Risk of stroke in SLE: a systematic review and meta-analysis Authors: Jingya Zhang, Weiman Shi, Yaling Xu, Dianming Li, Wu Chen, Lin Huang, Xuanlin Li, Qiao Wang Published: 2026 ABSTRACT AbstractThe association between SLE and composite stroke, ischaemic stroke and haemorrhagic stroke remains incompletely understood. This meta-analysis aims to assess the risk of stroke in patients with SLE.MethodsData sources included PubMed, Embase, the Cochrane Library and reference lists of included studies. This meta-analysis included cohort studies evaluating whether stroke risk is associated with SLE. The risk of bias was assessed using the Newcastle-Ottawa Quality Assessment Scale (NOS). Risk ratios (RRs) with 95% CIs were pooled using a random-effects model, and publication bias was assessed with funnel plots and Egger’s test.ResultsA total of 25 cohort studies involving 5 220 837 individuals were included in this meta-analysis, which were published between 2001 and 2026. The pooled analysis demonstrated a significantly increased risk of stroke in patients with SLE (RR of 2.60, 95% CI 2.21 to 3.05, I²=97.9%, p<0.001). The risk of composite stroke (RR of 2.83, 95% CI 2.25 to 3.57, I²=98.0%, p<0.001), ischaemic stroke (RR of 2.34, 95% CI 1.75 to 3.12, I²=97.6%, p<0.001) and haemorrhagic stroke (RR of 2.66, 95% CI 1.57 to 4.49, I²=96.2%, p<0.001) was also increased in SLE. Despite the large heterogeneity, the sensitivity analysis indicated that the results were robust, and there was little evidence of publication bias.ConclusionThe risk of composite stroke, ischaemic stroke and haemorrhagic stroke is increased in SLE.PROSPERO registration numberCRD420261294082. WHAT IS ALREADY KNOWN ON THIS TOPIC Patients with SLE are at increased cardiovascular risk due to chronic inflammation, endothelial dysfunction and antiphospholipid antibody-mediated hypercoagulability; however, the quantitative associations between SLE and specific stroke subtypes—particularly haemorrhagic stroke—remain incompletely characterised. Patients with SLE are at increased cardiovascular risk due to chronic inflammation, endothelial dysfunction and antiphospholipid antibody-mediated hypercoagulability; however, the quantitative associations between SLE and specific stroke subtypes—particularly haemorrhagic stroke—remain incompletely characterised. WHAT THIS STUDY ADDS This meta-analysis of 25 cohort studies involving over 5.2 million individuals demonstrates that SLE is an independent risk factor for all stroke subtypes, with significantly elevated risks of composite stroke (RR=2.83), ischaemic stroke (RR=2.34) and, notably, haemorrhagic stroke (RR=2.66)—the first meta-analysis to quantify this association. This meta-analysis of 25 cohort studies involving over 5.2 million individuals demonstrates that SLE is an independent risk factor for all stroke subtypes, with significantly elevated risks of composite stroke (RR=2.83), ischaemic stroke (RR=2.34) and, notably, haemorrhagic stroke (RR=2.66)—the first meta-analysis to quantify this association. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings support incorporating routine cerebrovascular risk assessment into SLE management and considering SLE as an independent stroke risk factor in clinical guidelines, while calling for mechanistic studies and age-stratified trials to optimise personalised prevention strategies. These findings support incorporating routine cerebrovascular risk assessment into SLE management and considering SLE as an independent stroke risk factor in clinical guidelines, while calling for mechanistic studies and age-stratified trials to optimise personalised prevention strategies. Introduction SLE is a chronic multisystem autoimmune disease characterised by loss of immune tolerance, massive production of autoantibodies and immune complex deposition, leading to widespread inflammation and tissue damage.1 2 The disease exhibits significant clinical heterogeneity, commonly affecting the skin, joints, kidneys, haematologic system and central nervous system.3 Persistent active inflammation can cause irreversible organ damage, posing a serious threat to patient survival. Notably, the female-to-male incidence ratio is 9:1, reflecting substantial disparities in gender and peak onset age.4 5 Beyond direct disease-related damage, long-term health risks for patients with SLE are increasingly recognised, with cardiovascular events being particularly prominent.6 Stroke, the second leading cause of death globally and the primary cause of adult disability, poses a major public health challenge due to its high incidence, disability rate, recurrence rate and substantial socioeconomic burden.7 Substantial evidence indicates that chronic inflammation-driven endothelial dysfunction, premature atherosclerosis and hypercoagulable states mediated by specific autoantibodies such as antiphospholipid antibodies collectively form the core pathophysiological basis for the markedly elevated stroke risk in patients with SLE.8 9 Patients with SLE, a chronic autoimmune disease characterised by widespread inflammation, are at an elevated risk for cerebrovascular events, including both ischaemic and haemorrhagic stroke.10 This meta-analysis aims to quantitatively evaluate the association between SLE and stroke incidence based on existing research evidence, thereby providing evidence-based medical support for early identification of high-risk patients, risk stratification and the development of targeted prevention strategies. Methods This meta-analysis is reported according to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 explanation and elaboration.11 The protocol of this systematic review and meta-analysis was registered with the International Prospective Register of Systematic Reviews (PROSPERO, CRD420261294082). Data sources Systematic retrieval of PubMed, Embase and Cochrane Library databases was conducted from the inception until 14 April 2026. Our search terms included ‘Lupus Erythematosus, Systemic’, ‘Stroke’ and their relevant synonyms. The complete search strategy is available in online supplemental tables 1-3. Inclusion criteria The inclusion criteria for the study were as follows: (1) patients: patients diagnosed with SLE and aged ≥18 years; (2) exposure: a definitive diagnosis of stroke, composite stroke, ischaemic stroke or haemorrhagic stroke; (3) comparator: healthy people or non-SLE sufferers; (4) outcomes: a presentation of quantitative point estimates (HRs, relative risks (RRs) or ORs) and a variance of the estimates of the association between SLE and the risk of stroke and also a description of adjustment for potential confounders; and (5) type of study: cohort study, either prospective or retrospective; (6) written in English. Exclusion criteria The exclusion criteria for the study were as follows: (1) duplicate records; (2) conference abstracts or letters to editors; (3) incomplete data or no interested outcome. Study selection and data extraction Two reviewers independently extracted data using a predesigned table. The variables extracted included, but were not limited to, the first author’s name, year of publication, country of origin, sample size, number of events, gender distribution and stroke type. After extracting the data, cross-checking was performed to ensure data accuracy. This process strictly adhered to standardised protocols, minimising the risk of errors and enhancing data reliability. Any discrepancies encountered during extraction were resolved by a third independent reviewer. Quality assessment This study used the Newcastle-Ottawa Quality Assessment Scale (NOS) to assess the quality of cohort studies. The NOS includes three dimensions—selection of study subjects, comparability between groups and measurement of outcomes—with a total of eight items, scoring from 0 to 9 points, with higher scores indicating better study quality. Specifically, the study population selection dimension (4 points) assesses the representativeness of the study sample, selection of the exposed population, selection of the unexposed population and the absence of outcome events at the start of the study; the comparability of groups dimension (2 points) focuses on whether the study controlled for important confounding factors; and the outcome measurement dimension (3 points) focuses on the reliability of outcome assessment, follow-up completeness and analysis of loss to follow-up rates. Based on the total score, included studies are categorised into three quality levels: low quality (0–3 points), moderate quality (4–6 points) and high quality (7–9 points). Statistical analysis All statistical analyses were performed using Stata software (V.14.0). ORs, HRs, rate ratios, standardised incidence ratios and standardised mortality ratios were considered equal estimates assuming rare occurrence and referred to as risk ratios throughout this publication. The most adjusted RR was used. Subarachnoid haemorrhage (SAH) and intracerebral haemorrhage (ICH) are the main subtypes of haemorrhagic stroke; in this study, both subtypes are classified as haemorrhagic stroke.12 Composite stroke is a composite endpoint of ischaemic stroke and haemorrhagic stroke.13 These data will be used in subsequent meta-analyses to quantify the risk of stroke in patients with SLE. By meticulously extracting and cross-checking data from each study, a reliable data foundation is established for subsequent analyses, ensuring a comprehensive and accurate assessment of the association between SLE and stroke. We assessed heterogeneity using the χ2 test and I² value, and p<0.1 or I²>50% indicated that heterogeneity was great, and thus the random-effects model was adopted. Otherwise, a fixed-effect model was employed according to a previous high-quality meta-analysis.14 The sensitivity analysis was performed to verify the robustness of the overall results and to explore sources of heterogeneity. The subgroup analysis was performed according to age, sex and area. Finally, funnel plots, Begg’s test and Egger’s regression test were used to detect publication bias.15 Results Literature search Through an initial search, a total of 12 672 relevant literature articles were obtained, all search records were imported into the NoteExpress literature management software for unified management. After a stepwise screening process, 25 cohort studies were ultimately included in the meta-analysis.16–40 The literature screening process complied with the PRISMA statement guidelines, with the specific screening process detailed in figure 1. Literature screening flowchart. Study characteristics A total of 25 cohort studies were included, with a total sample size of 5 220 837, which comprised 1 408 205 patients with SLE and 3 812 632 controls. There are 15 studies related to SLE developing composite stroke, 15 studies analysed SLE developing ischaemic stroke, 8 studies analysed SLE developing haemorrhagic stroke. These different results show that the relationship between SLE and different stroke is complex and diverse. The baseline data of the included subjects are shown in table 1. Seven studies are moderate quality, while 18 studies are high quality (table 2). Table 1Characteristics of studies included in the reviewReferencesCountrySample sizeFemale of SLE populationBaseline age (years)Diagnosis of SLEStroke typeHan (2025)16KoreaTotal: 21 360, SLE: 4272, no SLE: 17 08887.17%53.11; NRICD-10, M32.0Ischaemic strokeHuang (2024)17ChinaTotal: 1 010 461, SLE: 10 461, no SLE:000 00086.93%40.48; 40.48ICD-9-CM, 710.0Ischaemic strokeHuang (2023)18ChinaTotal: 24 930, SLE: 8310, no SLE: 16 62088.6%37.69; 38.34Code 710.0Composite stroke; Ischaemic stroke; Haemorrhagic strokeChang (2023)19ChinaTotal: 20 012, SLE: 10 006, no SLE: 10 00688.5%; 88.5%37.90; 38.53ICD-9-CM, 710.0 or ICD-10, M32Ischaemic stroke; Haemorrhagic strokeShams-Eldin (2022)20DanishTotal: 10 310, SLE: 2062, no SLE: 824887.7%50; 50ICD-8; ICD-10Ischaemic strokeBarbhaiya (2021)21USATotal: 57 300, SLE: 40 212, no SLE: 160 84892.45%40.34; 40.34ICD-9-CM, 710.0Composite strokeLai (2020)22ChinaTotal: 1106, SLE: 622, no SLE: 48474.2%; 31.3%53.5; 70.0ICD-9-CM, 710.0Ischaemic strokeBarbhaiya (2019)23USATotal: 65 88, SLE: 65 788, no SLE: NR93.1%47.04; NRICD-9-CM, 710.0Composite stroke; Ischaemic stroke; Haemorrhagic strokeCook (2018)24UKTotal: 2795, SLE: 559, no SLE: 223689%56; NRICD-10, M32.0Ischaemic strokeLim (2018)25KoreaTotal: 111 450, SLE: 18 575, no SLE: 92 87591%NRICD-10Composite strokeArkema (2017)26SwedenTotal: 2 120, SLE: 3390, no SLE: 167 3085.1%; 85.0%50; 49ICDComposite stroke; ischaemic stroke; haemorrhagic strokeZubeia (2017)27CanadaTotal: 54 920, SLE: 4972, no SLE: 49 94886%48.5; 48.6ICD-9; ICD-10Ischaemic strokeHermansen (2017)28DenmarkTotal: 8920, SLE: 1459, no SLE: 746187%47.9; 47.9ICD-8; ICD-10Composite strokeRees (2016)40UKTotal: 33 696, SLE: 7033, no SLE: 26 68386%48.1; 48.1Read codesComposite strokeLiou (2014)29ChinaTotal: 3105 SLE: 621, no SLE: 248488.9%;88.9%NRICD-9-CMComposite stroke; ischaemic strokeWu (2014)30ChinaTotal: 1,132 089, SLE: 1 132 089, no SLE: NR100%NRICD-9-CMComposite strokeRamagopalan (2013)39EnglandTotal: 25 576, SLE: 25 576, no SLE: NR86%NRICD-9; ICD-10Haemorrhagic strokeChiu (2012)31ChinaTotal: 69 822 SLE: 11 637, no SLE: 58 18589%56; NRICD-10Ischaemic strokeWang (2012)32ChinaTotal: 68 625, SLE: 13 869, no SLE: 54 75688.2%;88.2%35; 35ICD-9-CMComposite stroke; ischaemic stroke; haemorrhagic strokeZoller (2012)33SwedenTotal: 4179, SLE: 4179, no SLE: NR82%NRICD-7; ICD-8; ICD-9; ICD-10Ischaemic stroke; haemorrhagic strokeMok (2009)35ChinaTotal: 1 050 490, SLE: 490, no SLE: 1 050 00092%33; NRICD-9-CMComposite stroke; ischaemic stroke; haemorrhagic strokeHak (2009)34USATotal: 109 114, SLE: 148, no SLE: 108 966100%56.2; 56.2ACRComposite strokeBernatsky (2006)36MultinationalTotal: 9547, SLE: 9547, no SLE: NR90%NRICD-9-CMComposite strokeBjornádal (2004)37SwedenTotal: 4727, SLE: 4727, no SLE: NR79%NRICD-7; ICD-8; ICD-9Composite strokeEsdaile (2001)38CanadaTotal: 263, SLE: 263, no SLE: NR95.7%43; 43ACRComposite strokeACR, American College of Rheumatology; ICD, International Classification of Diseases; NR, not reported. Characteristics of studies included in the review ACR, American College of Rheumatology; ICD, International Classification of Diseases; NR, not reported. Table 2The quality assessment of cohort and case-control studiesStudy (year)SelectionComparabilityOutcomeTotalHan (2025)16*********9Huang (2024)17********8Huang (2023)18********8Chang (2023)19*********9Shams-Eldin (2022)20*********9Barbhaiya (2021)21********8Lai (2020)22********8Barbhaiya (2019)23******6Cook (2018)24*******7Lim (2018)25***/***6Arkema (2017)26*********9Zubeia (2017)27********8Hermansen (2017)28*********9Rees (2016)40********8Liou (2014)29*********9Wu (2014)30***/***6Ramagopalan (2013)39***/***6Chiu (2012)31********8Wang (2012)32*********9Zoller (2012)33***/***6Mok (2009)35********8Hak (2009)34*********9Bernatsky (2006)36***/***6Bjornádal (2004)37***/***6Esdaile (2001)38*******7"*", 1 point; "**", 2 point; "***", 3 point; "****", 4 point. The quality assessment of cohort and case-control studies "*", 1 point; "**", 2 point; "***", 3 point; "****", 4 point. Stroke Meta-analyses were performed for the following stroke outcomes: composite stroke (n=15), haemorrhagic stroke (n=8) and ischaemic stroke (n=15). SLE was associated with an increased risk of stroke, with a pooled RR=2.60, 95% CI 2.21 to 3.05, I²=97.9%, p<0.001 (figure 2). Forest plot for the risk of stroke in SLE. ES, effect size. Composite stroke Composite stroke was reported in 15 studies, which was identified by International Classification of Disease (ICD)-8, ICD-9 and ICD-10 codes in 12 of 15 studies. In the remaining studies, physician diagnosis or national insurance claims data confirmed case identification. SLE was associated with an increased risk of composite stroke, with a pooled RR of 2.83 (95% CI 2.25 to 3.57; I²=98.0%, p<0.001, figure 2). Due to significant heterogeneity, we performed sensitivity analysis by removing each study to explore the source of heterogeneity, which suggested that the results were stable after excluding any one study, suggesting that the results of our meta-analysis were robust (figure 3). Sensitivity analysis of risk of composite stroke in SLE. Ischaemic stroke Ischaemic stroke was reported in 15 studies, which was identified by ICD-8, ICD-9 and ICD-10 codes in 14 of 15 studies. In the remaining studies, physician diagnosis or national insurance claims data confirmed case identification. SLE was associated with an increased risk of ischaemic stroke, with a pooled RR of 2.34, 95% CI 1.75 to 3.12, I²=97.6%, p<0.001 (figure 2). Due to significant heterogeneity, we performed sensitivity analysis by removing each study to explore the source of heterogeneity, which suggested that the results were stable after excluding any one study, suggesting that the results of our meta-analysis were robust (figure 4). Sensitivity analysis of risk of ischaemic stroke in SLE. Haemorrhagic stroke Haemorrhagic stroke was reported in eight studies, which was identified by ICD-8, ICD-9 and ICD-10 codes in seven of eight studies. In the remaining studies, physician diagnosis or national insurance claims data confirmed case identification. SLE was associated with an increased risk of haemorrhagic stroke, with a pooled RR of 2.66, 95% CI 1.57 to 4.49; I²=96.2%; p<0.001 (figure 2). Due to significant heterogeneity, we performed sensitivity analysis by removing each study to explore the source of heterogeneity, which suggested that the results were stable after excluding any one study, suggesting that the results of our meta-analysis were robust (figure 5). Sensitivity analysis of risk of haemorrhagic stroke in SLE. Publication bias Visual inspection suggested that the funnel plot for stroke in SLE was symmetrical but in formal statistical tests, including Egger’s test (p=0.005) and Begg’s test (p=0.005) (online supplemental figure 1), publication bias was shown. The funnel plot is shown in figure 6. Funnel plot for stroke in SLE. LN, lupus nephritis; RR, risk ratio. However, after two iterations of trim-and-treat correction, no virtual results were ultimately generated (online supplemental figure 2). Nevertheless, trim-and-treat correction eliminated publication bias. The combined effect size remained unchanged post-trim-and-treat, suggesting that the results of existing meta-analyses are stable (online supplemental figure 3). Subgroup analysis The results for subgroups stratified according to the age stratification of the SLE were also consistent. After excluding cohort studies with incomplete age information, the risk of stroke was significantly increased when SLE patients are older than 40 years of age (RR of 2.38, 95% CI 1.88 to 3.02; I²=98.6%; p<0.001; table 3, online supplemental figure 4), when the patient is under 40 years of age (RR of 2.84, 95% CI 2.26 to 3.57; I²=93.2%; p<0.001; table 3, online supplemental figure 4). Table 3Subgroup analysis for the associations between stroke and SLESubgroupsIncluded studiesRR(95% CI)HeterogeneityI2 (%)P valueAge ≥40 years142.38 (1.88 to 3.02)98.60.000 <40 years42.84 (2.26 to 3.57)93.20.000Proportion of female ≥90%83.08 (2.33 to 4.07)98.70.000 <90%172.40 (2.10 to 2.73)89.70.000Regions Asia112.61 (2.20 to 3.10)91.70.000 North America53.25 (2.30 to 4.61)99.20.000 Europe82.31 (1.89 to 2.82)84.20.000 Subgroup analysis for the associations between stroke and SLE The results for subgroups stratified according to proportion of female in the cohort studies were also consistent. The risk of stroke was significantly increased when SLE more than 90% of the female in the cohort (RR of 3.08, 95% CI 2.33 to 4.07; I²=98.7%; p<0.001; table 3, online supplemental figure 5), when the proportion of female in the cohort study does not exceed 90% (RR of 2.40, 95% CI 2.10 to 2.73; I²=89.7%; p<0.001; table 3, online supplemental figure 5). The results for subgroups stratified according to classification of regions from cohort studies were also consistent. After excluding a multinational cohort study, the risk of stroke was significantly increased when SLE from Asia (RR of 2.61, 95% CI 2.20 to 3.10; I²=91.7%; p<0.001; table 3, online supplemental figure 6), North America (RR of 3.25, 95% CI 2.30 to 4.61; I²=99.2%; p<0.001; table 3, online supplemental figure 6) and Europe (RR of 2.31, 95% CI 1.89 to 2.82; I² = 84.2%; p=0.003; table 3, online supplemental figure 6). Discussion This systematic review and meta-analysis synthesise data from 25 cohort studies worldwide involving over 5 million individuals, providing the most comprehensive and up-to-date quantitative evidence to date on the association between SLE and stroke risk. Our core findings indicate that patients with SLE exhibit significantly elevated risks for composite stroke, ischaemic stroke and haemorrhagic stroke, with combined RRs of 2.83, 2.34 and 2.66, respectively. These results strongly confirm that SLE represents an independent and substantial risk factor for all stroke subtypes. The findings of this study align profoundly with the chronic systemic immune-inflammatory pathophysiological mechanisms of SLE. The markedly elevated risk ratio highlights the distinct role of SLE-specific pathological processes in stroke occurrence, as opposed to traditional risk factors. Persistent inflammation leads to endothelial dysfunction and accelerated atherosclerosis, while antiphospholipid antibodies (aPL), present in one-third to one-half of patients with SLE, directly enhance thrombin generation, inhibit fibrinolysis and activate platelets, creating a prothrombotic state that predisposes to arterial thromboembolism.41 42 Together, these form the core basis for increased ischaemic stroke risk. Notably, this analysis represents the first meta-analysis to quantify a significant association between SLE and haemorrhagic stroke risk. This finding suggests that, beyond anticoagulant-related bleeding, the elevated risk may be mediated by mechanisms such as vasculitis, hypertension secondary to renal involvement or thrombocytopenia.43–45 Although the relative risk is of similar magnitude to that observed for ischaemic stroke, it is important to recognise that the absolute incidence of haemorrhagic stroke is substantially lower than that of ischaemic stroke in both SLE populations and the general population. In the general population, ischaemic stroke accounts for approximately 85% of all strokes, whereas haemorrhagic stroke accounts for only 15%.46 Therefore, from a clinical standpoint, the population-attributable burden of ischaemic stroke in SLE remains considerably larger, and prevention strategies should prioritise ischaemic events while maintaining appropriate vigilance against bleeding, particularly in anticoagulated patients. Compared with previously published meta-analyses of similar studies, this research achieves significant advancement and innovation in several aspects.13 47 First, it incorporates the latest studies up to 2026, with a total sample size exceeding 5 million, making it one of the largest analyses of its kind, which greatly enhances statistical power and the generalisability of conclusions. Second, unlike previous studies that were limited by smaller sample sizes or an inability to reliably estimate haemorrhagic stroke risk, our analysis includes eight studies specifically addressing intracerebral and subarachnoid haemorrhage. This finding challenges the conventional view that the excess stroke risk in SLE is driven almost exclusively by ischaemia and highlights the need for clinical vigilance against bleeding events. Finally, although we employed a random-effects model and confirmed the robustness of our results, substantial statistical heterogeneity was observed across all endpoints, which is expected given the multifaceted nature of SLE. Rather than avoiding this issue, we addressed it through a preplanned random-effects model, comprehensive sensitivity analyses and trimming methods. This heterogeneity may reflect the inherent variability of SLE disease, differences across healthcare systems, racial disparities and the true diversity in stroke definitions. Our analysis provides a quantitative perspective on these factors. Several limitations must be acknowledged. First, although most included studies adjusted for major cardiovascular risk factors, residual confounding persists, especially regarding SLE-specific factors (disease activity scores, cumulative glucocorticoid dose, specific immunosuppressants, aPL status). Second, the definition of SLE and stroke relied primarily on administrative ICD codes, which may introduce misclassification, particularly for haemorrhagic stroke where coding accuracy is lower. Third, publication bias was suggested by Egger’s test; however, trim-and-fill analysis did not materially alter the pooled effect, and the asymmetry may reflect true heterogeneity rather than missing studies. Fourth, only eight studies reported on haemorrhagic stroke, limiting the precision of that estimate. Finally, we did not perform a subgroup analysis by lupus nephritis (LN) status because only two of the included studies reported LN-specific data; a reliable subgroup analysis would require substantially more studies. These findings have clear and actionable implications for clinical practice. First, stroke risk assessment should be integrated into the routine longitudinal care of all patients with SLE, regardless of age. Second, rheumatologists, neurologists and primary care physicians should collaborate to implement early vascular risk factor control (hypertension, dyslipidaemia, smoking) and consider tailored antithrombotic prophylaxis, particularly in aPL-positive patients. Third, the elevated haemorrhagic stroke risk calls for caution when prescribing anticoagulants or antiplatelet agents; a personalised approach balancing ischaemic and bleeding risks is essential, ideally guided by validated tools such as the adjusted Global Antiphospholipid Syndrome Score (GAPSS) or bleeding risk scores. Finally, our results support the inclusion of SLE as a risk enhancer in cardiovascular disease risk prediction models. Prospective cohort studies with standardised, adjudicated stroke subtyping and detailed collection of SLE activity, treatment and aPL profiles are urgently needed. Future meta-analyses should attempt individual participant data meta-analysis to explore effect modification by age, sex and LN status. Additionally, randomised controlled trials or large registry studies are required to determine the optimal primary and secondary prevention strategies for stroke in SLE, balancing efficacy and bleeding risk. Conclusion In summary, this meta-analysis provides robust evidence demonstrating that SLE is significantly associated with increased risks of stroke, ischaemic stroke and haemorrhagic stroke. These findings definitively position SLE as an independent risk factor for major vascular complications. This study not only consolidates existing knowledge but also expands the field’s understanding through large-scale data and quantification of haemorrhagic risk. We advocate integrating systemic vascular health management into standard SLE clinical pathways and call for further research to optimise risk stratification strategies and targeted preventive measures, ultimately improving long-term quality of life and prognosis for patients with SLE. Supplementary material