Predictors of Pulmonary Hypertension in Patients With End-Stage Heart Failure

Recently, a new definition of pulmonary hypertension (PH) has been proposed. PH is defined as an increase in mean pulmonary arterial pressure at rest ≥25 mm Hg as assessed by right heart catheterization.¹ In patients with heart failure, there are 2 major components responsible for PH: hydrostatic and vasoreactive mechanisms. The hydrostatic component occurs in all cases and is related to patients with heart failure (HF). It is the passive backward transmission of the elevated left ventricular (LV) end-diastolic pressure, and elevation is presumed to be the initial event in PH development. Therefore, pulmonary artery systolic pressure correlates tightly with pulmonary capillary wedge pressure. In this situation, the transpulmonary pressure gradient (TPG) and the pulmonary vascular resistance (PVR) are within the normal range. The vasoreactive component represents vasoconstriction and remodelling of pulmonary vasculature. The mechanism occurs due to an increase in the vasomotor tone of pulmonary arteries resulting from an imbalance in the local production of nitric oxide and endothelin with an elevation in TPG and PVR (which use to be reversible under acute pharmacologic vasodilator testing) and due to a structural proliferative remodelling (medial hypertrophy and intimal proliferation) of the pulmonary resistance vessels, which is not prone to respond to vasodilator testing.² This situation is also called reactive (out-of-proportion) PH. The threshold to define reactive PH has been established as the presence of a TPG level >12 mm gH.^1,3

Elevated PVR or TPG are risk factors for mortality following heart transplantation (HT) due to the inability of the grafted heart to adapt to pre-existing PH, resulting in right ventricular failure.^4–11 Patients with high fixed PVR or TPG (unresponsive to vasodilator therapy) are often not considered candidates for HT. Although the levels of pulmonary pressure above which mortality is increased are not clearly defined (severe PH), many institutions consider a PVR >3 Wood Units (WU) and a TPG >12 mm Hg as the reasonable limits beyond which mortality risk after HT is increased.^7,12–15

Risk factors for the presence of PH in patients with end-stage HF are not clearly defined. Moreover, mechanisms that lead to reactive reversible vasoconstrictive PH or the fixed obstructive components are not well understood. The aim of the present study is to analyze the influence of different variables on the risk of PH development, including reactive PH and severe PH.

Methods

A total of 419 patients with end-stage HF with an LV ejection fraction (LVEF) <50% measured by echocardiography were evaluated for HT at Hospital Doce de Octubre in Madrid, Spain, from November 1992 to April 2002. All patients went through a standard pretransplantation protocol for evaluation to determine cardiac function and exclude contraindications. Right heart catheterization and a transthoracic echocardiogram were performed in all patients who were included in the study, as well as cardiopulmonary exercise testing, if applicable, depending on the clinical situation of the patient.

We retrospectively reviewed the medical records of these patients and examined several variables that could indicate the presence of PH. We recorded sex, age, height, weight, body mass index (BMI; defined as weight in kilograms divided by height squared in centimeters), native heart disease creating the need for HT (ischemic heart disease, idiopathic dilated cardiomyopathy, valvular heart disease, and other causes), duration of disease (defined as the time in months between the diagnosis of HF and the evaluation for transplantation), cardiovascular risk factors (hypertension, diabetes mellitus, smoking and dyslipidemia), medical treatment at the time of evaluation (angiotensin-converting enzyme inhibitors or angiotensin II receptor antagonists, nitrates, and spironolactone), primary rhythm on electrocardiogram, and peak oxygen consumption assessed by a cardiopulmonary exercise testing. Transthoracic echocardiograms were analyzed for LV end-diastolic diameter, LVEF (measured using the Teichholz or Simpson’s method), and the grade of mitral regurgitation (MR) and tricuspid regurgitation (TR) (determined using regurgitant jet area by color-flow or continuous wave Doppler recordings). Right heart catheterization was performed in all patients, at rest after the echocardiogram was performed, with a Swan-Ganz pulmonary artery catheter (Baxter Health Care Corp, Edwards Division, Santa Ana, CA). Systolic, diastolic, and mean pulmonary artery pressure (PAP); mean pulmonary capillary wedge pressure; and cardiac output were measured, the latter by thermodilution technique. TPG was measured as mean PAP minus mean pulmonary capillary wedge pressure, and PVR was equal to TPG divided by cardiac output. To determine whether there were different risk factors depending on the presence and the severity of pulmonary hypertension, we divided patients into 4 groups: (1) non-PH, if the patient presented a mean pulmonary arterial pressure <25 mm Hg; (2) PH, if the patient presented a mean pulmonary arterial pressure ≥25 mm Hg; (3) reactive PH, which was defined as the presence of a TPG >12 mm Hg; and (4) severe PH, if the patient presented a PVR >3 WU and/or a mean PAP >40 mm Hg.

Continuous variables were compared using Student t test and are presented as mean±standard deviation. Categoric variables were compared using chi-square test or Fisher exact test and are expressed as absolute number and percentages. Univariate analysis was conducted to analyze statistically significant differences in baseline demographic and clinical characteristics between patients, depending on the presence and degree of PH. Risk factors for PH were evaluated by multivariate stepwise logistic regression analysis introducing clinically significant variables and statistically significant variables obtained in the univariate analysis. All statistical analysis tests were 2-tailed and P<.05 was considered statistically significant throughout the study. Statistical analysis was performed using SPSS 14.1 (SPSS, Inc, Chicago, IL). The study was approved by the institutional ethics of human research committee.

Results

The mean age was 51.8±10.5 years and 84.3% of the patients were men. The most frequent etiology of HF was ischemic heart disease (49.9%) followed by idiopathic dilated cardiomyopathy (27.4%). The mean time of evolution of disease at evaluation was 87.7±91 months. Mean LVEF was 25.9%± 10.5%, and 136 patients (32.4%) showed moderate or severe MR on echocardiogram. Mean TPG, PVR, and PAP were 9.6±5.3 mm Hg, 2.6±1.8 WU, and 30.2±12.3 mm Hg, respectively. Prevalence of PH, reactive PH, and severe PH was 62.2%, 23.8%, and 18.8%, respectively. Fifty-nine patients (14.2%) presented reactive and severe PH at the same time. Five patients (1.1%) were excluded from the study due to incomplete data of right heart catheterization or echocardiogram. Demographics, medical history, medication, clinical data, echocardiographic characteristics, and baseline hemodynamic parameters of all patients are summarized in Table I.

Seventy-six patients (18.1%) received different pharmacologic agents to reverse severe pulmonary hypertension. Inotropes (dopamine or dobutamine) were used to treat 32 patients, nonselective vasodilators (nitroglycerin or sodium nitroprusside) were used in 20 patients, and prostacyclin was used in 24 patients. The target hemodynamic goal in the vasoreactivity test was to achieve a TPG <12 mm Hg and/or a PVR <2.5 WU. A total of 36 patients (47.3%) were considered responders to the vasoreactivity test, and 274 patients (66%) received HR (66% of patients with PH, 60% of patients with reactive PH, 64% of patients with severe PH, and 63% of patients with reactive and severe PH).

Univariate analysis showed a number of statistically significant differences between the non-PH cohort and the group of patients in which PH (PH, severe PH, and reactive PH) was detected. The non-PH group had a lower proportion of men (79.3% vs 88%, 89.8%, and 93%, respectively), a shorter duration of the disease (71.5 vs 90.7, 104, and 91.6 months, respectively), and a lower prevalence of moderate-severe MR on echocardiography (18.1% vs 41.2%, 47.2%, and 47.8%, respectively). Moreover, in the non-PH group, the prevalence of moderate-severe TR on echocardiography was lower compared with the PH group (mean PAP >25 mm Hg) (6.9% vs 19.1%).

Multivariate regression analysis is presented in Table II. We found that the presence of moderate or severe MR (odds ratio [OR], 2.1; confidence interval [CI], 95%, 1.2–3.7), moderate or severe TR (OR, 2.9; CI 95%, 1.3–6.4), and a duration of disease >3 years (OR, 1.7; CI 95%, 1.1–2.7) were independent predictors of PH. Moreover, we found that the presence of moderate-severe MR and a duration of disease >3 years were also independent risk factors for reactive and severe PH.

We compared the group of patients with reactive PH with the rest of patients with PH. No statistically significant differences were noted in any of the baseline demographic, echocardiographic, and clinical characteristics (data not shown).

Discussion

Pulmonary hypertension is associated with a poor prognosis in patients with HF and it incorporates information on diastolic function, mitral regurgitation, and pulmonary circulation.^1–3 In our study, we found a prevalence of PH of 62%, which is similar to previous reports. Reactive PH has been recently defined as the presence of a TPG >12 mm Hg and denotes a physiopathologic change in pulmonary vasculature. We found a prevalence of reactive PH of 23.8%. Severe PH emerged as an independent prognostic factor for mortality after heart transplantation, primarily related to postoperative acute right ventricular failure.^6,16–18 The most generally accepted risk threshold at baseline or after a vasodilation test are the presence of TPG >12 mm Hg and/or PVR >3 WU, although there are still conflicting data regarding the degree of PH associated with an unfavorable prognosis. We based our definition of severe PH on previous reports and found a prevalence of 18.8%.^{5,7,11–15,19–21} We aimed to investigate risk factors for PH in patients with systolic heart failure in order to improve our knowledge of PH pathophysiology and to detect patients with a higher risk for PH development, suggesting a closer follow-up in these patients.

Initially, in patients with HF, PH development is produced due to an elevation in the left ventricular filling pressure, which results in a “passive” increase in pulmonary venous pressure. This reactive increase is secondary to pulmonary vasoconstriction and is readily reversed by vasodilators. Subsequently, a structural remodelling on the arterial wall due to abnormalities of the elastic fibers, intimal fibrosis, and medial hypertrophy is developed, leading to a fixed PH with reduced response to vasodilators.^1,22 Other myocardial changes related to HF such as left atrial enlargement, ventricular hypertrophy, or mitral regurgitation can also facilitate PH development. Although the association between all these changes and time is intuitive, data supporting it are scarce. Our results suggest that the timing of disease development plays an important role and is a risk factor for PH development. We found a wide range of disease duration due to the different development patterns associated with each etiology of HF (from <1 month in patients with myocarditis or HF related to acute coronary syndromes to >10 years in patients with advanced valvular disease). Therefore, the definition of “time of evolution” may not be very precise except in the case of ischemic heart disease, which usually begins at the moment of the first episode of acute myocardial infarction, which causes LV dysfunction. The remainder of the etiologies are detected when symptoms appear.

In the present study we found that the presence of moderate or severe MR was an independent predictor of PH. In a previous report, Enriquez-Sarano and colleagues²³ studied echocardiographic parameters that predict PH and described the association between the degree of MR and PH. Mitral regurgitation may induce LV volume overload, which, in turn, may affect diastolic filling, which is directly related to the degree of PH. We did not find any relation between PH and the degree of LVEF, which is congruent with previous studies.^23–25

There is little work in the literature examining the demographic and clinical determinants of PH in patients with end-stage HF and LV dysfunction. Detection of clinical predictors for PH was not the main focus of any of the previous studies and multivariate analysis was not performed. Goland and colleagues²⁶ evaluated 410 patients with end-stage HF, defining the threshold for severe PH as TPG >10 mm Hg and/or PVR >3 WU. They did not find differences by age, sex, BMI, presence of diabetes, serum creatinine, and etiology between patients with and without severe PH.²⁶ These findings are in agreement with our results. Klotz and colleagues²⁷ studied 151 patients with HF. Severe PH was defined as a TPG >12 mm Hg and/or PVR >2.5 WU. They compared both groups (with and without severe PH) for age, sex, etiology, peak oxygen consumption, LVEF, and medical treatment. They found statistically significant differences in peak oxygen consumption (14.8 vs 12.2 mL/ kg/min).²⁷ In the present study we did not find statistically significant differences in the peak oxygen consumption between groups in the univariate analysis, probably because cardiopulmonary exercise testing was performed in only 187 patients due to disease severity. However, the relationship between the degree of PH and the functional class has been previously described.^19,28 Lindelow and colleagues²⁹ found that ischemic heart disease was more frequent in patients with severe PH (defined as PVR >3 WU) than in patients with idiopathic dilated cardiomyopathy (50% vs 27%). In contrast to this report, we did not find any differences between etiologies. This could be related to the greater sample size of our study. Furthermore, we included all relevant causes of end-stage HF. Khush and colleagues²⁰ analyzed the Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness (ESCAPE) database and they found that patients with PH were older than those without PH (mean age, 59±12 vs 54±15 years; P=.022). There were no significant differences in baseline medical therapy; prevalence of hypertension, diabetes, or ischemic cardiomyopathy; BMI; dyspnea score (rated from 0 to 3); and physical signs of heart failure between these two groups.²⁰ These results are consonant with ours. To our knowledge this is the first study that analyzes risk factors of PH using the new criteria recently proposed.¹

A recent retrospective study analyzed risk factors of PH in patients with HF and preserved ejection fraction (LVEF ≥50%). Prevalence of PH was 52.5%. They found that an age older than 80 years, a BMI >40 kg/m², and the presence of atrial arrhythmias were found to be risk factors for PH development. Our data do not match with the results of the above-mentioned study, probably because the profile of patients with HF and preserved ejection fraction (advanced age, hypertension, diabetes, obesity, and other cardiovascular risk factors) differs from the profile of patients with systolic heart failure candidates for HT (eg, younger patients with lower weight due to cardiac cachexia).³⁰

Pulmonary arterial pressure and TPG threshold for defining different degrees of PH was based on previous reports.^1,12–15,20 Although it is known that there are different physiopathologic mechanisms and anatomical changes depending on the degree of PH,²² we did not find clinical or statistically significant differences between patients with different grades of PH (PH, reactive PH, and severe PH).

Limitations

Our study suffers from similar potential limitations of any retrospective analysis. We believe that this approach was appropriate, however, given the number of patients and the comprehensiveness and duration of recruitment. Caution is necessary in interpreting these results. Association does not imply causation, and for many of the significant associations that we have found, there is a reciprocal relationship.

Conclusions

The presence of moderate-severe MR and the time of evolution of disease are predictors of PH development in patients with HF and LV systolic dysfunction. Physicians should be aware of these factors during evaluation and treatment of end-stage HF patients and HT candidates.