Hyperuricemia in acute heart failure. More than a simple spectator?

European Journal of Internal Medicine 20 (2009) 74 – 79 www.elsevier.com/locate/ejim

Original article

Hyperuricemia in acute heart failure. More than a simple spectator?☆ Anna L. Alimonda, Julio Núñez ⁎, Eduardo Núñez, Oliver Husser, Juan Sanchis, Vicent Bodí, Gema Miñana, Rocio Robles, Luis Mainar, Pilar Merlos, Helene Darmofal, Ángel Llácer Servicio de Cardiología, Hospital Clínico Universitario, Universitat de Valencia, Valencia, Spain Received 4 October 2007; received in revised form 17 February 2008; accepted 27 April 2008 Available online 10 June 2008

Abstract Background: Hyperuricemia is a prevalent condition in chronic heart failure (CHF), describing increased oxidative stress and inflammation. Although there is evidence that serum uric acid (UA) predicts mortality in CHF, its role as a prognostic biomarker in acute heart failure (AHF) has not yet been well assessed. The aim of this study was to determine if UA levels predict all-cause mortality. Additionally, as a secondary endpoint we sought the clinical predictors of UA serum level in this population. Methods: We analyzed 560 consecutive patients with AHF admitted in a single university center. UA (mg/dl) was measured during early hospitalization. Patient survival status was followed up after discharge (median follow-up: 330 days). The independent association of UA level with all-cause mortality was analyzed using Cox regression analysis. Results: During follow-up 165 (29.5%) deaths were identified. Patients with UA levels above the median value (≥ 7.7 mg/dl) exhibited higher mortality rates (21.1 vs. 37.9%; p b 0.001). In multivariable analysis, after adjusting for recognized prognostic factors and potential confounders, UA ≥ 7.7 mg/dl and per change in 1 mg/dl of UA was associated with an increased risk of mortality (HR 1.45, CI 95% = 1.03–2.44; p = 0.03 and HR 1.08, CI 95% = 1.01–1.15; p = 0.03, respectively). Conclusion: UA serum levels is an independent predictor of all-cause mortality in an unselected patients admitted with AHF. © 2008 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved. Keywords: Acute heart failure; Mortality; Uric acid

1. Introduction Epidemiological studies have found an association between elevated serum uric acid (UA) with increased vascular event rate and mortality in patients with hypertension, diabetes and prior cardiovascular disease (CVD) [1–4]. Increased UA levels are also common in chronic heart failure (CHF), [5–7] with an increased prevalence as diseases progress. Even though the exact mechanism for this association is unknown, it has been postulated that hyperuricemia in heart failure obeys among others to: 1) a diminished renal function clearance, either due to renal function impairment and/or diuretics treatment, and 2) a ☆

This study was supported by Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III, RED HERACLES RD06/0009/1001 (Madrid, Spain). ⁎ Corresponding author. Avda. Blasco Ibáñez 17, CP 46010 Valencia, Spain. E-mail address: [email protected] (J. Núñez).

functional up-regulation of the xanthine oxidase (XO). This is a key enzyme in purine metabolism, and a large contributor to the generation of oxygen free radicals which ultimate leads to increased oxidative stress. Consequently, hyperuricemia has been associated with inflammation [5,8], impaired oxidative metabolism [6], vascular and endothelial dysfunction [9,10] and exercise intolerance [5] in CHF, conditions that have been implicated as the most likely explanation for the association found between elevated UA levels and poor prognosis in mild to moderate [11] as well as in moderate to severe CHF [12,13]. Nevertheless, in the setting of an acute heart failure (AHF), the prognostic role of increased UA has not been conclusively determined yet. Therefore, the objectives of this study were twofold: First, to determine the ability of UA serum levels to predict all-cause mortality in patients admitted for AHF, and second, to explore the clinical determinants of UA level in this type of patients.

0953-6205/$ - see front matter © 2008 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.ejim.2008.04.007

A.L. Alimonda et al. / European Journal of Internal Medicine 20 (2009) 74–79

75

2. Methods

Table 1 Baseline characteristics according uric acid median serum values (I)

2.1. Study population

Variables

We studied a cohort of 670 patients, consecutively admitted to our cardiology department for AHF from 1st January 2003 to 1st January 2006. AHF was diagnosed by trained cardiologists following current guidelines [14]. Patients with diagnosis of acute coronary syndrome, clinical evidence of cancer, severe hepatic disease, end-stage renal disease under dialysis were excluded from this registry. Additionally, patients taken pharmacological treatment for hyperuricemia were also excluded from this analysis (110 patients). Demographic data, medical history, vital signs, 12lead electrocardiogram and laboratory tests was routinely gathered. UA was collected in blood samples during the first admission days. A comprehensive two-dimensional echocardiogram (Agilent Sonos 5500-Phillips) coupled with Doppler flow studies was performed during hospitalization to determine underlying HF etiology, evaluate left ventricular systolic function, and quantify heart chamber dimensions. Serum levels of UA and other lab measurements were obtained during early hospitalization (48 ± 2 h) using a commercially available immunoassay kit (Roche Elecsys). Variables routinely collected in this registry are listed in Tables 1 and 2. Individual diagnostic techniques and/or treatments were conducted following established guidelines [14]. All patients received intravenous furosemide during the early course of hospitalization; further pharmacological treatment was individualized following the clinical criteria of the cardiologist in charge of the patients. The Charlson Index of Comorbidity [15] was used to evaluate the impact of concurrent disease on subsequent mortality using clinical information obtained from chart review. The index weights were used from the following disease states: myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic pulmonary disease, connective tissue disease, ulcer disease, liver disease, diabetes, hemiplegia, moderate or severe renal disease, diabetes with end-organ damage, any tumor, leukemia, lymphoma, metastatic solid tumor, and acquired immunodeficiency syndrome. Higher index scores on this scale correspond to more comorbid disease states. This study was approved by the local ethics committee in our institution. 2.2. Ascertainment of study endpoints Death from any cause was selected as the main endpoint and ascertained either during hospitalizations, by routine out-patient clinic visits or by phone contact with patient/family. Patient's follow-up was censored by having had valve replacement surgery or cardiac transplantation or lost to follow-up. 2.3. Statistical analysis Continuous variables are expressed as mean ± 1 standard deviation (SD) or median ± interquartile range when appropriate. Discrete variables are presented as percentages. Baseline characteristics were compared between median of UA levels. Univariate comparisons of the two groups were made using the

Uric acid ≤ 7.7 mg/dl (n = 280)

Uric acid N 7.7 mg/dl (n = 280)

p-value

72.9 ± 10.5 63.6 75 33.9 7.5 42.1 8.9

74.3 ± 10.3 46.8 75.4 31.4 11.8 35 20.4

0.107 b0.001 0.922 0.528 0.086 0.083 b0.001

34.6 26.4

40.4 26.1

0.163 0.923

64.3 26.8 8.2 0.7 33.2

68.9 21.8 7.5 1.8 45.4

0.244 0.168 0.753 0.450 0.349

Demographic and medical history Age, years a Female, % Hypertension, % Dyslipemia, % Current smoker, % Diabetes mellitus, % Charlson index N 2, % Etiology – Ischemic heart disease, % – Valvular heart disease, % Type of AHF – ADHF, % – Acute pulmonary oedema, % – Hypertensive–AHF, % – Shock, % Previous hospitalization for AHF, % Baseline NYHA Class III/IV, % Radiological pleural effusion, % Peripheral oedema, %

13.6 37.5

27.5 49.6

b0.001 0.004

51.1

59.3

0.051

Previous treatment Diuretics, % Betablockers, % ACE inhibitors, % Statins, %

53.6 16.8 41.1 23.3

73.2 19.5 42.9 19.1

b0.001 0.109 0.669 0.222

106 ± 30 156 ± 36

98 ± 28 146 ± 36

88 ± 21

81 ± 19

Vital signs Heart rate, bpm a Systolic blood pressure, mm Hg a Diastolic blood pressure, mm Hg a

0.002

Data are presented as the median ± interquartile range, unless otherwise specified; categorical variables as percentages. AHF: acute heart failure; ADHF: acute decompensate heart failure; NYHA: New York Heart Association; ACE: angiotensin converting enzyme. a Variable presented as mean ± SD.

two-sample t-test for continuous variables (or U Mann– Whitney test when the assumption of normality did not hold) and the χ2 test was used for discrete variables. Cumulative mortality rates for the population above and below the median were depicted with the Kaplan–Meier curve; differences between values were tested by the Log rank test. The independent association between UA and mortality was assessed with Cox proportional regression. UA was modelled as a both a continuous variable, and later a binary variable using the median value (7.7 mg/dl) as the cutpoint. Candidate covariates included in the multivariate analysis were those associated with mortality in a bivariate analysis (from all variables included in the registry), as well as those recognized as prognostic factors based on previous medical knowledge. From this initial model, a parsimonious, highly predictive model was derived by using backward stepdown selection.

76

A.L. Alimonda et al. / European Journal of Internal Medicine 20 (2009) 74–79

Table 2 Baseline characteristics according uric acid median serum values (II) Variables

Uric acid ≤ 7.7 mg/dl (n = 280)

Uric acid N 7.7 mg/dl (n = 280)

p-value

ECG Atrial fibrillation, % QRS N 120 ms, %

41.3 31.4

41.3 43.6

Laboratory Hemoglobin, g/dl a Serum creatinine, mg/dl GFR, ml/min/m2 Sodium, mEq/l WBC count, 103/ml Troponin I N 0.2 ng/ml, %

12.9 ± 1.8 1 (0.4) 64.1 (24.2) 139 (5) 9.5 (4.9) 23.5

12.8 ± 1.9 1.3 (0.6) 50.6 (25.5) 140 (5) 9.9 (4.8) 29.4

Echocardiography LVEF ≤ 50% LAD a, mm LVDD a, mm

43.6 43 ± 8 55 ± 10

45.7 45 ± 8 56 ± 10

0.610 0.037 0.212

Treatment Diuretics, % Betablockers, % ACE inhibitors, % ATII antagonists, % Statins, %

98.6 42.1 43.2 31.1 32.7

100 47.3 47.7 28.6 34.4

0.101 0.217 0.288 0.345 0.681

1 0.003

0.522 b0.001 b0.001 0.565 0.369 0.111

Data are presented as the median ± interquartile range, unless otherwise specified; categorical variables as percentages. GFR: glomerular filtration rate estimated by modification in diet formula; WBC: white blood cells; LVEF: left ventricular ejection fraction; LAD: left atrial diameter; LVDD: left ventricular diastolic diameter; ACE: angiotensin converting enzyme; ATII: angiotensin II. a Variable presented as mean ± SD.

Final multivarite model included the following covariates: age (years), gender, previous admission for acute HF, baseline NYHA class III/IV, valvular etiology; systolic blood pressure

(mm Hg), previous treatment with diuretics, Charlson index, ejection fraction b 50%, glomerular filtration rate ≤ 30 ml/min/ 1.73 m2 [16], serum sodium (mEq/L) hemoglobin (g/dl) and UA (exposure variable). The proportionality assumption for the hazard function over time was tested by means of the Schoenfeld residuals. The functional form of continuous variables in the log-hazard scale was examined by means of fractional polynomials and transformed when appropriate. The predictive ability of the Cox model was assessed by estimating the Harrell C statistics. As a sensitivity analysis, a Generalized Additive model (GAM) with Cox extension was used to support the use of the median cutpoint when dichotomizing UA serum levels. GAM uses a non-parametric algorithm with cubic smoothing splines in depicting the threshold by which UA can be stratified into two categories, above or below the point in which the risk for mortality was deemed higher or lowers, respectively [17,18]. Multiple linear regression analysis was used to select those variables that were significantly associated with UA serum levels. A parsimonious, although highly predictive model, was obtained by using a backward stepdown selection. Variables retained in the final model were ranked based on the magnitude of change in the R2. A 2-sided p-value of b 0.05 was considered to be statistically significant for all analyses. All statistical analyses were performed using STATA 9.2. 3. Results 3.1. Baseline characteristics and UA The median age of our study population was 73 ± 10 year; 55.2% were female and 53.4% exhibited preserved ejection fraction (LVEF N 50%). Table 1 shows the clinical characteristics of the study population according to UA median value (≥7.7 mg/

Fig. 1. Cumulative all-cause mortality risk stratified by uric acid median value. ⁎number of patients at risk and events at different moments of follow-up.

A.L. Alimonda et al. / European Journal of Internal Medicine 20 (2009) 74–79

dl). Patients with UA levels above to the median were more likely to be male and displayed higher proportion of NYHA class III/IV, Charlson index N 2, radiological pleural effusion, previous treatment with diuretics, wide electrocardiographic QRS complex, and a higher left atrial dimensions. Patients were also like to have a lower heart rate, systolic, diastolic blood pressure and estimated glomerular filtration rate. No other significant differences were found. In summary, patients with UA above median exhibited a higher risk clinical profile. 3.2. Variables associated with UA serum levels

77

Table 3 Hazard ratios for all-cause mortality according to UA serum levels Cox regression models

HR (95% CI)

P-value

Harrell's C statistics

Unadjusted hazard ratios UA, mg/dl (per 1-unit increase) UA N 7.7 mg/dla

1.18 (1.12–1.25) 2.03 (1.48–2.8)

b0.001 b0.001

0.639 0.611

Adjusted hazard ratiosb UA, mg/dl (per 1-unit increases) UA N 7.7 mg/dla

1.08 (1.01–1.15) 1.45 (1.03–2.05)

0.030 0.032

0.784 0.784

UA: uric acid; HR: hazard ratio; CI: confidence interval. UA N 7.7 mg/dl vs. UA ≤ 7 mg/dl. b Cox model adjusted by age (years), gender, previous admission for acute HF, baseline NYHA class III/IV, valvular etiology; systolic blood pressure (mm Hg), previous treatment with diuretics, Charlson index, ejection fraction b 50%, glomerular filtration rate ≤ 30 ml/min/1.73 m2, serum sodium (mEq/L) and hemoglobin (g/dl). a

In a multiple regression analysis, the following variables were significantly associated with mean of UA [presented with their respective β coefficients (standard errors) and p-values)]: male gender [0.910 (0.190); p b 0.001], prior treatment with diuretics [0.480 (0.201); p b 0.017], white blood cells count (×103 cells/ml) [0.004 (0.001); p = 0.049], age (years) [− 0.201 (0.100); p = 0.039), diabetes mellitus [−0.521 (0.191); p = 0.007], systolic blood pressure on admission (mm Hg)[(− 0.009 (0.002); p = 0.001] and estimated glomerular filtration rate (ml/min/mt2) [−0.05 (0.004); p b 0.001]. The estimated glomerular filtration rate ranked as the most important factor in the model (accounting for 67% of the total R2), followed by gender (18%) and systolic blood pressure on admission (7%). The rest of covariates accounted for the remaining 8% of the total model predictability. 3.3. UA and mortality After a median follow-up of 330 days (interquartile range = 120–600 days) a total of 165 deaths were identified

(29.5%). In unadjusted analyses, UA levels above median were strongly associated with all-cause mortality (Table 2), and showed a marked increase in the cumulative incidence of mortality (21.1% vs. 37.9%; p b 0.001) with established differences observed since the first months of follow-up, as shown in Fig. 1. In a sensitivity analysis, the best adjusted prognostic cutpoint depicted in the GAM plot (Fig. 1) and used to dichotomize the continuum of UA was similar to the median value of the study population (7.7 mg/dl), endorsing the use of this cutpoint for prognostic purposes (Fig. 2). In multivariate analyses, UA as continuous (per increments in 1 mg/dl) and above median, were independently associated with all-cause mortality during the follow-up with a HR = 1.08 (95% CI 1.01– 1.15; p = 0.030) and HR = 1.45 (95% CI 1.03–2.05; p = 0.032) respectively, after adjusting for well known prognostic covariates and potential confounders (Table 3). All potential and reasonable interactions between UA and candidates covariates were tested. Of note, no significant interaction was observed between UA (as continuous) and: estimated glomerular filtration rate ≤30 ml/min/1.73 m2 (p = 0.831), previous diuretic treatment (p = 0.786), left ventricular ejection fraction ≤ 50% (p = 0.546), and gender (p = 0.804). The C statistics of the final Cox models were 0.785 and 0.766 with and without including UA. 4. Discussion Similar to previous reports in HF patients, we found a positive and statistically significant association between elevated UA serum levels and mortality in HF patients. Our study extends this finding to an acute setting and an unselected population with HF.

Fig. 2. Optimal uric acid (UA) threshold suggested by the generalized additive model (GAM) plot. The functional form in the association between UA and the hazard rate for all-cause mortality. This association is adjusted simultaneously by age, gender, diabetes, stable New York Heart Association class III/IV, aetiology of valvular heart disease, systolic blood pressure, estimated glomerular filtration rate and haemoglobin. The dotted curves indicate approximate 95% CIs for the smoothed hazard. The arrow indicates the point in the continuum of UA that crosses the threshold between high and low risk for mortality (around 7.7 mg/dl).

4.1. Pathophysiology Serum concentration of UA is determined by a balance of production and excretion. Hyperuricemia in patients with HF, particularly in the acute setting, has been attributed to several underlying mechanisms including: a) decrease renal excretion:

78

A.L. Alimonda et al. / European Journal of Internal Medicine 20 (2009) 74–79

renal function impairment, diminish renal secretion due to diuretic treatment and competence with lactic acid at the renal proximal tubuli [19]; b) increase production: up-regulation of XO activity in patients with HF [20,21], increased conversion of xanthine dehydrogenases to XO or a higher abundance of purines due to an enhanced ATP breakdown into adenosine and hypoxanthine. During the enzymatic synthesis of UA from purines, XO also contributes to a large degree to the formation of reactive oxygen species [22]. The resulting oxidative stress plays an important role in endothelial dysfunction in CVD via an inactivation of endothelial nitric oxide [9,10]. This results in an attenuation of the beneficial effects of nitric oxide on e.g. endogenous vasodilation and platelet aggregation [23]. Among other effects linked to increased XO activity are: insulin resistance [24], tissue hypoxia [24] and inflammatory cytokine activation [5,6]. 4.2. Uric acid in heart failure — previous studies 4.2.1. Chronic heart failure There are several studies investigating the prognostic value of elevated serum UA in patients with CHF. Anker et al. [12] reported that high serum UA levels predicted mortality and emphasized the need for heart transplantation in patients with CHF. Serum concentrations of UA added important prognostic information alone and when combined with the assessment of cardiac function (left ventricular ejection fraction) and patient functional status (maximal oxygen consumption with exercise). In a recent study from by Sakai et al. [13] 150 patients with CHF were followed for a 3-year-period. High plasma UA levels were important predictors of mortality independent of traditional prognostic factors including glomerular filtration rate or brain natriuretic peptide (BNP). Interestingly, this study also demonstrated that UA is produced by the failing myocardium itself and that myocardial UA production increased with the severity of CHF. In a study in 119 consecutive patients with mild CHF, defined as a NYHA class of I–III, elevated serum levels, UA were strongly and independently associated with poor outcome correlated with high sensitive C-reactive protein levels and predicted intolerance to exercise [11]. 4.2.2. Acute heart failure In an early study performed in a small population, UA serum levels first appeared to have prognostic values in acute illnesses [25]. As recently reported by Cengel et al. 85 patients admitted with decompensate heart failure in NYHA functional class IV, serum UA levels and female gender were the only predictors of in-hospital mortality [26]. In this setting, our findings corroborate these preliminary reports and is the first who extend the prognostic value to: 1) a large non non-selected population with AHF (patients with similar baseline characteristics as those seen in large AHF registries, including a high proportion of females and preserved left ventricular ejection fraction and; 2) long-term follow-up. It is noteworthy that our results are adjusted by classical covariates known to have prognostic value in acute setting, and no significant interactions were found with gender, left ventricular ejection fraction, severe renal function impairment and previous diuretic treatment.

We reported a risk of intermediate magnitude in contrast with most studies that show higher UA serum levels attributable risk for mortality than our results. We speculate that these differences may be explained because: 1) population differs widely: most of the patients included in previous studies exhibited systolic dysfunction, advanced NYHA class, ischemic etiology and male gender; and 2) a thoroughly multivariate adjustment that has been performed in this study was introducing covariates as glomerular filtration rate and gender that have shown to be the most important predictors of UA levels in this setting. 4.3. Uric acid and xanthine oxidase as a therapeutic target Due to the deleterious effects of elevated serum UA levels and elevated XO activity on the cardiovascular system, the therapeutic approach of lowering uric UA levels (by uricosuric drugs) or inhibiting XO activity (via XO inhibitors) appears to be an appealing therapeutic strategy in HF. Indeed, XO inhibition in HF reduces myocardial oxygen consumption, improves endothelial function [27] and peripheral blood flow [28] and reduces the serum concentration of traditional HF markers like BNP [29]. In patients with HF due to idiopathic dilated cardiomyopathy XO, inhibition with intracoronary allopurinol improved the energetic efficiency of the failing myocardium. In a retrospective study it was shown that high dose allopurinol was associated with a lower all-cause mortality than low dose allopurinol in patients with CHF [30]. Recently, an abstract report of OPT-CHF clinical trial, showed that no prognostic benefit was observed when treatment with oxypurinol was administered, nevertheless, in a post hoc analysis a prognostic benefit was observed in the subgroup of patients with serum UA N 9.5 mg/dl [31]. 4.4. Limitations Some limitations need to be addressed: 1) inherent limitations of observational studies; 2) the lack of availability of the measurements of BNP in our registry does not permit a comparison of diagnostic accuracy between these two biomarkers and; 2) despite excluding from this analysis patients with previous or admission prescribed hypouricemiant drugs, we may not asses if during follow-up patients were to receive pharmacological therapy to reduce UA serum levels with it possible repercussion in prognosis. 4.5. Conclusions This is the first study to demonstrate the independent prognostic value of UA for predicting increased long-term mortality after an index admission for AHF. This study corroborates with previous findings in CHF, and extrapolates those results to an unselected sample of patients with AHF. Since UA measurements is easy, inexpensive and widely available, we advocate for future investigations in this field to establish the definitive role of UA as a potential biomarker for risk stratification and possible therapeutic target in AHF setting.

A.L. Alimonda et al. / European Journal of Internal Medicine 20 (2009) 74–79

Learning points • Hyperuricaemia is a common finding in patients with acute failure. • High UA serum levels have shown to be related with longterm mortality in this setting, despite adjusting for traditional risk factors, corroborating previous findings in chronic heart failure. • Future studies are needed in this field to establish the definitive role of UA as a potential biomarker for risk stratification and possible therapeutic target in AHF setting. References [1] Fang J, Alderman MH. Serum uric acid and cardiovascular mortality the NHANES I epidemiologic follow-up study, 1971–1992. National Health and Nutrition Examination Survey. JAMA 2000;283:2404–10. [2] 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. [3] Baker JF, Krishnan E, Chen L, Schumacher HR. Serum uric acid and cardiovascular disease: recent developments, and where do they leave us? Am J Med 2005;118:816–26. [4] Niskanen LK, Laaksonen DE, Nyyssonen K, Alfthan G, Lakka HM, Lakka TA, et al. Uric acid level as a risk factor for cardiovascular and all-cause mortality in middle-aged men: A prospective cohort study. Arch Intern Med 2004;164:1546–51. [5] Leyva F, Anker SD, Godsland IF, Teixeira M, Hellewell PG, Kox WJ, et al. Uric acid in chronic heart failure: a marker of chronic inflammation. Eur Heart J 1998;19:1814–22. [6] Leyva F, Anker S, Swan JW, Godsland IF, Wingrove CS, Chua TP, et al. Serum uric acid as an index of impaired oxidative metabolism in chronic heart failure. Eur Heart J 1997;18:858–65. [7] Hoeper MM, Hohlfeld JM, Fabel H. Hyperuricaemia in patients with right or left heart failure. Eur Respir J 1999;13:682–5. [8] Olexa P, Olexova M, Gonsorcik J, Tkac I, Kisel'ova J, Olejnikova M. Uric acid—a marker for systemic inflammatory response in patients with congestive heart failure? Wien Klin Wochenschr 2002;114:211–5. [9] 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. [10] Kanellis J, Kang DH. Uric acid as a mediator of endothelial dysfunction, inflammation, and vascular disease. Semin Nephrol 2005;25:39–42. [11] Jankowska EA, Ponikowska B, Majda J, Zymlinski R, Trzaska M, Reczuch K, et al. Hyperuricaemia predicts poor outcome in patients with mild to moderate chronic heart failure. Int J Cardiol 2007;115:151–5. [12] Anker SD, Doehner W, Rauchhaus M, Sharma R, Francis D, Knosalla C, et al. Uric acid and survival in chronic heart failure: Validation and application in metabolic, functional, and hemodynamic staging. Circulation 2003;107:1991–7. [13] Sakai H, Tsutamoto T, Tsutsui T, Tanaka T, Ishikawa C, Horie M. Serum level of uric acid, partly secreted from the failing heart, is a prognostic marker in patients with congestive heart failure. Circ J 2006;70:1006–11.

79

[14] Nieminen MS, Bohm M, Cowie MR, Drexler H, Filippatos GS, Jondeau G. Executive summary of the guidelines on the diagnosis and treatment of acute heart failure: the Task Force on Acute Heart Failure of the European Society of Cardiology. Eur Heart J 2005;26:384–416. [15] Romano PS, Roos LL, Jollis JG. Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: Differing perspectives. J Clin Epidemiol 1993;46:1075–9. [16] Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: A new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999;130:461–70. [17] Royston P, Sauerbrei W. A new measure of prognostic separation in survival data. Stat Med 2004;23:723–48. [18] Royston P, Altman DG, Sauerbrei W. Dichotomizing continuous predictors in multiple regression: a bad idea. Stat Med 2006;25:127–41. [19] Hare JM, Johnson RJ. Uric acid predicts clinical outcomes in heart failure: Insights regarding the role of xanthine oxidase and uric acid in disease pathophysiology. Circulation 2003;107:1951–3. [20] Cappola TP, Kass DA, Nelson GS, Berger RD, Rosas GO, Kobeissi ZA, et al. Allopurinol improves myocardial efficiency in patients with idiopathic dilated cardiomyopathy. Circulation 2001;104:2407–11. [21] Landmesser U, Spiekermann S, Dikalov S, Tatge H, Wilke R, Kohler C, et al. Vascular oxidative stress and endothelial dysfunction in patients with chronic heart failure: role of xanthine-oxidase and extracellular superoxide dismutase. Circulation 2002;106:3073–8. [22] Cai H, Harrison DG. Endothelial dysfunction in cardiovascular diseases: The role of oxidant stress. Circ Res 2000;87:840–4. [23] Dawson J, Walters M. Uric acid and xanthine oxidase: future therapeutic targets in the prevention of cardiovascular disease? Br J Clin Pharmacol 2006. [24] Leyva F, Chua TP, Anker SD, Coats AJ. Uric acid in chronic heart failure: A measure of the anaerobic threshold. Metabolism 1998;47:1156–9. [25] Woolliscroft JO, Colfer H, Fox IH. Hyperuricemia in acute illness: A poor prognostic sign. Am J Med 1982;72:58–62. [26] Cengel A, Turkoglu S, Turfan M, Boyaci B. Serum uric acid levels as a predictor of in-hospital death in patients hospitalized for decompensated heart failure. Acta Cardiol 2005;60:489–92. [27] Farquharson CA, Butler R, Hill A, Belch JJ, Struthers AD. Allopurinol improves endothelial dysfunction in chronic heart failure. Circulation 2002;106:221–6. [28] Doehner W, Schoene N, Rauchhaus M, Leyva-Leon F, Pavitt DV, Reaveley DA, et al. Effects of xanthine oxidase inhibition with allopurinol on endothelial function and peripheral blood flow in hyperuricemic patients with chronic heart failure: Results from 2 placebo-controlled studies. Circulation 2002;105:2619–24. [29] Gavin AD, Struthers AD. Allopurinol reduces B-type natriuretic peptide concentrations and haemoglobin but does not alter exercise capacity in chronic heart failure. Heart 2005;91:707–9. [30] Struthers AD, Donnan PT, Lindsay P, McNaughton D, Broomhall J, MacDonald TM. Effect of allopurinol on mortality and hospitalisations in chronic heart failure: a retrospective cohort study. Heart 2002;87:229–34. [31] Hare JM. Efficacy and safety study of oxypurinol added to standard therapy in patients with NYHA class 3–4 CHF (OPT-CHF). heart failure of America 2006 Scientific Meeting; Septemebr 13 2006. Seattle, WA.