH, and from 1.07 to 0.79 for IS). In terms of cohort effects, SAH showed the most substantial reduction across successive birth cohorts, while the decline in cohort effects for IS was more gradual (from 25.50 to 0.08 for SAH, and from 1.31 to 0.60 for IS). Despite women having a higher risk of tobacco-attributable stroke mortality in earlier periods or birth cohorts, both genders exhibited similar trends of monotonic decline in period and cohort effects. This finding suggests that women may experience a more favorable trend than men in the context of tobacco-attributable stroke mortality.

Fig. 5.

Period rate ratios (RRs) of tobacco–attributable stroke mortality for both sex (A), men (B), and women (C) in China. The RRs of the respective period compared with the reference period (2002–2007) adjusted for age and nonlinear cohort effects and the corresponding 95% CI. Abbreviations: ICH, intracerebral hemorrhage; IS, ischemic stroke; SAH, subarachnoid hemorrhage

Fig. 6.

Cohort rate ratios (RRs) of tobacco–attributable stroke mortality for both sex (A), men (B), and women (C) in China. The RRs of the respective cohort compared with the reference cohort (1947s) adjusted for age and nonlinear period effects and the corresponding 95% CI. Abbreviations: ICH, intracerebral hemorrhage; IS, ischemic stroke; SAH, subarachnoid hemorrhage

Discussion

In this study, we investigate the long-term trends in tobacco-attributable stroke mortality and its subtypes in China from 1990 to 2021, utilizing joinpoint analysis and the age-period-cohort framework with data from the GBD 2021 study. Our findings indicate that the ASMR for tobacco-attributable stroke and its subtypes in China are substantially higher than the global averages, with Male ASMR significantly exceeding those of females. From 1990 to 2021, the ASMR of all stroke subtypes demonstrated a downward trend in both China and globally, with SAH showing the greatest decline and IS exhibiting the least reduction. Mortality rates of all stroke types declined across nearly all age groups, with the most substantial decreases observed in middle-aged and older populations. Tobacco-related stroke mortality increased dramatically with age, but decreased significantly over time and across different birth cohorts. Moreover, in men, stroke mortality linked to tobacco was primarily driven by active smoking, while in women, second-hand smoke was the predominant factor.

Over the past 30 years, tobacco use has contributed to more than 200 million deaths, imposing a great economic burden on the family and society [25, 26]. As a primary risk factor for stroke, smoking increases the risk nearly twofold, with a clear dose–response relationship between the number of pack-years and the likelihood of stroke. Substantial evidence indicates that smoking cessation reduces stroke risk, with the Magnitude of risk reduction correlating positively with the duration of abstinence and the additional risk almost vanishing within 2 to 4 years after quitting [27–29]. Between 1990 and 2021, stroke mortality attributable to tobacco has also declined, highlighting the effectiveness of tobacco control policies in reducing stroke risk. According to the GBD 2019 study, since 1990, the global age-standardized prevalence of tobacco use has declined by 27.5% among Males aged 15 years and older, and by 37.7% among females. However, despite these reductions, the age-standardized prevalence of tobacco use remains high, with 32.7% of Males and 6.62% of females using tobacco in 2019 [30].

China faces a significantly higher stroke burden attributable to tobacco compared to global levels, which can be largely explained by the fact that China accounts for over a third of the world’s tobacco consumption, predominantly among men [31]. Since 2005, there has been a significant decline in the ASMRs of stroke attributable to tobacco in China. This trend May be attributed to the implementation of effective tobacco control policies, emphasis on stroke prevention, and the establishment of medical insurance programs. The Chinese government has placed significant emphasis on tobacco control, ratifying the WHO Framework Convention on Tobacco Control in 2005. Since then, tobacco control measures have been progressively implemented across the country [32]. In recent years, consecutive cross-sectional surveys observed a steady decline in smoking prevalence in China [31]. Emphasis on prevention and treatment of stroke also contributed to the decline of stroke mortality. World Stroke Day was established in 2004, marking a pivotal moment in the global recognition of the impact of stroke, prompting society to increasingly focus on its detrimental effects and allocate greater investments toward stroke prevention and treatment [33]. Additionally, the Chinese government funded the establishment and expansion of public medical insurance systems in recent years, including the New Rural Cooperative Medical Scheme for rural residents since 2002 and the Urban Residents'Basic Medical Insurance system since 2007 [34, 35]. The promotion of medical insurance has enhanced the affordability of stroke treatment, contributing to a better prognosis for stroke patients by ensuring timely and effective care [36].

We further extended our analysis to stroke subtypes using the most recent data from the GBD 2021 study, emphasizing the distinct trends in ASMR and relative risks (RRs) for IS, ICH, and SAH attributable to tobacco. Our findings indicate that the mortality risk for tobacco-attributable IS and ICH increased significantly with age, after controlling for cohort and period variations. Aging is the most significant non-modifiable risk factor for IS and ICH, with elderly stroke patients experiencing higher mortality, increased morbidity, and poorer functional recovery compared to their younger counterparts [37, 38]. Additionally, the cumulative damage caused by smoking exacerbates conditions such as hypertension, atherosclerosis, and vascular endothelial dysfunction, which are more pronounced in the elderly [39, 40]. As a result, the elderly are at a higher risk of developing IS and ICH [41]. Therefore, future smoking cessation initiatives should prioritize tailored interventions for elderly smokers, while health screenings for this demographic must integrate comprehensive cerebrovascular disease risk assessment. Period effects refer to the impact of medical advancements, diagnostic methods, and socio-economic or cultural changes within a specific time frame on the ASMR of tobacco-related stroke. The period effect decreased significantly in all stroke subtypes mortality, which can be attributed to the improvements in tobacco control policies and significant progress in the development of stroke centers in China [32, 42]. Of particular interest is the substantial reduction in period RRs for SAH attributable to tobacco. This trend reflects multifactorial determinants beyond tobacco control efforts. China's economic development and increased governmental healthcare investment have expanded imaging equipment accessibility, facilitating early detection and management of asymptomatic aneurysms and SAH [43]. Concurrently, advancements in neurointerventional techniques have improved SAH outcomes. Endovascular intervention demonstrates superior prognosis to traditional craniotomy for ruptured aneurysms, while prophylactic embolization effectively reduces SAH incidence in unruptured aneurysms [44, 45]. Collectively, these neurointerventional advances have reduced SAH occurrence and improved clinical outcomes. Additionally, hypertension constitutes a significant modifiable risk factor for SAH [46, 47]. China has implemented substantial initiatives targeting hypertension prevention and management, yielding measurable progress [48, 49]. Effective control of hypertension is critical for both mitigating SAH incidence and ameliorating its prognosis [50]. The period RRs for IS attributable to tobacco exhibit the smallest reduction, suggesting that greater focus should be placed on preventing IS. The cohort effect emphasizes the influence of socio-economic, behavioral, and environmental factors encountered earlier in life, and how these exposures shape risks across different birth cohorts. In our analysis, we found that the cohort effect on stroke mortality attributable to tobacco has decreased over time. Beyond the aging factor, this reduction is likely due to the higher levels of education and greater health awareness in later birth cohorts, which are less prevalent in earlier generations [51].

Our findings indicate that the ASMR for stroke attributable to tobacco is notably higher in males compared to females, which is consistent with other studies [52, 53]. This gender disparity can be primarily explained by the higher smoking prevalence among men, as reported in the GBD 2019 study [30]. Additionally, estrogen may offer neuroprotective effects, mitigating brain damage after stroke, particularly in premenopausal women [54]. Besides, mortality rates for smoking-attributable stroke subtypes declined substantially faster among women. This gender disparity arises from multiple factors. Female smoking prevalence is decreasing more rapidly than male prevalence, and male smokers consume greater quantities of tobacco, indicating heightened exposure [55, 56]. Secondhand smoke exposure predominantly drives female tobacco-related stroke mortality. Smoking cessation among males reduces secondhand smoke exposure to females, thereby contributing to decreased smoking-associated stroke mortality in women. Additionally, sociocultural norms in China result in lower acceptability of female smoking [55]. Therefore, tobacco control policies should prioritize male populations, due to their higher tobacco-related stroke mortality rates. We also noted that female tobacco related stroke mortality is mainly influenced by second-hand smoking. Indoor smoking could increase exposure to second-hand smoking and adversely affect indoor air quality [57, 58]. The total ban on indoor smoking could significantly reduce the hospitalization rate and mortality of myocardial infarction and stroke [59, 60]. Therefore, the smoking ban in public workplaces and indoors should be strengthened, and more attention should be given to the reduction of second-hand smoking exposure to women.

Our study has important public Health implications. We have outlined the temporal trends in tobacco-attributable stroke mortality in China from 1990 to 2021, highlighting the distinct patterns observed for IS, ICH, and SAH. Additionally, our results partially reflect the impact of tobacco control policies in China. Despite a reduction in tobacco consumption over the past three decades, the levels remain considerably higher than the global average [32]. Therefore, it is crucial to implement more comprehensive and targeted tobacco control measures in China to reduce the burden of stroke and other tobacco-related health conditions.

Several limitations in our study should also be acknowledged. Firstly, e-cigarettes is not included in our study, and e-cigarette use has remained prevalent in Many countries worldwide. Therefore, future research should pay attention to the effect of e-cigarette on stroke. Secondly, the data used in this article are derived from the latest GBD study, which May contain some discrepancies in the completeness and accuracy of stroke mortality data. Despite numerous adjustments and corrections made in the GBD 2021 to improve the accuracy and comparability of stroke mortality attributed to tobacco, some degree of data inaccuracy remains unavoidable. Thirdly, data from subnational regions in China are not included in the GBD 2021, thus preventing a clear assessment of trends across different provinces, states, or urban and rural areas. Fourthly, owing to limitations in the GBD 2021 methodology, our study could not differentiate between current and former smoking status and evaluate their differential impacts on stroke. Lastly, mortality data for individuals younger than 25 and older than 95 years were excluded from the analysis due to the scarcity of such data in the GBD database.

In conclusion, our findings provide compelling evidence that ASMR and the risk of stroke attributable to tobacco have declined over the past three decades in China; however, they remain significantly higher than global averages. It is strongly recommended that the government implement comprehensive tobacco prevention and cessation programs, alongside enforcing a smoke-free environment for all citizens across the country. Future smoking cessation efforts aimed at stroke prevention should focus on men and individuals at high risk for ischemic stroke.

Abbreviations

GBD

Global Burden of Disease

IS

Ischemic stroke

ICH

Intracerebral hemorrhage

SAH

Subarachnoid hemorrhage

ASMR

Age-standardized mortality rate

APC

Annual percentage change

AAPC

Average annual percentage change

Cis

Confidence intervals

RRs

Relative risks

Authors’ contributions

All authors contributed to the study conception and design. Conceptualization: W.C. and L.Z. Writing original draft preparation: J.Z. Literature search: H.J. Reviewing and editing: Y.X and J.S. All authors have read and agreed to the submitted version of the manuscript.

Funding

This work was funded by the National Natural Science Fund of China (No. 82104144), the Fifth Affiliated Hospital of Sun Yat-sen University of Outstanding young talents cultivation program (No. 3320104100322) and “Five Five” young talents program (No. 220904094231).

Data Availability

The Global Burden of Disease study 2021 is an open-access resource; data are available at http://ghdx.healthdata.org/gbd-results-tool.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.