Much Ado About Nothing, or Much to Do About Something? The Continuing Controversy Over the Role of Uric Acid in Cardiovascular Disease
To the Editor:
Recently, an update from the Framingham study could not find uric acid to be an independent risk factor for cardiovascular disease.1 While serum uric acid levels correlated significantly with the risk for cardiovascular events and mortality in women, this relationship became insignificant after factoring for 11 additional variables including hypertension, body mass index, and diuretic use.1 Both the authors1 and an accompanying editorial2 interpreted these findings as showing that uric acid is not a true risk factor for cardiovascular disease and that it should not be routinely measured to assess cardiovascular risk.
The careful analysis of the Framingham study is to be commended, but one must be cautious in the interpretation of the findings. While some epidemiologic studies such as the current one have not been able to show uric acid to be an independent risk factor for cardiovascular disease, other studies using multivariate analyses3 4 5 6 came to an opposite conclusion. Another recently completed study, the Worksite,7 also found uric acid to be an independent risk factor for cardiovascular events and mortality, especially in women. One might look for subtle explanations to account for the differences in these various studies, as Culleton et al1 have attempted, but most of the studies examined the very same variables.
A more central issue is whether one should interpret the finding that a risk factor is not statistically independent to mean that it should not be considered biologically important. We would argue that this is not true in several situations. First, if the risk factors are causally linked, then one may not be able to show that they are independent of each other. For example, although smoking is a risk factor for mortality, it might no longer be independent if it is controlled for chronic lung disease and lung cancer; major causes of death in smokers. Similarly, the impact of hypertension as a risk factor may be decreased when left ventricular hypertrophy is included in the multivariate model; or the importance of obesity as a risk factor may be decreased if hyperinsulinemia is also included in this model. Whether hyperuricemia is causally linked to other risk factors remains unclear. However, the observation that longstanding hyperuricemia in patients with gout is associated with intrarenal crystal deposition, tubulointerstitial injury and the development of hypertension provides a potential pathogenic mechanism linking an elevated uric acid with renal disease (even subclinical injury) and hypertension.8
It may also be very difficult to show a risk factor is independent if the linkage to other risk factors is strong. For example, in the Seven Countries Study, dietary intakes of saturated fat, trans-fatty acids, and cholesterol could not be separated as “independent” risk factors for 25-year coronary heart disease mortality because they are so highly interrelated.9 Nevertheless, all of these fats are believed to be pathophysiologically important, especially as they relate to producing adverse lipoprotein patterns.10 11 Another example is the remarkable association of obesity with hypertension, elevated serum insulin levels, and hyperuricemia (the metabolic syndrome or syndrome “X”).12 A multivariate model could well eliminate hyperuricemia as an independent risk factor even if it were contributing to the overall phenotypic risk of the syndrome. However, our recent angiographic study showed elevated uric acid to be an independent predictor for coronary atherosclerosis, even when hyperinsulinemia and other markers of the metabolic syndrome were included in the model.13
One might make the counterargument that multivariate analyses does take this latter issue into account and will provide a measure of the effect of just having one particular risk factor (such as elevated uric acid) in the absence of the other risk factors. Whether there are really adequate numbers of patients with hyperuricemia that would fit in this category is unclear. However, a concern from the recent Framingham analysis1 is that the analysis was performed using only a single baseline uric acid level. Since acute changes in uric acid (of 1 mg/mL or more) can be induced by alcohol, diet, or exercise,14 one wonders if some of the analysis may have involved patients with transient hyperuricemia who would presumably carry less risk for developing cardiovascular events. While it is not possible from the present study to separate individuals with transient hyperuricemia from chronic hyperuricemia, an earlier analysis from Framingham examined the risk for cardiovascular disease in patients with gout (who were more likely to have chronic hyperuricemia). This study did find a significant and independent risk for developing cardiovascular disease in males even when factored for numerous other risk factors.15
In the paper by Culleton et al,1 one of the major variables that eliminated uric acid as an independent risk factor was the use of diuretics, which are known to increase uric acid levels by reducing uric acid excretion.16 However, diuretics are used in the treatment of hypertension, in which they have been shown to reduce mortality and not to increase it.16 Therefore, one would not expect to see any additional effects from controlling for diuretics over the effect of factoring out hypertension on cardiovascular mortality and yet an effect was observed. Culleton et al suggested that perhaps the diuretics were more likely administered to the more severe cases of hypertension, but during the time period of the study, diuretics were generally considered first line therapy.16 In this regard, it is important that both epidemiologic and clinical studies, including the Hypertension Detection and Follow-up Program, the Worksite study, and the Systolic Hypertension in the Elderly program (SHEP) trial all found that the benefit of diuretics was blunted if hyperuricemia developed.6 8 17 In the SHEP trial, the entire benefit of diuretics on cardiovascular mortality was prevented if the uric acid levels increased by >1 mg/dL.17 This suggests that it is not the diuretics per se that increase our risk but whether or not they result in hyperuricemia. Perhaps we should reverse our thinking about the conclusions over which factor should be viewed as dependent or independent. In other words, we could view diuretic-associated mortality to be dependent on its associations with hypertension and hyperuricemia. Then an opposite conclusion could be made: hyperuricemia is an independent cause of mortality and can account for the excess mortality associated with diuretic use.
Given the concerns over the interpretation of the importance of an independent risk factor, one might posit that the best way to prove that a factor is pathogenically important is to show that blocking or treating the factor reduces cardiovascular risk. However, the timing may be critical. Just as quitting smoking may not have much effect in reducing mortality once a patient has developed lung cancer, the lowering of uric acid may or may not be helpful once a patient has established hypertension or renal injury. However, it is exciting that recent prospective and controlled studies of patients with gout suggest that lowering uric acid can improve renal function in patients with either minimal or moderate renal disease.18 19
We would argue that the role of uric acid as a risk factor remains unresolved and future studies examining pathogenesis and effects of treatment are badly needed. Regardless of whether uric acid is an independent risk factor, or even whether it has a pathogenic role in cardiovascular disease, the bottom line is that measuring uric acid is a useful test for the clinician, as it carries important prognostic information. An elevated uric acid is associated with increased risk for cardiovascular disease and mortality, especially in women.1 3 4 5 6 7 8 Hypertensive individuals with an elevated uric acid level are at particular risk for both cardiovascular and cerebrovascular events,20 and the benefit of diuretic use to lower blood pressure is thwarted if the uric acid levels remain high or increase.6 8 17
Freedman DS, Williamson DF, Gunter EW, Byers T. Relation of serum uric acid to mortality and ischemic heart disease. Am J Epidemiol. 1995;141:637–644.
Persky VW, Dyer AR, Idris-Soven E, Stamler J, Shekelle RB, Schoenburger JA, Berkson DM, Lindberg HA. Uric acid: a risk factor for coronary heart disease? Circulation. 1979;59:969–977.
Alderman MH, Cohen H, Madhavean S, Kivlighn S. Serum uric acid and cardiovascular events in successfully treated hypertensive patients. Hypertension. 1999;34:144–150.
Daan K, Allesandro M, Bloemberg B, Christ A, Blackburn H, Buzina R, Dontas AS, Fidanza F, Giampaoli S, Jansen A, Karvonen M, Katan M, Nissinen A, Nedeljkovic S, Pekkanen J, Pakkarinen M, Punsar S, Räsänen L, Simic B, Toshima H. Dietary saturated and trans fatty acids and cholesterol and 25-year mortality from coronary heart disease: The Seven Countries Study. Prev Med. 1995;24:308–315.
Tuttle KR, Short RA, Johnson RJ. Relationship between uric acid and coronary artery disease. J Am Soc Nephrol. 1999;10:373A. Abstract.
Papademetriou V. Diuretics in hypertension: clinical experiences. Eur Heart J. 1992;13:S92–S95.
Franse LV, Pahor M, DiBari M, Shorr RI, Wan JY, Somes GW, Applegate WB. Serum uric acid, its change with diuretic use and risk of cardiovascular events in the Systolic Hypertension in The Elderly Program (SHEP). Abstract presented at: Fourteenth Scientific Meeting of the American Society of Hypertension; May 16-20, 1999; New York, NY.
Perez-Ruiz F, Alonso-Ruiz A, Calabozo M, Herrero-Beites A, Garcia-Erauskin G, Ruiz-Lucca E. Efficacy of allopurinol and benzbromarone for the control of hyperuricemia: a pathogenic approach to the treatment of primary chronic gout. Ann Rheum Dis. 1998;57:545–549.
Perez-Ruiz F, Calabozo M, Fernandez-Lopez MJ, Herrero-Beites A, Ruiz-Lucea E, Garcia-Erasukin G, Duruelo J, Alonso-Ruiz A. Treatment of chronic gout in patients with renal function impairment: an open, randomized actively controlled study. J Clin Rheumatol. 1999;5:49–55.
Breckenridge A. Hypertension and hyperuricemia. Lancet. 1966; i:15–18.
The interest in serum uric acid as a potential cardiovascular disease (CVD) risk factor has ballooned in the last several years with numerous abstracts and research papersR1 R2 R3 R4 R5 R6 R7 R8 R9 and multiple editorials and review articles.R10 R11 R12 R13 R14 R15 There seems little doubt that serum uric acid is a risk factor for CVD, a perception that has persisted to varying degrees for nearly 50 years.R16 However, the central issue is not whether uric acid is associated with CVD, but whether uric acid plays a causal role in the development of CVD, ie, a causal risk factor. A recent editorialR12 is to be commended for considering this issue according to the main criteria for causation as described by Bradford Hill (Table⇑).R17 To avoid repetition, we will highlight several criteria and specifically address issues raised by Johnson and Tuttle in the accompanying commentary.
Johnson and Tuttle claim several studies demonstrate that uric acid is an “independent” risk factor for CVD.R1 R18 R19 R20 R21 Let us consider the evidence. First, the strength of association is generally weak or nonexistent in these reports. The Worksite project found a hazard ratio of 1.22 for CVD incidence in the top quartile of uric acid versus the lowest quartile.R1 Effects of this size are likely meaningful in randomized treatment trials but are less convincing in observational studies. Furthermore, on more careful analysis, the study by Bengtsson et alR19 found no association of uric acid with incident CVD. Data from the Chicago Heart Association Detection Project in Industry revealed no relation between uric acid and CVD in men.R20 In contrast to the claim by Johnson and Tuttle, Langford et alR21 concluded, “we doubt that uric acid has any causal role in increasing mortality.”
Examination of the temporal relation between uric acid and CVD also warrants specific comment. Clearly, if a factor is believed to be a cause of a disease, exposure to the factor must precede the development of the disease. In our study, individuals with clinically apparent CVD at baseline were excluded to determine whether a temporal relation between serum uric acid and CVD existed.R3 In contrast, 17% of subjects had known CVD at baseline in the Worksite report.R1 In the same report, uric acid was no longer predictive of incident CVD when subjects with prevalent CVD at baseline were excluded. Similarly, the Chicago Heart AssociationR20 and the Systolic Hypertension in the Elderly ProgramR7 also included subjects with prevalent CVD.
The biological plausibility for a causal role of uric acid in CVD is inconclusive. Intrarenal deposition of uric acid with subsequent tubulointerstitial injury and incident hypertensionR13 is possible but not established. Recently, attention has been focused on the role of uric acid as an antioxidant.R22 Increased oxidative stress and the oxidation of LDL in the arterial wall by peroxynitrite may play an important role in the development of atherosclerosis. Elevated serum uric acid may reflect a response to increased production of endogenous reactive oxygen species because uric acid is a potent scavenger of peroxynitrite.R23 This can be demonstrated clinically as well. Hyperuricemia occurs during interruption of limb arterial flow,R24 after coronary angioplasty,R25 during coronary artery bypass surgery,R26 and in other hypoxic statesR27 R28 R29 perhaps reflecting oxidative stress, but by no means does such evidence implicate uric acid as a cause of atherosclerosis.
We would agree with Johnson and Tuttle’s claim that “it may also be very difficult to show a risk factor is independent if the linkage to other risk factors is very strong.” The associations of uric acid with hypertension, dyslipidemia, insulin resistance, obesity, and alcohol abuse are well described. Multicollinearity in statistical models exploring the relations of these factors with incident CVD is a difficult problem. Ideally, one would examine the impact of a range of uric acid values on CVD incidence in a cohort of individuals homogeneous for other CVD risk factors; however, such a theoretical study sample does not exist. We therefore adjust, or control for, these factors using accepted statistical techniques. The choice of which variables to include in the multivariable model is based largely on the knowledge of established risk factors for the outcome of interest. Obviously, one does not adjust for lung cancer when determining the mortality risk associated with smoking!
Even after adjusting for known CVD risk factors, the possibility of residual confounding must be considered. As highlighted in the accompanying commentary, one of the major variables that eliminated uric acid as an independent risk factor in our reportR3 was the use of diuretics. We feel diuretic therapy is an important confounder of the uric acid–CVD relationship and must be considered in analyses. In addition to inducing hyperuricemia, diuretic therapy is a marker for more severe hypertension. Although diuretics were generally considered first line therapy during the time of uric acid collection for our investigation, a large majority of hypertensives in the United States at that time were not receiving any drug therapy.R30 Subjects receiving antihypertensive treatment had more severe hypertension than those not receiving therapy.R31 Contrary to Johnson and Tuttle’s claim, one does see an additional effect from controlling for diuretic use over and above the effect of factoring out hypertension alone and that is precisely what we observed.
In conclusion, results from our extensive and carefully conceived analyses revealed no evidence to support uric acid as a causal CVD risk factor. Similarly, other studies meet few, if any, of Bradford Hill’s criteria for causation.
Alderman MH, Cohen H, Madhavan S, Kivlighn S. Serum uric acid and cardiovascular events in successfully treated hypertensive patients. Hypertension. 1999;34:144–150.
Anker SD, Leyva F, Poole-Wilson PA, Coats AJS. Uric acid as independent predictor of impaired prognosis in patients with chronic heart failure. J Am Coll Cardiol. 1998;31:154. Abstract.
Culleton BF, Larson MG, Kannel WB, Levy D. Serum uric acid and risk for cardiovascular disease and death: the Framingham Heart Study. Ann Intern Med. 1999;131:7–13.
Lehto S, Niskanen L, Ronnemaa T, Laakso M. Serum uric acid is a strong predictor of stroke in patients with non-insulin-dependent diabetes mellitus. Stroke. 1998;29:635–639.
Alderman MH, Cohen H, Kivlighn S, Madhavan S. Does treatment-mediated increase in uric acid reduce the cardioprotective effect of diuretics in hypertensive patients? Am J Hypertens. 1998;11:16. Abstract.
Franse LV, Pahor M, Di Bari M, Shorr RI, Wan JY, Somes GW, Applegate WB. Serum uric acid, its change with diuretic use and risk of cardiovascular events in the Systolic Hypertension in the Elderly Program (SHEP). Am J Hypertens. 1999;12:13A. Abstract.
Liese AD, Hense HW, Lowel H, Doring A, Keil U. Association of serum uric acid with incident myocardial infarction, all cause and CVD mortality in the MONICA Augsburg cohort. Circulation. 1998;97:822. Abstract.
Tuttle KR, Short RA, Johnson RJ. Relationship between uric acid and coronary artery disease. J Am Soc Nephrol. 1999;10:373A. Abstract.
Burnier M, Brunner HR. Is hyperuricemia a predictor of cardiovascular risk? Curr Opin Nephrol Hypertens. 1999;8:167–172.
Johnson RJ, Kivlighn SD, Kim Y-G, Suga S, Fogo A. Reappraisal of the pathogenesis and consequences of hyperuricemia in hypertension, cardiovascular disease, and renal disease. Am J Kidney Dis. 1999;33:225–234.
Vaccarino V, Krumholz HM. Risk factors for cardiovascular disease: one down, many more to evaluate. Ann Intern Med. 1999;131:62–63.
Gertler MM, Garn SM, Levine SA. Serum uric acid in relation to age and physique in health and in coronary heart disease. Ann Intern Med. 1951;34:1421–1431.
Freedman DS, Williamson DF, Gunter EW, Byers T. Relation of serum uric acid to mortality and ischemic heart disease: the NHANES I epidemiologic follow-up study. Am J Epidemiol. 1995;141:637–644.
Bengtsson C, Lapidus L, Stendahl C, Waldenstrom J. Hyperuricaemia and risk of cardiovascular disease and overall death: a 12-year follow-up of participants in the population study of women in Gothenburg, Sweden. Acta Med Scand. 1988;224:549–555.
Persky VW, Dyer AR, Idris-Soven E, Stamler J, Shekelle RB, Schoenberger JA, Berkson DM, Lindberg HA. Uric acid: a risk factor for coronary heart disease? Circulation. 1979;59:969–977.
Langford HG, Blaufox MD, Borhani NO, Curb JD, Molteni A, Schneider KA, Pressel S. Is thiazide-produced uric acid elevation harmful? Analysis of data from the hypertension detection and follow-up program. Arch Intern Med. 1987;147:645–649.
Hooper DC, Spitsin S, Kean RB, Champion JM, Dickson GM, Chaudhry I, Koprowski H. Uric acid, a natural scavenger of peroxynitrite, in experimental allergic encephalomyelitis and multiple sclerosis. Proc Natl Acad Sci USA. 1998;95:675–680.
Lazzarino G, Raatikainen P, Nuutinen M, Nissinen J, Tavazzi B, Di Pierro D, Giardina B, Peuhkurinen K. Myocardial release of malondialdehyde and purine compounds during coronary bypass surgery. Circulation. 1994;90:291–297.
Anker SD, Leyva F, Poole-Wilson PA, Kox WJ, Stevenson JC, Coats AJ. Relation between serum uric acid and lower limb blood flow in patients with chronic heart failure. Heart. 1997;78:39–43.
Leyva F, Anker S, Swan JW, Godsland IF, Wingrove CS, Chua TP, Stevenson JC, Coats AJ. Serum uric acid as an index of impaired oxidative metabolism in chronic heart failure. Eur Heart J. 1997;18:858–865.
Morbidity and Mortality: Chartbook on Cardiovascular, Lung, and Blood Diseases. Bethesda, MD: U. S. Dept of Health and Human Services; 1998.
Shea S, Cook EF, Kannel WB, Goldman L. Treatment of hypertension and its effect on cardiovascular risk factors: data from the Framingham Heart Study. Circulation. 1985;71:22–30.