Pulmonary hypertension is frequently observed in patients with dilated cardiomyopathy and heart failure. This results from passive transmission of elevated left ventricle end-diastolic pressure, and eventually reactive pulmonary vasoconstriction occurs [1]. It is considered a relative contraindication to cardiac transplantation listing, if irreversible [2], and is associated with higher mortality and morbidity in heart transplant candidates, recipients [3], and other patients [4,5].

According to Costard-Jackle et al. [6], and Drakos et al. [7] vasoreactivity testing with intravenous sodium nitroprusside in heart transplant candidates, identifies a subgroup of patients with excellent post-transplant prognosis.

Inhaled nitric oxide (iNO), more specific pulmonary vasodilator, increases the concentration of cyclic guanosine monophosphate (cGMP) in vascular smooth muscle cells, resulting in pulmonary arteries vasorelaxation [8]. INO administered to patients with heart failure, in contrast to intravenous sodium nitroprusside, reduces selectively pulmonary vascular resistance without influence on systemic arterial pressure or systemic resistance [9].

Sildenafil citrate, a phosphodiesterase-5 inhibitor, leads to accumulation of cGMP, and it has been proven to be selective for pulmonary circulation [10]. A single oral dose of sildenafil has also been proven to be as effective pulmonary vasodilator as iNO, and their combination synergistically increases cGMP level [11].

Several papers have concerned the use of various drugs for testing of pulmonary hypertension reversibility in a pre-transplant evaluation [12-14]. The results of these studies were inconsistent. Some favored either prostaglandin or prostacyclin, and others favored iNO. It is of note that the major limitation of these studies was the small number of enrolled patients. Moreover, the issue of the predictive value of vasodilator response in patients with dilated cardiomyopathy complicated with pulmonary hypertension remains unanswered.

Therefore, we designed and conducted this prospective cohort study to examine the utility of vasodilator response to three different agents and a combination thereof (oral sildenafil and iNO) in predicting long-term survival in patients with dilated cardiomyopathy and pulmonary hypertension.

Of 150 consecutive patients with dilated cardiomyopathy (idiopathic or post inflammatory) referred for management of heart failure, 111 patients were subject to routine procedure of periodical hemodynamic assessment as a part of heart transplantation evaluation between May 2005 and February 2009. We identified 79 patients with pulmonary postcapillary hypertension defined as mPAP>25 mmHg and PWP>15 mmHg. Of these patients, we selected 29 patients (6 women) with TPG>12 mm Hg (24 patients) and/or PVRI>5 WU/m² (5 patients) for enrollment into the study (Table I) [15]. All of the patients underwent right heart catheterization and a single session of pulmonary artery vasoreactivity testing with sodium nitroprusside, iNO (120 ppm), and oral sildenafil (50 mg) and repeat inhalation of nitric oxide.

Table 1: Demographic, echocardiographic and baseline hemodynamic (B-L) characteristics of patients stratified by vital status at completion of follow-up.

Inclusion criteria were dilated cardiomyopathy (history of heart failure longer than six months), left ventricle ejection fraction LVEF<35%, mPAP at least 25 mmHg supine at rest and no angiographic evidence of coronary artery disease. All patients were examined for significant valvular/congenital heart disease, requiring surgical repair - none of the enrolled patients had the echocardiographic evidence of significant valvular pathology. All patients were on optimal pharmacological treatment for heart failure (Table 2), and the treatment was not modified within a month before enrollment into the study. The exclusion criteria were: previous use of sildenafil, recent heart failure worsening, and diseases leading to pulmonary hypertension such as severe obstructive lung disease, sleep apnea syndrome, or pulmonary embolism. In the follow-up, patients were evaluated every three to six months on an ambulatory basis for 30 months. The study endpoint was death in the follow-up. The patients who underwent cardiac transplantation during follow-up were withdrawn as alive cases from further observation (censored observation).

Table 2: Baseline medical treatment stratified by vital status at completion of follow-up.

The protocol was approved by the Ethics Committee of the Medical University of Silesia. Written, informed consent was obtained from all enrolled patients before screening.

Right heart catheterization

Patients underwent right heart catheterization via right jugular access under local anesthesia in a supine position at rest. A 7F introducer (Balton, Warsaw, Poland) was inserted and then flow-directed Swan-Ganz catheter (Edwards Lifesciences, Irvine, USA) was advanced to right heart and pulmonary artery under fluoroscopy. After 10 minutes of stabilization of circulation parameters pulmonary wedge pressure (PWP), systolic pulmonary artery pressure (sPAP), diastolic pulmonary artery pressure (dPAP) and right atrium pressure (RA) were measured. Cardiac output (CO) was measured by thermodilution using rapid bolus injection of 10 cc of cold saline. Then cardiac output was indexed by body surface area and expressed as cardiac index (CI). Systolic (sABP) and diastolic (dBAP) systemic arterial pressure were measured non-invasively. For each examined drug PWP was measured two times, before and after CO measurements and systemic arterial pressure was measured three times: before, during third and after fifth CO measurements. Other hemodynamic parameters were acquired five times during each CI measurement-mean values were used for final evaluation. Acquired data enabled calculation of mean pulmonary artery pressure (mPAP) and mean systemic arterial pressure (mABP), transpulmonary gradient (TPG), pulmonary vascular resistance index (PVRI) and systemic vascular resistance index (SVRI):

•mPAP (mm Hg) equals the sum of dPAP and one third of a subtraction of sPAP and dPAP in pulmonary artery (mPAP=dPAP+[sPAP-dPAP]/3)
•TPG (mm Hg) equals difference of mPAP and PWP (TPG=mPAP-PWP)
•PVRI (WU/m²) equals quotient of TPG and CI (PVRI=TPG / CI)
•mABP (mm Hg) equals the sum of diastolic arterial blood pressure (dABP) and one third of a subtraction of sABP and Dabp (mABP=dABP+[sABP-dABP]/3)
•SVRI (WU/m²) equals quotient of subtraction of mABP, RAP and CI ([mABP-RAP]/CI).
Blood pressure parameters were expressed in millimeters of mercury (mm Hg), CI as liters per minute per square meter (l/min/m²), heart rate as number of heart beats per minute. Measured parameters of resistance (PVRI, SVRI) were expressed in WU/m².

The ratio of pulmonary vascular resistance index to systemic vascular resistance index (PVRI/SVRI) served as a measure of pulmonary selectivity of examined drugs. In brief, drug leading to reduction of the ratio was assumed to be selective for pulmonary circulation.

Protocol of vasoreactivity testing

The hemodynamic measurements were made at baseline (NTP B-L), and then, via the jugular sheath, an intravenous infusion of maximum tolerated dose of sodium nitroprusside (1-4 μg/kg/min) was started. NTP solution was prepared immediately before administration and we used infusion pumps with light resistant amber syringes, and IV lines to prevent NTP breakdown. The infusion of NTP was interrupted when TPG decreased below 12 mmHg or systemic systolic arterial blood pressure dropped below 85 mm Hg, or when reaching a maximum dose of 4 μg/kg/min. Hemodynamic data were monitored continuously, and the data at the peak nitroprusside activity (NTP) were recorded for further analysis. The duration of nitroprusside infusion did not exceed 10 min throughout the study. Ten minutes were allowed for washout of sodium nitroprusside, and repeat hemodynamic measurements were taken (NO B-L). Subsequently, patients were administered iNO 120 parts per million (ppm). iNO (INO_max 400 ppm mol/mol, INO Therapeutics AB, Lidingö, Sweden) was administered via a face mask from a source tank equipped with a microflow gauge (Linde France SA, Malmaison, France). Ten minutes of iNO inhalation were allowed before consecutive hemodynamic recordings were performed (NO). After measurements, iNO inhalation was terminated, and 10 min were allowed for NO washout. Its half-life is approximately 3 min [11]. Afterward, third baseline measurements (and fifth consecutive) were recorded (SIL B-L), and 50 mg of sildenafil citrate (Viagra-Pfizer) was administered orally. Fifty minutes were allowed for sildenafil to start its activity, and consecutive hemodynamic measurements were recorded (SIL). Afterward, the final NO inhalation at 120 ppm was administered for 10 min with a final (seventh consecutive) hemodynamic measurement (SIL/NO) recording [16,17]. All measurements were performed during end expiration. At each measuring point, five results were averaged, and an arithmetic mean was recorded. It was assumed that pulmonary hypertension was reversible with a successful reduction of PVRI<5 WU/m² and TPG<12 mmHg. Such an order of vasodilator drug testing enabled us, in our opinion, to avoid unintentional drug interaction. The entire right heart catheterization and vasoreactivity testing took no longer than 2 h. The feasibility and safety of such a protocol was assessed in a pilot study and reported previously [18].

Statistical analysis

Continuous variables are presented as the mean ± standard deviation and were compared with Mann-Whitney U test. Categorical data are presented as absolute numbers and percentages and were compared using χ² tests. Survival analysis of all death and censored cases, based on Kaplan-Meier curves and log-rank tests were used to assess the event-free survival between responders and non-responders. Cox proportional hazards regression analysis was applied to identify the variables associated with long-term outcome in these patients. All demographic, clinical, echocardiographic, and hemodynamic variables and whether the patient was a vasodilator responder were included in a univariate Cox analysis, but only variables with a value of P ≤ 0.1 at univariate analysis were included in the multivariate model. Post-hoc power calculations for the primary endpoint were computed either. A two-tailed P value<0.05 was considered significant. Statistical analysis was performed with Statistica 8.0 software (Statsoft Inc., Tulsa, USA).

Table 1 presents the clinical, echocardiographic and baseline hemodynamic data stratified according to participants’ vital status at the completion of follow-up. There were eight deaths (27.6%) in the follow-up: one out of a group of five patients with TPG<12, seven in group of patients with TPG ≥ 12 mm Hg (difference is not significant P=0.8944, χ² with Yates correction). The all cohort of decedents had a significantly lower left ventricle ejection fraction, and there were significantly fewer responders to inhaled nitric oxide and a combination of oral sildenafil and inhaled nitric oxide in this cohort. Four patients had chronic atrial fibrillation: three out of survivors (14.3%) and one out of decedents (12.5%, the difference is not significant P=0.6328, χ² with Yates correction). There were five patients with implanted ICD, 1 patient with implanted CRTP and 1 patient with implanted CRTD in the cohort of survivors. There were 3 patients with implanted CRTP in the decedent’s cohort. In this regard the cohorts did not differ.

Both cohorts of survivors and decedents were treated similarly at baseline (Table 2), and none of the patients received sildenafil citrate in the follow-up.

Table 3 presents hemodynamic measurements obtained during right heart catheterization and vasoreactivity testing with studied drugs. Sodium nitroprusside was administered intravenously at a mean dose of 1.83 ± 0.8 μg/kg/min. Eight patients had an initial mABP below 85 mmHg, and nitroprusside infusion led to a decrease of mABP below 85 mmHg in another 18 patients. Twenty patients (69.0%) met the reversibility criteria.

Table 3: Hemodynamic measurements obtained during right heart catheterization and vasoreactivity testing with studied drugs.

During inhalation of nitric oxide nineteen patients (65.5%) were classified as responders. Oral sildenafil disclosed eighteen responders (62.1%). The combination of oral sildenafil and iNO proved to have a mixed effect on hemodynamic measurements. INO abolished the CI increase and SVRI reduction caused by sildenafil and led to a massive consecutive reduction of TPG and PVRI/SVRI ratio to values comparable to iNO alone (Table 3, columns NO and SIL/NO). Nineteen patients (65.5%) were classified as responders to the combination of sildenafil and iNO. The right heart catheterization with vasoreactivity challenge was well tolerated, and none of the patients developed any adverse effects.

Stratification of patients’ baseline hemodynamic measurements according to vasodilator response (Table 4) revealed that the cohort of responders to each of the studied drug had a significantly lower TPG and PVRI. Responders to studied drugs had similar clinical and echocardiographic characteristics (data not shown).

Table 4: Baseline hemodynamic measurements stratified by vasodilator response to studied.

There were eight deaths observed in the follow-up (the mean follow-up was 1.61 ± 0.99 years). Kaplan-Meier event-free survival curves of responders versus non-responders to studied drugs are shown in Figure 1. Survival analysis demonstrates improved survival for responders to iNO and a combination of oral sildenafil and iNO (Figure 1, panels B and D). The following clinical, echocardiographic, hemodynamic variables: patients’ sex, age, duration of the disease, LVEDD, LVEDV, LVEF, NTproBNP, creatinine, 6-MWT, NYHA classification, mABP, mPAP, TPG, PAWP, PVRI, and SVRI were used in the univariate Cox proportional hazard analysis. Moreover, the information on whether the patient was a responder to a study drug (NTP non-responder, iNO non-responder, SIL non-responder, SIL/NO non-responder) was included in the univariate proportional hazards model (Table 5). In the final multivariate model, left ventricle ejection fraction, creatinine level, NYHA class, being non-responder to inhaled nitric oxide and being a non-responder to a combination of oral sildenafil and iNO were included. Being a responder to iNO is the only variable that predicted improved survival according to the multivariate Cox proportional hazard analysis. Contrary to this, non-responders to iNO have a death hazard ratio of 11.77 with a 95% CI 1.117-123.9 (P=0.04). The post-hoc power computations yielded the power of 0.56 for the smaller group of iNO nor-responders (n=10) and power of 0.83 for the larger group of responders to iNO (n=19).

Figure 1: Panel A: Kaplan-Meier death free survival in 30-months long follow-up depending on pulmonary hypertension reversibility tested with nitroprusside. Panel B: Kaplan-Meier death free survival in 30-months long follow-up depending on pulmonary hypertension reversibility tested with nitric oxide. Panel C: Kaplan- Meier death free survival in 30-months long follow-up depending on pulmonary hypertension reversibility tested with sildenafil. Panel D: Kaplan-Meier death free survival in 30-months long follow-up depending on pulmonary hypertension reversibility tested with sildenafil and repeat inhalation of nitric oxide.

Table 5: Univariate and multivariate Cox proportional hazard analysis. Data with P ≤ 0,1 at univariate analysis were included into multivariate model.

Patients with implanted electrotherapy devices (ICD, CRTP, CRTD) did not show improved survival in our study. In the follow-up, six patients underwent successful cardiac transplantation, and as per protocol, they were considered alive and they were withdrawn from further observation.

Figure 2 shows detailed scheme of vasoreactivity testing along with study follow-up. It summarizes that different studied drugs allowed identifying different patient populations. Of note is a case of female patient, non-responder to all of the studied drugs, who underwent heart transplantation. The results of her vasoreactivity testing were as follows: baseline TPG 12.61 mmHg, PVRI 7.11WU/m²; after nitroprusside TPG 13.41 mmHg, PVRI 4,38 WU/m²; after iNO TPG 10.03 mmHg, PVRI 7.02 WU/m²; after SIL TPG 14.41 mmHg, PVRI 5.57 WU/m², and after SIL/NO TPG 11.87 mmHg, PVRI 6.3 WU/m². As seen above, none of the studied drugs proved to be effective in simultaneous reduction of both parameters to the level defined as reversible pulmonary hypertension.

Figure 2: Detailed scheme of vasoreactivity to different drugs along with outcomes that occurred in the follow-up.

In our prospective cohort study of patients with dilated cardiomyopathy and mixed pulmonary hypertension, we proved that acute responders to vasodilator challenge with iNO have improved survival. The administration of neither intravenous sodium nitroprusside nor oral sildenafil was predictive of survival in our cohort of patients. All of the drugs used in our study have demonstrated similar efficacy in pulmonary hypertension reversibility, although each of the drugs tested has unique vascular activity and hemodynamic properties. Two recent reports on the ability of iNO to predict improved prognosis in adult patients with chronic thromboembolic pulmonary hypertension [19] and in a large cohort of Dana Point class 1, 3, 4, and 5 pulmonary hypertension patients [20] are available. Therefore, the results of our study are mutually supplementary with their observations and support the evidence that acute response to iNO predicts improved survival in pulmonary hypertension patients, independently of its etiology.

The common denominator shared by our study and the observations of Krasuski et al. [20], acute vasodilator response to inhaled nitric oxide, requires further explanation. The first issue is the reversibility of pulmonary hypertension, and the second issue is iNO by itself. Chronically increased left ventricle end-diastolic and left atrial pressures are the driving force for a vicious cycle of continuous remodelling and functional derangements, which eventually leads to increases in pulmonary resistance and reduced pulmonary arterial compliance [21].

The significantly increased baseline TPG and PVRI in a cohort of non-responders to studied drugs in our experiment (Table 4) support the evidence of the progressive nature of pulmonary hypertension in left heart disease and indicate that preserved reversibility of hypertension is of paramount importance in improved survival. Moreover, Krasuski et al. [20] reported a similar distribution of baseline PVR values after stratification of the entire study population to responders, non-responders, survivors and decedents. This observation might suggest some other factors, beyond preserved reversibility, contribute to an increased risk of mortality in patients with pulmonary hypertension and that we have not controlled for these factors in our experiment. Literature data provide evidence on right ventricle function [22], plasma big endothelin-1 level and natriuretic peptides [23-25] as other prognostic factors that might be suitable for precise risk stratification in patients with heart failure and pulmonary hypertension. Although, Wierzbicki et al. [25] reported significant correlations between pulmonary artery pressures and blood levels of NTproBNP in a cohort of patients scheduled for heart transplantation, and Fonarow et al. [24] indicated, that BNP level on admission is positively and linearly associated with in-hospital mortality, we failed to prove the prognostic ability of NTproBNP to predict mortality in our cohort. Szymik et al. [26] have reported on prognostic ability of left ventricle ejection fraction, NTproBNP, plasma glucose and bilirubin levels to predict three year survival. The discrepancies reported in the above mentioned publications and in our study may be explained by different patient populations studied. Recently a report on prognostic value of soluble ST2 protein in predicting composite endpoint of cardiovascular death, heart failure hospitalization, increase in NYHA class and increase in diuretics use was published [27].

Ghio et al. [28] have reported that right ventricle function has prognostic significance in patients with chronic heart failure. Thus we might speculate that the vasodilative properties of iNO to increase PWP and secondary decrease of TPG and PVRI are the major predictors of reversible reactive component of pulmonary hypertension. The advantage of iNO over other tested drugs may result from redistribution of blood from precapillary to postcapillary pulmonary capacitance vessels with a consequent increase in left ventricle end-diastolic volume and filling pressure. This would have caused an increase in left ventricle stroke volume in normal but not in severily depressed left ventricle function and iNO might be considered as a drug inducing inotropic negative effect on failing left ventricle [29].

Inhaled nitric oxide is the only drug, out of the three drugs tested in our experiment, found to be predictive of improved survival in patients with dilated cardiomyopathy and pulmonary hypertension. D’Alto et al. [30] reported on the reversibility testing with epoprostenol infusion in patients with Eisenmenger syndrome. They accepted clinical worsening as a study end point, and they did not prove epoprostenol per se was predictive of improved patient prognosis. The authors concluded that the only independent predictor of clinical worsening in a group of their patients was change in PVRI (ΔPVRI).

We have not performed any further analyses that could explain this unique property of inhaled nitric oxide, but it is apparent that inhaled nitric oxide at 120 ppm has the highest pulmonary specificity in comparison with sodium nitroprusside and oral sildenafil. Its pulmonary specificity might result from its extreme short half-life time and the fact that it is degraded after passing pulmonary circulation. It is also of note that the dose of 120 ppm used in our experiment is the highest of those reported previously [31]. Such a hemodynamic vasodilator response is in contrast with that reported by Lepore et al. [32] for 80 ppm iNO.

Our results indicate the equal acute efficacy of sildenafil to reverse the pulmonary hypertension in comparison with infused sodium nitroprusside and inhaled nitric oxide but fail to indicate sildenafil prognostic ability. This lack of prognostic properties in our experiment may be explained by alternative mechanism of sildenafil action. It keeps the high level of cGMP, but earlier the natural stimuli have to activate its production. Nevertheless, sildenafil is of great hope for the management of chronic heart failure. The drug has been proven to have beneficial effects on hemodynamics, clinical status, cardiac cachexia, and on improved peri-transplant survival in chronic treatment of patients with advanced heart failure and severe pulmonary hypertension [33,34].

Our study has demonstrated that vasoreactivity testing with multiple drugs is feasible and safe, but from a prognostic perspective, there is no benefit of continuing the examination beyond intravenous infusion of sodium nitroprusside and inhaled nitric oxide. Administration of oral sildenafil prolongs the examination and yields no additional predictive value.

We are aware of some limitations of our study. We have reported the results of a single center study, and thus, we had a relatively small number of patients. Moreover, the small amount of patients is the result of selection of patients with idiopathic or postinflammatory dilated cardiomyopathy complicated with venous pulmonary hypertension only. This pre selection allowed us to avoid bias resulting from the influence of myocardial ischaemia and/or congenital or acquired valvular disease on the prognosis.

Vasodilator response to inhaled nitric oxide predicts longer survival in patients with dilated cardiomyopathy complicated with pulmonary hypertension. We postulate that nitroprusside and inhaled nitric oxide should be routinely used in vasoreactivity testing: nitroprusside for its established prognostic significance in post transplantation period and nitric oxide for its predictive value in pre-transplantation period.

All of the authors contributed equally to the design of the study, collection and analysis of data, edition and preparation of final version of manuscript.

This research was supported by Medical University of Silesia grants KNW-085-08 and NN-1-063/07, granted to WJ.