Serum Uric Acid and Outcome after Acute Ischemic Stroke: PREMIER StudyChiquete E.a · Ruiz-Sandoval J.L.b · Murillo-Bonilla L.M.c · Arauz A.d · Orozco-Valera D.R.b · Ochoa-Guzmán A.b · Villarreal-Careaga J.e · León-Jiménez C.f · Barinagarrementeria F.g · Ramos-Moreno A.h · Cantú-Brito C.a
aDepartment of Neurology and Psychiatry, Instituto Nacional de Ciencias Médicas y Nutrición ‘Salvador Zubirán', Mexico City, bDepartment of Neurology, Hospital Civil de Guadalajara ‘Fray Antonio Alcalde', Guadalajara, cDepartment of Neurology, Universidad Autónoma de Guadalajara, Zapopan, dStroke Clinic, Instituto Nacional de Neurología y Neurocirugía, Mexico City, eDepartment of Neurology, Hospital General de Culiacán, Culiacán, fDepartment of Neurology, Hospital Regional ‘Dr. Valentín Gómez Farías', ISSSTE, Zapopan, gDepartment of Neurology, Hospital Ángeles de Querétaro, Querétaro, hMedical Research Area, Sanofi-Aventis, Mexico City, Mexico Corresponding Author
Background: Current evidence shows that uric acid is a potent antioxidant whose serum concentration increases rapidly after acute ischemic stroke (AIS). Nevertheless, the re-lationship between serum uric acid (SUA) levels and AIS outcome remains debatable. We aimed to describe the prognostic significance of SUA in AIS. Methods: We studied 463 patients (52% men, mean age 68 years, 13% with glomerular filtration rate <60 ml/min at hospital arrival) with AIS pertaining to the multicenter registry PREMIER, who had SUA measurements at hospital presentation. Multivariate models were constructed to analyze the association of SUA with functional outcome as assessed by the modified Rankin scale (mRS) at 30-day, 3-, 6- and 12-month follow-up. A mRS 0-1 was regarded as a very good outcome. Results: Mean SUA concentration at hospital arrival was 6.1 ± 3.7 mg/dl (362.8 ± 220.0 μmol/l). Compared with cases with higher SUA levels at hospital admission, patients with ≤4.5 mg/dl (≤267.7 μmol/l; the lowest tertile of the sample) had more cases of a very good 30-day outcome (30.5 vs. 18.9%, respectively; p = 0.004). SUA was not associated with mortality or functional dependence (mRS >2) at 30 days, or with any outcome measure at 3, 6 or 12 months poststroke. After adjustment for age, gender, stroke type and severity (NIHSS <9), time since event onset, serum creatinine, hypertension, diabetes and smoking, a SUA ≤4.5 mg/dl (≤267.7 μmol/l) was positively associated with a very good short-term outcome (odds ratio: 1.76, 95% confidence interval: 1.05-2.95; negative predictive value: 81.1%), but not at 3, 6 or 12 months of follow-up. When NIHSS was entered in the multivariate model as a continuous variable, the independent association of SUA with outcome was lost. Compared with cases with higher levels, patients with SUA ≤4.5 mg/dl (≤267.7 μmol/l) were more frequently younger than 55 years, women, with mild strokes, with normal serum creatinine and fewer had hypertension. The time since event onset to hospital arrival was not significantly associated with AIS severity or SUA levels; nevertheless, a nonsignificant tendency was observed for patients with severe strokes and high SUA levels arriving in <24 h. Conclusions: A low SUA concentration is modestly associated with a very good short-term outcome. Our findings support the hypothesis that SUA is more a marker of the magnitude of the cerebral infarction than an independent predictor of stroke outcome.
Copyright © 2013 S. Karger AG, Basel
Raised serum uric acid (SUA) is a modest risk factor for stroke and cardiovascular disease, especially among patients with hypertension or diabetes mellitus [1,2,3,4,5]. Uric acid is a potent endogenous antioxidant whose serum concentrations increases within the first hours after acute ischemic stroke (AIS), showing a decrease to basal levels in the following days after the event . Basic and clinical evidence points to a function of uric acid as an antioxidant that acts chronically during inflammation and long-standing ischemia [6,7,8], as well as acutely in tissue infarction [9,10,11,12,13,14]. However, the relationship between SUA and stroke outcome remains controversial [15,16,17,18,19], and it is not clear whether this association is either causal or circumstantial. As a consequence, some investigators suggest that anti-hyperuricemic therapy may be beneficial [16,20,21,22], while others currently investigate the pharmacological administration of uric acid as a free-radical scavenger in AIS treatment, particularly as an adjuvant to thrombolysis [15,23,24,25]. We sought to examine the role of SUA at hospital admission as a prognostic marker of functional outcome in patients with AIS. Our hypothesis was that SUA levels are independently associated with stroke severity and outcome. This study supports the notion that SUA elevation is proportional to the magnitude of the cerebral infarction, but at the other extreme of the phenomenon, that low SUA levels are associated with better outcomes.
We analyzed data of patients with AIS, for whom SUA was measured at hospital admittance, from different geographic regions of Mexico, included in the PREMIER (Primer Registro Mexicano de Isquemia Cerebral) study [26,27,28]. Briefly, PREMIER was a prospective, hospital-based multicenter registry of consecutive patients with AIS or transient ischemic attack (TIA) in Mexico, performed in 59 urban hospitals by 77 physicians [26,27]. A central Institutional Review Board and the local Committee of Ethics of each participating center approved the protocol. Signed informed consent was required for all patients or their legal proxies.
Consecutive patients with AIS or TIA aged ≥18 years were registered. All patients received medical care within 7 days of stroke onset. Data collection was prospectively performed during one year (in 6 different medical visits) by using a standardized structured questionnaire (clinical research format, CRF) outlined in a manual of definitions and procedures. All participating physicians were formally instructed on current stroke guidelines, stroke classification, evaluation, treatment and prevention. Stroke subtypes were registered according to the Trial of ORG 10172 in Acute Stroke (TOAST) classification. AIS severity was assessed by the National Institutes of Health Stroke Scale (NIHSS) at baseline, and the modified Rankin scale (mRS) was used to evaluate the functional outcome at hospital discharge, and at 1, 3, 6, 12 and 18 months after AIS. Demographic, clinical, laboratory, anthropometric and neuroimaging variables were systematically registered in a standardized format. Other variables registered were about management of comorbidities, preventive measures, hospitalizations, vascular events and stroke recurrences during follow-up and functional outcome at each visit. Information of the CRF was saved on independent electronic files in data capture software developed and revised periodically by a contract research organization (CRO; INNOVAL Co.). Members of the CRO analyzed information on every patient for completeness and plausibility. Missing or implausible variables were referred to the investigators for clarification. Data quality was ascertained by periodic statistical reports and onsite visits by CRO monitors [26,27].
Among 1,376 participants with AIS or TIA registered in the PREMIER study , 463 AIS patients with complete clinical and laboratory evaluations, as well as complete 12-month follow-up information were selected for this analysis. TIA cases were not included here. A mRS = 0-1 was regarded as a very good outcome. Clinical raters of the mRS were not blinded to SUA levels; nonetheless, they were not aware of the objective of the present report, since the purpose of this analysis arose when the registry was terminated.
Parametric continuous variables are expressed as geometric means and standard deviations (SD), or minimum and maximum. Nonparametric continuous variables are expressed as medians. Categorical variables are expressed as percentages. To compare quantitative variables distributed between two groups, Student's t test and Mann-Whitney U test were performed in distributions of parametric and nonparametric variables, respectively. Chi-square statistics (i.e. Pearson's χ2 or Fisher's exact test, as corresponded) were used to compare nominal variables in bivariate analyses, among two or more nominal categories. SUA was analyzed as tertiles, quartiles and quintiles to find a rational cutoff to dichotomize SUA levels in prediction analyses, to facilitate its clinical application. Sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), positive likelihood ratio (LR+) and negative likelihood ratio (LR-) was calculated for the selected SUA cutoff that had any statistically significant association with an outcome measure. Multivariate analyses were used to assess the independent relationship of SUA levels and outcome (mRS = 0, mRS = 0-1, mRS ≥2, mRS ≥3 and mRS = 6) at 30- and 90-day, 6- and 12-month follow-up. Known factors that may influence SUA (i.e. gender, renal function, time since stroke onset, stroke severity, type of stroke) were taken into account. Multivariate analyses were constructed by forward stepwise binary logistic regression. Input variables were those that resulted significantly associated with AIS outcome in bivariate analyses. Adjusted odds ratios (OR) with 95% confidence intervals (CI) are provided. The fitness of the models was evaluated by using the Hosmer-Lemeshow goodness-of-fit test, which was considered as reliable if p > 0.2. All p values are two-sided and considered significant when p < 0.05. SPSS v 17.0 software was used for all statistical calculations.
A total of 463 patients were analyzed (52% men, mean age 68 years; range 21-104 years) (table 1); 61 (13%) patients had glomerular filtration rate <60 ml/min at hospital arrival. Mean SUA at hospital admittance was 6.1 ± 3.7 mg/dl (362.8 ± 220.0 μmol/l), higher in men than in women [6.6 ± 3.9 vs. 5.5 ± 3.5 mg/dl (392.6 ± 232.0 vs. 327.1 ± 208.2 μmol/l); p = 0.002] (fig. 1).
|Table 1. Main characteristics of the patients according to 30-day outcome|
|Fig. 1. SUA tertiles as a function of gender.|
At 30 days after AIS, a very good outcome occurred more frequently in patients with SUA ≤4.5 mg/dl (≤267.7 μmol/l; the lowest tertile of the sample) than in those with higher levels (30.5 vs. 18.9%, respectively; p = 0.004; univariate Mantel-Haenszel OR: 1.88, 95% CI: 1.21-2.92). Neither low nor high SUA levels were associated with stroke mortality or functional dependence (mRS >2, or mRS >3) at 30-day follow-up. Moreover, we could not find any significant association between SUA and any outcome at 3-, 6-, or 12-month follow-up, either in univariate or in multivariate analyses.
In a multivariate analysis adjusted for relevant cofactors (table 2), SUA ≤4.5 mg/dl (≤267.7 μmol/l) was associated with 30-day very good outcome (mRS = 0-1). However, the significance of this association was lost when NIHSS was entered as a continuous variable. SUA ≤4.5 mg/dl (≤267.7 μmol/l) had a sensitivity, specificity, PPV, NPV, LR+ and LR- for predicting a very good 30-day outcome, of 47.7% (95% CI: 38.4-57.0), 67.4% (95% CI: 62.4-72.1), 30.5% (95% CI: 24.1-37.9), 81.1% (95% CI: 76.2-85.1), 1.46 (95% CI: 1.14-1.87) and 0.78 (95% CI: 0.64-0.94), respectively. The time since AIS onset to hospital arrival was not significantly associated with AIS severity (scoring of the NIHSS), or SUA levels (fig. 2); nevertheless, a nonsignificant tendency was observed for patients with severe strokes and high SUA levels arriving earlier (fig. 2). Low SUA levels did not occur more frequently among patients with NIHSS <9 points (fig. 3), but milder strokes (i.e. NIHSS <5 points) occurred more commonly among cases with SUA ≤4.5 mg/dl (≤267.7 μmol/l) (fig. 3). Furthermore, variables significantly associated with low SUA levels were: female gender, young age, normal serum creatinine and (inversely) hypertension (table 3).
|Table 2. Factors positively associated with a very good outcome (modified Rankin scale 0-1) at 30-day follow-up: a multivariate logistic regression model*|
|Table 3. Factors positively associated with a SUA ≤4.5 mg/dl (≤268 μmol/l) at hospital admittance|
|Fig. 2. Hospital arrival according to stroke severity by the NIHSS (a), and SUA tertiles (b).|
|Fig. 3. Stroke severity by the NIHSS as a function of SUA tertiles, dividing the NIHSS distribution with wider (a) and closer (b) score ranges.|
In this study we found that a low SUA at hospital admission is modestly associated with a very good short-term outcome. However, we could not demonstrate the opposite, that severe strokes or adverse outcomes are associated with higher concentrations of SUA. Our results may contrast with some previous findings [9,17,20,21,29,30], but are in line with the report of Nardi and Milia  showing that a low SUA is associated with an excellent functional state after AIS. Moreover, the factors that we found as associated with SUA concentrations have also been described in similar settings . Nonetheless, although from a different perspective, our findings support the concept that SUA elevation is proportional to the magnitude of the brain ischemia [12,13,18], with low SUA levels indicating a good outcome possibly through mild strokes. The present results offer relevant information on the factors that may confound and partially explain the variation of SUA among AIS cohorts.
The time of blood sampling appears to be a crucial issue that may explain much of the variation of results among publications [9,12,13]. If SUA is a consumptive and rapidly changing marker of the antioxidant response elicit by the brain ischemia, then higher SUA levels should be observed during certain moments in larger strokes to offer a metabolic response for scavenging oxygen free radicals excessively produced during blood deprivation. Patients failing to mount such a reaction may have a bad outcome. In the end, large strokes may originate in part the SUA response, but may also be the consequence of a failing antioxidant reaction. This subject needs more exploration in basic and clinical studies. The considerable debate on the significance of SUA as a marker of AIS outcome may be due to different interpretations of the same phenomenon .
Cells and tissue preparations exposed to hypoxia/ischemia exhibit an increased expression of xanthine oxidase, the rate-limiting enzyme in the conversion of hypoxanthine to xanthine and xanthine to uric acid . Xanthine oxidase is the only enzyme capable of catalyzing the formation of uric acid in humans. In other mammals, urate oxidase can metabolize uric acid to allantoin, a potent antioxidant, but this enzyme activity is lost in primates. It is possible that uric acid production may have evolved as a compensatory mechanism in primates that cannot produce other potent organic antioxidants . As such, uric acid production by means of xanthine oxidase activity is possibly the most potent acute antioxidant mechanism in response to ischemia in humans, and hence, it represents a marker of tissue infarction [22,32,33]. However, a limitation of the models that examine the SUA systemic reaction to focal ischemia (i.e. AIS) is that uric acid production in situ may not be adequately measured with peripheral blood sampling.
Here we could not demonstrate a particular pattern of SUA concentration as a function of time from stroke onset to hospital arrival, possibly because patients with mild strokes tended to arrive later than patients with profound neurologic deficits. However, although SUA levels were measured at hospital admittance, we did not standardize precisely the moment of blood sampling or laboratory analyses, and no serial SUA measurements were undertaken; therefore, our data set is not adequate to evaluate the SUA time response as a function of stroke severity. This is a very interesting topic to be examined in future studies.
The main limitation of this study is the sample size that might not be large enough to detect a more robust difference between SUA levels across the whole range of stroke severity, and to find that patients with severe strokes and with higher SUA levels arrive earlier than their counterparts. No volumetric analyses were performed to demonstrate that SUA is associated with the size of the cerebral necrosis, and no serial measurements were performed to demonstrate that SUA levels change over time, as a function of stroke volume, clinical severity, renal function, age and gender. Also, the previous use of thiazides and other uric acid-modifying drugs was not registered in our study, although indeed, none of the patients was treated with thiazides during the poststroke hospitalization.
In conclusion, these findings support the hypothesis that SUA is more a marker of the magnitude of the cerebral infarction than a strong independent predictor of AIS outcome. In other words, SUA is an acute indicator of stroke severity if this response is proportional to the infarction size. This idea should be consistently confirmed in further studies.
The PREMIER study received unrestricted financial support from conception to execution by Sanofi, Mexico. The pharmaceutical company was involved in the design of the study, but had no role in the selection of patients, data analysis, the preparation of this article or the decision for submission to publication.
Department of Neurology and Psychiatry, INCMNSZ
Vasco de Quiroga 15
Tlalpan, Mexico City 14000 (Mexico)
Received: July 10, 2012
Accepted: December 18, 2012
Published online: February 22, 2013
Number of Print Pages : 7
Number of Figures : 3, Number of Tables : 3, Number of References : 33
Vol. 35, No. 2, Year 2013 (Cover Date: March 2013)
Journal Editor: Hennerici M.G. (Mannheim)
ISSN: 1015-9770 (Print), eISSN: 1421-9786 (Online)
For additional information: http://www.karger.com/CED