Long-term outcomes of Eisenmenger syndrome in the era of targeted pulmonary vasodilator therapy

INTRODUCTION

Eisenmenger syndrome (ES) is pulmonary hypertension at or near the systemic level with a reversed or bidirectional shunt between the pulmonary and systemic circulation and pulmonary vascular resistance (PVR) above 10 Wood units.¹ Long-term follow-up studies showed fair intermediate-term survival and poor long-term survival in ES.² These observations were made before the era of targeted pulmonary vasodilator therapy. Pulmonary vasodilators such as endothelin receptor antagonists and phosphodiesterase inhibitors showed improved exercise capacity and haemodynamics without compromising peripheral oxygen saturation in patients with ES.³•⁴ Newer studies on the survival of patients with ES in the era of targeted therapy have shown improved survival associated with disease modifier therapy. No randomized controlled studies have been done to prove the mortality benefit of targeted pulmonary vasodilator therapy; existing treatment recommendations are from retrospective studies. There is limited data from studies that include baseline haemodynamic data and patients with complex cyanotic lesions with ES physiology. We studied the survival pattern,•mortality predictors and outcome of targeted pulmonary vasodilator therapy in patients with ES.

METHODS

This was a single-centre retrospective study conducted at a tertiary care referral hospital. All consecutive patients with clinical, echocardiographic, and/or cardiac catheterization diagnoses of ES between January 2000 and January 2020 were considered for inclusion. Those who underwent palliative surgical shunts, stenotic and/or had tortuous and/or non-uniform pulmonary artery segmental branching, and/or unknown mean PVR of major aortopulmonary collaterals in patients with pulmonary atresia were excluded. Shunt lesions were classified as pre-tricuspid, post-tricuspid and complex defects.⁵ Demo-graphic data, the WHO functional class at diagnosis and follow-up, resting oxygen saturation at diagnosis and on follow-up, performed through pulse oximetry or from arterial blood gas analysis and 6-minute walk distance were retrieved. PVR was estimated from cardiac catheterization data using Fick’s principle. The use of pulmonary hypertension-specific medications (phosphodiesterase-5 inhibitor, endothelin receptor antagonist) and drug compliance were assessed during follow-up. Follow-up clinical outcomes studied included right ventricular dysfunction, clinical heart failure defined according to American Heart Association (AHA) 2018 criteria for congenital heart disease patients onstage C and D,⁶ arrhythmias (documented atrial or ventricular arrhythmias on electrocardiogram [ECG] or Holter study), embolic episodes, brain abscess, infective endocarditis and all-cause and cause-specific mortality.

RESULTS

A total of 206 patients with ES were identified and constituted the study cohort (136 females, 66%; Fig. 1). Haemodynamic data confirming systemic pulmonary artery pressure were available for 156 patients, and the remaining 50 patients were adjudged to have ES physiology from clinical assessment and echocardiography. The age of presentation varied from 3 months to 65 years (mean 23.4 [14.3] years). Those with pre-tricuspid shunts developed ES later (34.2 [12.3] years) as compared to post-tricuspid shunts (20.3 [12.1] years) and those with complex defects (10.7 [7.6] years; p=0.001). Overall, post-tricuspid shunts accounted for nearly half of the patients of ES (98, 47.6%), followed by pre-tricuspid shunts (34.4%) and complex defects (18%). However, the most common individual cause of ES was atrial septal defect, in 68 cases (33%), followed by ventricular septal defect, in 55 cases (26.6%) and patent ductus arteriosus in 34 cases (16.5%; Fig. 2).

Close

FIG 1.

Flow chart showing the total number of patients enrolled and followed up

Export to PPT

Close

FIG 2.

Cardiac defects associated with Eisenmenger syndrome. Y-axis shows number of patients in each group. ASD atrial septal defect VSD ventricular septal defect PDA patent ductus arteriosus AVCD atrioventricular septal defect SV single ventricle APW aortopulmonary window VSD PA MAPCA ventricular septal defect pulmonary atresia with major aortopulmonary collaterals dependent pulmonary circulation with documented elevated pulmonary vascular resistance TGA transposition of great arteries TAPVC total anomalous pulmonary venous connection TRUNCUS truncus arteriosus

Export to PPT

Effort intolerance, palpitation, haemoptysis, pedal oedema and syncope were the common presenting symptoms of ES: 64% had clinical cyanosis and 40% had cardiomegaly (Table 1).

The mean (SD) resting oxygen saturation at the time of diagnosis for all patients was 87.7 (6.3)%, pre-tricuspid shunt 88.1 (5.12)%, post-tricuspid shunt 89.6 (5.1)% and complex defects 81.7 (8.3)%. The lowest resting baseline oxygen saturation was noted in those with complex defects (p=0.001). One hundred and seventy-one (83%) patients were in WHO functional classes I and II and 35 (17%) in WHO functional classes III and IV at the time of diagnosis. Patients were followed up for a median duration of 6 years (1.5-11.5 years). We observed significant desaturation and worsening of functional class on follow-up and erythrocytosis as a response to worsening hypoxia (Table 2). Cardiac catheterization data showed aortic saturation <88% was predictive of PVR of >10 Wood units with 95% sensitivity and 88% specificity, which indicates inoperability in pre-tricuspid and post-tricuspid level shunts, excluding complex cyanotic diseases.

One hundred and thirty-six patients (76%) were on disease modifier therapy, 28.6% received combination therapy (phospho-diesterase-5 inhibitor and endothelin receptor antagonist). Patients who received disease modifier therapy had significantly better functional class at a mean follow-up of 85 months as compared to those who were treatment naïve (p=0.001). However, the change in oxygen saturation was not significantly different in either group (1.62% in patients who received disease modifier therapy as compared to 2.11% in treatment naïve; p=0.612). Disease modifier therapy was associated with significantly better survival at a mean of 85 months after diagnosis of ES (62.7%) as compared to treatment naïve patients (21.4%; p=0.001).

Complications observed during follow-up included (Table 2) had right ventricular (RV) dysfunction (32%), and it was more common in ES secondary to pre-tricuspid shunts (p=0.001). Atrial arrhythmia burden was also most common in pre-tricuspid shunt ES (p=0.001). The mean (SD) age of onset of atrial arrhythmia in the pre-tricuspid ES group was 43.6 (8.9) years. The incidence of brain abscess tended to be more common in ES secondary to complex defects (p=0.21). 39% of patients developed congestive heart failure after establishing ES. More than half the patients in the pre-tricuspid and complex groups developed clinical heart failure on follow-up.

Fifty-six (31.5%) patients died during follow-up. Heart failure was the most common cause of death, followed by haemoptysis (Fig. 3).

Close

FIG 3.

Cause-specific mortality shows a percentage of total deaths

Export to PPT

The actuarial survival of ES by the Kaplan-Meier analysis at the end of 1,3, 5, 10, and 20 years of diagnosis was 96%, 89%, 85%, 77% and 69%, respectively. Post-tricuspid shunt lesions had the best long-term survival since diagnosis of ES as compared to pre-tricuspid lesions and those with complex defects (log-rank p<0.001; Fig. 4a). While 89% of patients with post-tricuspid shunts and ES survived 5 years after the diagnosis of ES, this statistic was only 79% with either pretricuspid shunts or complex defects. The 10-year survival of post-tricuspid ES was 82% as opposed to 65% and 68% with pre-tricuspid lesions and complex defects, respectively. The survival curves of pre-tricuspid lesions with ES and complex defects with ES tended to be similar from the time of diagnosis of ES (log-rank p=0.37). The mean age at death was lowest for ES with complex defect (17.4 years), as compared to posttricuspid shunts (33 years) and pre-tricuspid shunts (42 years; p<0.001).

Close

FIG 4.

(a) Survival curve of Eisenmenger patients based on shunt type. (b) Kaplan-Meier analysis of propensity-matched group which received disease modifier treatment and treatment naïve patients

Export to PPT

Propensity score regression adjustment for baseline clinical differences showed the survival benefit of disease modifier therapy (Fig. 4b). Survival at 5 and 10 years in the treatment group was 84% and 69%, respectively and in the treatment naïve group was 66% and 44%, respectively.

A Cox multivariate logistic regression was performed on clinical variables selected as predictors for mortality. A systemic oxygen saturation <80% (95% sensitive and 65% specific for mortality), the presence of atrial arrhythmias, and the absence of disease modifier therapy were identified as independent predictors of mortality in ES (Tables 3 and 4).

DISCUSSION

We examined the patterns of survival of ES in the era of targeted pulmonary vasodilator therapy. The availability of cardiac catheterization data of the patients in our study constitutes one of its distinct advantages. The survival curves of ES in our study are better than the reported survival of patients with primary pulmonary hypertension.⁷ However, the overall long-term prognosis was still dismal, with a 23% attrition rate 10 years from the diagnosis of ES. ES secondary to complex defects had the worst prognosis among all subsets.

Single or dual-targeted pulmonary vasodilator therapy was associated with better survival at intermediate and longterm follow-ups. Propensity score analysis adjusted for baseline clinical differences showed the survival benefit of disease modifier therapy. This observation is in line with various registry data.^8-10 We did not identify an incremental survival benefit with dual-targeted pulmonary vasodilators in ES. This finding has been reported earlier as well, but interpretations have varied.¹⁰ However, our study was not primarily designed for comparing the efficacy between monotherapy and dual therapy. Escalation of therapy has been observed to result in clinical stabilization and improved effort tolerance in ES.¹¹ We did not observe any detrimental effect with dual therapy as compared to monotherapy. The sicker patients likely had earlier escalation to dual therapy which led to a survival similar to those on monotherapy.^12-14

We observed a poor intermediate and long-term survival pattern in ES patients with pre-tricuspid shunts and complex lesions, whereas post-tricuspid shunts had comparatively better survival patterns. Pre-tricuspid shunt patients develop ES later in life compared to post-tricuspid shunt patients. The onset of ES physiology in pre-tricuspid and complex lesions has a similar downhill clinical course and poor prognosis. Loss of foetal phenotype,¹⁵ volume-overloaded sub-pulmonary ventricle and incapability of the dysfunctional RV to adapt to high after load could lead to heart failure and mortality.¹⁶ The lowest systemic saturation observed in the complex shunt lesion can also contribute to the development of ventricular dysfunction.

Heart failure (59%) was the most common cause of death, followed by haemoptysis (16%), as reported in other studies in the disease modifier therapy era. This observation is in contrast to pre-disease modifier therapy era studies from the subcontinent, where the common causes of death in ES were sudden cardiac death (33%) followed by heart failure (29%) and haemoptysis (2%). Five and ten-year survival comparison in all ES patients in the disease modifier therapy era showed similar short and intermediate-term survival at 85% and 77%, respectively compared to 87% and 79%, respectively of patients in the pre-disease modifier therapy era. The lack of survival benefit in all patients with ES is likely due to the predominant proportion of pre-tricuspid shunt and the inclusion of complex defects with ES; both groups portray a poor long-term survival. The inclusion of patients who were not on disease modifier treatment for the survival analysis could also contribute to this finding. However, the propensity-matched analysis showed 5 years to 10 years of survival in the treatment group was 84%, and 69%, respectively as compared to 66% and 44%, respectively, in the treatment naïve group.

Resting oxygen saturation <80%, the presence of atrial arrhythmia and the absence of disease modifier therapy were associated with mortality in ES. Low oxygen saturation reflects the extent of right-to-left shunting, low cardiac output and advanced stage of pulmonary vascular disease and has been noted to predict outcomes in ES. At the same time, the development of atrial arrhythmia predicted adverse outcomes in ES, especially in pre-tricuspid shunt lesions.