Uric acid is the final oxidation product of purine metabolism and a water-soluble antioxidant and radical scavenger in humans (1, 2). Despite the antioxidant properties of uric acid, epidemiologic studies have shown that hyperuricemia is related to an increased risk of cardiovascular events, stroke, and their risk factors (3–6). On the other hand, growing evidence has been emerging that elevated uric acid concentration is associated with better functional outcomes of ischemic stroke (IS) (7–11). Moreover, the exogenous administration of uric acid exerts robust neuroprotective effect on the outcome of IS patients in clinical trials (12, 13).

Obesity, a metabolic disorder, has a wide range of negative effects on health consequences (14). There is consistent evidence that obesity is an established independent risk factor for incidence of IS (15). However, many studies, but not all, suggest a better prognosis in overweight and obese patients after IS (16–19). The improved survival and functional outcomes, and stroke recurrence in overweight and obese IS patients are termed as “obesity-stroke paradox” in the literature (18, 19). Although the obesity-stroke paradox related factors and metabolic pathways have been documented in previous studies, the underpinning mechanisms are still not fully understood (20, 21).

Oxidative stress is a major contributor to brain damage in patients with IS (22). Uric acid, a powerful antioxidant, is highly and positively correlated with body weight (6). To date, the joint effect of uric acid and obesity on the IS outcomes has not been reported in previous studies. The working hypothesis of this study is that uric acid concentration contributes to the obesity-stroke outcome paradox in patients with IS. The current study aims to investigate the synergistic effect of uric acid and body mass index (BMI) on clinical functional outcomes after IS in a 1-year follow-up study cohort.

Table 1 summarizes characteristics of study variables by sex groups and in the total sample. Males were younger and had greater prevalence of smokers and alcohol drinkers, lower prevalence of anti-hypertensive treatment and lower glucose, NIHSS than females. Males had lower BMI but higher UA and eGFR than females. The difference in incidence rates of poor outcome (mRS ≥ 3) between sex groups was significant, with males having a lower rate than females.

Table 2 shows the relationship of uric acid, BMI, age, sex, and diabetes. Female sex was significantly associated with lower uric acid, but higher BMI. Older age was significantly associated with higher uric acid and lower BMI. IS patient with diabetes had lower uric acid, but higher BMI. Of note, BMI was significantly and positively correlated with uric acid.

Figure 1 presents incidence rates of poor outcome (mRS ≥ 3) in 1-year follow-up by uric acid and BMI groups. Low and high BMI was defined by the threshold of 24, and low and high uric acid by the median. As shown in Panel A, the lower uric acid group showed a higher incidence rate of poor outcome than the higher uric acid group (31.6 vs. 25.0%, p = 0.001); the lower BMI group showed a higher incidence rate of poor outcome than the higher BMI group (31.4 vs. 26.3%, p = 0.013). Panel B showed incidence rates of poor outcome by groups of combined BMI/uric acid. Incidence of the poor outcome was 34.5, 29.4, 27.7, and 23.5% in the BMI/uric acid groups of low/low, high/low, low/high and high/high, respectively, with p = 0.001 for trend.

Table 3 shows the association of BMI and uric acid with outcome, adjusting for covariates. Older age was significantly associated with higher risk of worse outcome. Patients with higher NIHSS were 1.26 times more likely to have worse outcome. Higher blood glucose was significantly associated with higher risk of poor outcome. Higher levels of BMI and uric acid had significantly protective effects on the outcome. The interactions of uric acid were significant with sex (p = 0.001) and BMI (p = 0.045), but not significant with age (p = 0.313), systolic blood pressure (p = 0.487) and diabetes (p = 0.974).

Figure 2 presents odds ratio (OR) and 95% confidence interval (CI) of lower uric acid for poor outcome (mRS ≥ 3) by BMI quartile (panel A) and sex (panel B) groups, with quartile 4 of uric acid (the highest) as reference. The association between lower uric acid and poor outcome was significant in lower BMI groups (OR = 1.36, p = 0.006 in quartile 1; OR = 1.28, p = 0.021 in quartile 2), but the ORs were not significant in BMI quartile 3 and 4 groups. The interaction between uric acid and BMI was significant, with p = 0.038 for the decreasing trend in ORs across increasing BMI quartiles. The association between low uric acid and poor outcome was significant in males (OR = 1.31, p < 0.001), but not significant in females (OR = 0.96, p = 0.618), with p = 0.006 for sex difference.

Figure 3 presents OR and 95% CI of lower uric acid for poor outcome by sex/BMI groups, with quartile 4 of uric acid (the highest) as reference. The adverse effect of lower uric acid measured as OR was significant in males, but not in females, with p = 0.004 for the trend. Of note, the difference in ORs between low and high BMI groups was not significant in males (p = 0.512 for interaction) and females (p = 0.335 for interaction).

Obesity is an escalating pandemic worldwide, representing an emerging threat to public health (26, 27). The obesity association with cardiovascular and cerebrovascular diseases is well-demonstrated. Studies have concluded that obesity is strongly associated with increased stroke risk in both men and women (28, 29). Although the adverse health consequences of obesity in the general population have been well-documented, strong evidence suggests that obesity is associated with improved survival rate and functional outcomes and lower stroke recurrence in patients with IS, indicating the existence of an “obesity-stroke paradox” (16–19). However, some studies supported that no obesity paradox exists in patients with IS after adjustment for initial neurological severity (30–32). The existence of obesity paradox is still controversial. In the current study, the higher BMI (≥24) group had a significantly lower incidence of poor functional outcome (mRS ≥ 3) than the lower BMI (<24) group (26.3 vs. 31.4%, p = 0.013) in a 1-year follow-up of IS patients; BMI was significantly and negatively associated with the risk of poor functional outcome in multivariable logistic regression models, adjusting for covariates. Despite strong observational data indicating a survival benefit of obese patients after stroke, methodological concerns exist (16, 33), and the biologic mechanisms contributing to pathways in metabolic imbalance underlying the obesity-stroke paradox are largely unknown (20, 21).

Data from most observational studies, systematic reviews, and meta-analyses support the concept that elevated uric acid concentration is associated with better functional outcomes of IS (7–11). Patients with acute IS have a 12% increase in the odds of good clinical outcome for each milligram per deciliter increase of serum uric acid (34). Moreover, a meta-analysis of 10 studies found that high serum uric acid level was associated better outcome after acute IS as compared to low serum uric acid level (9). Furthermore, one prospective study suggested that the increased uric acid levels are associated with better outcome in IS patients treated with reperfusion therapy (35). Similarly, another study demonstrated that early elevation of uric acid during or shortly after IS onset presented significant protection against neurological deficit in acute IS patients treated with recombinant tissue plasminogen activator (rt-PA) (36). The ACROSS-China study reported that lower serum uric acid levels strongly predicted short-term poor functional outcome in acute stroke with normoglycaemia (11). In addition, evidence for the beneficial effect of exogenous administration of uric acid on the outcome of IS patients in clinical trials has been emerging in recent years (12, 13). We found in this study that lower uric acid levels were significantly associated with poor functional outcome of IS patients, and this effect was exacerbated by lower BMI. The joint effect of uric acid and obesity status on the IS outcome has not been reported in previous studies. The detrimental impact of lower uric acid observed in the present study may be one of metabolic mechanisms of the obesity-stroke outcome paradox.

Obesity, a metabolic disorder, is well-known to correlate with uric acid. The Bogalusa Herat Study has shown that BMI levels were significantly, positively associated with uric acid in both children and adults (6). It is also noted in current study cohort that lower BMI was significantly correlated with decreased levels of uric acid in IS patients. Incidence of poor outcome (mRS ≥ 3) was the highest in the low BMI-low uric acid group and was the lowest in the high BMI-high uric acid group, with p for trend = 0.001 (Figure 1B). Uric acid concentrations decrease significantly over time in stroke patients, and the plasma antioxidant capacity has been inversely correlated with the volume of cerebral infarction and the severity of neurological impairment (37). IS patients with lower BMI tended to have lower uric acid which was associated with decreased antioxidant capacity. The highest OR of low uric acid for poor outcome in IS in the lowest BMI quartile group (Figure 2A) indicated a synergistic effect of lower BMI and lower uric acid on the poor outcome. The findings on the interaction effect between BMI and uric acid on the IS outcome suggested that lower BMI levels might be associated with poor outcome in IS, at least in part, through lower uric acid concentrations. Studies with specifically sophisticated design in other populations are needed to confirm our findings. Further investigations should be performed in acute IS patient treated with rt-PA to further explore whether same results can be observed.

Another interesting finding in this study is the sex-specific association between uric acid and the outcome in IS. Despite the significant difference in the incidence of poor outcome (males < females), the association between lower uric acid and the poor outcome was significant in male patients, but not in female patients. The male-low BMI group showed the highest OR (1.43), and the female-high BMI group showed the lowest OR (0.89), with p for trend = 0.004 (Figure 3B). In a prospective study of the China Antihypertensive Trial in Acute Ischemic Stroke, the association between serum uric acid and primary outcomes of IS measured as death and major disability (mRS ≥ 3) at 3 months was modified by sex (P for interaction n = 0.007) in 3284 acute IS patients. Elevated serum uric acid was significantly and positively associated with the primary outcome in men, but not in women (38). In contrast, the administration of uric acid reduced infarct growth in patients with acute IS treated with alteplase in women, but not in men in a clinical trial in Spain (39). As we know, uric acid level is different between sex, and men usually have a higher uric acid level than women. Women with higher uric acid levels were significantly associated with the development of hypertension and metabolic syndrome, however, it was not found in men. We hypothesized that exogenous administration of uric acid may have a greater benefit for male patients with low uric acid level in acute ischemic stroke with or without reperfusion therapies. To date, data on the sex-specific association between uric acid and outcomes in IS are still limited. Prospective double blind randomized controlled trials in this regard are required to investigate the effect of uric acid on the outcomes in male and female IS patients.

There are several limitations of our study include: First, the result of this study could not generalized to western population, because there is much difference between western population and Chinese population in the distribution of BMI. Second, our study has significant missing data about serum uric acid and the included population has higher proportion of diabetes, lipid-lowering medication, and higher BMI level compared with excluded population. Thus, popularizing the results to general population is limited. Further investigations in other larger populations are needed to confirm the findings. Third, the telephone follow-up for outcome assessment is sub-optimal. Fourth, this study is a hospital-based cohort, which may lead to selection bias. A large population-based study is needed to confirm our findings. Fifth, the cohort was recruited between 2008 and 2009, therefore, it was considered not contemporary. Sixth, the lack of information on use of acute reperfusion therapies in this study may hampered further investigation on the relationship between uric acid and obesity-stroke paradox in patients treated with reperfusion therapies.

In summary, this observational prospective study highlighted that across BMI categories, uric acid was differentially associated with functional outcome after stroke. Of note, a synergistic effect of uric acid and BMI was found that the association between low levels of uric acid and the poor clinical functional outcome was strengthened by lower BMI in IS patients. The detrimental effect of low uric acid on the outcome was observed to be significant in male, but not in female IS patients. The findings of the present study have implications in understanding the metabolic mechanisms of the obesity-stroke outcome paradox, help identify high-risk patients of IS and provide additional evidence for the administration of uric acid in clinical trial studies. Given the practical implications that could be derived in the field of prevention of poor outcome in IS, larger studies would be required to confirm these encouraging results.

The datasets generated for this study are available on request to the corresponding author.

YW had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design were done by HT, JM, and ZC. Acquisition, analysis, or interpretation of data were done by LD, AC, and JX. Drafting of the manuscript was done by HT, JM, and ZC. Statistical analysis was done by AW. Study supervision was the responsibility of YW.