O R I G I N A L P A P E R

Serum Uric Acid Levels Correlate With Filling Pressures in SystolicHeart Failure

H eart failure (HF) is a generalhealth problem with an increas-

ing prevalence. It is characterized byimpaired organ perfusion and elevatedventricular filling pressure. As a systemicdisease, HF has different effects on mul-tiple organs such as the kidney and liver,resulting in their malfunctioning, withtypical disturbances in kidney and liverfunction tests.

Uric acid (UA) is a marker of oxida-tive stress and myocardial damage, andincreasing serum levels can be a mani-festation of toxic myocardium.1,2 It hasa pathophysiologic significance in theprogression of HF3 and is predictive ofhigh New York Heart Association(NYHA) class and is independentlyassociated with poor prognosis andincreased mortality.4,5 The inhibition ofxanthine oxidase by oxypurinol reducesUA levels and improves left ventricular(LV) ejection fraction (LVEF).6 Arestrictive LV filling pattern is thoughtto be associated with elevated UA lev-els, and it correlates significantly withdiastolic abnormality and right atrialpressure (RAP) and mean pulmonaryarterial pressure (PAP).6,7 A goodknowledge of the associated clinical andhemodynamic problems is important inthe determination of disease severityand prognosis.

In this study, we sought to assess therelationship between UA serum levelsand liver function test abnormalitiesand clinical and hemodynamic profilesin patients with chronic systolic HF(LVEF<35% by echocardiography).

Methods and MaterialsPatient Selection. This single-centercase series randomly enrolled 50 HFpatients with an LVEF <35% who werereferred to the Heart Failure Clinic of

Rajaee Cardiovascular, Medical andResearch Center between December2009 and June 2010.

For all the patients, a thorough historywas obtained, followed by a completephysical examination. Additionally, clin-ical signs of HF such as rales, edema, andparoxysmal nocturnal dyspnea, as well asstandard NYHA functional class weredetermined.

Exclusion criteria comprised uncon-trolled diabetes, known previous liverdisease (chronic viral ⁄autoimmune ⁄drug-related hepatitis, liver malignancy,and known biliary tract disease), signifi-cant congenital heart disease, significantrheumatismal heart disease, known

hematologic disease other than anemiaof chronic disease (hemoglobinopa-thies and hemolytic states), continuousconsumption of xanthine oxidase inhibi-tor, and a history of gout, kidney stone,and renal dysfunction (creatinine >2or glomerular filtration rate [GFR]<30 mL ⁄min ⁄m2).

Fifty age- and sex-matched controlswere enrolled from the referrals for aroutine check-up who had nonanginalchest pain and normal echocardiography(LVEF>55%).

Right Heart Catheterization. Thepatients were evaluated by standard rightheart catheterization using the Edwards

The authors studied the relationship between liver function tests and serum uric acid levelwith clinical and hemodynamic profiles in heart failure. Fifty patients (aged 44�15 years;74.5% men) with an ejection fraction (EF)<35% were enrolled and clinical assessmentwas performed. Hemodynamic indices (including pulmonary arterial pressure [PAP], pul-monary capillary wedge pressure [PCWP], and cardiac index were studied by standardEdwards Lifesciences Swan-Ganz catheters, and liver function tests and serum uric acidlevel were measured simultaneously. Fifty age- and sex-matched controls with normal EFwere also studied. A total of 73% of patients had ischemic cardiomyopathy. Mean uricacid level was 7.2�3.8 mg ⁄dL and was significantly higher than in the control group(P value<.001). In multivariate analysis, uric acid correlated significantly with PAP (r=.5,P<.001) and PCWP (r=.4, P=.002) and was also associated with clinical signs of rales,edema, paroxysmal nocturnal dyspnea (r=.5, P=.01), and New York Heart Associationclass (r=.4, P=.005). Uric acid level was also correlated inversely with left ventricular EF(r=.27, P=.006). Elevated uric acid levels in patients with systolic heart failure isassociated with impaired clinical and hemodynamic profile and might be used as anoninvasive indicator of elevated left ventricular filling pressures. Congest Heart Fail.2011;17:79–83. �2011 Wiley Periodicals, Inc.

Ahmad Amin, MD; Farveh Vakilian, MD; Majid Maleki, MD, FACCFrom the Rajaee Cardiovascular, Medical & Research Center, Heart Failure &Transplantation, Tehran, Iran

Address for correspondence:Farveh Vakilian, MD, Rajaee Heart Center, Vali Asr street, Nyayesh Cross,Tehran 12345, IranE-mail: dr.vakilian@yahoo.comManuscript received December 11, 2010; accepted December 20, 2010

doi: 10.1111/j.1751-7133.2010.00205.x

filling pressures in systolic heart failure march • april 2011 79

Lifescience Swan-Ganz catheter (Irvine,CA) placed in their right pulmonaryarteries. Pressures were read through afluid-filled system, including RAP, pul-monary arterial pressure (PAP), rightventricular pressure (RVP), and pulmo-nary capillary wedge pressure (PCWP).Cardiac output and cardiac index weredetermined via the standard ther-modilution method. PAP and PCWPwere measured in 3 cycles at end-expira-tion and then averaged. Pulmonaryvascular resistance was calculated asmean PAP ) PCWP ⁄cardiac output.The patients received no intravenousmedication during catheterization.

Laboratory Data. Blood samples weredrawn at the time of catheterizationfrom the subclavian vein, and the fol-lowing indices were measured via thePraetorius and Poulson method8 (KimiaPazhouhan kit, Iran): aspartate amino-transferase (AST), alanine aminotrans-ferase (ALT), bilirubin (total anddirect), alkaline phosphatase (ALP),and serum UA.

According to the references and labo-ratory data, the normal cutpoints forliver function tests were as follows: ASTand ALT: 1–25 mg ⁄dL, ALP: 0–140mg ⁄dL, total bilirubin: 0–1.5 mg ⁄dL,and direct bilirubin: 0–0.5 mg ⁄dL. Asfor serum UA level, different valueshave been proposed as a normal cut-point, eg, 6 mg ⁄dL, 7 mg ⁄dL, and8 mg ⁄dL due to the dependence of UAon body surface area, physical activity,diet, and renal function.9 In order tohave a reliable comparison, a controlgroup of 50 persons with normal cardiacfunction, matched for sex, age, diet,habits, and race, was studied for serumUA levels. The mean UA level in thecontrol group was 4.26�0.9 mg ⁄dL.The normal values for right heart cathe-terization were 25 ⁄9 mm Hg for PAP,24 ⁄0 mm Hg to 4 mm Hg for RVP,5 mm Hg for RAP, and 8 mm Hg to15 mm Hg for PCWP.10

Statistical Analysis. The data are pre-sented as mean � standard deviationwhen normally distributed and as fre-quencies and percentages for categoric

variables. Differences between baselinevariables were evaluated via the Studentt test. The Kruskal–Wallis test wasapplied for evaluation of correlationbetween NYHA functional class andliver function test items. The WhitneyU test was used for investigating therelationship between clinical variables(rales, edema, and paroxysmal nocturnaldyspnea) with mean serum UA level.Finally, the relationship between hemo-dynamic profile and liver function testitems (including serum UA) was ana-lyzed by initial linear regression analysis.A P value <.05 was considered signifi-cant. Statistical analyses were performedusing SPSS 15 (SPSS, Inc, Chicago,IL).

ResultsTable I depicts the baseline characteris-tics of the patient population, who hada mean age of 44�15 years. Menaccounted for 74.5% of the study popu-lation, and 73% had ischemic cardiomy-opathy. The respective mean systolicPAP and diastolic PAP was40�16.5 mm Hg and 20�9.5 mm Hg,the respective mean RAP and PCWPwas 9�4.6 mm Hg and 18�9.5 mm

Hg, and the mean cardiac output was3.8�0.9 L ⁄m2 ⁄min.

Assuming the above-mentioned nor-mal values,10 72.2% of the patients hada high systolic PAP, 92.5% had a highdiastolic PAP, 66.6% had a high RAP,83.3% had a high mean PAP, and57.4% had a high PCWP. The meanUA serum level was 7.2�3.8 mg ⁄dL,which was significantly higher than thatin the control group (P<.001).

Considering the standard NYHAclassification, 34% of the patients werein class IV, 50% in class III, and 15.5%in class II. Additionally, 51.5% of thestudy population had rales, 68.2%edema, and 66.4% paroxysmal noctur-nal dyspnea according to their physicalexamination. The mean values of UAlevel were 8.2 mg ⁄dL, 7.4 mg ⁄dL, and5.9 mg ⁄dL in NYHA class IV, III, andII patients, respectively. The mean UAlevel was 8�2.2 mg ⁄dL in patients withrales, 6.8�2.3 mg ⁄dL in those withedema, and 7.9�2.2 mg ⁄dL in thosewith paroxysmal nocturnal dyspnea.These values were significantly higherthan those in patients without thesesigns (P=.01) (Table II). All patientsreceived oral furosemide mostly at daily

Table I. Baseline Characteristics According to Different Hemodynamic and ClinicalProfiles

MEAN MINIMUM MAXIMUM

Age, y 44�15.22 11 72Sex, female ⁄ male 13 ⁄ 38 25.5 ⁄ 74.5PAP (systolic) 40.18�16.5 15 84PAP (diastolic) 20.7�16.5 6 40PCWP 18.41�9.5 4 40RAP 9.33�4.6 3 24RVP 40.6�15.9 15 84PVR 5.7�5.7 0.4 26CO 3.88�0.92 2 6.3CI 2.39�0.6 1.3 3.9AST 42.39�72.06 9 435ALT 44.8 � 66.6 5 440ALP 99.3�47.9 38 286TBILI 2.17�1.27 0.8 6.5DIRBILI 0.67�0.4 0.1 2Uric acid 7.06�2.2 3 12Control group (uric acid) 4.26�0.96 3.1 6.4DCM, No. (%) 14 (26.4)ICM, No. (%) 36 (73.6)

Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferases; AST, aspartateaminotransferase; CI, cardiac index; CO, cardiac output; DCM, dilated cardiomyopathy;DIRBILI, direct bilirubin; ICM, ischemic cardiomyopathy; PAP, pulmonary arterial pressure;PCWP, pulmonary capillary wedge pressure; PVR, pulmonary valvular resistance; RAP, rightatrial pressure; RVP, right ventricular pressure; TBILI, total bilirubin.

filling pressures in systolic heart failure march • april 201180

doses of 40 mg to 80 mg; 21.3% of thepatients received>80 mg daily.

The mean LVEF was 19%�6.5% onechocardiography. In the evaluation ofright ventricular function with respectto the echocardiographic data, a tissueDoppler–derived systolic motion of theright ventricle at the lateral tricuspidring in the 4-chamber view (Sm) of<7 cm ⁄ s and an M mode–derived tri-cuspid annular plane systolic excursion(Tapse) of <10 mm were consideredsevere right ventricular systolic function,and Sm >10 cm ⁄ s and Tapse >17 mmwere regarded as normal right ventricu-lar systolic function.11 Furthermore,15% of the patients had severe rightventricular systolic function and 60%had mild to moderate systolic dysfunc-tion, with no significant difference interms of PAP, PCWP, and UA levelsbetween these groups.

Table III summarizes the result of thelinear regression analysis of the liverfunction tests and UA levels and pre-sents the hemodynamic profile of thestudy population.

In the multivariable analysis, a signifi-cant correlation was found between UAlevel and systolic pressure, diastolic pres-sure, and mean PAP (r=.4, .5, and .5,respectively; P <.001) and also rightventricular pressure (r=.4, P=.004).Interestingly, there was a significantcorrelation between serum UA leveland PCWP (r=.4, P =.002). Figure 1demonstrates the regression analysis of

the relationship between UA level andPCWP.

There was no significant correlationbetween cardiac output, cardiac index,RAP, pulmonary vascular resistance,and UA levels.

Regarding NYHA functional classand clinical signs, a significant correla-tion was found between UA levels andNYHA class (r=.4, P=.005) and thepresence of rales, edema, and paroxys-mal nocturnal dyspnea (r=.5, P=.01).UA level was also correlated inverselywith LVEF on echocardiography (r=).27, P=.006).

Table II. Prevalence of Different ClinicalSigns and Serum Uric Acid Levels

NUMBER PERCENT

URIC ACID

(MEAN) P VALUE

RalesYes 26 51.8 8.09�2.2 .01No 24 48.2 6.88�2.33

EdemaYes 34 68.2 7.9�2.2 .01No 16 31.8 6.69�2.3

PNDYes 33 66.4 7.9�2.2 .01No 17 33.6 6.69�2.3

NYHAII 8 15.5 5.9�2.1 .05III 25 50 7.4�2.1IV 17 34.5 8.2�2.3

Abbreviations: NYHA, New York HeartAssociation classification; PND,paroxysmal nocturnal dyspnea.

Table III. Univariate Regression Analysis for Different Liver Function Tests and Uric Acid in 50 Heart Failure Patientsa

AST ALT ALP TBILI DIRBILI URIC ACID

PAP (systolic) 0.11 ⁄ .44 0.08 ⁄ .57 0.04 ⁄ .79 0.45 ⁄ .003 0.34 ⁄ .02 0.44 ⁄ .002PAP (diastolic) 0.07 ⁄ .58 0.08 ⁄ .56 0.08 ⁄ .62 0.38 ⁄ .01 0.34 ⁄ .02 0.52 ⁄<.001PAP (mean) 0.59 ⁄ .52 0.44 ⁄ .58 0.27 ⁄ .68 0.04 ⁄ .004 0.01 ⁄ .02 0.09 ⁄<.001RAP )0.16 ⁄ .24 )0.13 ⁄ .35 )0.02 ⁄ .9 0.33 ⁄ .03 0.22 ⁄ .15 0.22 ⁄ .13RVP 0.10 ⁄ .45 0.07 ⁄ .60 0.03 ⁄ .82 0.43 ⁄ .004 0.33 ⁄ 0.03 0.41 ⁄ .004PCWP 0.05 ⁄ .72 0.09 ⁄ .49 0.04 ⁄ .79 0.36 ⁄ .01 0.32 ⁄ .03 0.43 ⁄ .002CI )0.07 ⁄ .58 0.15 ⁄ .28 )0.19 ⁄ .9 )0.3 ⁄ .05 )0.37 ⁄ .01 )0.22 ⁄ .13PVR )0.1 ⁄ .43 )0.1 ⁄ .47 0.07 ⁄ .67 0.18 ⁄ .26 0.29 ⁄ .06 0.13 ⁄ .36EF )4.4 ⁄ .15 )4.24 ⁄ .08 )2.3 ⁄ .01 )0.04 ⁄ .002 )0.01 ⁄ .001 )0.09 ⁄ .006

Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferases; AST, aspartate aminotransferase; CI, cardiac index; DCM,dilated cardiomyopathy; DIRBILI, direct bilirubin; EF, ejection fraction; PAP, pulmonary arterial pressure; PCWP, pulmonary capillary wedgepressure; PVR, pulmonary valvular resistance; RAP, right atrial pressure; RVP, right ventricular pressure; TBILI, total bilirubin. aValues arepresented as Pearson correlation ⁄ P value.

Figure 1. Regression analysis of significant correlation between pulmonary capillary wedgepressure (PCWP) and uric acid serum levels.

filling pressures in systolic heart failure march • april 2011 81

DiscussionWith the progression of systolic HF andthe concomitant decline in the maximalcardiac output generated for any givencardiac filling pressure (ie, flattening ofthe Frank Starling curve), the heart hasto increase its filling pressure at the costof worse pulmonary congestive symp-toms in order to achieve a reasonablestroke volume. The fact that manyparaclinical parameters are affected bythis disturbed hemodynamic state canexplain their correlation with congestivesigns and symptoms in patients with sys-tolic HF.

Hyperuricemia has been shown to bean independent predictor of all-causemortality in patients with systolic HF.3,7

Anker and colleagues,12 reporting thathyperuricemia predicted mortality, high-lighted the need for heart transplant inpatients with HF independent of renalfunction, serum sodium, serum urea,diuretic usage, and patient age.

To our knowledge, this is the firststudy of its kind to demonstrate the cor-relation between serum UA level withPAP and PCWP, which can also predictthe significant correlation of serum UAwith congestive signs and symptoms.

Pathophysiology. The presence of liverfunction abnormalities and UA eleva-tion in the context of HF has long beenrecognized. Some studies claim thatalmost 70% of HF patients have elevatedUA levels >6 mg ⁄dL.13,14 The VanDeuren and Jankowska studies showedthat serum UA elevation could predict ahigher NYHA class and a lower LVEF inmild to moderate HF.15,16

UA is the end point of the metabo-lism of purine compounds, produced inthe liver from the degradation of dietaryand endogenously synthesized purinecompounds via the xanthine oxidasereaction, which irreversibly oxidizesxanthase to UA.1 Xanthine oxidase isthe product of the proteolytic cleavageor sulfhydryl modification of a 150 kDaprotein encoded by xanthine oxidore-ductase gene (Fig. 2). Xanthine oxidaseis produced in the liver and gut and isalso expressed in cardiomyocytes.2 UAis eliminated in the gut (33%) and

kidneys (67%). Xanthine oxidase activ-ity participates in both mechani-coenergetic uncoupling and vasculardysfunction in failing myocardium.

Serum urate concentration is about6 mg ⁄dL in healthy adults. UA itselfmay have a role in cardiovascular andrenal metabolism.4 UA can be protec-tive itself and by promoting superoxidedismutase activity, it can promote vascu-lar smooth muscle cell proliferationthrough the stimulation of mitogen-acti-vated protein kinases, cyclooxygenase-2,and platelet-derived growth factor.13

Both UA and xanthine oxidase mayplay a role in nitric oxide signaling, andUA itself can impair nitric oxide pro-duction. Moreover, we know that nitricoxide plays a key role in modulating car-diac excitation-contraction coupling, sothis evidence suggests that xanthine oxi-dase and UA have a pathophysiologicrole in HF patients.1,2,4

Elevated UA levels are a marker ofinflammation, metabolic disturbances,oxidative stress (free oxygen radicals),vascular and endothelial dysfunction,and perhaps myocardial injury.1,3,4

Cachectic HF patients have higher UAlevels.17 Hyperuricemia can be a result ofrenal hypoperfusion and reduced urateexcretion (shown to be related to highrenin expression and high endothelinand aldosterone levels) and also long-term diuretic use. It can also be begotten

by local tissue ischemia or hypoxia. UAserum levels show xanthine oxidaseactivation that is upregulated in HF.3,8

Xanthine oxidase contributes to the for-mation of superoxide anions, which areable to inactivate nitric oxide and giverise to the formation of peroxynitrite.Peroxynitrite is a strong oxidant thatinhibits endothelium-dependent vasore-laxation, limits the favorable effects ofnitric oxide on platelet aggregation andvascular smooth muscle proliferation,and causes the oxidation of deoxynucleo-tide acid and lipids. There is an associa-tion between increased xanthine oxidaseactivity and insulin resistance, tissuehypoxia, and inflammatory cytokineactivation.2,4 It is thought that UA is abystander and not a player.18

Serum UA levels add important prog-nostic information alone and whencombined with the assessment of cardiacfunction and the patient’s functionalstatus. In recent studies, UA was animportant predictor of in-hospital andlong-term mortality independent ofGFR and LVEF or other factors.3,5,7

All of our patients received loopdiuretics with nearly the same doses asin other studies. The correlationbetween UA serum levels and hemody-namic parameters is independent ofdiuretic use. Consequently, it cannot bea significant confounding factor in thisregard.15 In addition, with regard to

Figure 2. Production and metabolism of uric acid in human body.

filling pressures in systolic heart failure march • april 201182

right ventricular systolic function, therewere few patients with severe rightventricular systolic function and therewas no significant difference in terms ofUA levels, PAP, and PCWP betweenthe different right ventricular functiongroups in our study. It is also noteworthythat patients with severe tricuspid regur-gitation were excluded from this study.

High UA levels correlate well directlywith increasing PAP, RVP, and PCWPand clinical congestive signs. There is astudy showing a correlation between B-type natriuretic peptide, C-reactive pro-tein, and UA levels with the hemody-namic profile19 and another studyrevealing the relationship between dia-stolic function parameters in echocardi-ography and UA.6 Therefore, measuringserum UA level in patients with systolicHF may be of prognostic benefit as a sur-rogate for LV filling pressures.

In the present study, we showed thatliver function test abnormalities andincreased serum UA levels are observedfrequently in patients with HF. More-over, severe HF symptoms in terms of

NYHA functional class and signs andlower LVEF denote higher levels ofserum UA. Regarding the pulmonaryarterial catheterization–derived parame-ters, serum UA levels showed a promis-ing correlation with PAP and PCWP.

LimitationsThis study was performed in stablepatients receiving optimal HF medica-tions, and findings cannot be extendedto patients with unstable and decom-pensated conditions. Also, the findingswould be more valuable if the changesin serum UA levels and hemodynamicindices had been assessed during treat-ment. Many patients had been fastingbefore right heart catheterization,although not for a long time (the con-trol patients were sampled while fastingfor routine blood chemistry tests). Themultivariate regression analysis regardingthe prediction of PCWP from UAserum levels yielded significant but weakcorrelation results. Further studies with alarger sample size may be helpful in thisregard.

ConclusionsLiver function abnormalities and UAelevations are frequently observed inHF.13 Severity of clinical signs, NYHAclass, and elevated levels of mean PAPand particularly PCWP had a signifi-cant correlation with serum UA levelsin our study. Therefore, measuring UAlevels with a simple biochemical testcould be of great value in conjunctionwith all other clinical and paraclinicalparameters in a better evaluation andmanagement of patients with systolicHF and might be considered a usefulroutine evaluation in the future. Futurestudies to assess whether changes inserum UA levels in response to HFtherapy can be predictive of theclinical outcome and can be of greatvalue.

Acknowledgment: We would like tothank our colleagues in the Heart Failureand Transplantation Department and in par-ticular Drs Sepideh Taghavi, NasimNaderi, Hooman Bakhshandeh,and Pedram Amouzadeh.

REFERENCES

1 Becker MA. Uric acid balance. http://www.uptodate.com/patients/content/topic.do?topicKey=~rarUEbeFY6It5hO. AccessedJuly 2010.

2 Cimbaro CJP, Perna ER, Macin SM, et al.Elevated serum uric acid levels in outpatientswith chronic heart failure: side effect ormarker of myocardial damage? J Card Fail.2003;9(5):S91.

3 Ferriera S, Azevodo A, Frioes F, BettencourtP. Uric acid and prognosis in HF patients. EurJ Intern Med. 2003;14:151–159.

4 Alimonda A, Nunez J, Husser O, et al.Hyperuricemia in acute heart failure. Morethan a simple spectator? Eur J Intern Med.2009;20:74–79.

5 Dawson J, Walter M. Uric acid and xan-thine oxidase: future therapeutic targets inthe prevention of cardiovascular disease. Br JClin Pharmacol. 2006;62(6):633–644.

6 Cingolani HE, Plastino FJA, Escudero EM. Theeffect of xanthine oxidase inhibition uponejection fraction in heart failure patients: LaPlata study. J Card Fail. 2006;12(7):491–498.

7 Cicoira M, Zanolla L, Rossi A, et al. Elevatedserum uric acid levels are associated with

diastolic dysfunction in patients with DCM.Am Heart J. 2002;143:1107–1111.

8 Uric acid PAP fluid 5 + 1enzymatic in vitrotest for the quantitative determination of uricacid in human serum and plasma [Packageinsert]. Worms, Germany: AXIOM Gesell-schaft fur Diagnostica und Biochemica mbH;2010, http://www.axiom-online.net. Acce-ssed July 2010.

9 Hare JM, Johnson RJ. Uric acid predictsclinical outcomes in heart failure. Circulation.2003;107:1951–1953.

10 Davidson CJ, Bonow RO. Cardiac catheteri-zation. In: Braunwald’s Heart Disease: A TextBook of Cardiovascular Medicine. 8th ed.:New York, NY: Saunders; 2008:439–462.

11 Rudski LG, Lai WW, Afilalo J, et al. Newguidelines for the echocardiographic assess-ment of the right heart in adults: a report fromthe American Society of Echocardiography.J Am Soc Echocardiogr. 2010;23:685–713.

12 Leyva F, Chua TP, Anker SD, Coats AJS. Uricacid in chronic HF. A measurement of anaer-obic threshold. Metabolism. 1998;47:1156–1159.

13 Rieclim T, Potachi M, Breidthordt T, et al.Diagnistic and prognostic value of uric acid in

patients with dyspnea. Am J Med. 2009;122:1054e7–1054e14.

14 Friedman LS. Congestive hepatopathy. UpTo-Date. http://www.uptodate.com/patients/content/topic.do?topicKey=~HUq7qbIyepB8NO. Accessed July 2010.

15 van Deursen VM, Damman K, Hillege HL,et al. Abnormal liver function in relation tohemodynamic profile in HF patients. J CardFail. 2010;16(1):84–90.

16 Jankowska EA, Panikowska B, Majda J, et al.Hyperuricemia predicts poor outcome inpatients with mild to moderate HF. Int JCardiol. 2007;115:151–155.

17 Doehner W, Rouchlance M, Florea VG, et al.Uric acid in cachectic and noncachecticpatients with chronic HF. Am Heart J.2001;141:792–800.

18 Duan X, Ling F. Is uric acid itself a player or abystander in the pathophysiology of chronicheart failure? Med Hypotheses. 2008;70:578–581.

19 Zhang YH, Lu R, Zhao XY, et al.. Associationof uric acid, BNP, CRP with invasive hemo-dynamic monitoring in heart failure patients.Zhonghua Xin Xue Guan Bing Za Zhi.2009;37(2):126–129.

filling pressures in systolic heart failure march • april 2011 83