Stroke is the major cause of mortality and disability worldwide, imposing a substantial challenge to human health. Despite considerable efforts dedicated to stroke prevention and treatment, its prevalence and disability rates remain high, signaling ongoing gaps in understanding and management strategies for stroke. Serum urate (SUA), the final output of purine metabolism, possesses pro-oxidant and antioxidant dual properties and may exert complex biological effects in our body. Observational studies showed that high SUA levels may indirectly contribute to stroke by inducing inflammatory responses, promoting oxidative stress, and triggering endothelial dysfunction (1–3). Additionally, SUA is also implicated in increasing thrombosis risk, which may impact the occurrence of cardiovascular disease (4). However, some studies have argued that the link between SUA and stroke is still a “pseudo-association” as hyperuricemia is closely associated with other stroke risk factors such as hypertension and obesity; thus, whether or not SUA affects stroke is controversial (5, 6). Furthermore, interventions aimed at reducing SUA did not prevent the onset and progression of cardiovascular disease, which further raises doubts about a direct link between SUA and stroke (7). On the other hand, there are conflicting findings regarding the influence of SUA on stroke prognosis (8–10). Therefore, it is necessary to conduct a more in-depth investigation into the causal association of SUA with stroke risk and prognosis as well as to unravel their precise mechanisms.

Conventional epidemiologic studies are susceptible to unaccounted confounders, excessive adjustment for mediators, or reverse causality, potentially leading to bias in the established causal associations. Mendelian randomization (MR) is an emerging approach for causal inference assessment that cleverly exploits the random distribution of genetic variants as instrumental variables (IVs) at conception to simulate the “random assignment” of participants in randomized controlled trials (RCTs) and aims to identify the differential risk of disease between genetic variant carriers and non-carriers (11). Due to genetic variation being inherited at birth and remaining stable throughout our lifespan, associations obtained from MR are less susceptible to causal inversions and unaccounted confounders compared to traditional observational studies. This study employed a two-sample MR to identify causal associations of SUA with stroke, post-stroke functional outcome, and motor recovery. Additionally, we further performed a two-step MR to explore the potential mediating roles of mediators in the identified causal pathways, which may contribute to enhancing our understanding of the mechanisms linking SUA to stroke risk and prognosis.

To investigate the causal associations and mechanisms of SUA on stroke, post-stroke functional outcome, and motor recovery, we performed this MR study utilizing large-scale, publicly available GWAS summary data. Four critical findings emerged from our analysis: (1) Genetically predicted elevated SUA levels were significantly causally associated with an increased risk of AS and IS, and these associations remained significant even after adjusting for BMI and eGFR. (2) SBP and DBP were identified as critical mediators in the causal pathways linking SUA to AS and IS. (3) Elevated SUA levels were found to lower poor functional outcomes after 90d of ischemic stroke and also contributed to motor recovery within 24 months after stroke. (4) There was no evidence of a significant causal association between SUA and ICH. Additionally, sensitivity analyses suggested consistent estimates and effect directions for almost all outcomes, and no horizontal pleiotropy was found.

Previous observational studies have explored the link between SUA and stroke but yielded conflicting conclusions (6, 20, 21). Epidemiologic studies often struggle to fully exclude confounders and address reverse causality, which may be key drivers of inconsistent conclusions. MR using genetic variants as IVs to proxy for different phenotypes contributes to maximizing control for confounders and avoids reverse causality interference, thus providing more reliable causal inferences. A recent MR study suggested that higher SUA levels would increase AS risk, with each 1-SD (80.3 μmol/L) increase in genetically predicted SUA levels being associated with an 11% increase in the risk of developing AS (22). Our MR analyses using GWAS data from different sources yielded similar results. The subsequent meta-analysis of the IVW estimates derived from these analyses obtained a more significant OR estimate and a narrower 95% confidence interval. IS is the primary subtype of AS, and evidence from two previous MR studies involving SUA and IS did not support a causal association among them (23, 24). These were contradictory to our findings. Given that our MR extracted more eligible IVs (310–314 vs. 28 SNPs) and employed GWAS summary data only from European ancestry, this enhances the reliability of our findings. Additionally, the study further indicated that there was no causal association between SUA and ICH. Therefore, a previous observational study reported that a positive relationship of stable high SUA with ICH risk but not with IS may have been coincidental (21).

Currently, there is a lack of consensus on the exact role of SUA in stroke prognosis (25, 26). In a recent MR study, SUA was not found to be causally associated with poor functional prognosis (mRS > 2) at 90d after ischemic stroke (27). However, our MR analysis, using GWAS data defining mRS > 1 as poor functional outcome, identified a significant and negative causal association between SUA and poor functional prognosis in ischemic stroke patients at 90d. Prior meta-analysis suggested a non-linear (U-shaped) association between SUA and the risk of poor functional outcomes after ischemic stroke (28). This complex relationship between SUA and post-stroke functional prognosis may be the key factor contributing to this contradictory result. Additionally, our MR analyses found a positive causal association between SUA and motor recovery within 24 months after stroke. This is consistent with previous findings that SUA exerted antioxidant properties and potential neuroprotective effects (29). However, further studies are still required to fully grasp the complexity and potential non-linear relationship between SUA and post-stroke prognosis.

Through this study, we have found that SUA has dual effects, with elevated SUA being a risk factor for AS and IS, while also exerting a protective effect on post-stroke functional outcome and motor recovery. However, the exact mechanism between them remains ambiguous. Numerous studies suggest that hyperuricemia plays an important role in the etiology of hypertension, which is the leading risk factor for stroke (30, 31). Our further mediation analyses identified SBP and DBP as key mediators lying on the SUA to AS and IS are causal pathways, which is consistent with the result previously reported by Chaudhary NS et al. (31). Additionally, blood pressure is the primary driver of maintaining cerebral perfusion. Sustaining appropriate blood pressure in the early stages of ischemic stroke is beneficial for cerebral perfusion and aids in neurological function recovery. These factors, in turn, positively influence stroke prognosis (32). Of note, although we failed to detect the mediating effect of blood pressure in the pathway of SUA on stroke prognosis, the influence of SUA on post-stroke functional outcome and motor recovery became non-significant after adjusting for SBP and DBP using MVMR. This evidence provided insights into the role of SUA in the pathogenesis and rehabilitation of stroke and may benefit future research and clinical practice.

Overweight is also a well-known risk factor for stroke. Our analyses revealed that BMI did not act as a mediator or attenuate the impact of SUA on AS and IS, suggesting that SUA increases the stroke risk and may be independent of BMI. It is important to note that being overweight is not always detrimental. Some studies have suggested that being overweight indicated the presence of more nutritional reserves, which could potentially help counteract post-stroke hypermetabolic depletion, thus translating into a protective factor against poor functional outcomes (33, 34). This might be a potential explanation for our finding that the causal effect of SUA on post-stroke functional outcome and motor recovery became non-significant after adjusting for BMI. Additionally, impaired kidney function is an independent risk factor for stroke and is linked to more severe stroke and worse outcomes (35). In our analysis, we also found that eGFR attenuated the causal effect and significance of SUA on IS. Although further mediation analysis failed to yield a significant mediating effect, it still implied that the improvement of post-stroke functional outcome and the promotion of motor recovery by SUA may rely on favorable renal function.

Our study holds several strengths as follows. First, the study employed MR to investigate the causal association between SUA and stroke risk and prognosis, which minimized bias from confounders and reverse causality, thus providing credible causal inferences. Second, our efforts identified the dual effects of SUA on stroke risk and prognosis, and a series of replication analyses, GWAS meta-analyses, and sensitivity analyses observed similar results, ensuring the robustness of the findings. Finally, potential mechanisms of stroke risk factors in causal pathways were explored by MVMR and mediation analyses, and this insight may guide the development of targeted intervention and prevention strategies.

The study also has some limitations. First, the potential pleiotropy is an inherent limitation of MR analysis. Despite performing strict selection criteria for genetic IVs and examining for outliers and horizontal pleiotropy, pleiotropy may still be present and could introduce bias into our results. Second, it is necessary to point out the genetic variation differences that exist among different races, which may introduce heterogeneity into the causal estimates. Although the incidence of stroke is also high in the Asian region (36–38), considering that our study mainly involves participants of European descent, this may limit the generalization of our findings to other races and populations. Third, in this study, we noted a 43.6% sample overlap in SUA and eGFR GWAS datasets, potentially impacting the accuracy of MVMR and mediation analysis. Due to the absence of independent GWAS datasets excluding UK Biobank, we evaluated the robustness of our findings by estimating potential bias and type I error rate. The finding indicates that a 43.6% overlap might cause a bias ranging from 0.007 to 0.008 and increase type I error rates between 0.19 and 0.27. Although the bias suggests a somewhat reliable result, the elevated type I error rates, substantially above the 5% standard threshold, indicate a heightened risk of false positives. Future studies with non-overlapping samples are required to corroborate and reinforce our conclusions. Finally, limited by the GWAS data, we were unable to stratify SUA levels or explore trend relationships between SUA and stroke risk and prognosis. In future studies, a more comprehensive collection of SUA data and more in-depth analysis will contribute to further insight into the potentially complex associations between them.

In conclusion, our MR study supports a dual role for genetically predicted SUA in increasing the risk of stroke, especially ischemic stroke, and in improving post-stroke poor functional outcome and motor recovery. Moreover, we provide credible genetic evidence that SBP and DBP mediate considerable proportions of the causal effects of SUA on AS and IS. These findings contribute to a deeper understanding of the underlying mechanisms that link SUA to stroke risk and prognosis.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

SC: Conceptualization, Funding acquisition, Methodology, Writing – original draft. ZC: Data curation, Visualization, Writing – original draft. QX: Data curation, Writing – original draft. XJ: Formal analysis, Validation, Writing – original draft. CL: Formal analysis, Writing – original draft. JJ: Conceptualization, Funding acquisition, Methodology, Writing – original draft.