Author + information
- Received September 5, 2008
- Revision received December 2, 2008
- Accepted December 8, 2008
- Published online April 7, 2009.
- ↵*Reprint requests and correspondence:
Dr. Marion Delcroix, Department of Pneumology, Universitaire Ziekenhuizen Leuven, Herestraat 49, B-3000 Leuven, Belgium
Objectives Our aim was to investigate in a prospective study a potential role of C-reactive protein (CRP) in predicting the outcome in pulmonary arterial hypertension (PAH) and chronic thromboembolic pulmonary hypertension (CTEPH).
Background CRP is a well-known marker of inflammation and tissue damage, widely recognized as a risk predictor of cardiovascular and coronary heart diseases.
Methods Plasma levels of CRP have been measured in consecutive patients diagnosed with PAH and CTEPH, at the time of right heart catheterization.
Results Circulating CRP levels were increased in CTEPH and PAH patients compared with those in control subjects (4.9 mg·l−1, 95% confidence interval [CI]: 3.9 to 6.2 mg·l−1; 4.4 mg·l−1, 95% CI: 3.5 to 5.4 mg·l−1; and 2.3 mg·l−1, 95% CI: 1.9 to 2.7 mg·l−1, respectively; p < 0.0001). In PAH patients, CRP levels correlated with New York Heart Association functional class (r = 0.23), right atrial pressure (r = 0.25), and 6-min walking distance (r = −0.19) and were significantly higher in nonsurvivors than in survivors (p = 0.003). All PAH, idiopathic PAH, and patients naive for disease-specific medication with CRP levels >5.0 mg·l−1had a significantly lower survival rate (p = 0.02, p = 0.009, and p < 0.05, respectively). In CTEPH patients, circulating CRP levels significantly decreased 12 months after pulmonary endarterectomy (n = 23, 4.0 mg·l−1, 95% CI: 2.8 to 5.8 mg·l−1, to 1.6 mg·l−1, 95% CI: 2.2 to 3.0 mg·l−1; p = 0.004). PAH patients normalizing their CRP levels under treatment (n = 29), assigned as responders, had a significantly higher survival rate (p < 0.05). The proportion of patients treated with a parenteral prostacyclin-analogue was significantly higher among the responders than the nonresponders (55% vs. 17%, p = 0.002).
Conclusions This is the first evidence of a role of an inflammatory marker, such as CRP, in predicting outcome and response to therapy in PAH.
Pulmonary hypertension is a rare and severe disorder, characterized by an elevated pulmonary arterial pressure (PAP). On the basis of common clinical features, 5 categories of pulmonary hypertension have been identified (1). Among them, pulmonary arterial hypertension (PAH) is characterized by an intrinsic distal vessel arteriopathy and chronic thromboembolic pulmonary hypertension (CTEPH) by proximal vessel occlusions occasionally followed by distal vessel remodeling. The pathophysiological processes commonly involved in PAH include vasoconstriction, vascular remodeling, and thrombosis (2). Although inflammatory cell infiltrates had been observed around the distal pulmonary arteries (3), the potential role of inflammation in the pathogenesis of PAH has only recently been investigated. Increased plasma levels of interleukin-1, interleukin-6, and CX3C chemokine fractalkine (4,5) and increased expression of CCL3 chemokine macrophage inflammatory protein-1 alpha in lung biopsies (6) have been reported. In CTEPH, dysregulated thrombosis and/or thrombolysis, as causes of intraluminal thrombus organization and fibrous vessel obliteration, have been inconsistently evidenced (7). Recent works also suggest the importance of systemic inflammation in the genesis of the disease. The CCL2 chemokine monocyte chemoattractant protein-1 is up-regulated in plasma and in large pulmonary arteries of CTEPH patients (8). An increased prevalence of inflammatory disease in patients with CTEPH versus acute pulmonary embolism has been reported (9).
C-reactive protein (CRP) is a marker of inflammation and tissue damage. It has been identified as a predictor of risk for cardiovascular events (10). Besides its commonly accepted role of bystander and marker of clinical risk in various inflammatory diseases, CRP may also be an active player in the physiopathology of the vascular wall, as demonstrated in atherosclerosis (11).
The role of CRP in pulmonary hypertension is completely unknown. The aim of the present study was to investigate a potential role of CRP as a predictor of pulmonary hypertension severity and outcome.
This prospective study included all consecutive patients diagnosed with PAH and CTEPH, as defined by the Venice classification (1), who underwent right heart catheterization at the University Hospital of Leuven between April 1, 2004, and April 30, 2008. Healthy subjects were recruited out of screening programs from the departments of urology and gynecology. The study protocol has been approved by the Institutional Ethics Committee of the University Hospitals Leuven, and participants gave informed consent.
Medical records of the patients were reviewed, and the following data were extracted: sex, age, weight and size, modified New York Heart Association (NYHA) functional class (1), 6-min walking distance (6MWD), current disease-specific drug use (prostacyclin analogues, endothelin receptor antagonists, and phosphodiesterase inhibitors) at the time of catheterization, date of symptom onset, and date of diagnosis. At study entry, patients and healthy control subjects completed a questionnaire that provided information about smoking habits, and history of diabetes mellitus and hypertension. The right heart catheterization included measurements of mean right atrial pressure (RAP), PAP, pulmonary vascular resistance (PVR), and cardiac index. PAH and CTEPH patients were stratified into subgroups of low versus high disease severity according to NYHA functional class (I to II orIII to IV). For survival and event-free survival, the observation period was from study entry until death or June 25, 2008. Death, lung transplantation, and start of parenteral prostacyclin analogues were assigned as clinical worsening events.
Blood samples were collected on ethylenediaminetetra-acetic acid at the time of right heart catheterization, and plasma was prepared. CRP levels (Tina-quant C-Reactive Protein, Roche Diagnostics, Vilvoorde, Belgium) were determined in the University Hospital routine laboratory. The upper limit of normal (ULN) was 5 mg·l−1.
Database management and statistical analyses were performed using SAS software version 9.1 (SAS Inc., Cary, North Carolina) and GraphPad Prism version 4.01 (GraphPad Software Inc., La Jolla, California). Continuous and normally distributed values were expressed as mean ± SD. Values not normally distributed were log-transformed to normalize their distribution and expressed as a geometric mean with a 95% confidence interval (CI). Differences between the 3 groups were analyzed using analysis of variance followed by post-hoc tests (Bonferroni) for continuous variables and a chi-square test for categorical variables. Differences between 2 groups were compared using a Student ttest. Differences within groups before and after treatment were analyzed using a paired ttest. Associations between variables were investigated using Pearson correlation. Survival curves for patients were contrasted with baseline levels of CRP above and below the ULN by Kaplan-Meier survival function estimate and log-rank test. The sensitivity and specificity of CRP to predict survival was assessed by a receiver-operator characteristic (ROC) analysis. Cox regression was applied to model the relation between failure time (time to death for PAH, and time to death or persistent pulmonary hypertension for CTEPH, defined by a PAP >35 mm Hg at 3 days post-surgery) and the plasma CRP concentration at baseline, adjusting for other explanatory variables including sex, age, body mass index, smoking status (current and past), PAP, RAP, PVR, cardiac index, NYHA functional class, etiology, and 6MWD. We identified covariates by a stepwise regression procedure with the p values for variables to stay in the model set at 0.10. Covariates considered for entry in the model were age, sex, body mass index, smoking status, PAP, RAP, PVR, cardiac index, NYHA functional class, 6MWD, etiology, and logCRP. All p values were for 2-sided tests. A value of p < 0.05 was considered statistically significant.
Characteristics of the study population
One hundred and four patients with PAH, 79 patients with CTEPH, and a control group of 95 healthy subjects were included in the study. Patient demographics and clinical characteristics at study entry are reported in Table 1.Among the PAH patients, 50 (49%) had idiopathic pulmonary arterial hypertension (IPAH) and 54 (51%) had PAH associated to other diseases including connective tissue disease (24%), congenital heart disease (7%), portal hypertension (10%), myeloproliferative disorders (5%), pulmonary veno-occlusive disease (3%), drug abuse (1%), and human immunodeficiency virus (1%). Sixty-seven (68%) PAH patients were naive for disease-specific medication. Among the CTEPH patients, 58 (75%) had a history of acute venous thromboembolism and 53 (66%) had at least 1 thrombophilic disorder. During the observation period (mean 786 days, range 4 to 1,554 days), 27 PAH and 8 CTEPH patients died. Forty-four (56%) CTEPH patients underwent a pulmonary endarterectomy (PEA). Nine PAH patients underwent a lung transplantation, and 11 started intravenous or subcutaneous prostacyclin analogues.
CRP and severity in PAH
Circulating CRP levels were significantly higher in CTEPH and PAH patients (4.9 mg·l−1, 95% CI: 3.9 to 6.2 mg·l−1and 4.4 mg·l−1, 95% CI: 3.5 to 5.4 mg·l−1) compared with those seen in control subjects (2.3 mg·l−1, 95% CI: 1.9 to 2.7 mg·l−1, analysis of variance p < 0.0001) (Fig. 1).CRP levels were not statistically different between the different subcategories of PAH.
In PAH patients, CRP levels correlated with NYHA functional class (r = 0.23, p = 0.02) and RAP (r = 0.25, p = 0.01) and inversely correlated with 6MWD (r = −0.19, p = 0.04). CRP levels were significantly higher in NYHA functional class III to IV compared with those seen in NYHA functional class I to II (5.9 mg·l−1, 95% CI: 4.4 to 7.9 mg·l−1vs. 3.2 mg·l−1, 95% CI: 2.3 to 4.3 mg·l−1, p = 0.004) and in nonsurvivors compared with those seen in survivors (8.3 mg·l−1, 95% CI: 5.0 to 13.6 mg·l−1vs. 3.6 mg·l−1, 95% CI: 2.9 to 4.5 mg·l−1, p = 0.0008). This last observation was also found in IPAH patients (8.0 mg·l−1, 95% CI: 3.3 to 19.2 mg·l−1vs. 3.6 mg·l−1, 95% CI: 2.7 to 4.8 mg·l−1, p = 0.02) and in treatment-naive PAH patients (8.5 mg·l−1, 95% CI: 4.3 to 16.7 mg·l−1vs. 3.6 mg·l−1, 95% CI: 2.7 to 4.9 mg·l−1, p = 0.03). The nonsurvivors also had a lower 6MWD (233 ± 150 m vs. 371 ± 152 m, p = 0.0001) and a higher RAP (12 ± 6 mm Hg vs. 8 ± 5 mm Hg, p = 0.0002) and NYHA functional class (2.8 ± 0.7 vs. 2.4 ± 0.8, p = 0.01). Severity did not affect CRP levels in CTEPH patients.
CRP as a predictor of outcome in PAH
Survival Kaplan-Meier analyses showed that all PAH, IPAH, and treatment-naive PAH patients with CRP levels higher than ULN had a significantly lower survival rate (2-year survival of 65% vs. 82%, p = 0.02 [Fig. 2A]; 61% vs. 96%, p = 0.009 [Fig. 2B]; and 63% vs. 87%, p < 0.05 [Fig. 2C], respectively). In accordance, all PAH, IPAH, and treatment-naive PAH patients with CRP > ULN had a higher NYHA functional class (p = 0.02, p = 0.02, and p = 0.003, respectively) and a lower 6MWD (p = 0.0001, p = 0.01, and p = 0.0009, respectively), without any significant differences in the hemodynamic profile.
Similarly, event-free survival Kaplan-Meier analyses showed that all PAH, IPAH, and treatment-naive PAH patients with CRP > ULN had a significantly lower event-free survival rate (2-year event-free survival of 57% vs. 76%, p = 0.01 [Fig. 2D]; 54% vs. 95%, p = 0.009 [Fig. 2E]; and 63% vs. 78%, p = 0.04 [Fig. 2F], respectively). Accordingly, the significantly lower CRP levels observed in event-free PAH patients (3.6 mg·l−1, 95% CI: 2.8 to 4.6 mg·l−1vs. 6.3 mg·l−1, 95% CI: 4.3 to 9.4 mg·l−1, p = 0.01) and in event-free treatment-naive PAH patients (3.5 mg·l−1, 95% CI: 2.5 to 4.9 mg·l−1vs. 6.5 mg·l−1, 95% CI: 3.9 to 10.7 mg·l−1, p = 0.04) were accompanied by a significantly lower NYHA functional class (p = 0.01 and p = 0.007, respectively) and RAP (p = 0.005 and p = 0.01, respectively) and a higher 6MWD (p = 0.0003 and p = 0.006, respectively).
All PAH and IPAH patients with a CRP > ULN had a significantly lower time to death (1.7 ± 1.2 years vs. 2.3 ± 1.3 years, p = 0.04; 1.8 ± 1.2 years vs. 2.7 ± 1.3 years, p = 0.02, respectively), without any significant change in time to clinical worsening.
According to ROC analyses, CRP plasma levels of 4.75, 4.9, and 5.0 mg·l−1, respectively, were the best cutoff values for all PAH, IPAH, and treatment-naive PAH patients. Sensitivity was 63% (95% CI: 42% to 81%) and specificity 64% (95% CI: 52% to 74%) for all PAH patients, 70% (95% CI: 35% to 93%) and 69% (95% CI: 51% to 81%) for IPAH patients, and 62% (95% CI: 35% to 85%) and 63% (95% CI: 48% to 76%) for treatment-naive PAH patients.
The univariate analysis showed that age, etiology, NYHA functional class, 6MWD, RAP, and CRP significantly predicted mortality of all PAH and treatment-naive PAH patients (Table 2).Only 6MWD and CRP significantly predicted mortality of IPAH patients (Table 2). Both before (Fig. 2) and after adjustment for potential confounding variables, plasma CRP predicted mortality of all PAH, IPAH, and treatment-naive PAH patients (Table 2). Independent of the other covariates, a doubling of the plasma CRP level was associated with a 41% (95% CI: 9% to 81%, p = 0.008) increase in the risk for a fatal event. The corresponding hazard ratios for IPAH and treatment-naive PAH patients were 100% (95% CI: 6% to 290%, p = 0.033) and 48% (95% CI: 10% to 99%, p = 0.009), respectively (Table 2). CRP levels did not correlate with the elapsed time since the onset of symptoms or since diagnosis. In CTEPH patients, PAP was predictive for adverse outcome (hazard ratio: 1.05, 95% CI: 1.01 to 1.10, p = 0.02), whereas CRP was not.
Effect of medical treatment in PAH
A second evaluation was performed in 24 PAH patients, mean 28 months (range 1 to 82 months) after start of disease-specific medication. In 33 of the 37 PAH patients who were already treated at inclusion, the baseline evaluation was retrospectively retrieved from the reference center database. In the whole group of 57 PAH patients, an average period of 40-month treatment (range 9 to 82 months) resulted in significant decreases in PVR (−134 ± 397 dyn·s·cm−5, p = 0.003) and NYHA functional class (−0.3 ± 0.8, p = 0.01) and increase in cardiac index (0.38 ± 0.97 l·min−1·m2, p = 0.0001), without any significant change in RAP, PAP, 6MWD, and CRP levels. However, when patients were stratified according to treatment effect on CRP levels, patients normalizing CRP levels under ULN (responders, n = 29, −3.7 ± 6.8, p < 0.0001) displayed a concomitant decrease in NYHA functional class (−0.3 ± 0.8, p < 0.05) and increase in cardiac index (0.29 ± 0.78 l·min−1·m2, p = 0.002). Survival Kaplan-Meier analysis showed that responders had a higher survival rate (84% vs. 61% at 3 years, p < 0.05) (Fig. 3).
Twenty-nine PAH patients were treated with endothelin receptor antagonists, 17 with prostacyclin analogues, 5 with phosphodiesterase-5 inhibitors, 5 with a combination of 2 of them, and 1 with calcium-channel blockers. The proportion of patients receiving prostacyclin analogues was significantly higher among the responders compared with the nonresponders (55% vs. 17%, p = 0.002). Patients receiving prostacyclin analogues displayed significantly lower CRP levels (2.7 mg·l−1, 95% CI: 1.8 to 3.9 mg·l−1vs. 5.1 mg·l−1, 95% CI: 3.5 to 7.4 mg·l−1; p = 0.01).
Effect of surgical treatment in CTEPH
A second evaluation was performed in 23 CTEPH patients, mean 12 months (range 2 to 22 months) after PEA. Circulating CRP levels were significantly decreased (4.0 mg·l−1, 95% CI: 2.8 to 5.8 mg·l−1vs. 5.1 mg·l−1, 95% CI: 3.5 to 7.4 mg·l−1; p = 0.004) (Fig. 4),with concomitant decreases in RAP (−3.1 ± 5.4 mm Hg, p = 0.01), PAP (−15 ± 12 mm Hg, p < 0.0001), PVR (−522 ± 433 dyn·s·cm−5, p < 0.0001), and NYHA functional class (−0.9 ± 1.0, p = 0.0002) and increase in cardiac index (0.59 ± 0.55 min−1·m2, p = 0.0002) and 6MWD (95 ± 99 m, p = 0.0003).
The aim of the present study was to investigate a potential role of CRP in predicting the outcome in patients with PAH and CTEPH. The results revealed that CRP levels were higher in patients with pulmonary hypertension compared with those in control subjects. Severe pulmonary hypertension was associated with increased circulating CRP levels in PAH. CRP predicted mortality and clinical worsening of PAH patients. An increase in CRP levels under disease-specific medication also predicted mortality of PAH patients. The novelty of the present study is a potential role of CRP as a predictor of adverse outcome and response to therapy in PAH.
Inflammation in PAH: circulating CRP as the main indicator
Several studies have already argued for a role of both systemic and local inflammation in PAH (3–6). This fits with our findings demonstrating an increase in circulating CRP levels in PAH patients compared with those in control subjects. Elevated levels of CRP predict the risk of recurrent ischemia and death among patients with known atherosclerotic disease. Elevated CRP has been recently proposed as a predictor of pulmonary hypertension in Gaucher disease (12). In chronic obstructive pulmonary disease, also associated with systemic inflammation, CRP has been identified as a marker of prognosis (13). Chronic obstructive pulmonary disease patients with pulmonary hypertension display higher plasma levels of CRP (14). The present study is the first that evidenced circulating CRP as an independent predictor of mortality and prognosis in patients with PAH. In accordance, besides supranormal CRP, poor survival was also associated with higher RAP and NYHA functional class and lower 6MWD, known as surrogate markers of survival in PAH (15,16). Nowadays, NYHA functional class and 6MWD are widely recognized as important and useful primary end points for clinical trials (17,18). Our findings are even strengthened by the remaining role of CRP as an independent risk factor in predicting adverse outcome in more homogeneous groups of patients like IPAH or PAH patients without any disease-specific treatment.
The N-terminal fragment of brain natriuretic peptide (NT-proBNP) is commonly used as a marker for diagnosis, severity, and prognosis of patients with congestive heart failure (19). NT-proBNP is so far the only biomarker currently used in the practice as a prognostic parameter (20) and to stratify disease severity (21) in patients with pulmonary hypertension. According to our present findings, we may consider including CRP as an additional biomarker in the evaluation of PAH.
CRP and medical treatment in PAH
In accordance with previous reports highlighting the importance of improvement in exercise capacity, NYHA functional class, and cardiac index in evaluating effects of treatment in PAH (22,23), we have observed that a 40-month period of specific disease treatment resulted in a decrease in PVR and NYHA functional class and an increase in cardiac index. Any association between the inflammation status of PAH patients and treatment effect has been established so far. Our present findings concerning 57 PAH-treated patients highlight a potential role of CRP in predicting response to therapy. In addition, the treated patients normalizing CRP under ULN (5 mg·l−1) had a significantly better survival accompanied by a decrease in NYHA functional class and an increase in cardiac index. The better survival observed in treated patients normalizing CRP under ULN was also associated with a higher proportion under prostacyclin analogues, in agreement with previous observations reporting an enhanced survival in patients under long-term intravenous epoprostenol infusion (23,24).
NT-proBNP decreases together with an improvement in hemodynamics in treated PAH patients (25). Consequently, dosing NT-proBNP and CRP may help to orientate therapeutic options.
CRP and surgical treatment in CTEPH
In our population of CTEPH patients who underwent a PEA, a significantly decrease in circulating CRP levels has been observed 12 months after surgery, suggesting an improvement in their inflammatory status. Similarly, Langer et al. (26) have found elevated tumor necrosis factor-alpha levels in CTEPH patients before PEA, which dropped after the surgery. Besides the decrease in tumor necrosis factor-alpha and CRP after surgical treatment in CTEPH, NT-proBNP, also used as a noninvasive marker of the severity of right ventricular dysfunction in CTEPH (27), is decreasing under oral endothelin receptor antagonist (28) or subcutaneous prostacyclin analog therapy (29) in inoperable CTEPH patients. Although CRP could not be identified as a prognostic factor to predict adverse outcome of PEA in our population of operated CTEPH patients, CRP may be useful to evaluate both medical and surgical treatment efficacy in CTEPH.
Relevance of the study
An interesting finding and unexplored aspect of the study is the predictive value of CRP for the outcome of PAH patients and its sensitivity to disease-specific medication. The noninvasive feature associated with the low cost of the measurement of CRP minimizes risks for the patients and deserves consideration to include it in the evaluation of patients with pulmonary hypertension and to orientate the therapeutic options.
The present study was exploratory and performed on consecutive patients coming to the reference center for right heart catheterization. At first line, our prospective study only included 24 PAH patients who had a second evaluation after an average period of 28 months under disease-specific medication. In order to increase this population and to strengthen our results, we have retrospectively retrieved the baseline evaluation concerning 33 patients who were already treated at inclusion. To properly evaluate the effect of treatment on inflammatory markers, we should extend these measurements to a larger number of patients evaluated at a fixed time after starting disease-specific medications or PEA, although the present results are promising.
ROC analysis indicated relatively low sensitivity and specificity for CRP in predicting survival, compared with NT-proBNP (20). However, we must emphasize that in PAH: 1) ROC analysis has never been performed for inflammatory biomarkers; and 2) we do not expect CRP to be the unique predictor of death.
In the present study, CRP did not correlate strongly with hemodynamics in PAH and CTEPH patients. Circulating monocyte chemoattractant protein-1 has been found to correlate with PVR in CTEPH patients (8), which we did not confirm in our CTEPH population. A plausible explanation could be that inflammation is not a linear evolving process in the course of the disease (i.e., early peaks of inflammation may occur without direct consequences on the hemodynamics). However, we were not able to demonstrate any relationship between the length of the disease and CRP.
The present study clearly shows that circulating CRP may predict the severity and the outcome in PAH and that its sensitivity to disease-specific medication may orientate therapeutic options.
The authors thank Peter Bynens, Viviane De Broyer, Lynn Decoster, Frederick Guns, Petra Janssens, Pieter Kelber, Erna Ruers, Hans Scheers, Evi Smeyers, Siska Van Damme, Ellen Vandevelde, and Wim Wuyts for their excellent logistical and technical support.
Dr. Delcroix received research grants and speaker fees from Actelion, Encysive, Pfizer, and LungRx.
Dr. Delcroix is holder of the Actelion chair for pulmonary hypertension at the Katholieke Universiteit Leuven. Dr. Nawrot is the recipient of a fellowship from the Fonds voor Wetenschappelijk Onderzoek-Vlaanderen.
- Abbreviations and Acronyms
- confidence interval
- C-reactive protein
- chronic thromboembolic pulmonary hypertension
- idiopathic pulmonary arterial hypertension
- N-terminal fragment of brain natriuretic peptide
- New York Heart Association
- pulmonary arterial hypertension
- mean pulmonary arterial pressure
- pulmonary endarterectomy
- pulmonary vascular resistance
- mean right atrial pressure
- receiver-operator characteristic
- upper limit of normal
- 6-min walking distance
- Received September 5, 2008.
- Revision received December 2, 2008.
- Accepted December 8, 2008.
- American College of Cardiology Foundation
- Galie N.,
- Torbicki A.,
- Barst R.,
- et al.
- Scirica B.M.,
- Morrow D.A.,
- Verma S.,
- et al.
- Hoeper M.M.,
- Markevych I.,
- Spiekerkoetter E.,
- Welte T.,
- Niedermeyer J.
- Sitbon O.,
- Humbert M.,
- Nunes H.,
- et al.
- McLaughlin V.V.,
- Shillington A.,
- Rich S.
- Andreassen A.K.,
- Wergeland R.,
- Simonsen S.,
- Geiran O.,
- Guevara C.,
- Ueland T.