Hyperuricemia and risk of stroke: A systematic review and meta-analysis

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/26329540

Hyperuricemia and Risk of Stroke: A Systematic Review and Meta-Analysis ARTICLE in ARTHRITIS & RHEUMATOLOGY · JULY 2009 Impact Factor: 7.76 · DOI: 10.1002/art.24612 · Source: PubMed

CITATIONS

READS

161

69

6 AUTHORS, INCLUDING: James P Guevara

Daniel F Heitjan

The Children's Hospital of Philadelphia

Southern Methodist University

63 PUBLICATIONS 1,792 CITATIONS

212 PUBLICATIONS 9,657 CITATIONS

SEE PROFILE

SEE PROFILE

Available from: Albert Daniel Retrieved on: 05 February 2016

NIH Public Access Author Manuscript Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

NIH-PA Author Manuscript

Published in final edited form as: Arthritis Rheum. 2009 July 15; 61(7): 885–892. doi:10.1002/art.24612.

Hyperuricemia and Risk of Stroke: A Systematic Review and Meta-analysis Seo Young Kim, MD1,2, James P Guevara, MD, MPH2,3, Kyoung Mi Kim, MD4, Hyon K Choi, MD, DrPH5, Daniel F. Heitjan, PhD2,6, and Daniel A Albert, MD7 1Division of Rheumatology, Department of Medicine, University of Pennsylvania 2Center

for Clinical Epidemiology and Biostatistics, University of Pennsylvania

3Division 4Pusan

of General Pediatrics, Department of Pediatrics, Children’s Hospital of Philadelphia

National University, Pusan, South Korea

5Division

NIH-PA Author Manuscript

of Rheumatology, Department of Medicine, University of British Columbia, Vancouver, BC, Canada 6Department 7Division

of Biostatistics and Epidemiology, University of Pennsylvania

of Rheumatology, Dartmouth Hitchcock Medical Center, Lebanon, NH

Abstract BACKGROUND—Hyperuricemia is hypothesized to be a risk factor for stroke and other cardiovascular disease, but to date results from observational studies are conflicting. METHODS—We conducted a systematic review and meta-analysis to assess the association between hyperuricemia and risk of stroke incidence and mortality. Studies were identified by searching major electronic databases using the Medical Subject Headings and keywords without restriction in languages. Only prospective cohort studies were included if they had data on stroke incidences or mortalities related to serum uric acid levels in adults. Pooled risk ratios (RRs) for the association of stroke incidence and mortality with serum uric acid levels were calculated.

NIH-PA Author Manuscript

RESULTS—A total of 16 studies including 238,449 adults were eligible and abstracted. Hyperuricemia was associated with a significantly higher risk of both stroke incidence [N=6 studies, RR 1.41, 95% confidence interval (CI): 1.05–1.76] and mortality [N=6 studies, RR 1.36, 95% CI: 1.03–1.69] in our meta-analyses of unadjusted study estimates. Subgroup analyses of studies adjusting for known risk factors such as age, hypertension, diabetes, and cholesterol still showed that hyperuricemia was significantly associated with both stroke incidence [N=4 studies, RR 1.47, 95% CI: 1.19–1.76] and mortality [N=6 studies, RR 1.26, 95% CI: 1.12–1.39]. The pooled estimate of multivariate RRs did not differ much by gender. CONCLUSION—Our study suggests that hyperuricemia may modestly increase the risks of both stroke incidence and mortality. Future research is needed to determine whether lowering uric acid level has any beneficial effects on stroke.

Corresponding author and Reprint requests: Seo Young Kim, M.D., Division of Rheumatology, University of Pennsylvania, 504 Maloney, 3600 Spruce Street, Philadelphia PA 19104, Tel) 215-662-2350, Fax) 215-615-4312, E-mail: [email protected]. Financial supports or conflicts disclosure S Kim - NIH T32 (AR07442) Training Program in Rheumatic Disease HK Choi - Holds the Mary Pack Arthritis Society of Canada Chair in Rheumatology, served on the advisory board for TAP and Savient Pharmaceuticals JP Guevara and DA Albert – None

Kim et al.

Page 2

Keywords

NIH-PA Author Manuscript

hyperuricemia; stroke; systematic review; meta-analysis

INTRODUCTION

NIH-PA Author Manuscript

Uric acid is the end-product of purine metabolism in humans. There is no universally accepted definition for hyperuricemia, but it is usually defined as serum urate concentration in excess of 6.8 mg/dL, which is the limit of urate solubility in serum 1, 2. Currently, no urate-lowering therapy is indicated in asymptomatic hyperuricemia 3. High serum uric acid level with or without gout is associated with cardiovascular diseases such as hypertension, coronary heart disease, peripheral vascular disease, and stroke 4–10. However, the role of high serum uric acid level as an independent risk factor for cardiovascular disease including stroke has been controversial 11–16. Several patho-physiological mechanisms through endothelial dysfunction, oxidative metabolism, platelet adhesiveness and aggregation, related to hyperuricemia in cardiovascular disease have been suggested 17–19. According to a recent report 20, about 780,000 Americans experience a new or recurrent stroke each year, on average, one stroke every 40 seconds. Preliminary data from 2005 indicate that stroke accounted for about 1 of every 17 deaths in the United States 20. If asymptomatic hyperuricemia has a deleterious effect on serious morbidity and mortality related to stroke, hyperuricemia may become a new target for more comprehensive risk factor management in the primary prevention of stroke. Accordingly, we conducted a meta-analysis of prospective cohort studies to determine the association between hyperuricemia and the risk of stroke.

METHODS Data Sources We searched three major electronic databases — MEDLINE (1950-July 2008), EMBASE (1980-July 2008), and the Cochrane library — for studies of the association between serum uric acid levels and stroke incidence and/or mortality. We also searched bibliographies of identified reports and review articles for additional references. Our search strategy is described in Figure 1. Study Eligibility

NIH-PA Author Manuscript

To be eligible for inclusion, we only considered (1) prospective cohort studies of adult patients, (2) longer than one year of follow-up, (3) with sample size of at least fifty subjects, and (4) inception cohort free of stroke. We imposed no geographic or language restrictions. Studies reporting interventional and secondary prevention trials were excluded. Selection of Studies Two authors (S. Kim and K. Kim) independently screened each of the potential titles, abstracts, and/or full-texts to determine inclusion. Areas of disagreement or uncertainty were resolved by consensus. When multiple articles were published from a single study, we selected the report that contained the most complete and relevant data on the association between hyperuricemia and stroke. The electronic search retrieved 566 potentially relevant studies. Non-electronic search identified 3 additional studies. On initial screening, 504 were excluded based on title. Of the 65 screened abstracts, 22 studies were retrieved for detailed evaluation. Of those, two studies were based on the same patient population, so they were considered as one study in the meta-analysis 21, 22. One study was excluded because it reported insufficient data on stroke outcome 23. Four studies that reported data on carotid

Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 3

intima-media thickness only were excluded 24–27. Eventually 16 studies were included in this meta-analysis (Figure 2).

NIH-PA Author Manuscript

Data Abstraction and Quality Assessment All data were independently abstracted in duplicate by two authors (S. Kim and K. Kim) using a data abstraction form. Discrepancies were resolved by consensus. When necessary, the original authors were contacted for additional information. Data on the study characteristics such as author name, year of publication, language, sample size, mean age, uric acid level and number of outcome were collected. The Newcastle–Ottawa Scale was used to assess the quality of studies 28 (Table 1). A quality score was calculated on the basis of three major components: selection of the study groups (0–4 points), quality of the adjustment for confounding (0–2 points) and ascertainment of the outcome of interest in the cohorts (0–3 points). A higher score represents better methodological quality. Adjustment for known stroke risk factors, duration of follow-up of at least 5 years, and adequate followup rate were criteria of higher quality. Statistical Analysis

NIH-PA Author Manuscript

Some studies included in our meta-analysis used the International system (SI) of units (µmol per liter) to report levels of serum uric acid. We therefore converted those to the conventional units (milligram per deciliter), using a conversion rate of 16.81 (1 mg/dL = 59.48 µmol/L) 29. Pooled estimates of risk ratios (RRs) were calculated using the DerSimonian and Laird random effects model 30, 31 for stroke incidence and mortality. This statistical technique weighs individual studies by sample size and variance (both within- and between-study variance) and yields a pooled point estimate and a 95% confidence interval. The DerSimonian and Laird technique was considered an appropriate pooling technique because of the relative heterogeneity of the source population in each study. We also evaluated the presence of heterogeneity across trials by using the I2 statistic, which quantifies the percentage of variability that can be attributed to between-study differences 32. To assess the potential for publication bias, we performed the Begg’s test and the Egger’s test and constructed funnel plots to visualize a possible asymmetry 33. All the statistical analyses were done in Stata 10 (Stata Corp, College Station, TX). We followed the Meta-analyses of Observational Studies in Epidemiology (MOOSE) guidelines 34 in the report of this meta-analysis.

RESULTS Study Characteristics

NIH-PA Author Manuscript

Sixteen prospective cohort studies representing data from 238,449 participants were included in the meta-analysis. The characteristics of the studies and of their participants are presented in Table 2. Of the 16 trials, 2 were conducted primarily in the United States although the study by Kagan et al 35 was based on a Hawaiian Japanese cohort. Eight studies were done in Asian countries and six studies were from European countries. The number of participants ranged from 153 in a study by Tofuku et al 36 to 83,683 in the Vorarlberg Health Monitoring and Promotion Program cohort study by Strasak et al 37. Nine studies 14, 15, 21, 22, 36, 38–42 included both men and women. Four studies used a lower cut-off value to define hyperuricemia for women compared to men 39, 40, 42, 43. Six studies 35, 37, 44–47 included only men, and one study 43 only women. Seven studies 35, 37, 39, 41, 43–45 reported gender-specific outcome for ischemic stroke. Six studies 15, 35, 37, 39, 41, 43 provided only adjusted risk estimates.

Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 4

Hyperuricemia and Stroke Incidence

NIH-PA Author Manuscript

The pooled estimate of unadjusted RRs for stroke based on six studies 14, 36, 38–40, 46 was 1.41 (95% CI: 1.05–1.76) among subjects with hyperuricemia, compared with those with normouricemia. The heterogeneity test was not significant (I2 =23.8%, p=0.25). The pooled multivariate RR based on four studies14, 39–41 fully adjusting for known risk factors of stroke was 1.47 (95% CI: 1.19–1.76). The heterogeneity test for this analysis was not significant (I2=0.0%, p=0.96). The pooled estimate of unadjusted RRs among men based on two studies 40, 46 was 1.82, (95% CI: 0.84–2.80), whereas only one study reported unadjusted RR of 2.55 among women (95% CI: 0.84–4.27)40. The pooled estimate of multivariate RRs based on three studies39– 41 was 1.42 (95% CI: 1.03–1.80) among women and 1.42 among men (95% CI: 0.94–1.90). The forest plot of multivariate RRs and 95% CIs for stroke incidence and hyperuricemia are shown in Figure 3 (top). Hyperuricemia and Stroke Mortality

NIH-PA Author Manuscript

The pooled estimate of unadjusted RR based on 6 studies 21, 22, 36, 42, 44, 45, 47 was 1.36 (95% CI: 1.03–1.69) for patients with hyperuricemia, compared with those with normouricemia. There was no statistically significant heterogeneity among the studies (I2= 0.0%, p=0.62). The pooled multivariate RR based on six studies 15, 37, 42–44 fully adjusting for known risk factors of stroke was 1.26 (95% CI: 1.12–1.39) with a nonsignificant heterogeneity test (I2= 0.0%, p=0.50). The pooled estimate of unadjusted RRs among men based on 5 studies 21, 42, 44, 45, 47 was 1.34 (95% CI: 1.01–1.67). The pooled unadjusted RR among women based on only two studies 21, 42 was 4.75, but it was not statistically significant (95% CI: 0.53–8.98). The pooled multivariate RRs were significantly higher for both men [N=4 studies 37, 42, 44, 45, RR 1.20, 95% CI: 1.05–1.35] and women [N=2 studies 42, 43, RR 1.35, 95% CI: 1.04– 1.66], compared to those with normouricemia. The forest plot of multivariate RRs and 95% CIs for stroke mortality and hyperuricemia are shown in Figure 3 (bottom). Publication Bias Assessment The funnel plots for both stroke incidence and mortality were visually examined (Figure 4). There was no statistical evidence of publication bias among studies for stroke incidence or mortality by using Egger’s (p=0.80; 0.25 respectively) and Begg’s (p=0.70; 0.48 respectively) tests.

NIH-PA Author Manuscript

Sensitivity Analyses Meta-regression was performed to further investigate the effect of three study-level characteristics (year of publication, race, and gender) on the risk of stroke. None of the regression coefficients was statistically significant (Appendix 1). A linear regression model (Appendix 2) showed that studies that adjusted for more confounding variables had lower risk estimates than studies that adjusted for fewer or no potentially confounding variables. For both stroke outcomes, the risk estimates remained above 1 even after adjusting for more than ten potential risk factors although the confidence intervals crossed 1.

DISCUSSION Our systematic review and meta-analysis of 16 prospective cohort studies finds that the elevated serum uric acid level in adults is associated with a modest but statistically significant increased risk of stroke incidence and mortality. Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 5

NIH-PA Author Manuscript

There has been considerable debate whether uric acid is neuro-protective as an antioxidant or neuro-toxic as a pro-oxidant 12, 48, 49. A pathogenetic role for uric acid in cardiovascular disease also remains to be elucidated, although recent experimental studies have shown that hyperuricemia is associated with endothelial dysfunction, local oxidant generation, elevated circulating levels of systemic inflammatory mediators such as monocyte chemoattractant protein-1, NF-κB, interleukin-1β, interleukin-6, and tumor necrosis factor-α, and vascular smooth muscle proliferation 18, 19, 50–53. Hypertension, one of the most common causes of stroke, is closely related to hyperuricemia 19. Several experimental animal models showed the development of systemic hypertension in hyperuricemic rats 54–57. In a recent randomized, placebo-controlled, crossover trial 58 involving 30 hyperuricemic adolescents with newly diagnosed hypertension, allopurinol treatment was associated with significant reductions in casual and 24-hour ambulatory blood pressure compared to placebo. More clinical trials with longer follow-up periods are needed to determine the safety and the generalizability of urate-lowering therapy such as allopurinol in hypertension.

NIH-PA Author Manuscript

Interestingly, many previous studies that investigated the role of the uric acid level on either cardiovascular disease or all-cause mortality observed a J-curve relationship 11, 16, 59–61. Similar results were noted in some of the studies included in our analysis 14, 15, 35, 42, 44. It has been postulated that a low uric acid level is associated with a higher mortality because uric acid may play an antioxidant role 19. The limitations of this meta-analysis fall into two categories: those attributable to the data available for analysis and those attributable to the techniques generally used to perform the meta-analysis. Our analysis is based on observational studies which are subject to confounding and other biases and cannot prove causality. Randomized clinical trials were, however, neither generally useful to evaluate etiological hypotheses over a long period of follow-up 31, nor available in the literature to examine our study questions. We selected only large prospective studies with inception cohort free of disease, which helped increase precision of estimates while minimizing heterogeneity. There were different definitions of hyperuricemia across the studies; therefore, we chose the category nearest to 6.8 mg/dl in each study for the hyperuricemia group. Although studies differed by geographic location, age, race, sex distribution and size, meta-regression analysis did not reveal any significant association with these factors. Nonetheless, we cannot rule out other potential sources for heterogeneity such as clinical features.

NIH-PA Author Manuscript

All meta-analyses are inherently vulnerable to publication bias. We attempted to minimize this bias by searching three major electronic databases with no language restriction. Three different statistical tests to examine the issue of publication bias were performed revealing no statistical evidence for significant publication bias. Lastly, because we did not have access to the raw data on adjustments in each study, we utilized the best adjusted RRs per individual study. The use of medications such as diuretics was adjusted for in only a few studies 14, 39, 41, 42. For the stroke outcome data, there is a possibility of misclassification bias because most of the papers we included used death certificates or diagnostic codes to define their outcomes. Our study has several important strengths. This meta-analysis is based on large prospective cohort studies over a long follow-up period in many different countries. To our knowledge, this is the first systematic review and meta-analysis on hyperuricemia and the risk of stroke. We assessed the quality of individual studies using the Newcastle–Ottawa Scale 28. The majority of the included studies had adequate sizes, follow-up lengths, and adjustments for other risk factors. We furthermore performed subgroup analyses of the studies with full

Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 6

adjustments for known stroke risk factors as the degree of adjustments for potential confounders was still different in each study.

NIH-PA Author Manuscript

In conclusion, our meta-analysis of published prospective studies suggests that high serum uric acid levels may modestly increase the risk of stroke incidence and mortality. Future research should focus on confirming the pathogenetic mechanisms of hyperuricemia as well as examining the role of urate-lowering therapy in stroke.

Acknowledgments Seo Young Kim is funded, in part, by a National Institutes of Health T32 training grant for training program in rheumatic disease.

Appendix 1 Effect of study variables by meta-regression

NIH-PA Author Manuscript

Stroke

Incidence

Stroke

Mortality

Coefficient

95 % CI

Coefficient

95 % CI

Gender

−0.09

−0.33-0.14

0.21

−0.11-0.53

Year

0.003

−0.02-0.03

−0.06

−0.12-0.01

Race

0.25

−0.14-0.65

0.27

−0.14-0.69

CI; confidence interval

Appendix 2 Linear regression plot for the relative risks (RR) of stroke against the number of adjusted risk factors in each study

NIH-PA Author Manuscript

CI: confidence interval

REFERENCES 1. Wilcox W, Khalaf A, Weinberger A, et al. Solubility of uric acid and monosodium urate. Med Biol Eng. 1972; 10:522–531. [PubMed: 5074854] 2. Becker M, Schumacher HJ, Wortmann R, et al. Febuxostat compared with allopurinol in patients with hyperuricemia and gout. N Engl J Med. 2005 December.353:2450–2461. [PubMed: 16339094]

Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 7

NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

3. Wortmann R. Gout and hyperuricemia. Curr Opin Rheumatol. 2002; 14:281–286. [PubMed: 11981327] 4. Baker J, Krishnan E, Chen L, et al. Serum uric acid and cardiovascular disease: recent developments, where do they leave us? Am J Med. 2005; 118:816–826. [PubMed: 16084170] 5. Becker M, Jolly M. Hyperuricemia and associated diseases. Rheum Dis Clin N Am. 2006; 32:275– 293. 6. Wannamethee S. Serum uric acid and risk of coronary heart disease. Curr Pharm Des. 2005; 11:4125–4132. [PubMed: 16375733] 7. Tseng C. Independent association of uric acid levels with peripheral arterial disease in Taiwanese patients with Type 2 diabetes. Diabet Med. 2004; 21:724–729. [PubMed: 15209765] 8. Baker J, Schumacher H, Krishnan E. Serum uric acid level and risk for peripheral arterial disease: analysis of data from the multiple risk factor intervention trial. Angiology. 2007; 58(4):450–457. [PubMed: 17875958] 9. Kim S, De Vera M, Choi H. Gout and mortality. Clin Exp Rheumatol. 2008; 26 Suppl. 51(5):S115– S119. [PubMed: 19026153] 10. Feig D, Kang D-H, Johnson R. Uric Acid and Cardiovascular Risk. NEJM. 2008 October; 359(17): 1811–1821. [PubMed: 18946066] 11. Culleton B, Larson M, Kannel W, et al. Serum uric acid and risk of cardiovascular disease and mortality: the Framingham Heart Study. Ann Intern Med. 1999; 31:7–13. [PubMed: 10391820] 12. Kanellis J, Johnson R. Elevated uric acid and ischemic stroke: accumulating evidence that it is injurious and not neuroprotective. Stroke. 2003; 34:1956–1958. [PubMed: 12843345] 13. Weir C, Muir S, Walters M, et al. Serum urate as an independent predictor of poor outcome and future vascular events after acute stroke. Stroke. 2003; 34:1951–1957. [PubMed: 12843346] 14. Lehto S, Niskanen L, Ronnemaa T, et al. Serum uric acid is a strong predictor of stroke in patients with non- insulin-dependent diabetes mellitus. Stroke. 1998 Mar; 29(3):635–639. [PubMed: 9506605] 15. Mazza A, Pessina A, Pavei A, et al. Predictors of stroke mortality in elderly people from the general population: The CArdiovascular STudy in the ELderly. European J Epidemiol. 2002; 17(12):1097–1104. [PubMed: 12530768] 16. Fang J, Alderman M. Serum uric acid and cardiovascular mortality: the NHANES I epidemiologic follow-up study, 1971–1992. JAMA. 2000; 283:2404–2410. [PubMed: 10815083] 17. Butler R, Morris A, Belch J, et al. Allopurinol normalizes endothelial dysfunction in type 2 diabetics with mild hypertension. Hypertension. 2000; 35:746–751. [PubMed: 10720589] 18. Kang D, Nakagawa T, Feng L, et al. A role for uric acid in the progression of renal disease. J Am Soc Nephrol. 2002; 13(12):2888–2897. [PubMed: 12444207] 19. Johnson R, Kang D, Feig D, et al. Is there a pathogenetic role for uric acid in hypertension and cardiovascular and renal disease? Hypertension. 2003; 41(6):1183–1189. [PubMed: 12707287] 20. Rosamond W, Flegal K, Furie K, et al. Heart Disease and Stroke Statistics—2008 Update. Circulation. 2008 January; 117(4):e25–e27. [PubMed: 18086926] 21. Bae J, Hyun D, Kwon T, et al. Serum uric acid is associated with cardiovascular events in patients with coronary artery disease. Korean Circ J. 2007 Apr; 37(4):161–166. 22. Hyun D, Kim K, Yoon H, et al. Gender differences in the role of serum uric acid for predicting cardiovascular events in patients with coronary artery disease. Korean Circ J. 2007 May; 37(5): 196–201. 23. Longo-Mbenza B, Luila E, Mbete P. Is hyperuricemia a risk factor of stroke and coronary heart disease among Africans? Int J Cardiol. 1999 Sep 30; 71(1):17–22. [PubMed: 10522560] 24. Joensuu T, Salonen R, Winblad I, et al. Determinants of femoral and carotid artery atherosclerosis. J Intern Med. 1994; 236(1):79–84. [PubMed: 8021577] 25. Viazzi F, Parodi D, Leoncini G, et al. Serum uric acid and target organ damage in primary hypertension. Hypertension. 2005 May; 45(5):991–996. [PubMed: 15781669] 26. Iribarren C, Folsom A, Eckfeldt J, et al. Correlates of uric acid and its association with asymptomatic carotid atherosclerosis: the ARIC Study. Atherosclerosis Risk in Communities. Ann Epidemiol. 1996 Jul; 6(4):331–340. [PubMed: 8876844]

Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 8

NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

27. Kawamoto R, Tomita H, Oka Y, et al. Association between uric acid and carotid atherosclerosis in elderly persons. Intern Med. 2005 Aug; 44(8):787–793. [PubMed: 16157974] 28. Wells, G.; Shea, A.; O'Connell, D., et al. The Newcastle-Ottawa Scale (NOS) for Assessing the Quality of Nonrandomised Studies in Meta-Analyses. 2008. http://www.ohri.ca/programs/clinical_epidemiology/oxford.htm 29. Young, D.; Huth, E. SI Units for Clinical Measurement. Philadelphia, Pa: American College of Physicians. 1st ed.. Philadelphia: American College of Physicians; 1998. 30. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986; 7(3):177–188. [PubMed: 3802833] 31. Egger, M.; Smith, G.; Altman, D. Systematic reviews in health care. Second ed.. London: BMJ Publishing Group; 2001. 32. Higgins J, Thompson S. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002; 21:1539– 1558. [PubMed: 12111919] 33. Sterne J, Egger M. Funnel plots for detecting bias in meta-analysis: Guidelines on choice of axis. J Clin Epidemiol. 2001; 54:1046–1055. [PubMed: 11576817] 34. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of Observational Studies in Epidemiology: A Proposal for Reporting. JAMA. 2000; 283(15):2008–2012. [PubMed: 10789670] 35. Kagan A, Popper JS, Rhoads GG, et al. Dietary and other risk factors for stroke in Hawaiian Japanese men. Stroke. 1985 May–Jun; 16(3):390–396. [PubMed: 4002255] 36. Tofuku Y, Kuroda M, Takeda R. Hyperuricemia associated with hypertension. A 4-year follow-up study of hyperuricemic hypertensives. Jpn Circ J. 1978 Jul; 42(7):871–879. [PubMed: 713025] 37. Strasak A, Ruttmann E, Brant L, et al. Serum uric acid and risk of cardiovascular mortality: A prospective long-term study of 83 683 Austrian men. Clin Chem. 2008 Feb; 54(2):273–284. [PubMed: 18039719] 38. Baba T, Amasaki Y, Soda M, et al. Fatty liver and uric acid levels predict incident coronary heart disease but not stroke among atomic bomb survivors in Nagasaki. Hypertens Res. 2007 Sep; 30(9): 823–829. [PubMed: 18037775] 39. Bos MJ, Koudstaal PJ, Hofman A, et al. Uric acid is a risk factor for myocardial infarction and stroke: The Rotterdam Study. Stroke. 2006 Jun; 37(6):1503–1507. [PubMed: 16675740] 40. Chien KL, Hsu HC, Sung FC, et al. Hyperuricemia as a risk factor on cardiovascular events in Taiwan: The Chin-Shan Community Cardiovascular Cohort Study. Atherosclerosis. 2005 Nov; 183(1):147–155. [PubMed: 16154134] 41. Hozawa A, Folsom AR, Ibrahim H, et al. Serum uric acid and risk of ischemic stroke: the ARIC Study. Atherosclerosis. 2006 Aug; 187(2):401–407. [PubMed: 16239005] 42. Sakata K, Hashimoto T, Ueshima H, et al. Absence of an association between serum uric acid and mortality from cardiovascular disease: NIPPON DATA 80, 1980–1994. Eur J Epidemiol. 2001; 17(5):461–468. [PubMed: 11855580] 43. Strasak A, Kelleher C, Brant L, et al. Serum uric acid is an independent predictor for all major forms of cardiovascular death in 28,613 elderly women: A prospective 21-year follow-up study. Int J Cardiol. 2008 Apr; 125(2):232–239. [PubMed: 18237790] 44. Gerber Y, Tanne D, Medalie J, et al. Serum uric acid and long-term mortality from stroke, coronary heart disease and all causes. Eur J Cardiovasc Prev Rehabil. 2006 Apr; 13(2):193–198. [PubMed: 16575272] 45. Jee SH, Lee SY, Kim MT. Serum uric acid and risk of death from cancer, cardiovascular disease or all causes in men. Eur J Cardiovasc Prev Rehabil. 2004 Jun; 11(3):185–191. [PubMed: 15179098] 46. Takagi M. Serum uric acid as a risk factor for stroke in a fishing village of rural southern Japan. Jpn Circ J. 1982 Feb; 46(2):131–136. [PubMed: 7057612] 47. Tomita M, Mizuno S, Yamanaka H, et al. Does hyperuricemia affect mortality? A prospective cohort study of Japanese male workers. J Epidemiol. 2000; 10(6):403–409. [PubMed: 11210110] 48. Cutler R. Urate and ascorbate: their possible roles as antioxidants in determining longevity of mammalian species. Arch Gerontol Geriatr. 1984; 3:321–348. [PubMed: 6532339] 49. Proctor P. Uric acid: neuroprotective or neurotoxic? Stroke. 2008; 39:e88. [PubMed: 18369163]

Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 9

NIH-PA Author Manuscript NIH-PA Author Manuscript

50. Kanellis J, Feig D, Johnson R. Does asymptomatic hyperuricaemia contribute to the development of renal and cardiovascular disease? An old controversy renewed. Nephrology (Carlton). 2004; 9(6):394–399. [PubMed: 15663643] 51. Kanellis J, Watanabe S, Li J, et al. Uric acid stimulates monocyte chemoattractant protein-1 production in vascular smooth muscle cells via mitogen-activated protein kinase and cyclooxygenase-2. Hypertension. 2003; 41(6):1287–1293. [PubMed: 12743010] 52. Kang D, Han L, Ouyang X, et al. Uric acid causes vascular smooth muscle cell proliferation by entering cells via a functional urate transporter. Am J Nephrol. 2005; 25(5):425–433. [PubMed: 16113518] 53. Rao G, Corson M, Berk B. Uric acid stimulates vascular smooth muscle cell proliferation by increasing platelet-derived growth factor A-chain expression. J Bio Chem. 1991; 266(13):8604– 8608. [PubMed: 2022672] 54. Mazzali M, Kanellis J, Han L, et al. Hyperuricemia induces a primary arteriolopathy in rats by a blood pressure-independent mechanism. Am J Physiol Renal Physiol. 2002; 282:F991–F997. [PubMed: 11997315] 55. Watanabe S, Kang D, Feng L, et al. Uric acid, hominoid evolution, and the pathogenesis saltsensitivity. Hypertension. 2002; 40:355–356. [PubMed: 12215479] 56. Mazzali M, Hughes J, Kim Y, et al. Elevated uric acid increases blood pressure in the rat by a novel crystal-independent mechanism. Hypertension. 2001; 38:1101–1106. [PubMed: 11711505] 57. Sanchez-Lozada L, Tapia E, Avila-Casado C, et al. Mild hyperuricemia induces glomerular hypertension in normal rats. Am J Physiol Renal Physiol. 2002; 283:F1105–F1110. [PubMed: 12372787] 58. Feig D, Soletsky B, Johnson R. Effect of Allopurinol on Blood Pressure of Adolescents With Newly Diagnosed Essential Hypertension. JAMA. 2008; 300(8):924–932. [PubMed: 18728266] 59. Hsu S-P, Pai M-F, Peng Y-S, et al. Serum uric acid levels show a ‘J-shaped’ association with allcause mortality in haemodialysis patients. Nephrol Dial Transplant. 2004; 19(2):457–462. [PubMed: 14736974] 60. Mazza A, Zamboni S, Rizzato E, et al. Serum uric acid shows a J-shaped trend with coronary mortality in non-insulin-dependent diabetic elderly people. The CArdiovascular STudy in the ELderly (CASTEL). Acta Diabetol. 2007; 44(3):99–105. [PubMed: 17721747] 61. Suliman M, Johnson R, García-López E, et al. J-shaped mortality relationship for uric acid in CKD. Am J Kidney Dis. 2006; 48(5):761–771. [PubMed: 17059995]

NIH-PA Author Manuscript Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 10

NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 1.

Search strategy

NIH-PA Author Manuscript Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 11

NIH-PA Author Manuscript NIH-PA Author Manuscript Figure 2.

Selection of studies included in the analysis.

NIH-PA Author Manuscript Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 12

NIH-PA Author Manuscript

Figure 3.

Random effects analysis of fully adjusted studies for the association between hyperuricemia and stroke Points (dot) and overall (diamond) estimates are given as risk ratios (RR) with 95% CI. The size of each box represents the weight of the corresponding study in our meta-analysis; Combined: studies which did not have gender specific data, ES: effect size, CI: confidence interval.

NIH-PA Author Manuscript NIH-PA Author Manuscript Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 13

NIH-PA Author Manuscript Figure 4.

Begg’s funnel plot for publication bias in studies for stroke incidence and mortality RR; risk ratio, s.e.; standard error

NIH-PA Author Manuscript NIH-PA Author Manuscript Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

Kim et al.

Page 14

Table 1

Quality assessment of included studies based on Newcastle-Ottawa Scale

NIH-PA Author Manuscript NIH-PA Author Manuscript

Author

Selection

Comparability

Outcome

Tofuku 36

3

0

2

Takagi 46

4

0

3

Kagan 35

4

1

3

Lehto 14

3

2

3

Chien 40

4

2

3

Bos 39

4

2

3

Hozawa 41

4

2

3

Baba 38

3

0

3

Tomita 47

4

1

3

Sakata 42

4

2

3

Mazza 15

4

2

3

Jee 45

4

2

3

Gerber 44

4

2

3

Bae/Hyun 21, 22

3

0

2

Strasak 37

4

2

3

Strasak 43

4

2

3

NIH-PA Author Manuscript Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.

NIH-PA Author Manuscript 314 men in the fishing village (100) 7,895 Japanese Hawaiian men (100)

1,017 subjects with NIDDM (54)

3,602 subjects in the ChinShan Community (47) 4,385 subjects in the Rotterdam study (35)

13,413 subjects in the ARIC study (46)

2,024 atomic bomb survivors (39)

49,413 male railroad workers (100) 8,172 subjects in the National Cardiovascular study (44)

Kagan 35 1985 USA

Lehto 14 1998 Finland

Chien 40 2005 Taiwan

Bos 39 2006 Netherlands

Hozawa 412006 USA

Baba 38 2007 Japan

Tomita 47 2000 Japan

Sakata 42 2001 Japan

153 hypertensive patients (75)

Tofuku*36 1978 Japan

Takagi 46 1982 Japan

Participants (% male)

Author, year of publication, country

Arthritis Rheum. Author manuscript; available in PMC 2010 July 15. 50

25–60

63

45–64

63–76

35+

45–64

45–68

50–79

16–77

Age range or mean (yr)

13.3

5.4

8

12.6

8.4

11

7.2

10

8

3.7

Hyperuric emia definition (mg/dl)

163 (79 ischemic, 65 hemorrhagic and 19 unknown) 114

155

381 (205 ischemic, 46 hemorrhagic, and 130 unknown) 381

84 (63 ischemic and 21 hemorrhagic)

≥ 7.2 (M)

≥ 5 (C)

≥ 7.7 (M) ≥ 6.6 (W) ≥ 6.4 (M) ≥ 5.4 (W)

≥ 6.9 (M) ≥ 6.9 (W)

≥ 7 (C)

127 174

≥ 8.5 (M) ≥ 6.5 (M) ≥ 5.0 (W)

STROKE MORTALITY

30

4

No. of total outcomes

≥ 7.5 (C)

≥ 7 (C)

STROKE INIDENCE

Follow-up (yr)

NIH-PA Author Manuscript

Characteristics of included cohort studies.

Any type of stroke based on ICD-9 codes

Any type of stroke based on ICD-9 codes

Any type of stroke based on clinical presentation and CT/ MRI

Ischemic stroke based on ICD-9 codes and CT/MRI

All stroke, ischemic and hemorrhagic stroke based on hospital records and CT/MRI

Any type of stroke based on clinical manifestation, hospital records, and death certificates

Any stroke except subarachnoid hemorrhage based on ICD-9 codes and hospital records

All stroke, ischemic, and hemorrhagic stroke based on hospital records

Any stroke

Any stroke

Outcome definition

Age, BMI, SBP, anti- HTN meds, cholesterol, serum creatinine, glucose, smoking, alcohol intake, LVH

Age

--

Age, sex, race, education, SBP, DM, anti-HTN meds, smoking, alcohol intake, serum albumin, vWF, and HDL

Age, sex, SBP, cholesterol, HDL, DM, smoking, use of diuretics, and waist/hip ratio

Age, SBP, BMI, LDL, HDL, smoking, alcohol intake, LVH, and AF

Age, gender, smoking, cholesterol, HTN, BMI, total triglycerides, HDL, glucose, use of diuretics, duration of DM, and historical stroke †

Age

--

--

Variables controlled

NIH-PA Author Manuscript

Table 2 Kim et al. Page 15

660 subjects with coronary artery disease (57) 83,683 men in the risk factor surveillance program (100)

28,613 women in the risk factor surveillance program (0)

Bae/Hyun 21, 22 2007 Korea

Strasak 37 2008 Austria

Strasak 43 2008 Austria

62.3

41.6

59

49

30–77

65+

13.6

12.4

2.3

23

6.5

14

Follow-up (yr)

12 645 (147 ischemic, 147 hemorrhagic, and 351 unknown) 776 (211 ischemic, 114 hemorrhagic, and 451 unknown)

≥ 6.1 (M) ≥ 5.1 (W) ≥ 6.7 (M)

≥ 5.4 (W)

192

≥ 7.01(M)

292

170

≥ 6.4 (C)

≥ 5.6 (M)

No. of total outcomes

Hyperuric emia definition (mg/dl)

All stroke, ischemic, and hemorrhagic stroke based on ICD-9 codes and autopsy

All stroke, ischemic, and hemorrhagic stroke based on ICD-9 codes and autopsy

Any type of stroke based on hospital records

Any type of stroke based on ICD-9 codes

Any type of stroke based on hospital records and ICD-9 codes

Any type of stroke based on ICD-9 codes

Outcome definition

Age, BMI, SBP, DBP, cholesterol, triglycerides, GGT, glucose, smoking, occupation, and year of examination

Age, BMI, SBP, DBP, cholesterol, triglycerides, GGT, glucose, smoking, and year of examination

-

Age, BMI, SBP, DM, cholesterol, smoking, and LVH

Age, DM, HTN, cholesterol, and smoking

Age, historical stroke †, CAD, HTN, SBP, pulse pressure, AF, LVH, smoking, serum potassium and sodium

Variables controlled

previous stroke in less than 10 % of the study population; CAD: coronary artery disease; AF: atrial fibrillation; LVH: left ventricular hypertrophy; BMI: body mass index; NIDDDM: non-insulin dependent diabetes mellitus; ARIC: Atherosclerosis Risk in Communities study; KMIC: Korea Medical Insurance Corporation

†

Included for both stroke mortality and incidence; C: combined; M: men; W: women; -- : unadjusted; SBP: systolic blood pressure; DBP: diastolic blood pressure; DM: diabetes mellitus; HTN: hypertension; vWF, von Willebrand factor; HDL: high density lipoprotein; LDL: low density lipoprotein;

*

9,125 middle-aged male workers (100)

22,698 subjects in the KMIC (100)

Jee 45 2004 Korea

Gerber 44 2006 Israel

3,282 elderly subjects (39)

NIH-PA Author Manuscript

Mazza 15 2002 Italy

Age range or mean (yr)

NIH-PA Author Manuscript

Participants (% male)

NIH-PA Author Manuscript

Author, year of publication, country

Kim et al. Page 16

Arthritis Rheum. Author manuscript; available in PMC 2010 July 15.