Ce

Ra

b

c

d

a

ARRAA

KICLNHC

1

iiaaiMrAsrm[

0d

Atherosclerosis 205 (2009) 319–324

Contents lists available at ScienceDirect

Atherosclerosis

journa l homepage: www.e lsev ier .com/ locate /a therosc leros is

-reactive protein predicts functional status and correlates with left ventricularjection fraction in patients with chronic stable angina

amón Arroyo-Espligueroa,1, Pablo Avanzasb,1, Juan Quilesc, Juan Carlos Kaskid,∗

Division of Cardiology, Hospital General Universitario, Guadalajara, SpainDivision of Cardiology, Hospital Universitario Central de Asturias, Oviedo, SpainDivision of Cardiology, Hospital San juan, Alicante, SpainDivision of Cardiac and Vascular Sciences, St. George’s, University of London, Cranmer Terrace, London SW17 0RE, United Kingdom

r t i c l e i n f o

rticle history:eceived 13 March 2008eceived in revised form 3 December 2008ccepted 8 December 2008vailable online 14 December 2008

eywords:nflammation-reactive proteineft ventricular ejection fraction

a b s t r a c t

C-reactive protein (CRP) is a marker for cardiovascular risk but may also participate in the pathogenesisof atherosclerosis and myocardial injury. We sought to investigate the relationship among CRP, left ven-tricular ejection fraction (LVEF) and symptoms of congestive heart failure (CHF) in patients with chronicstable angina (CSA) pectoris.Methods: We studied 841 patients (63 ± 10 years, 72% men) with CSA undergoing coronary angiography.Symptoms of CHF were assessed using the New York Heart Association (NYHA) functional classification.CRP measurements were performed using a high sensitivity (hs-) immunoassay at the time of diagnosticcoronary angiography.Results: Baseline serum hs-CRP levels showed a significant correlation with LVEF (r = −0.11; P = 0.004),

YHA functional classeart failurehronic stable angina

and prevalence of moderate-to-severe CHF correlated with serum hs-CRP quartiles (Ptrend < 0.0001). Afteradjustment, age (P = 0.004), female gender (P = 0.03), body mass index (P < 0.0001) and hs-CRP (OR 2.2[1.3–3.6] CI 95%; P = 0.002) were independent predictors of NYHA functional classes III–IV irrespective ofLVEF and angiographic severity of CAD. A CRP value of 3.2 mg/L had a sensitivity of 72%, a specificity of75%, and a negative predictive value of 96% for detecting an impaired functional class.Interpretation: Hs-CRP serum concentrations showed an inverse correlation with LVEF and were an inde-

A fun

pendent predictor of NYH

. Introduction

Proinflammatory cytokines have been reported to be elevatedn patients with congestive heart failure (CHF) [1–3] and have beenmplicated in the pathogenesis of the disease [4,5]. Cytokines suchs tumor necrosis factor (TNF)-� and interleukin (IL)-6 have beenssociated with certain pathophysiological aspects of CHF, includ-ng myocyte apoptosis and left ventricular (LV) remodeling [6,7].

oreover, both cytokines predict mortality in patients with CHFegardless of age, gender, etiology of heart failure, New York Heartssociation (NYHA) functional class, LV ejection fraction (LVEF) anderum sodium levels [8]. It has been suggested that rather than rep-

esenting an epiphenomenon of disease worsening, inflammatoryechanisms can promote LV dysfunction and worsening of CHF

3,8].

∗ Corresponding author. Tel.: +44 208 725 5901; fax: +44 208 725 3328.E-mail address: jkaski@sgul.ac.uk (J.C. Kaski).

1 These authors contributed equally to the study.

021-9150/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.atherosclerosis.2008.12.018

ctional class in patients with CSA.© 2009 Elsevier Ireland Ltd. All rights reserved.

Despite concerns regarding the proatherogenic and proinflam-matory effects of human C-reactive protein (CRP) in animal modelsand the relatively modest prediction of coronary events by CRPand other inflammatory markers [9], recent reports suggest thatCRP may be implicated in the genesis of myocardial damage [10]and serum CRP concentrations predict the development of CHF [8].The aim of our study was to assess whether an association existsbetween serum CRP concentrations, LVEF and clinical symptoms ofCHF in patients with chronic stable angina (CSA).

2. Methods

2.1. Patients

We studied 841 stable angina patients (63 ± 10 years, 72% men)who were prospectively entered on our institutional database at

the time of diagnostic coronary angiography from November 1994to May 1996. CSA was defined as typical chest pain brought onby exertion and relieved by rest, sublingual nitrates or both, withsymptoms stable for at least 3 months before study entry. All CSApatients had a positive ECG exercise stress test response (>1 mm

3 theros

Smpw

tldpiCat

aahiqte(cw

2

tpEsnc1owTa

2

tascgclameoa

2

entAwte

20 R. Arroyo-Espliguero et al. / A

T-segment depression) or reversible perfusion defects duringyocardial perfusion scintigraphy. The clinical management of all

atients was left to the discretion of the managing cardiologist whoas unaware of serum CRP concentrations.

Patients with recent (<12 weeks) myocardial infarction (MI), life-hreatening arrhythmias, cardiac valve disease, acute or chroniciver disease, infectious diseases or other conditions likely to causeeath within 1 year were not entered in the study. None of theatients included in the study had ongoing systemic or cardiac

nflammatory processes, with the exception of angina pectoris andHF. All patients gave written informed consent before study entrynd the study was approved by the Local Research Ethics Commit-ee.

During baseline evaluation, patient’s functional status wasssessed using the NYHA functional classification. The patients’bility to perform a variety of every day tasks, i.e. personal care,ousework, occupational and recreational activities was systemat-

cally assessed by the managing cardiologist by using a normalizeduestionnaire and consequently assigned to a specific NYHA func-ional class, as appropriate. Cardiac medications taken at studyntry, specifically aspirin, 3-hydroxy-3-methylglutaryl coenzyme AHMG-CoA) reductase inhibitors (statins), �-blockers, angiotensinonverting enzyme (ACE) inhibitors, digoxin, diuretics and nitratesere recorded.

.2. High sensitivity C-reactive protein measurements

Fasting blood samples were obtained from CSA patients at theime of diagnostic coronary angiography. CRP measurements wereerformed on the COBAS Integra (Roche Diagnostics Limited, Lewes,ast Sussex, U.K.) using the CRP-Latex assay in both the high sen-itivity (hs-) application (analytical range 0.2–12 mg/L) and theormal application (analytical range 2–160 mg/L). Analytical pre-ision of the high sensitivity CRP-Latex assay was 7.6% at a level of.02 mg/L, 3.3% at 1.79 mg/L and 1.3% at a level of 4.36 mg/L. Samplesutside the analytical range of the high sensitivity CRP-Latex assayere analyzed by the CRP-Latex assay in the normal application.

he analytical precision of the normal CRP-Latex assay was 2.4% atlevel of 29.5 mg/L and 1.3% at a level of 113 mg/L.

.3. Angiographic analysis

Images of the coronary tree were obtained in all patients withhe digital Philips Integris 3000 system (Philips, Holland), using anutomated quantitative coronary artery (QCA) stenosis assessmentystem. Two experienced cardiologists, unaware of the patients’linical characteristics and biochemical results reviewed all angio-raphic images to assess severity of CAD. Coronary arteries wereonsidered to have significant stenoses in the presence of ≥75%umen diameter narrowing. The percentage of the coronary lumenalrea affected by atheroma, an index of global coronary athero-atous burden, was assessed and scored according to Sullivan

t al. (“extension score”) [11]. This score refers to the proportionf the coronary artery tree showing angiographically detectabletheroma.

.4. Statistical analysis

Results of normally distributed continuous variables arexpressed as the mean value ± S.D. Continuous variables with aon-normal distribution are presented as median value (interquar-

ile interval) and qualitative variables are presented as frequencies.nalysis of normality of the continuous variables was performedith the Kolmogorov–Smirnov test. The Kruskal–Wallis test for

wo-way analysis of variance (ANOVA) was used to evaluate differ-nces in hs-CRP levels among NYHA functional classes. Moreover,

clerosis 205 (2009) 319–324

for the purpose of analysis (and also for clinical relevance), NYHAfunctional class was dichotomized into binary predictors: I–II vs.III–IV. Differences between NYHA functional classes (I–II vs. III–IV)were assessed by univariate logistic regression for continuous vari-ables and �2 testing was used for discrete variables. Correlationsbetween continuous variables were analyzed with two-way Pear-son or Spearman correlation tests, as appropriate. Multivariateanalysis used binary logistic regression analysis for a dependentdiscrete variable and linear regression analysis for a continuousvariable. Backward stepwise selection was used to derive the finalmodel for which significance levels of 0.1 and 0.05 were chosento exclude and include terms, respectively. Variables included inmultivariate analysis were age, gender, body mass index (BMI),previous history of MI and percutaneous coronary intervention(PCI), cardiovascular risk factors such as diabetes mellitus, hyper-tension, dyslipidemia, smoking and family history of CAD, use ofaspirin, statins, �-blockers, ACE inhibitors, digoxin, diuretics ornitrates at study entry, LVEF, number of diseased coronary ves-sels with ≥75% lumen diameter narrowing and serum CRP levels.The relationship between hs-CRP levels and NYHA functional classwas not of linear nature. Therefore hs-CRP was logarithmicallytransformed before entering the logistic regression analysis. In thesame way, in order to fulfill the statistical requirement for multi-ple regression analysis (normal distribution and constant varianceof the residuals) a logarithmical transformation of hs-CRP levelswas performed. Receiver operating characteristic (ROC) curve anal-ysis (calculation of the area under the curve (AUC)) was carriedout to evaluate the ability of hs-CRP values to correctly discrimi-nate between those with and without an impaired NYHA functionalclass. From the ROC curve, the optimum diagnostic cut-off pointfor the population tested was chosen to maximize clinical sensitiv-ity and specificity. Differences were considered to be statisticallysignificant if the null hypothesis could be rejected with >95%confidence. All P values are two-tailed. The SPSS 13.0 statisticalsoftware package (SPSS Inc., Chicago, Illinois) was used for all cal-culations.

3. Results

Baseline characteristics of the 841 patients with CSA includedin the study are shown in Table 1. One hundred and twenty-eightpatients (15%) were in NYHA functional class III or IV. The relation-ship between NYHA functional classes and serum hs-CRP levelsis depicted in Fig. 1. Baseline serum hs-CRP levels raised along-side the severity of symptoms as assessed by NYHA functionalclasses (Ptrend < 0.0001). Univariate analysis showed that patientsin NYHA classes III–IV were older compared to those in classesI–II (P = 0.04). A larger proportion of women, compared to men,were in NYHA functional classes III–IV at study entry (37% vs. 27%;P = 0.03). Patients who were on NYHA functional classes III–IV hada high BMI (P < 0.0001), a previous history of PCI (P = 0.04) and areduced LVEF (P = 0.006) compared to those in classes I–II. Patientswho were in NYHA functional classes III–IV required more diuretics(P < 0.0001), nitrates (P = 0.009) and ACE inhibitors (P = 0.09), com-pared to patients in classes I–II. (Table 2) NYHA functional classwas not significantly associated with the angiographic severity ofCAD (P = 0.3) or the extension score (P = 0.24). After adjustmentusing backward stepwise binary logistic regression, age (P = 0.004),female gender (P = 0.03), BMI (P < 0.0001), LVEF (P = 0.02) and serumhs-CRP levels (OR 2.2 [1.3–3.6] CI 95%; P = 0.002) were indepen-dent predictors of markedly impaired NYHA functional classes (i.e.

III–IV) (Table 3).

Serum hs-CRP levels showed a weak, albeit significant, cor-relation with LVEF (r = −0.11; P = 0.004). After adjustment usingbackward linear multiple regression analysis where age, gender,BMI, previous history of MI and PCI, cardiovascular risk factors,

R. Arroyo-Espliguero et al. / Atherosclerosis 205 (2009) 319–324 321

Table 1Baseline characteristics of 841 patients with chronic stable angina included in thestudy.

Age (years) 63 ± 10Gender (men), n (%) 602 (72%)Systolic BP (mmHg) 135 ± 21Diastolic BP (mmHg) 80 ± 11BMI (kg/m2) 27 ± 4Previous MI, n (%) 271 (32%)Previous PCI, n (%) 59 (7%)

Cardiovascular risk factorsDiabetes mellitus, n (%) 51 (6%)Dyslipidemia, n (%) 354 (42%)Hypertension, n (%) 286 (34%)Current or ex-smokers, n (%) 578 (69%)Family history <60 years, n (%) 451 (54%)

Number of diseased vessels ≥75% (n) 1.0 [1.0–2.0]LVEF, (%) 66 ± 12NYHA functional classes (III–IV), n (%) 128 (15%)

TreatmentAspirin, n (%) 659 (78%)�-Blockers, n (%) 457 (54%)Nitrates, n (%) 456 (54%)ACE inhibitors, n (%) 146 (17%)Digoxin, n (%) 24 (3%)Diuretics, n (%) 142 (17%)HMG-CoA reductase inhibitors, n (%) 126 (15%)

BiochemistryCholesterol (mmol/L) 5.9 ± 1.1LDL-cholesterol (mmol/L) 4.1 ± 1.4HDL-cholesterol (mmol/L) 1.2 ± 0.8Triglycerides (mmol/L) 1.6 ± 1.2CRP (mg/L) 2.3 [1.1–4.8]

Data are expressed as mean ± standard deviation for normally distributed data,median [interquartile range] for non-normally distributed data or number (%) forcategorical variables. ACE: angiotensin converting enzyme; BMI: body mass index;BP: blood pressure; CRP: C-reactive protein; HDL: high-density lipoprotein. HMG-ClA

ne(h

Fpd

Table 2Univariate logistic regression analysis: comparison between patients with and with-out NYHA functional class III or IV.

NYHA functional class P value

I–II (n = 713) III–IV (n = 128)

Age (years) 63 ± 10 65 ± 9 0.04Female gender, n (%) 192 (27%) 47 (37%) 0.03BMI (kg/m2) 27 ± 4 29 ± 5 <0.0001Previous MI, n (%) 223 (33%) 48 (39%) 0.18Previous PCI, n (%) 44 (6%) 15 (12%) 0.04

Cardiovascular risk factorsDiabetes mellitus, n (%) 41 (6%) 10 (8%) 0.42Dyslipidemia, n (%) 307 (43%) 47 (37%) 0.21Hypertension, n (%) 238 (33%) 48 (38%) 0.36Current or ex-smokers, n (%) 488 (68%) 90 (70%) 0.76Family history <60 years, n (%) 372 (52%) 79 (62%) 0.05

Number of diseased vessels ≥75% (n) 1.0 [1.0–2.0] 2.0 [1.0–3.0] 0.3LVEF (%) 66 ± 11 62 ± 14 0.006

Treatment at study entryAspirin, n (%) 560 (79%) 99 (77%) 0.82�-Blockers, n (%) 396 (56%) 61 (48%) 0.1Nitrates, n (%) 373 (52%) 83 (65%) 0.009ACE inhibitors, n (%) 117 (16%) 29 (23%) 0.09Digoxin, n (%) 19 (3%) 5 (4%) 0.39Diuretics, n (%) 105 (15%) 37 (29%) <0.0001HMG-CoA reductase inhibitors, n (%) 103 (14%) 23 (18%) 0.35

BiochemistryCholesterol (mmol/L) 5.9 ± 1.2 6.0 ± 1.1 0.49LDL-cholesterol (mmol/L) 4.0 ± 1.4 4.1 ± 1.3 0.67HDL-cholesterol (mmol/L) 1.2 ± 0.8 1.2 ± 0.7 0.62Triglycerides (mmol/L) 1.6 ± 1.1 1.7 ± 1.3 0.44

CRP (mg/L) 2.2 [1.0–4.2] 4.4 [1.8–7.4] <0.0001

Data are expressed as mean ± standard deviation for normally distributed data,median [interquartile range] for non-normally distributed data or number (%) forcategorical variables. ACE: angiotensin converting enzyme; BMI: body mass index;CRP: C-reactive protein; HDL: high-density lipoprotein; HMG-CoA: 3-hydroxy-3-methylglutaryl coenzyme A; LDL: low-density lipoprotein; LVEF: left ventricular

oA: 3-hydroxy-3-methylglutaryl coenzyme A; LDL: low-density lipoprotein. LVEF:

eft ventricular ejection fraction; MI: myocardial infarction; NYHA: New York Heartssociation; PCI: percutaneous coronary intervention.

umber of coronary arteries with ≥75% reduction in lumen diam-

ter and hs-CRP levels were assessed, only a previous history of MIP < 0.0001), number of diseased coronary arteries (P < 0.0001) ands-CRP levels (P = 0.02) were independent predictors of LVEF.

ig. 1. Distribution of high sensitivity C-reactive protein (hs-CRP) levels among 841atients with chronic stable angina according to NYHA functional classes. Box plotsemonstrate median, 25th, and 75th percentile values for hs-CRP.

ejection fraction; MI: myocardial infarction; NYHA: New York Heart Association;PCI: percutaneous coronary intervention.

The optimum diagnostic cut-off point for hs-CRP levels to pre-dict impaired NYHA functional class in this study population wasfound to be 4.2 mg/L by ROC analysis (Fig. 2), with an AUC of 0.65(0.62–0.68; P < 0.0001). Based on the results of the multivariateanalysis in which age, BMI and gender were independently asso-

ciated to an impaired NYHA functional class and given the fact thatthese variables could negatively affect functional status not nec-essarily associated with a cardiovascular origin, patients with anage >75 years, postmenopausal women and overweight patients

Table 3Multivariate predictors of NYHA functional class in 841 patients with chronic stableangina included in the study.

OR P 95% confidence interval

Lower bound Upper bound

Age 1.04 0.004 1.01 1.1Gender (females) 1.7 0.03 1.04 2.7BMI 1.1 <0.0001 1.1 1.2LVEF 0.1 0.02 0.01 0.7CRP 2.2 0.002 1.3 3.6

BMI: body mass index; CRP: C-reactive protein; LVEF: left ventricular ejection frac-tion; NYHA: New York Heart Association. Variables such as previous MI (P = 0.5),previous PCI (P = 0.28), diabetes mellitus (P = 0.35), dyslipidemia (P = 0.64), hyperten-sion (P = 0.81), current or ex-smokers (P = 0.76), family history (P = 0.22), �-blockers(P = 0.1), aspirin (P = 0.9), ACE inhibitors (P = 0.82), HMG-CoA reductase inhibitors(P = 0.1), digoxin (P = 0.93), diuretics (P = 0.29), nitrates (P = 0.1), number of diseasedvessels ≥75% (P = 0.98) were not independent predictors of baseline NYHA functionalclass.

322 R. Arroyo-Espliguero et al. / Atherosclerosis 205 (2009) 319–324

Fig. 2. The receiver operating characteristic (ROC) curves and the associated area under the curve (AUC) for the different models assessed as predictors for an impaired NYHAfunctional class. Model A encompasses the global study population (n = 841). In model B, patients >75 years, postmenopausal women and patients with BMI >30 kg/m2 weree sal w 2

a

wdcsahenF

4

rf

4

CLsiaapiad

xcluded form the analysis (n = 406). In Model C, patients >75 years, postmenopaunalysis (n = 267).

ith BMI >30 kg/m2 were excluded form the ROC analysis. The pre-ictive ability of this model increased to 0.72 (P < 0.0001) with aut-off point of 3.2 mg/L. Furthermore, if patients who currentlymoked at the time of study entry were also excluded from thenalysis, AUC increased to 0.82 (P < 0.0001). However, despite theigh negative predictive value (PV) showed by the different mod-ls, positive PV was extremely low (24–28%). Sensitivity, specificity,egative and positive PV of the different models are depicted inig. 2.

. Discussion

The present study showed that elevated CRP concentrations cor-elate with LVEF and are also an independent predictor of NYHAunctional class in patients with CSA.

.1. CRP and functional status

Proinflammatory cytokines are upregulated in patients withHF and may be implicated in the progressive impairment ofVEF and functional class in patients with CHF [1–3]. Cytokinesuch as TNF-� and IL-6 have been shown to trigger apoptosis insolated cardiomyocytes [12], to impair contractile function [4,5]nd to induce LV remodeling and dilatation [4,5]. In fact, TNF-�

nd their soluble receptors (sTNFR1 and sTNFR2) as well as IL-6,redict mortality in patients with CHF [7]. Whether these find-

ngs simply indicate a relationship between inflammatory markersnd LV dysfunction or a causal mechanism has not been eluci-ated.

omen, patients with BMI >30 kg/m and current smokers were excluded from the

In our study, CRP levels correlated with LVEF and serum CRP con-centrations were higher in patients with NYHA functional classesIII–IV compared to patients in classes I–II. These findings suggestthat inflammation is linked to LV dysfunction in patients with CSAand inflammatory biomarkers may represent a clinical tool for theprediction of the development of heart failure in CSA patients.Our results are in agreement with previous reports that CRP maybe elevated early in the course of LV dysfunction and predict theoccurrence of subsequent overt CHF [8]. Moreover, elevated serumCRP concentrations have been associated with a higher hospitaliza-tion and readmission rate, probably reflecting a worse therapeuticresponse, and increased mortality in patients with CHF [13].

The question however remains as to whether high CRP levelsare simply a marker of CHF and CAD “severity”, or a pathogenicmechanism. Serum CRP elevation may be the consequence of theupregulated expression of TNF-�, IL-6, and other proinflamma-tory cytokines in patients with CHF, but CRP has been implicatedin the pathogenesis of LV dilation and remodeling, as it directlyenhances inducible nitric oxide (NO) synthase (iNOS) expressionand NO synthesis in cardiac myocytes, which can exert negativeinotropic and cytotoxic effects on these cells [10]. Moreover, bystimulating endothelial production of IL-6 and neutrophil sheddingof the soluble IL-6 receptor (sIL-6R), CRP may markedly exaggeratethe action of IL-6 at the level of the endothelium [14,15]. BecauseIL-6 is a potent stimulus for the production of CRP, increased vas-

cular production may represent a positive feedback mechanism forthe continued production of CRP by the liver. In fact, the risk ofdeveloping CHF has been reported to be 4.1-fold higher in patientswith elevated levels of both CRP and IL-6 [8]. Another pathogenicmechanism in which CRP may be also implicated is endothelial

theros

dinta

4

hdHsrPdfsatf[dsibmCtdte[

4

Ntltctt“wobcmcpfhic

ilifataal

[

[

[

[

[

[

[

[

[

[

[

[

R. Arroyo-Espliguero et al. / A

ysfunction [16]. Flow-mediated vasodilation of conduit arteriess reduced in CHF patients and several experiments have shown itsegative impact on exercise capacity [17]. Despite the above data,he pathogenic role of CRP in LV dysfunction is purely speculativend beyond the scope of our study.

.2. Clinical and therapeutic implications

Our results may have therapeutic implications. We and othersave previously reported that hs-CRP levels predict future car-iovascular events in patients with cardiovascular disease [18,19].s-CRP may also be clinically useful to predict worsening functional

tatus and may help to identify CSA patients with a deterio-ated quality of life [20]. However, despite the high negativeV, as well as sensitivity and specificity, showed by hs-CRP toetect impaired NYHA functional class, positive PV was low. There-ore, additional testing is recommended after an initial positivecreening test if a low positive PV can be expected. Perhaps

multimarker approach that includes the measurement of N-erminal pro-brain natriuretic peptide (NT-proBNP), as suggestedor risk stratification in patients with acute coronary syndromes21], could represent a more effective option [22]. The indepen-ent predictive value reported for both NT-proBNP and hs-CRPuggests that these may be the markers of different processesn the pathogenesis of CHF. Elevated NT-proBNP levels may thuse a marker for ventricular strain, whereas elevated CRP levelsay be an indicator of the inflammatory process [22,23]. Hs-

RP, as well as NT-proBNP, may be potentially useful to evaluatehe efficacy of treatment and help physicians decide when toischarge patients from hospital more accurately [24]. Whetherherapeutic strategies aimed to lower CRP and NT-proBNP lev-ls will affect patient outcome warrants future investigation20,25–27].

.3. Limitations

The potential limitations of these data merit consideration.YHA functional classification was originally developed to predict

he impact of cardiovascular disease on the day-to-day patient’sifestyle. It is however a difficult variable to measure and quan-ify given its rather subjective nature. Indeed, NYHA functionallass is affected not only by the physical disability triggered byhe disease, but also by the individual’s perception of their symp-oms, comorbidities and psychological factors, among others. Anobjective” assessment of functional capacity and performanceith instruments such as the Webber classification of maximal aer-

bic capacity (VO2max) or the 6-min walk-test distance, may beetter measurements for the assessment of limitations to physi-al activity imposed by cardiovascular disease. However, “objectiveeasurements” of functional outcomes such as maximal aerobic

apacity (VO2max) may not be widely available and require specificersonnel and infrastructure. Moreover, the validity of the NYHAunctional classification has been demonstrated in studies thatave reported moderate correlations with VO2max and a strong

ndependent prognostic impact in a wide variety of cardiovascularlinical settings [28].

Another limitation is that only 17% and 15% of the patientsncluded in the study were receiving ACE inhibitors and lipid-owering medications, respectively, at study entry. Thus, treatmentn our study is not representative of current management strategiesor patients with stable CAD. However, as no significant associ-

tions were found among CRP levels, ACE inhibitor and statinreatment and, furthermore, the relationship between CRP levelsnd clinical symptoms was assessed using multivariable analysesdjusted for the use of ACE inhibitors and statins, our results areikely to represent a true reflection of the importance of inflamma-

[

clerosis 205 (2009) 319–324 323

tory mechanisms in the gradual development of cardiac failure inpatients with CSA.

5. Conclusions

High serum hs-CRP concentrations correlate with lower LVEFand impaired NYHA functional class in patients with CSA and maybe a useful clinical marker of LV dysfunction in this setting. Whethertherapeutic strategies aimed to lower CRP levels will affect patientoutcome warrants future investigation.

References

[1] Mann DL. Recent insights into the role of tumor necrosis factor in the failingheart. Heart Fail Rev 2001;6:71–80.

[2] Wollert KC, Drexler H. The role of interleukin-6 in the failing heart. Heart FailRev 2001;6:95–103.

[3] Torre-Amione G, Kapadia S, Benedict C, et al. Proinflammatory cytokine lev-els in patients with depressed left ventricular ejection fraction: a reportfrom the studies of left ventricular dysfunction (SOLVD). J Am Coll Cardiol1996;27:1201–6.

[4] Bozkurt B, Kribbs S, Clubb FJ, et al. Pathophysiology relevant concentrations oftumor necrosis factor-� promote progressive left ventricular dysfunction andremodeling in rats. Circulation 1998;97:1382–91.

[5] Murray DR, Freeman GL. Tumor necrosis factor-� induces a biphasic effect onmyocardial contractility in conscious dogs. Circ Res 1997;81:627–35.

[6] Deswal A, Petersen NJ, Feldman AM, et al. Cytokines and cytokines receptors inadvanced heart failure: an analysis of the cytokine database from the vesnari-none trial (VEST). Circulation 2001;103:2055–9.

[7] Kelly RA, Smith TW. Cytokines and cardiac contractile function. Circulation1997;95:778–81.

[8] Vasan RS, Sullivan LM, Roubenoff R, et al. Inflammatory markers and risk ofheart failure in elderly subjects without a prior myocardial infarction: theFramingham Study. Circulation 2003;107:1486–91.

[9] Pepys MB. CRP or not CRP? That is the question. Arterioscler Thromb Vasc Biol2005;25:1091–4.

10] Ikeda U, Maeda Y, Yamamoto K, et al. C-reactive protein augments induciblenitric oxide synthase expression in cytokine-stimulated cardiac myocytes. Car-diovasc Res 2002;56:86–92.

[11] Sullivan DR, Marwick TH, Freedman SB. A new method of scoring coronaryangiograms to reflect extent of coronary atherosclerosis and improve correla-tion with major risk factors. Am Heart J 1990;119(6):1262–7.

12] Krown KA, Page MT, Nguyen C, et al. Tumor necrosis factor alpha-induced apop-tosis in cardiac myocytes: involvement of the sphingolipid signaling cascade incardiac cell death. J Clin Invest 1996;98:2854–65.

13] Alonso-Martínez JL, Llorente-Diez B, Echegaray-Agara M, et al. C-reactive pro-tein as a predictor of improvement and readmission in heart failure. Eur J HeartFail 2002;4:331–6.

14] Jones SA, Novick D, Horiuchi S, et al. C-reactive protein: a physiological activatorof interleukin-6 receptor shedding. J Exp Med 1999;189:599–604.

15] Verma S, Li SH, Badiwala MV, et al. Endothelin antagonism and interleukin-6inhibition attenuate the proatherogenic effects of C-reactive protein. Circula-tion 2002;105:1890–6.

16] Fichtlscherer S, Rosenberger G, Walter DH, et al. Elevated C-reactive proteinlevels and impaired endothelial vasoreactivity in patients with coronary arterydisease. Circulation 2000;102:1000–6.

[17] Hambrecht R, Fiehn E, Weigl C, et al. Regular physical exercise corrects endothe-lial dysfunction and improves exercise capacity in patients with chronic heartfailure. Circulation 1998:2709–15.

18] Lindahl B, Toss H, Siegbahn A, et al. Markers of myocardial damage and inflam-mation in relation to long-term mortality in unstable coronary artery disease.N Engl J Med 2000;343:1139–47.

19] Arroyo-Espliguero R, Avanzas P, Cosín-Sales J, et al. C-reactive protein eleva-tion and disease activity in patients with coronary artery disease. Eur Heart J2004;25:401–8.

20] Anand IS, Latini R, Florea VG, et al. C-reactive protein in heart failure. Prognosticvalue and the effect of valsartan. Circulation 2005;112:1428–34.

21] Sabatine MS, Morrow DA, de Lemos JA, et al. Multimarker approach to riskstratification in non-ST elevation acute coronary syndromes: simultaneousassessment of troponin I, C-reactive protein, and B-type natriuretic peptide.Circulation 2002;105:1760–3.

22] Hartmann F, Packer M, Coats AJ, et al. Prognostic impact of plasma N-terminal pro-brain natriuretic peptide in severe chronic congestive heartfailure: a substudy of the Carvedilol Prospective Randomized Cumulative Sur-vival (COPERNICUS) trial. Circulation 2004;110:1780–6.

23] Campbell DJ, Woodward M, Chalmers JP, et al. Prediction of heart failure

by amino terminal-pro-B-type natriuretic peptide and C-reactive pro-tein in subjects with cerebrovascular disease. Hypertension 2005;45:69–74.

24] Bettencourt P, Azevedo A, Pimenta J, et al. N-terminal-pro-brain natriureticpeptide predicts outcome after hospital discharge in heart failure patients.Circulation 2004;110:2168–74.

3 theros

[

[

24 R. Arroyo-Espliguero et al. / A

25] Parissis JT, Adamopoulos S, Farmakis D, et al. Effects of serial levosimendaninfusions on left ventricular performance and plasma biomarkers of myocardialinjury and neurohormonal and immune activation in patients with advancedheart failure. Heart 2006;92:1768–72.

26] Krum H, McMurray JJ. Statins and chronic heart failure: do we need a large-scaleoutcome trial? J Am Coll Cardiol 2002;39:1567–73.

[

[

clerosis 205 (2009) 319–324

27] Go AS, Lee WY, Yang J, et al. Statin therapy and risks for death and hospitalizationin chronic heart failure. JAMA 2006;296:2105–11.

28] Bennett JA, Riegel B, Bittner V, et al. Validity and reliability of the NYHA classesfor measuring research outcomes in patients with cardiac disease. Heart Lung2002;31:262–70.