Knowledge From the Afrikan World, For the World

PMC Articles

Global, regional, and national burden of stroke and its risk factors, 1990–2021: a systematic analysis for the Global Burden of Disease Study 2021

PMCID: PMC12254192

PMID: 39304265

Abstract

Summary Background Up-to-date estimates of stroke burden and attributable risks and their trends at global, regional, and national levels are essential for evidence-based health care, prevention, and resource allocation planning. We aimed to provide such estimates for the period 1990–2021. Methods We estimated incidence, prevalence, death, and disability-adjusted life-year (DALY) counts and age-standardised rates per 100 000 people per year for overall stroke, ischaemic stroke, intracerebral haemorrhage, and subarachnoid haemorrhage, for 204 countries and territories from 1990 to 2021. We also calculated burden of stroke attributable to 23 risk factors and six risk clusters (air pollution, tobacco smoking, behavioural, dietary, environmental, and metabolic risks) at the global and regional levels (21 GBD regions and Socio-demographic Index [SDI] quintiles), using the standard GBD methodology. 95% uncertainty intervals (UIs) for each individual future estimate were derived from the 2·5th and 97·5th percentiles of distributions generated from propagating 500 draws through the multistage computational pipeline. Findings In 2021, stroke was the third most common GBD level 3 cause of death (7·3 million [95% UI 6·6–7·8] deaths; 10·7% [9·8–11·3] of all deaths) after ischaemic heart disease and COVID-19, and the fourth most common cause of DALYs (160·5 million [147·8–171·6] DALYs; 5·6% [5·0–6·1] of all DALYs). In 2021, there were 93·8 million (89·0–99·3) prevalent and 11·9 million (10·7–13·2) incident strokes. We found disparities in stroke burden and risk factors by GBD region, country or territory, and SDI, as well as a stagnation in the reduction of incidence from 2015 onwards, and even some increases in the stroke incidence, death, prevalence, and DALY rates in southeast Asia, east Asia, and Oceania, countries with lower SDI, and people younger than 70 years. Globally, ischaemic stroke constituted 65·3% (62·4–67·7), intracerebral haemorrhage constituted 28·8% (28·3–28·8), and subarachnoid haemorrhage constituted 5·8% (5·7–6·0) of incident strokes. There were substantial increases in DALYs attributable to high BMI (88·2% [53·4–117·7]), high ambient temperature (72·4% [51·1 to 179·5]), high fasting plasma glucose (32·1% [26·7–38·1]), diet high in sugar-sweetened beverages (23·4% [12·7–35·7]), low physical activity (11·3% [1·8–34·9]), high systolic blood pressure (6·7% [2·5–11·6]), lead exposure (6·5% [4·5–11·2]), and diet low in omega-6 polyunsaturated fatty acids (5·3% [0·5–10·5]). Interpretation Stroke burden has increased from 1990 to 2021, and the contribution of several risk factors has also increased. Effective, accessible, and affordable measures to improve stroke surveillance, prevention (with the emphasis on blood pressure, lifestyle, and environmental factors), acute care, and rehabilitation need to be urgently implemented across all countries to reduce stroke burden. Funding Bill & Melinda Gates Foundation.

Full Text

Evidence from the Global Burden of Disease, Injuries, and Risk Factors Study (GBD) suggests that prevalent cases of total cardiovascular disease (including stroke) nearly doubled from 271 million (95% uncertainty interval [UI] 257–285) in 1990 to 523 million (497–550) in 2019. Moreover, despite a consistent decline in age-standardised cardiovascular disease (including stroke) mortality rates globally in the second half of the 20th century, there has been a subsequent deceleration in the decline and an overall flattening of the decline in the past few years. Since 2010, age-standardised cardiovascular disease (including stroke) mortality rates have even increased in many locations (eg, Mexico, the UK, and the USA),1, 2 and the age-standardised incidence of stroke in individuals younger than 55 years has increased substantially in high-income countries.3, 4 The previous GBD study on stroke burden and risks covered the period 1990–2019, and identified stroke as the second leading cause of death in the world. The most recent GBD stroke burden project has estimated an almost doubling of disability-adjusted life-years (DALYs), deaths, and cost due to stroke from 2020 to 2050. Globally, the age-standardised prevalence of cardiovascular disease (including stroke) risk factors (including hypertension, overweight, and diabetes) are also increasing. There has been a rapid increase in the number of people who died or remained disabled from stroke over the past 30 years, with a trend towards increasing incidence rates in people younger than 55 years, and increased prevalence of major risk factors for stroke (elevated blood pressure, overweight, and diabetes) over the past 10–15 years. These findings necessitate timely updated data on the most recent changes in stroke burden and risks across the globe to inform adequate health-care planning, resource allocation, and priority setting for stroke and to assess the success or failure of measures to reduce stroke burden.
Details of the GBD 2021 methods for stroke burden and risk factors estimates remained the same as for the latest GBD estimates and are described elsewhere8, 9, 10 (appendix pp 61–99). Stroke was defined according to the clinical WHO criteria and categorised into three pathological types (ischaemic stroke, intracerebral haemorrhage, and subarachnoid haemorrhage). To simplify the stroke modelling process and to ensure that all major pathological types were estimated correctly, vital registration and surveillance data were used to separately produce independent acute and chronic stroke models for ischaemic stroke, intracerebral haemorrhage, and subarachnoid haemorrhage type (appendix pp 75–76). As in previous GBD stroke burden estimates, we modelled first-ever-in-a-lifetime ischaemic stroke, intracerebral haemorrhage, and subarachnoid haemorrhage from the day of stroke onset to 28 days, and separately modelled survival (prevalence) beyond 28 days.
Cause of Death Ensemble modelling (CODEm) was used to estimate deaths due to overall stroke and stroke pathological types. For non-fatal disease modelling (incidence and prevalence of stroke), we used the DisMod-MR 2.1 tool, a Bayesian modelling software that uses data on various disease parameters and the epidemiological relationships between these parameters. In the GBD study, the incidence rate represents new events in a given year, whereas the death rate represents those that occurred in that year regardless of when the stroke occurred.
Stroke incidence, mortality, prevalence, and DALY estimates are presented in absolute numbers and as age-standardised rates per 100 000 population (with 95% UIs) and are stratified by age, sex, 21 GBD regions, and seven GBD super-regions (appendix pp 202–203). Countries and territories were also grouped into quintiles of high, high-middle, middle, low-middle, and low Socio-demographic Index (SDI; a summary indicator of geometric mean of normalised values of a location's lag-distributed income per capita, the average years of schooling in the population aged 15 years or older, and the total fertility rate in females younger than 25 years), on the basis of their 2021 values. Expressed on a scale from 0 to 1, a location with an SDI of 0 would have a theoretical minimum level of development relevant to health, whereas a location with an SDI of 1 would have a theoretical maximum level.
To analyse the attributable burden of stroke and its three pathological types due to 23 risk factors currently available for such analysis in GBD 2021, we calculated population attributable fractions (PAFs) of DALYs (appendix pp 31–43), using the exposure level for each risk factor and theoretical minimum risk exposure level (TMREL) that minimises risk for each individual in the population as the reference variable. We analysed data on the prevalence of exposure to a risk and derived relative risks for any risk–outcome pair for which we found sufficient evidence of a causal relationship. Adjustments for mediation were applied to account for relationships involving risk factors that act indirectly on outcomes via intermediate risks, as described elsewhere. Relative risk data were pooled using meta-regression of cohort, case–control, or intervention studies. From the prevalence and relative risk results, PAFs were estimated relative to the TMREL. The PAF represents a proportion of the stroke DALYs that would be decreased if the exposure to the risk factor in the past had been at the counterfactual level of the TMREL.
The risks included in the analysis were ambient particulate matter pollution; household air pollution from solid fuels; low ambient temperature (daily temperatures below the TMREL); high ambient temperature (daily temperatures above the TMREL); lead exposure; diet high in sodium; diet high in red meat; diet high in processed meat; diet low in fruits; diet low in vegetables; diet low in wholegrains; alcohol use (any alcohol dosage consumption); diet high in sugar-sweetened beverages; diet low in fibre; diet low in omega-6 polyunsaturated fatty acids; low physical activity (only for ischaemic stroke burden); smoking; second-hand smoke; high BMI; high fasting plasma glucose; high systolic blood pressure; high LDL cholesterol (only for ischaemic stroke burden); and kidney dysfunction, as measured by low glomerular filtration rate (not assessed for subarachnoid haemorrhage burden). We set the TMREL to zero for all harmful dietary risk factors with monotonically increasing risk functions (eg, processed meat intake), excluding sodium. For protective risks with monotonically declining risk functions with exposure (eg, fruit intake), we first determined the 85th percentile of exposure in the cohorts or trials used in the meta-regression of each outcome that was associated with the risk. Then, we determined the TMREL by weighting each risk–outcome pair by the relative global magnitude of each outcome.
As with causes, GBD organises risk factors into four levels, from the broadest (level 1: environmental risks, behavioural risks, and metabolic risks) to the most specific (level 4; 23 individual risk factors). The PAFs of risk factor groups took into account mediation between risk factors included in the group, as explained elsewhere. Percentages and number of DALYs are not mutually exclusive. The crude sum of the PAF of the risk factors might exceed 100% because the effects of many of these risk factors are mediated partly or wholly through another risk factor or risk factors. Definitions of risk factors and risk groups and further details of risk factors are in the appendix (pp 31–43). Changes in the modelling of stroke for GBD 2021 are presented in the appendix (pp 44–47). Analyses were also done by cluster of risk factors. The air pollution cluster includes ambient PM2·5 pollution and household air pollution. The behavioural risks cluster includes smoking (including second-hand smoking), dietary risks (diet high in sodium, diet high in processed meat diet, high in red meat, diet high in sugar-sweetened beverages, diet low in omega-6 polyunsaturated fatty acids, diet low in fruits, diet low in vegetables, and diet low in wholegrains), alcohol use, and low physical activity. The dietary risks cluster includes diet high in sodium, diet high in processed meat, diet high in red meat, diet high in sugar-sweetened beverages, diet low in omega-6 polyunsaturated fatty acids, diet low in fruits, diet low in vegetables, and diet low in wholegrains. The environmental risks cluster includes the air pollution cluster, low ambient temperature, high ambient temperature, and lead exposure. The metabolic risks cluster includes high fasting plasma glucose, high LDL cholesterol, high systolic blood pressure, high BMI, and kidney dysfunction. Finally, the tobacco smoke cluster includes smoking and second-hand smoking.
In 2021, there were 93·8 million (95% UI 89·0–99·3) stroke survivors, 11·9 million (10·7–13·2) new stroke events, 7·3 million (6·6–7·8) deaths from stroke, and 160·5 million (147·8–171·6) DALYs from stroke, comprising 10·7% (9·8–11·3) of all deaths and 5·6% (5·0–6·1) of all DALYs from all causes, the third leading cause of deaths (after ischaemic heart disease and COVID-19) and the fourth leading cause of DALYs (after COVID-19, ischaemic heart disease, and neonatal disorders; table 1; appendix p 204).
In 2021, 83·3% incident, 76·7% prevalent, and 87·2% fatal strokes, and 89·4% stroke-related DALYs occurred in all low-income and middle-income countries (LMICs) combined (appendix pp 48–49). We also observed geographical differences in age-standardised stroke incidence, with the lowest in Luxembourg (57·7 [95% UI 53·5–62·1] per 100 000) and highest in the Solomon Islands (355·0 [332·7–378·1] per 100 000); prevalence, with the highest in Ghana (2045·8 [1977·3–2120·1] per 100 000) and lowest in Cyprus (521·5 [495·7–553·5] per 100 000); deaths, with the lowest in Singapore (14·2 [12·3–15·6] per 100 000) and highest in North Macedonia (277·4 [235·5–321·2] per 100 000); and in DALY rates, with the lowest in Switzerland (333·3 [291·0–368·8] per 100 000) and highest in Nauru (6100·0 [4917·8–7576·1] per 100 000). Overall, the highest stroke burden (as measured by age-standardised incidence, prevalence, death, and DALY rates) in 2021 was observed in east Asia, central Asia, and sub-Saharan regions and lowest in high-income North America, Australasia, and Latin America regions, with the majority of the stroke burden in middle SDI, high-middle, and low-middle SDI regions (figure 1; appendix pp 50–85).
Ischaemic stroke constituted the largest proportion of all incident strokes (7·8 million [95% UI 6·7–8·9], or 65·3% [62·4–67·7] of all strokes), followed by intracerebral haemorrhage (3·4 million [3·1–3·8] incident events, or 28·8% [28·3–28·8] of all strokes). However, the absolute number of DALYs due to intracerebral haemorrhage (79·5 million [72·7–85·2], or 49·6% [49·3–49·8] of total DALYs due to stroke) was greater than the number of DALYs due to ischaemic stroke (70·4 million [64·1–76·0], or 43·8% [43·5–44·3]). In 2021, subarachnoid haemorrhage occurred in 0·7 million (0·6–0·8) people (5·8% [5·7–6·0] of all strokes), and there were 10·6 million (9·4–12·1) DALYs due to subarachnoid haemorrhage (6·6% of DALYs from all strokes combined). Similar to total stroke, differences were observed for age-standardised rates for the three pathological types of stroke and their trends from 1990 to 2021 globally and by SDI (appendix pp 50–85, 208): rates of incident and fatal stroke were highest for ischaemic stroke (92·4 [79·8–105·8] per 100 000 and 44·2 [39·5–47·8] per 100 000, respectively) followed by intracerebral haemorrhage (40·8 [36·2–45·2] per 100 000 and 39·1 [35·4–42·6] per 100 000, respectively) and subarachnoid haemorrhage (8·3 [7·3–9·5] per 100 000 and 4·2 [3·7–4·8] per 100 000, respectively).
In 2021, there were large variations in the proportion of ischaemic stroke and intracerebral haemorrhage between high-income countries and LMICs (appendix pp 100–101). Whereas in high-income countries ischaemic stroke constituted 74·9% (95% UI 72·3−84·1) and intracerebral haemorrhage constituted 17·8% (17·3−17·9) of all incident strokes, in all LMICs combined these stroke subtypes constituted 63·4% (53·6−73·7) and 31·1% (30·2−31·3), respectively. Among all LMICs combined, the proportion of intracerebral haemorrhage was highest in low-income countries (36·9% [36·5−37·1]). The proportion of subarachnoid haemorrhage in high-income countries (7·3% [7·2−8·3]) was higher than that in all LMICs combined (5·5% [5·4−5·7]).
Among 11·9 million new strokes in 2021, 6·3 million (95% UI 5·6 to 7·0; or 52·6% [52·4 to 53·1]) occurred in males and 5·7 million (5·1 to 6·3; or 47·4% [47·3 to 47·6]) in females; the corresponding sex distribution of prevalent stroke was 51·0% (47·8 million [45·3 to 50·6]) for males and 49·0% (46·0 million [43·5 to 48·8]) for females; that for deaths from stroke was 52·1% (3·8 million [3·4 to 4·1]) for males and 47·9% (3·5 million [3·1 to 3·8] for females; and that for stroke-related DALYs was 55·0% (88·3 million [80·6 to 97·2]) for males and 45·0% (72·2 million [65·6 to 78·2) for females (table 1; appendix pp 138–139). From 1990 to 2021, the age-standardised incidence, prevalence, death, and DALY rates (table 1) of stroke and its pathological types were reduced virtually across all World Bank country income levels (except for ischaemic stroke incidence and prevalence in upper-middle-income countries, where the rates were increased by 1% [–4 to 5] for ischaemic stroke incidence and 11% [8 to 14] for ischaemic stroke prevalence). Although there was a trend towards lower age-standardised stroke burden rates (incidence, prevalence, deaths, and DALYs) across all quintiles of the SDI, there was a stagnation in the reduction of incidence rates from 2015 onwards, and even some increase in the prevalence rates in high-middle SDI countries from 2020 to 2021 (appendix p 209). Similar trend patterns were observed in seven GBD super-regions, with more prominent increases in age-standardised incidence and prevalence rates after 2015 in southeast Asia, east Asia, and Oceania (appendix p 206).
Although from 1990 to 2021 there was a decrease in the age-standardised incidence (–21·8% [95% UI–23·7 to –19·8]), prevalence (–8·5% [–9·7 to –7·3]), death (–39·4% [–44·0 to –34·6]), and DALY (–38·7% [–43·4 to –34·0]) stroke rates, increases were seen over that period in the numbers of people who had a new stroke (70·2% [65·9 to 74·6]), survived stroke (86·1% [83·0 to 89·4]), died from stroke (44·1% [32·3 to 56·0]), and who died or remained disabled from stroke (as measured by DALYs; 32·2% [21·7 to 42·7]; table 1; appendix pp 100–101). The percentage decline in age-standardised stroke incidence rates in the 2019–2021 period (–1·8% [–2·8 to –0·6]) was smaller than that for the overall 2010–21 period (–3·1% [–4·2 to –2·0]).
Although all-age (not age-standardised) stroke incidence, death, and DALY rates were substantially reduced in people aged 70 years or older between 1990 and 2021 (–18·2% [95% UI –21·3 to – 14·6] incidence rate, –34·2% [–39·4 to –29·3] death rate, and –35·6% [–40·2 to –30·8] DALY rate), and all-age prevalence rate in this age group did not change over this period (–1·0 [–3·1 to 1·2]), all-age incidence increased by 4·1% (0·9 to 7·6), prevalence increased in people younger than 70 years by 14·8% (13·1 to 16·8), and death and DALY rates were reduced in people younger than 70 years by 17·4% (–25·0 to –8·9) and 19·0% (–26·0 to –11·6), respectively (appendix p 140). Similar patterns were observed for all-age incidence, prevalence, death, and DALY rates of ischaemic stroke for both age groups (<70 years and ≥70 years). Of the three pathological types of stroke, only all-age subarachnoid haemorrhage prevalence rates increased in people younger than 70 years, by 3·4% (1·3 to 5·5), whereas all-age incidence, prevalence, death, and DALY rates of intracerebral haemorrhage and incidence, death, and DALY rates of subarachnoid haemorrhage were reduced in both age groups (appendix pp 88, 210–211).
Globally, the total number of stroke-related DALYs due to risk factors increased substantially from 1990 (100·1 million [95% UI 92·7 to 107·8]) to 2021 (135·0 million [122·0 to 147·7]), but there was no substantial change in the age-standardised stroke DALYs attributable to risk factors (–0·5% [–2·4 to 1·1]). In 2021, 84·1% (77·8 to 88·8) of DALYs from stroke were attributed to the 23 risk factors analysed (table 2), with the largest proportions of attributable risks for total stroke, ischaemic stroke, intracerebral haemorrhage, and subarachnoid haemorrhage observed in eastern Europe, Asia, and sub-Saharan Africa (appendix p 212).
At level 1 of the GBD risk factors hierarchy (table 2, appendix pp 141–253), metabolic risk factors contributed most to the stroke-related DALYs (range 66–70%) across all World Bank country income levels, followed by the environmental risk cluster in low-income, lower-middle-income, and upper-middle-income countries (range 35–53%), and behavioural risks (range 30–38%) across different income level countries. Stroke burden associated with the environmental or occupational risks was lowest in high-income countries (18·1% [95% UI 13·6 to 22·7]). Similarly, regions with higher SDI (appendix pp 141–143) had a larger contribution of metabolic and behavioural risks to stroke-related DALYs, whereas environmental risks most prominently contributed to stroke-related DALYs in lower SDI quintiles. From 1990 to 2021, the age-standardised proportion of stroke DALYs attributable to risk factors increased in north Africa and the Middle East (6·8% [4·1 to 11·5]) and sub-Saharan Africa (3·3% [1·8 to 5·3]), but did not change in south Asia (0·4% [–1·0 to 1·9]) and southeast Asia, east Asia, and Oceania (–0·9% [–4·2 to 1·9]), and decreased in central Europe, eastern Europe, and central Asia (–2·0% [–3·8 to –0·7]) and Latin America and the Caribbean (–5·0% [–9·1 to –2·3]), as well as high-income GBD regions (–7·3% [–10·1 to –5·0]).
Globally, of the 23 risk factors analysed, 14 individually significant risk factors for stroke were high systolic blood pressure (56·8% [95% UI 42·5–68·0] attributable DALYs), ambient particular matter (16·6% [11·5–20·9]), smoking (13·8% [2·5–26·0]), high LDL cholesterol (13·1% [4·6–21·3]), household air pollution (11·2% [6·4–19·3]), diet high in sodium (10·6% [2·8–22·8]), high fasting plasma glucose (10·3% [8·1–12·6]), kidney disfunction (9·3% [6·8–11·8]), diet low in fruits (5·9% [0·4–10·4]), high alcohol use (5·2% [1·3–9·8]), high BMI (4·7% [0·4–9·8]), second-hand smoking (4·4% [1·0–7·9]), low physical activity (2·1% [0·5–3·9]), and diet low in vegetables (1·6% [0·4–2·6]; figure 2).
Stroke attributable to metabolic risks constituted 68·8% (95% UI 57·6 to 77·5) of all strokes, environmental risks constituted 36·7% (29·0 to 44·2), and behavioural risks constituted 35·2% (26·9 to 44·7). Although the proportion of stroke DALYs attributable to metabolic risks increased from 1990 to 2021 by 6·7% (3·8 to 10·0; mainly because of the increase in the burden attributable to high BMI, high fasting plasma glucose, and high systolic blood pressure), proportions of stroke DALYs attributable to behavioural risks decreased by 8·0% (–13·6 to –3·4) and those due to environmental risks by 14·8% (–21·6 to –8·7), mainly because of the decrease in the burden attributable to diet high in processed meat, diet low in vegetables, diet low in fibre, low ambient temperature, particulate matter pollution, diet low in fruits, and smoking (figure 3). However, from 1990 to 2021, there was a substantial increase in the stroke DALYs attributable to high ambient temperature, high fasting plasma glucose, diet high in sugar-sweetened beverages, low physical activity, diet high in red meat, lead exposure, and diet low in omega-6 polyunsaturated fatty acids. There were noticeable geographical and regional variations in the PAF of the risk factors for ischaemic stroke, intracerebral haemorrhage, subarachnoid haemorrhage, and all stroke types combined (appendix p 264), as well as in the ranking of PAFs of age-standardised stroke DALYs attributable to risk factors by 21 GBD regions (figure 4). For the PAF of risk factors by pathological type of stroke, SDI, 21 GBD regions, and 204 countries and territories were used (appendix pp 89–201). Unlike the PAF of risk factors for total stroke, high alcohol use was not associated with ischaemic stroke-related DALYs (appendix pp 92–95, 126–149, 199), and diet low in fruits and vegetables and high BMI were not associated with intracerebral haemorrhage-related DALYs (appendix pp 92–95, 150–173, 200). Unlike ischaemic stroke and intracerebral haemorrhage, non-optimal ambient temperature appeared to be associated with the subarachnoid haemorrhage-related DALYs, with the greater contribution of low ambient temperature (4·5% [3·8 to 5·3]) than high ambient temperature (1·1% [0·2 to 2·5]). Other substantial risk factors for subarachnoid haemorrhage (appendix pp 151−153) were second-hand smoking (4·7% [3·2 to 6·2]), diet high in sodium (8·9% [2·0 to 19·8]), household air pollution from solid fuels (10·3% [5·5 to 17·4]), ambient particulate matter pollution (14·2% [9·8 to 18·0]), smoking (14·5% [2·7 to 27·2]), and high systolic blood pressure (51·6% [38·0 to 62·6]).
In 2021, stroke remained the second most common cause (after ischaemic heart disease) of death and the third most common cause of DALYs (after ischaemic heart disease and neonatal disorders) among non-communicable disorders (NCDs) globally. However, at level 3 of the GBD all-cause hierarchy, stroke was the third most common cause of death (after ischaemic heart disease and COVID-19) and the fourth most common cause of DALYs (after COVID-19, ischaemic heart disease, and neonatal disorders), with the bulk of the stroke burden in LMICs and countries with lower SDI. Consistent with previous studies,5, 17 this study showed disparities in stroke burden (including almost two times greater proportion of intracerebral haemorrhage in LMICs than in high-income countries) and risk factors by GBD region, country, country income level, and SDI quintiles, as well as an overall trend towards decreasing age-standardised stroke incidence, prevalence, and DALY rates from 1990 to 2021. Given the leading role of arterial hypertension in the occurrence of intracerebral haemorrhage, the greater prevalence and poorer control of hypertension in LMICs than in high-income countries are the most likely causes of the differences in the proportion of intracerebral haemorrhage in those countries. Similar proportions of subarachnoid haemorrhage in high-income countries and LMICs are likely to be related to the significant reduction in tobacco smoking prevalence that we observed since 2010 in high-income countries, with almost no change in tobacco smoking prevalence in LMICs. However, the current study documented a stagnation in the reduction of age-standardised incidence rates from 2015 onwards, and even some increase in the age-standardised stroke incidence, death, prevalence, and DALY rates in southeast Asia, east Asia, and Oceania, and countries with lower SDI from 2015 onwards. Globally, there was also an increase in all-age incidence and prevalence rates in people younger than 70 years, whereas there was a reduction in all-age stroke incidence, prevalence, death, and DALY rates in people aged 70 years or older. A trend towards increasing incidence and prevalence rate of cardiovascular diseases (including stroke) in people aged 15–39 years globally and stroke incidence rates in people younger than 55 years versus older people has also been shown in previous systematic reviews and is likely to be related to the increase in prevalence of arterial hypertension23, 24 (including poorly controlled and uncontrolled hypertension), overweight or obesity,26, 27, 28 and type 2 diabetes in young adults, especially in LMICs.24, 26, 27, 28, 29 This situation is complicated by the fact that a large proportion of young adults with vascular risk factors, arterial hypertension, and dyslipidaemia remain under-treated owing at least partly to the widespread use of absolute cardiovascular disease risk treatment thresholds. The observed slowing down of the percentage of decline of age-standardised stroke incidence rates in 2019–21 is likely to be related to the decreased hospital admissions of patients with acute stroke that was observed over the COVID-19 period in many countries.
Apart from population growth and ageing,1, 34 other factors responsible for the increased burden of stroke, in terms of absolute numbers in the world, are likely to be related to the insufficient effectiveness of the currently used primary stroke and cardiovascular disease prevention strategies35, 36 as well as the disparities and major gaps in stroke service provision and accessibility, and workforce of stroke care providers in many countries (especially LMICs).17, 37, 38 Although stroke is highly preventable, globally there were substantial increases in DALYs attributable to high BMI, high ambient temperature, high fasting plasma glucose, diet high in sugar-sweetened beverages, low physical activity, high systolic blood pressure, and diet low in polyunsaturated omega-6 fatty acids, suggesting the growing role of these environmental and behavioural risks in the stroke burden. However, from 1990 to 2021, we also observed a reduction of PAF due to diet high in processed meat, diet low in vegetables, diet low in fibre, low ambient temperature, ambient particulate matter pollution, diet low in fruits, and smoking, suggesting effectiveness of the strategies towards reduction of the exposure to these risk factors. The observed increase in the age-standardised proportion of stroke DALYs attributable to risk factors in north Africa and the Middle East and sub-Saharan Africa regions might reflect a failure in the control of stroke risk factors. However, in central Europe, eastern Europe, central Asia, Latin America and the Caribbean, and high-income GBD regions, this might reflect a success in the control of stroke risk factors.
This study is, to our knowledge, the first to show the large contribution of ambient particulate matter pollution and household air pollution from solid fuels to subarachnoid haemorrhage DALYs, with a similar PAF to that of smoking. A close relationship between ambient air pollution and subarachnoid haemorrhage mortality was found in some studies.39, 40, 41 Air pollution in 2021 appeared to be highly important to other types of stroke and also caused 11·9% (95% UI 10·0–13·8) of total deaths from all causes, making it the second largest cause of deaths from all causes globally (after high systolic blood pressure) and the second leading cause of DALYs (8·2% [6·9–9·6]) from all causes (after malnutrition). These findings are in line with research showing that rises in ambient temperature (including heatwaves) and climate change are associated with increased stroke morbidity and mortality.43, 44 Because ambient air pollution is reciprocally associated with the ambient temperature and climate change, all of which synergistically influence cardiovascular disease (including stroke) occurrence44, 45, 46 and overall health,47, 48 the importance of urgent climate actions and measures to reduce ambient air pollution cannot be overestimated.47, 48 Experts have recommended that governments increase implementation of a clean-energy economy, promote unprocessed plant-based food choices, and globally phase out industrialised animal farming.
Every member state of the UN has committed to meeting the Sustainable Development Goals (SDGs), but currently few countries are on target to achieve SDG 3.4, which is to reduce by a third premature mortality from NCDs through prevention and treatment and promote mental health and wellbeing by 2030. By implementing and monitoring the World Stroke Organization–Lancet Neurology Commission's recommendations, the global burden of stroke would be reduced drastically this decade and beyond. Not only would this substantial reduction enable SDG 3.4, as well as other key SDGs, to be met, it would improve brain health and the overall wellbeing of millions of people across the globe. One of the most common problems in implementing stroke prevention and care recommendations is the scarcity of funding. The World Stroke Organization–Lancet Neurology Commission on stroke recommends introducing legislative regulations and taxation of unhealthy products by each government in the world. Such taxation would not only reduce consumption of these products and, therefore, lead to the reduction of burden from stroke and other major NCDs,50, 51, 52 but also generate a large revenue sufficient to fund prevention programmes and services for stroke and other major disorders, reduce poverty and inequality in health service provision, improve wellbeing of the population, and boost local economies.
The main strength of this study is the extended number of data sources included in the analysis that allowed us to generate more accurate and up-to-date stroke burden and risk factor estimates. This allows evidence-based health-care planning and resource allocation by health policy makers on the national, regional, and global levels. However, good-quality stroke epidemiological studies are still scarce in most countries, which prevented us from including in the analysis many other important risk factors, such as sickle cell disease and HIV, which are particularly important for sub-Saharan Africa. Differences in health-care systems and completeness and accuracy of stroke case ascertainment might play a part in the observed between-country differences (eg, very high stroke prevalence in Ghana compared with neighbouring countries). Although the GBD methods for estimating attributable burden of stroke due to risk factors accounts for a cumulative effect of multiple risk factors, it might not fully account for all potential confounders. Moreover, some new risk factors, such as high ambient temperature, might require further validation and examination to confirm their impact on stroke burden. Furthermore, more granular data analysis is needed. For example, stroke burden variation by race and ethnicity within countries, which can mask disparities in stroke incidence, risk factors, and outcomes among different population groups, and analysis of attributable effects of different levels of exposure to smoking, alcohol, and so on. We expect such analysis will be done in future GBD iterations.
In summary, our study findings continue to point out that currently used stroke prevention strategies are not sufficiently effective to halt, let alone reduce, the fast-growing stroke burden. Additional, more effective stroke prevention strategies (with the emphasis on population-wide measures, task shifting from doctors to nurses or health volunteers, and the wider use of evidence-based mobile and telehealth platforms) and pragmatic solutions to address the critical gaps in stroke service delivery, along with development of context-appropriate workforce capacity building and epidemiological surveillance systems, need to be urgently implemented across all countries. Without scaling up these innovative evidence-based strategies and policies that target local, national, regional, and global stroke prevention and care disparities, the burden of stroke will continue to grow, thus threatening the sustainability of health systems worldwide.

Sections

"[{\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib1\", \"bib1\", \"bib1\", \"bib1\", \"bib2\", \"bib3\", \"bib4\", \"bib5\", \"bib6\", \"bib6\", \"bib1\", \"bib7\", \"bib5\"], \"section\": \"Introduction\", \"text\": \"Evidence from the Global Burden of Disease, Injuries, and Risk Factors Study (GBD) suggests that prevalent cases of total cardiovascular disease (including stroke) nearly doubled from 271 million (95% uncertainty interval [UI] 257\\u2013285) in 1990 to 523 million (497\\u2013550) in 2019. Moreover, despite a consistent decline in age-standardised cardiovascular disease (including stroke) mortality rates globally in the second half of the 20th century, there has been a subsequent deceleration in the decline and an overall flattening of the decline in the past few years. Since 2010, age-standardised cardiovascular disease (including stroke) mortality rates have even increased in many locations (eg, Mexico, the UK, and the USA),1, 2 and the age-standardised incidence of stroke in individuals younger than 55 years has increased substantially in high-income countries.3, 4 The previous GBD study on stroke burden and risks covered the period 1990\\u20132019, and identified stroke as the second leading cause of death in the world. The most recent GBD stroke burden project has estimated an almost doubling of disability-adjusted life-years (DALYs), deaths, and cost due to stroke from 2020 to 2050. Globally, the age-standardised prevalence of cardiovascular disease (including stroke) risk factors (including hypertension, overweight, and diabetes) are also increasing. There has been a rapid increase in the number of people who died or remained disabled from stroke over the past 30 years, with a trend towards increasing incidence rates in people younger than 55 years, and increased prevalence of major risk factors for stroke (elevated blood pressure, overweight, and diabetes) over the past 10\\u201315 years. These findings necessitate timely updated data on the most recent changes in stroke burden and risks across the globe to inform adequate health-care planning, resource allocation, and priority setting for stroke and to assess the success or failure of measures to reduce stroke burden.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib8\", \"bib9\", \"bib10\", \"sec1\", \"bib11\", \"bib12\", \"sec1\", \"bib5\"], \"section\": \"Overview\", \"text\": \"Details of the GBD 2021 methods for stroke burden and risk factors estimates remained the same as for the latest GBD estimates and are described elsewhere8, 9, 10 (appendix pp 61\\u201399). Stroke was defined according to the clinical WHO criteria and categorised into three pathological types (ischaemic stroke, intracerebral haemorrhage, and subarachnoid haemorrhage). To simplify the stroke modelling process and to ensure that all major pathological types were estimated correctly, vital registration and surveillance data were used to separately produce independent acute and chronic stroke models for ischaemic stroke, intracerebral haemorrhage, and subarachnoid haemorrhage type (appendix pp 75\\u201376). As in previous GBD stroke burden estimates, we modelled first-ever-in-a-lifetime ischaemic stroke, intracerebral haemorrhage, and subarachnoid haemorrhage from the day of stroke onset to 28 days, and separately modelled survival (prevalence) beyond 28 days.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib13\", \"bib5\"], \"section\": \"Overview\", \"text\": \"Cause of Death Ensemble modelling (CODEm) was used to estimate deaths due to overall stroke and stroke pathological types. For non-fatal disease modelling (incidence and prevalence of stroke), we used the DisMod-MR 2.1 tool, a Bayesian modelling software that uses data on various disease parameters and the epidemiological relationships between these parameters. In the GBD study, the incidence rate represents new events in a given year, whereas the death rate represents those that occurred in that year regardless of when the stroke occurred.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"sec1\", \"bib14\"], \"section\": \"Overview\", \"text\": \"Stroke incidence, mortality, prevalence, and DALY estimates are presented in absolute numbers and as age-standardised rates per 100\\u2008000 population (with 95% UIs) and are stratified by age, sex, 21 GBD regions, and seven GBD super-regions (appendix pp 202\\u2013203). Countries and territories were also grouped into quintiles of high, high-middle, middle, low-middle, and low Socio-demographic Index (SDI; a summary indicator of geometric mean of normalised values of a location's lag-distributed income per capita, the average years of schooling in the population aged 15 years or older, and the total fertility rate in females younger than 25 years), on the basis of their 2021 values. Expressed on a scale from 0 to 1, a location with an SDI of 0 would have a theoretical minimum level of development relevant to health, whereas a location with an SDI of 1 would have a theoretical maximum level.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"sec1\", \"bib9\", \"bib15\", \"bib9\"], \"section\": \"Attributable burden of stroke due to risk factors\", \"text\": \"To analyse the attributable burden of stroke and its three pathological types due to 23 risk factors currently available for such analysis in GBD 2021, we calculated population attributable fractions (PAFs) of DALYs (appendix pp 31\\u201343), using the exposure level for each risk factor and theoretical minimum risk exposure level (TMREL) that minimises risk for each individual in the population as the reference variable. We analysed data on the prevalence of exposure to a risk and derived relative risks for any risk\\u2013outcome pair for which we found sufficient evidence of a causal relationship. Adjustments for mediation were applied to account for relationships involving risk factors that act indirectly on outcomes via intermediate risks, as described elsewhere. Relative risk data were pooled using meta-regression of cohort, case\\u2013control, or intervention studies. From the prevalence and relative risk results, PAFs were estimated relative to the TMREL. The PAF represents a proportion of the stroke DALYs that would be decreased if the exposure to the risk factor in the past had been at the counterfactual level of the TMREL.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib5\"], \"section\": \"Attributable burden of stroke due to risk factors\", \"text\": \"The risks included in the analysis were ambient particulate matter pollution; household air pollution from solid fuels; low ambient temperature (daily temperatures below the TMREL); high ambient temperature (daily temperatures above the TMREL); lead exposure; diet high in sodium; diet high in red meat; diet high in processed meat; diet low in fruits; diet low in vegetables; diet low in wholegrains; alcohol use (any alcohol dosage consumption); diet high in sugar-sweetened beverages; diet low in fibre; diet low in omega-6 polyunsaturated fatty acids; low physical activity (only for ischaemic stroke burden); smoking; second-hand smoke; high BMI; high fasting plasma glucose; high systolic blood pressure; high LDL cholesterol (only for ischaemic stroke burden); and kidney dysfunction, as measured by low glomerular filtration rate (not assessed for subarachnoid haemorrhage burden). We set the TMREL to zero for all harmful dietary risk factors with monotonically increasing risk functions (eg, processed meat intake), excluding sodium. For protective risks with monotonically declining risk functions with exposure (eg, fruit intake), we first determined the 85th percentile of exposure in the cohorts or trials used in the meta-regression of each outcome that was associated with the risk. Then, we determined the TMREL by weighting each risk\\u2013outcome pair by the relative global magnitude of each outcome.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib16\", \"bib5\", \"sec1\", \"sec1\"], \"section\": \"Attributable burden of stroke due to risk factors\", \"text\": \"As with causes, GBD organises risk factors into four levels, from the broadest (level 1: environmental risks, behavioural risks, and metabolic risks) to the most specific (level 4; 23 individual risk factors). The PAFs of risk factor groups took into account mediation between risk factors included in the group, as explained elsewhere. Percentages and number of DALYs are not mutually exclusive. The crude sum of the PAF of the risk factors might exceed 100% because the effects of many of these risk factors are mediated partly or wholly through another risk factor or risk factors. Definitions of risk factors and risk groups and further details of risk factors are in the appendix (pp 31\\u201343). Changes in the modelling of stroke for GBD 2021 are presented in the appendix (pp 44\\u201347). Analyses were also done by cluster of risk factors. The air pollution cluster includes ambient PM2\\u00b75 pollution and household air pollution. The behavioural risks cluster includes smoking (including second-hand smoking), dietary risks (diet high in sodium, diet high in processed meat diet, high in red meat, diet high in sugar-sweetened beverages, diet low in omega-6 polyunsaturated fatty acids, diet low in fruits, diet low in vegetables, and diet low in wholegrains), alcohol use, and low physical activity. The dietary risks cluster includes diet high in sodium, diet high in processed meat, diet high in red meat, diet high in sugar-sweetened beverages, diet low in omega-6 polyunsaturated fatty acids, diet low in fruits, diet low in vegetables, and diet low in wholegrains. The environmental risks cluster includes the air pollution cluster, low ambient temperature, high ambient temperature, and lead exposure. The metabolic risks cluster includes high fasting plasma glucose, high LDL cholesterol, high systolic blood pressure, high BMI, and kidney dysfunction. Finally, the tobacco smoke cluster includes smoking and second-hand smoking.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"tbl1\", \"sec1\"], \"section\": \"Stroke incidence, prevalence, death, and DALYs by geographical location\", \"text\": \"In 2021, there were 93\\u00b78 million (95% UI 89\\u00b70\\u201399\\u00b73) stroke survivors, 11\\u00b79 million (10\\u00b77\\u201313\\u00b72) new stroke events, 7\\u00b73 million (6\\u00b76\\u20137\\u00b78) deaths from stroke, and 160\\u00b75 million (147\\u00b78\\u2013171\\u00b76) DALYs from stroke, comprising 10\\u00b77% (9\\u00b78\\u201311\\u00b73) of all deaths and 5\\u00b76% (5\\u00b70\\u20136\\u00b71) of all DALYs from all causes, the third leading cause of deaths (after ischaemic heart disease and COVID-19) and the fourth leading cause of DALYs (after COVID-19, ischaemic heart disease, and neonatal disorders; table 1; appendix p 204).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"sec1\", \"fig1\", \"sec1\"], \"section\": \"Stroke incidence, prevalence, death, and DALYs by geographical location\", \"text\": \"In 2021, 83\\u00b73% incident, 76\\u00b77% prevalent, and 87\\u00b72% fatal strokes, and 89\\u00b74% stroke-related DALYs occurred in all low-income and middle-income countries (LMICs) combined (appendix pp 48\\u201349). We also observed geographical differences in age-standardised stroke incidence, with the lowest in Luxembourg (57\\u00b77 [95% UI 53\\u00b75\\u201362\\u00b71] per 100\\u2008000) and highest in the Solomon Islands (355\\u00b70 [332\\u00b77\\u2013378\\u00b71] per 100\\u2008000); prevalence, with the highest in Ghana (2045\\u00b78 [1977\\u00b73\\u20132120\\u00b71] per 100\\u2008000) and lowest in Cyprus (521\\u00b75 [495\\u00b77\\u2013553\\u00b75] per 100\\u2008000); deaths, with the lowest in Singapore (14\\u00b72 [12\\u00b73\\u201315\\u00b76] per 100\\u2008000) and highest in North Macedonia (277\\u00b74 [235\\u00b75\\u2013321\\u00b72] per 100\\u2008000); and in DALY rates, with the lowest in Switzerland (333\\u00b73 [291\\u00b70\\u2013368\\u00b78] per 100\\u2008000) and highest in Nauru (6100\\u00b70 [4917\\u00b78\\u20137576\\u00b71] per 100\\u2008000). Overall, the highest stroke burden (as measured by age-standardised incidence, prevalence, death, and DALY rates) in 2021 was observed in east Asia, central Asia, and sub-Saharan regions and lowest in high-income North America, Australasia, and Latin America regions, with the majority of the stroke burden in middle SDI, high-middle, and low-middle SDI regions (figure 1; appendix pp 50\\u201385).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"sec1\"], \"section\": \"Burden by pathological type of stroke\", \"text\": \"Ischaemic stroke constituted the largest proportion of all incident strokes (7\\u00b78 million [95% UI 6\\u00b77\\u20138\\u00b79], or 65\\u00b73% [62\\u00b74\\u201367\\u00b77] of all strokes), followed by intracerebral haemorrhage (3\\u00b74 million [3\\u00b71\\u20133\\u00b78] incident events, or 28\\u00b78% [28\\u00b73\\u201328\\u00b78] of all strokes). However, the absolute number of DALYs due to intracerebral haemorrhage (79\\u00b75 million [72\\u00b77\\u201385\\u00b72], or 49\\u00b76% [49\\u00b73\\u201349\\u00b78] of total DALYs due to stroke) was greater than the number of DALYs due to ischaemic stroke (70\\u00b74 million [64\\u00b71\\u201376\\u00b70], or 43\\u00b78% [43\\u00b75\\u201344\\u00b73]). In 2021, subarachnoid haemorrhage occurred in 0\\u00b77 million (0\\u00b76\\u20130\\u00b78) people (5\\u00b78% [5\\u00b77\\u20136\\u00b70] of all strokes), and there were 10\\u00b76 million (9\\u00b74\\u201312\\u00b71) DALYs due to subarachnoid haemorrhage (6\\u00b76% of DALYs from all strokes combined). Similar to total stroke, differences were observed for age-standardised rates for the three pathological types of stroke and their trends from 1990 to 2021 globally and by SDI (appendix pp 50\\u201385, 208): rates of incident and fatal stroke were highest for ischaemic stroke (92\\u00b74 [79\\u00b78\\u2013105\\u00b78] per 100\\u2008000 and 44\\u00b72 [39\\u00b75\\u201347\\u00b78] per 100\\u2008000, respectively) followed by intracerebral haemorrhage (40\\u00b78 [36\\u00b72\\u201345\\u00b72] per 100\\u2008000 and 39\\u00b71 [35\\u00b74\\u201342\\u00b76] per 100\\u2008000, respectively) and subarachnoid haemorrhage (8\\u00b73 [7\\u00b73\\u20139\\u00b75] per 100\\u2008000 and 4\\u00b72 [3\\u00b77\\u20134\\u00b78] per 100\\u2008000, respectively).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"sec1\"], \"section\": \"Burden by pathological type of stroke\", \"text\": \"In 2021, there were large variations in the proportion of ischaemic stroke and intracerebral haemorrhage between high-income countries and LMICs (appendix pp 100\\u2013101). Whereas in high-income countries ischaemic stroke constituted 74\\u00b79% (95% UI 72\\u00b73\\u221284\\u00b71) and intracerebral haemorrhage constituted 17\\u00b78% (17\\u00b73\\u221217\\u00b79) of all incident strokes, in all LMICs combined these stroke subtypes constituted 63\\u00b74% (53\\u00b76\\u221273\\u00b77) and 31\\u00b71% (30\\u00b72\\u221231\\u00b73), respectively. Among all LMICs combined, the proportion of intracerebral haemorrhage was highest in low-income countries (36\\u00b79% [36\\u00b75\\u221237\\u00b71]). The proportion of subarachnoid haemorrhage in high-income countries (7\\u00b73% [7\\u00b72\\u22128\\u00b73]) was higher than that in all LMICs combined (5\\u00b75% [5\\u00b74\\u22125\\u00b77]).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"tbl1\", \"sec1\", \"tbl1\", \"sec1\", \"sec1\"], \"section\": \"Trends in stroke burden by age, sex and geographical location\", \"text\": \"Among 11\\u00b79 million new strokes in 2021, 6\\u00b73 million (95% UI 5\\u00b76 to 7\\u00b70; or 52\\u00b76% [52\\u00b74 to 53\\u00b71]) occurred in males and 5\\u00b77 million (5\\u00b71 to 6\\u00b73; or 47\\u00b74% [47\\u00b73 to 47\\u00b76]) in females; the corresponding sex distribution of prevalent stroke was 51\\u00b70% (47\\u00b78 million [45\\u00b73 to 50\\u00b76]) for males and 49\\u00b70% (46\\u00b70 million [43\\u00b75 to 48\\u00b78]) for females; that for deaths from stroke was 52\\u00b71% (3\\u00b78 million [3\\u00b74 to 4\\u00b71]) for males and 47\\u00b79% (3\\u00b75 million [3\\u00b71 to 3\\u00b78] for females; and that for stroke-related DALYs was 55\\u00b70% (88\\u00b73 million [80\\u00b76 to 97\\u00b72]) for males and 45\\u00b70% (72\\u00b72 million [65\\u00b76 to 78\\u00b72) for females (table 1; appendix pp 138\\u2013139). From 1990 to 2021, the age-standardised incidence, prevalence, death, and DALY rates (table 1) of stroke and its pathological types were reduced virtually across all World Bank country income levels (except for ischaemic stroke incidence and prevalence in upper-middle-income countries, where the rates were increased by 1% [\\u20134 to 5] for ischaemic stroke incidence and 11% [8 to 14] for ischaemic stroke prevalence). Although there was a trend towards lower age-standardised stroke burden rates (incidence, prevalence, deaths, and DALYs) across all quintiles of the SDI, there was a stagnation in the reduction of incidence rates from 2015 onwards, and even some increase in the prevalence rates in high-middle SDI countries from 2020 to 2021 (appendix p 209). Similar trend patterns were observed in seven GBD super-regions, with more prominent increases in age-standardised incidence and prevalence rates after 2015 in southeast Asia, east Asia, and Oceania (appendix p 206).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"tbl1\", \"sec1\"], \"section\": \"Trends in stroke burden by age, sex and geographical location\", \"text\": \"Although from 1990 to 2021 there was a decrease in the age-standardised incidence (\\u201321\\u00b78% [95% UI\\u201323\\u00b77 to \\u201319\\u00b78]), prevalence (\\u20138\\u00b75% [\\u20139\\u00b77 to \\u20137\\u00b73]), death (\\u201339\\u00b74% [\\u201344\\u00b70 to \\u201334\\u00b76]), and DALY (\\u201338\\u00b77% [\\u201343\\u00b74 to \\u201334\\u00b70]) stroke rates, increases were seen over that period in the numbers of people who had a new stroke (70\\u00b72% [65\\u00b79 to 74\\u00b76]), survived stroke (86\\u00b71% [83\\u00b70 to 89\\u00b74]), died from stroke (44\\u00b71% [32\\u00b73 to 56\\u00b70]), and who died or remained disabled from stroke (as measured by DALYs; 32\\u00b72% [21\\u00b77 to 42\\u00b77]; table 1; appendix pp 100\\u2013101). The percentage decline in age-standardised stroke incidence rates in the 2019\\u20132021 period (\\u20131\\u00b78% [\\u20132\\u00b78 to \\u20130\\u00b76]) was smaller than that for the overall 2010\\u201321 period (\\u20133\\u00b71% [\\u20134\\u00b72 to \\u20132\\u00b70]).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"sec1\", \"sec1\"], \"section\": \"Trends in stroke burden by age, sex and geographical location\", \"text\": \"Although all-age (not age-standardised) stroke incidence, death, and DALY rates were substantially reduced in people aged 70 years or older between 1990 and 2021 (\\u201318\\u00b72% [95% UI \\u201321\\u00b73 to \\u2013 14\\u00b76] incidence rate, \\u201334\\u00b72% [\\u201339\\u00b74 to \\u201329\\u00b73] death rate, and \\u201335\\u00b76% [\\u201340\\u00b72 to \\u201330\\u00b78] DALY rate), and all-age prevalence rate in this age group did not change over this period (\\u20131\\u00b70 [\\u20133\\u00b71 to 1\\u00b72]), all-age incidence increased by 4\\u00b71% (0\\u00b79 to 7\\u00b76), prevalence increased in people younger than 70 years by 14\\u00b78% (13\\u00b71 to 16\\u00b78), and death and DALY rates were reduced in people younger than 70 years by 17\\u00b74% (\\u201325\\u00b70 to \\u20138\\u00b79) and 19\\u00b70% (\\u201326\\u00b70 to \\u201311\\u00b76), respectively (appendix p 140). Similar patterns were observed for all-age incidence, prevalence, death, and DALY rates of ischaemic stroke for both age groups (<70 years and \\u226570 years). Of the three pathological types of stroke, only all-age subarachnoid haemorrhage prevalence rates increased in people younger than 70 years, by 3\\u00b74% (1\\u00b73 to 5\\u00b75), whereas all-age incidence, prevalence, death, and DALY rates of intracerebral haemorrhage and incidence, death, and DALY rates of subarachnoid haemorrhage were reduced in both age groups (appendix pp 88, 210\\u2013211).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"tbl2\", \"sec1\"], \"section\": \"Contribution of risk factors to stroke-related DALYs\", \"text\": \"Globally, the total number of stroke-related DALYs due to risk factors increased substantially from 1990 (100\\u00b71 million [95% UI 92\\u00b77 to 107\\u00b78]) to 2021 (135\\u00b70 million [122\\u00b70 to 147\\u00b77]), but there was no substantial change in the age-standardised stroke DALYs attributable to risk factors (\\u20130\\u00b75% [\\u20132\\u00b74 to 1\\u00b71]). In 2021, 84\\u00b71% (77\\u00b78 to 88\\u00b78) of DALYs from stroke were attributed to the 23 risk factors analysed (table 2), with the largest proportions of attributable risks for total stroke, ischaemic stroke, intracerebral haemorrhage, and subarachnoid haemorrhage observed in eastern Europe, Asia, and sub-Saharan Africa (appendix p 212).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"tbl2\", \"sec1\", \"sec1\"], \"section\": \"Contribution of risk factors to stroke-related DALYs\", \"text\": \"At level 1 of the GBD risk factors hierarchy (table 2, appendix pp 141\\u2013253), metabolic risk factors contributed most to the stroke-related DALYs (range 66\\u201370%) across all World Bank country income levels, followed by the environmental risk cluster in low-income, lower-middle-income, and upper-middle-income countries (range 35\\u201353%), and behavioural risks (range 30\\u201338%) across different income level countries. Stroke burden associated with the environmental or occupational risks was lowest in high-income countries (18\\u00b71% [95% UI 13\\u00b76 to 22\\u00b77]). Similarly, regions with higher SDI (appendix pp 141\\u2013143) had a larger contribution of metabolic and behavioural risks to stroke-related DALYs, whereas environmental risks most prominently contributed to stroke-related DALYs in lower SDI quintiles. From 1990 to 2021, the age-standardised proportion of stroke DALYs attributable to risk factors increased in north Africa and the Middle East (6\\u00b78% [4\\u00b71 to 11\\u00b75]) and sub-Saharan Africa (3\\u00b73% [1\\u00b78 to 5\\u00b73]), but did not change in south Asia (0\\u00b74% [\\u20131\\u00b70 to 1\\u00b79]) and southeast Asia, east Asia, and Oceania (\\u20130\\u00b79% [\\u20134\\u00b72 to 1\\u00b79]), and decreased in central Europe, eastern Europe, and central Asia (\\u20132\\u00b70% [\\u20133\\u00b78 to \\u20130\\u00b77]) and Latin America and the Caribbean (\\u20135\\u00b70% [\\u20139\\u00b71 to \\u20132\\u00b73]), as well as high-income GBD regions (\\u20137\\u00b73% [\\u201310\\u00b71 to \\u20135\\u00b70]).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"fig2\"], \"section\": \"Contribution of risk factors to stroke-related DALYs\", \"text\": \"Globally, of the 23 risk factors analysed, 14 individually significant risk factors for stroke were high systolic blood pressure (56\\u00b78% [95% UI 42\\u00b75\\u201368\\u00b70] attributable DALYs), ambient particular matter (16\\u00b76% [11\\u00b75\\u201320\\u00b79]), smoking (13\\u00b78% [2\\u00b75\\u201326\\u00b70]), high LDL cholesterol (13\\u00b71% [4\\u00b76\\u201321\\u00b73]), household air pollution (11\\u00b72% [6\\u00b74\\u201319\\u00b73]), diet high in sodium (10\\u00b76% [2\\u00b78\\u201322\\u00b78]), high fasting plasma glucose (10\\u00b73% [8\\u00b71\\u201312\\u00b76]), kidney disfunction (9\\u00b73% [6\\u00b78\\u201311\\u00b78]), diet low in fruits (5\\u00b79% [0\\u00b74\\u201310\\u00b74]), high alcohol use (5\\u00b72% [1\\u00b73\\u20139\\u00b78]), high BMI (4\\u00b77% [0\\u00b74\\u20139\\u00b78]), second-hand smoking (4\\u00b74% [1\\u00b70\\u20137\\u00b79]), low physical activity (2\\u00b71% [0\\u00b75\\u20133\\u00b79]), and diet low in vegetables (1\\u00b76% [0\\u00b74\\u20132\\u00b76]; figure 2).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"fig3\", \"sec1\", \"fig4\", \"sec1\", \"sec1\", \"sec1\", \"sec1\"], \"section\": \"Contribution of risk factors to stroke-related DALYs\", \"text\": \"Stroke attributable to metabolic risks constituted 68\\u00b78% (95% UI 57\\u00b76 to 77\\u00b75) of all strokes, environmental risks constituted 36\\u00b77% (29\\u00b70 to 44\\u00b72), and behavioural risks constituted 35\\u00b72% (26\\u00b79 to 44\\u00b77). Although the proportion of stroke DALYs attributable to metabolic risks increased from 1990 to 2021 by 6\\u00b77% (3\\u00b78 to 10\\u00b70; mainly because of the increase in the burden attributable to high BMI, high fasting plasma glucose, and high systolic blood pressure), proportions of stroke DALYs attributable to behavioural risks decreased by 8\\u00b70% (\\u201313\\u00b76 to \\u20133\\u00b74) and those due to environmental risks by 14\\u00b78% (\\u201321\\u00b76 to \\u20138\\u00b77), mainly because of the decrease in the burden attributable to diet high in processed meat, diet low in vegetables, diet low in fibre, low ambient temperature, particulate matter pollution, diet low in fruits, and smoking (figure 3). However, from 1990 to 2021, there was a substantial increase in the stroke DALYs attributable to high ambient temperature, high fasting plasma glucose, diet high in sugar-sweetened beverages, low physical activity, diet high in red meat, lead exposure, and diet low in omega-6 polyunsaturated fatty acids. There were noticeable geographical and regional variations in the PAF of the risk factors for ischaemic stroke, intracerebral haemorrhage, subarachnoid haemorrhage, and all stroke types combined (appendix p 264), as well as in the ranking of PAFs of age-standardised stroke DALYs attributable to risk factors by 21 GBD regions (figure 4). For the PAF of risk factors by pathological type of stroke, SDI, 21 GBD regions, and 204 countries and territories were used (appendix pp 89\\u2013201). Unlike the PAF of risk factors for total stroke, high alcohol use was not associated with ischaemic stroke-related DALYs (appendix pp 92\\u201395, 126\\u2013149, 199), and diet low in fruits and vegetables and high BMI were not associated with intracerebral haemorrhage-related DALYs (appendix pp 92\\u201395, 150\\u2013173, 200). Unlike ischaemic stroke and intracerebral haemorrhage, non-optimal ambient temperature appeared to be associated with the subarachnoid haemorrhage-related DALYs, with the greater contribution of low ambient temperature (4\\u00b75% [3\\u00b78 to 5\\u00b73]) than high ambient temperature (1\\u00b71% [0\\u00b72 to 2\\u00b75]). Other substantial risk factors for subarachnoid haemorrhage (appendix pp 151\\u2212153) were second-hand smoking (4\\u00b77% [3\\u00b72 to 6\\u00b72]), diet high in sodium (8\\u00b79% [2\\u00b70 to 19\\u00b78]), household air pollution from solid fuels (10\\u00b73% [5\\u00b75 to 17\\u00b74]), ambient particulate matter pollution (14\\u00b72% [9\\u00b78 to 18\\u00b70]), smoking (14\\u00b75% [2\\u00b77 to 27\\u00b72]), and high systolic blood pressure (51\\u00b76% [38\\u00b70 to 62\\u00b76]).\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib5\", \"bib17\", \"bib18\", \"bib19\", \"bib20\", \"bib21\", \"bib22\", \"bib3\", \"bib23\", \"bib24\", \"bib25\", \"bib26\", \"bib27\", \"bib28\", \"bib29\", \"bib24\", \"bib26\", \"bib27\", \"bib28\", \"bib29\", \"bib30\", \"bib31\", \"bib32\", \"bib3\", \"bib33\"], \"section\": \"Discussion\", \"text\": \"In 2021, stroke remained the second most common cause (after ischaemic heart disease) of death and the third most common cause of DALYs (after ischaemic heart disease and neonatal disorders) among non-communicable disorders (NCDs) globally. However, at level 3 of the GBD all-cause hierarchy, stroke was the third most common cause of death (after ischaemic heart disease and COVID-19) and the fourth most common cause of DALYs (after COVID-19, ischaemic heart disease, and neonatal disorders), with the bulk of the stroke burden in LMICs and countries with lower SDI. Consistent with previous studies,5, 17 this study showed disparities in stroke burden (including almost two times greater proportion of intracerebral haemorrhage in LMICs than in high-income countries) and risk factors by GBD region, country, country income level, and SDI quintiles, as well as an overall trend towards decreasing age-standardised stroke incidence, prevalence, and DALY rates from 1990 to 2021. Given the leading role of arterial hypertension in the occurrence of intracerebral haemorrhage, the greater prevalence and poorer control of hypertension in LMICs than in high-income countries are the most likely causes of the differences in the proportion of intracerebral haemorrhage in those countries. Similar proportions of subarachnoid haemorrhage in high-income countries and LMICs are likely to be related to the significant reduction in tobacco smoking prevalence that we observed since 2010 in high-income countries, with almost no change in tobacco smoking prevalence in LMICs. However, the current study documented a stagnation in the reduction of age-standardised incidence rates from 2015 onwards, and even some increase in the age-standardised stroke incidence, death, prevalence, and DALY rates in southeast Asia, east Asia, and Oceania, and countries with lower SDI from 2015 onwards. Globally, there was also an increase in all-age incidence and prevalence rates in people younger than 70 years, whereas there was a reduction in all-age stroke incidence, prevalence, death, and DALY rates in people aged 70 years or older. A trend towards increasing incidence and prevalence rate of cardiovascular diseases (including stroke) in people aged 15\\u201339 years globally and stroke incidence rates in people younger than 55 years versus older people has also been shown in previous systematic reviews and is likely to be related to the increase in prevalence of arterial hypertension23, 24 (including poorly controlled and uncontrolled hypertension), overweight or obesity,26, 27, 28 and type 2 diabetes in young adults, especially in LMICs.24, 26, 27, 28, 29 This situation is complicated by the fact that a large proportion of young adults with vascular risk factors, arterial hypertension, and dyslipidaemia remain under-treated owing at least partly to the widespread use of absolute cardiovascular disease risk treatment thresholds. The observed slowing down of the percentage of decline of age-standardised stroke incidence rates in 2019\\u201321 is likely to be related to the decreased hospital admissions of patients with acute stroke that was observed over the COVID-19 period in many countries.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib1\", \"bib34\", \"bib35\", \"bib36\", \"bib17\", \"bib37\", \"bib38\"], \"section\": \"Discussion\", \"text\": \"Apart from population growth and ageing,1, 34 other factors responsible for the increased burden of stroke, in terms of absolute numbers in the world, are likely to be related to the insufficient effectiveness of the currently used primary stroke and cardiovascular disease prevention strategies35, 36 as well as the disparities and major gaps in stroke service provision and accessibility, and workforce of stroke care providers in many countries (especially LMICs).17, 37, 38 Although stroke is highly preventable, globally there were substantial increases in DALYs attributable to high BMI, high ambient temperature, high fasting plasma glucose, diet high in sugar-sweetened beverages, low physical activity, high systolic blood pressure, and diet low in polyunsaturated omega-6 fatty acids, suggesting the growing role of these environmental and behavioural risks in the stroke burden. However, from 1990 to 2021, we also observed a reduction of PAF due to diet high in processed meat, diet low in vegetables, diet low in fibre, low ambient temperature, ambient particulate matter pollution, diet low in fruits, and smoking, suggesting effectiveness of the strategies towards reduction of the exposure to these risk factors. The observed increase in the age-standardised proportion of stroke DALYs attributable to risk factors in north Africa and the Middle East and sub-Saharan Africa regions might reflect a failure in the control of stroke risk factors. However, in central Europe, eastern Europe, central Asia, Latin America and the Caribbean, and high-income GBD regions, this might reflect a success in the control of stroke risk factors.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib39\", \"bib40\", \"bib41\", \"bib42\", \"bib43\", \"bib44\", \"bib44\", \"bib44\", \"bib45\", \"bib46\", \"bib47\", \"bib48\", \"bib47\", \"bib48\", \"bib44\", \"bib49\"], \"section\": \"Discussion\", \"text\": \"This study is, to our knowledge, the first to show the large contribution of ambient particulate matter pollution and household air pollution from solid fuels to subarachnoid haemorrhage DALYs, with a similar PAF to that of smoking. A close relationship between ambient air pollution and subarachnoid haemorrhage mortality was found in some studies.39, 40, 41 Air pollution in 2021 appeared to be highly important to other types of stroke and also caused 11\\u00b79% (95% UI 10\\u00b70\\u201313\\u00b78) of total deaths from all causes, making it the second largest cause of deaths from all causes globally (after high systolic blood pressure) and the second leading cause of DALYs (8\\u00b72% [6\\u00b79\\u20139\\u00b76]) from all causes (after malnutrition). These findings are in line with research showing that rises in ambient temperature (including heatwaves) and climate change are associated with increased stroke morbidity and mortality.43, 44 Because ambient air pollution is reciprocally associated with the ambient temperature and climate change, all of which synergistically influence cardiovascular disease (including stroke) occurrence44, 45, 46 and overall health,47, 48 the importance of urgent climate actions and measures to reduce ambient air pollution cannot be overestimated.47, 48 Experts have recommended that governments increase implementation of a clean-energy economy, promote unprocessed plant-based food choices, and globally phase out industrialised animal farming.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib38\", \"bib38\", \"bib50\", \"bib51\", \"bib52\", \"bib50\"], \"section\": \"Discussion\", \"text\": \"Every member state of the UN has committed to meeting the Sustainable Development Goals (SDGs), but currently few countries are on target to achieve SDG 3.4, which is to reduce by a third premature mortality from NCDs through prevention and treatment and promote mental health and wellbeing by 2030. By implementing and monitoring the World Stroke Organization\\u2013Lancet Neurology Commission's recommendations, the global burden of stroke would be reduced drastically this decade and beyond. Not only would this substantial reduction enable SDG 3.4, as well as other key SDGs, to be met, it would improve brain health and the overall wellbeing of millions of people across the globe. One of the most common problems in implementing stroke prevention and care recommendations is the scarcity of funding. The World Stroke Organization\\u2013Lancet Neurology Commission on stroke recommends introducing legislative regulations and taxation of unhealthy products by each government in the world. Such taxation would not only reduce consumption of these products and, therefore, lead to the reduction of burden from stroke and other major NCDs,50, 51, 52 but also generate a large revenue sufficient to fund prevention programmes and services for stroke and other major disorders, reduce poverty and inequality in health service provision, improve wellbeing of the population, and boost local economies.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib7\"], \"section\": \"Discussion\", \"text\": \"The main strength of this study is the extended number of data sources included in the analysis that allowed us to generate more accurate and up-to-date stroke burden and risk factor estimates. This allows evidence-based health-care planning and resource allocation by health policy makers on the national, regional, and global levels. However, good-quality stroke epidemiological studies are still scarce in most countries, which prevented us from including in the analysis many other important risk factors, such as sickle cell disease and HIV, which are particularly important for sub-Saharan Africa. Differences in health-care systems and completeness and accuracy of stroke case ascertainment might play a part in the observed between-country differences (eg, very high stroke prevalence in Ghana compared with neighbouring countries). Although the GBD methods for estimating attributable burden of stroke due to risk factors accounts for a cumulative effect of multiple risk factors, it might not fully account for all potential confounders. Moreover, some new risk factors, such as high ambient temperature, might require further validation and examination to confirm their impact on stroke burden. Furthermore, more granular data analysis is needed. For example, stroke burden variation by race and ethnicity within countries, which can mask disparities in stroke incidence, risk factors, and outcomes among different population groups, and analysis of attributable effects of different levels of exposure to smoking, alcohol, and so on. We expect such analysis will be done in future GBD iterations.\"}, {\"pmc\": \"PMC12254192\", \"pmid\": \"39304265\", \"reference_ids\": [\"bib38\"], \"section\": \"Discussion\", \"text\": \"In summary, our study findings continue to point out that currently used stroke prevention strategies are not sufficiently effective to halt, let alone reduce, the fast-growing stroke burden. Additional, more effective stroke prevention strategies (with the emphasis on population-wide measures, task shifting from doctors to nurses or health volunteers, and the wider use of evidence-based mobile and telehealth platforms) and pragmatic solutions to address the critical gaps in stroke service delivery, along with development of context-appropriate workforce capacity building and epidemiological surveillance systems, need to be urgently implemented across all countries. Without scaling up these innovative evidence-based strategies and policies that target local, national, regional, and global stroke prevention and care disparities, the burden of stroke will continue to grow, thus threatening the sustainability of health systems worldwide.\"}]"

Metadata

"{}"