Article Text

Download PDFPDF

Original research article
Pulmonary vasodilator therapy is associated with greater survival in Eisenmenger syndrome
  1. Clare Arnott1,2,
  2. Geoff Strange3,4,
  3. Andrew Bullock5,6,
  4. Adrienne C Kirby7,
  5. Clare O’Donnell8,9,
  6. Dorothy J Radford10,11,
  7. Leeanne E Grigg12,13,
  8. David S Celermajer1,2
  1. 1 Department of Cardiology, Royal Prince Alfred Hospital, Sydney, New South Wales, Australia
  2. 2 Sydney Medical School, University of Sydney - Camperdown and Darlington Campus, Sydney, New South Wales, Australia
  3. 3 Pulmonary Hypertension Society ANZ, Sans Souci, New South Wales, Australia
  4. 4 Faculty of Medicine, University of Notre Dame, Sydney, New South Wales, Australia
  5. 5 Department of Cardiology, Royal Perth Hospital, Perth, Western Australia, Australia
  6. 6 Children’s Cardiac Centre, Princess Margaret Hospital for Children, Perth, Western Australia, Australia
  7. 7 National Health and Medical Research Council Clinical Trials Centre, Biostatistics, Camperdown, New South Wales, Australia
  8. 8 Department of Paediatric Cardiology, Auckland City Hospital, Auckland, New Zealand
  9. 9 School of Medicine, University of Auckland, Auckland, New Zealand
  10. 10 Adult Congenital Heart Unit, The Prince Charles Hospital, Brisbane, Queensland, Australia
  11. 11 Department of Medicine, University of Queensland, Brisbane, Queensland, Australia
  12. 12 Faculty of Medicine, University of Melbourne, Melbourne, Victoria, Australia
  13. 13 Department of Cardiology, Royal Melbourne Hospital, Melbourne, Victoria, Australia
  1. Correspondence to Dr Clare Arnott, Department of Cardiology, Royal Prince Alfred Hospital, Camperdown, 2050, NSW, Australia; carn8379{at}


Objective Eisenmenger syndrome (ES) is a severe form of pulmonary hypertension in adults with congenital heart disease (CHD) and has a poor prognosis. We aimed to understand factors associated with survival in ES and particularly to assess the potential benefits of advanced pulmonary vasodilator therapy (AT).

Methods From January 2004, when AT became generally available for patients with ES, we followed 253 ES adults from 12 adult congenital heart disease centres across Australia and New Zealand. Demographic, medical and outcome data were collected and analysed prospectively and retrospectively.

Results The patients with ES were predominantly female (60%), aged 31 (SD 12) years. At diagnosis of ES, 64% were WHO functional class ≥3. The most common underlying lesion was ventricular septal defect (33%) with 21% having ‘complex’ anatomy. Over a median follow-up time of 9.1 years, the majority (72%) had been prescribed at least one AT (49% single agent), mostly bosentan (66%, 168 patients). The mean time on AT was 6 (SD 3.6) years. Those on AT were more functionally impaired at presentation (69% WHO ≥3 vs 51%, p=0.007) and more likely to have been prescribed anticoagulation (47% vs 27%, p=0.003). The risk of death/transplant was 4.8 %/year in AT exposed versus 8.4% in those never exposed. On multivariable analysis, exposure to AT was independently associated with greater survival (survival HR 2.27, 95% CI 1.49 to 3.45; p<0.001). WHO ≥3 at presentation was associated with a worse prognosis (mortality HR 1.82, 95% CI 1.19 to 2.78; p=0.006).

Conclusion Treatment with AT was independently associated with greater survival in patients with ES, even though they were comparatively sicker prior to treatment.

  • Congenital Heart Disease
  • Pulmonary Vascular Disease
  • Complex Congenital Heart Disease

Statistics from


Eisenmenger syndrome (ES) occurs in patients with a structurally abnormal heart where intracardiac or extracardiac systemic to pulmonary communications result in severe pulmonary vascular disease thence right to left shunting, leading to chronic cyanosis. As the prevalence of severe congenital heart disease (CHD) in adults continues to increase in developed nations,1 so does the need to better understand the ES population, their outcomes and their therapeutic requirements.2

ES leads to a number of systemic complications related to chronic hypoxaemia and secondary erythrocytosis. Such complications include clotting abnormalities with haemorrhage, cerebrovascular complications, predisposition to serious infections, malignant cardiac arrhythmias and sudden death.3–6

Prior to the mid-2000s, the main therapeutic options for those with ES were supportive, aimed at addressing the associated complications. In 2004, advanced pulmonary vasodilators (advanced therapy (AT)) first became readily available across Australia and New Zealand for pulmonary arterial hypertension (PAH) via compassionate access schemes.7 In 2006, the Bosentan Randomized Trial of Endothelin Antagonist Therapy-5 (BREATHE-5) study demonstrated the safety and efficacy of an AT, bosentan, in those with ES, leading to more widespread adoption of such therapies for patients with ES.8 Access to such medications almost always required 6 monthly follow-up for continued government funding.

In 2010, we established a registry study of patients with PAH-CHD across Australia and New Zealand and published our preliminary analyses in 2012.7 9 Now, 5 years later, we aimed to update these survival analyses and to concentrate on that subset of patients with ES, in whom we now have had both greater experience and longer follow-up times.

Accordingly, the primary aim of this study was to determine whether exposure to AT might influence survival. Secondary endpoints were to assess the effect of gender, age at presentation, WHO functional status, oxygen saturations, underlying cardiac lesion, anticoagulation and the presence of Down syndrome on survival.


Patient population

Data were collected retrospectively (prior to 2010) and thence prospectively both directly from the Binational Adult Congenital Heart Registry7 9 and from the local records of the 12 contributing adult congenital heart disease (ACHD) sites across Australia and New Zealand. The Registry enrolled all CHD patients with PAH seen at least once since 1 January 2000 (360 patients). Of these, 263 patients had ES and 253 were alive and seen at least once since 1 January 2004. This paper reports on those 253 patients. Ethics committee approval and patient consent were obtained for this study at each site.

Entry criteria for inclusion in the study was age greater than 16 years, a confirmed diagnosis of ES by an ACHD specialist and review in an ACHD centre at least once since 1 January 2004. This date marked the earliest time point that the majority of patients had potential access to AT (on compassionate grounds or in clinical trials initially, government funded thereafter). An analysis of the effects of AT was performed in a subset of 233 individuals who were alive as of 1 January 2004 and seen at least twice since that time point and in whom complete data on duration of AT therapy was known (figure 1).

Figure 1

Study design and timeline. AT, advanced therapy; CHD, congenital heart disease; ES, Eisenmenger syndrome; PAH, pulmonary arterial hypertension.

The underlying congenital lesion was defined as ‘simple’ or ‘complex’. ‘Simple’ included solitary large atrial septal defects (ASDs, usually >20 mm), solitary ventricular septal defects (VSDs), atrioventricular septal defects, persistent ductus arteriosus (PDA), aortopulmonary window or ‘multiple left to right shunts’ for those with multiple septal defects or septal defects combined with PDA or anomalous pulmonary venous drainage. ‘Complex’ referred to those with (shunted) tetralogy of Fallot, univentricular hearts, transposition of the great arteries or hypoplastic left or right heart syndromes. The diagnosis was made by an ACHD specialist using a combination of clinical review, transthoracic echocardiogram and right heart catheterisation, as clinically appropriate in each individual case.7

Medical therapies

Duration of exposure to AT was recorded as was cessation of either AT or anticoagulant and reason for cessation. The government-funded Pharmaceutical Benefits Scheme in Australia has specific guidelines regarding both who is eligible to receive specific ATs and the requirements and frequency of their follow-up (at least every 6 months). Bosentan was the first government-funded AT available to those with ES in Australia, first listed for those with WHO functional class (FC) 3 or 4 disease in March 2008. Thus, while treatment was at physician’s discretion, overwhelmingly the initial therapy used was bosentan. Additional or alternative AT in the event of an adverse reaction was at the discretion of the treating physician.

Statistical analysis

Data were analysed using SPSS V.22.0 and SAS V.9.3. Descriptive data are expressed as number and percentages, medians with IQR or means with SD, as appropriate. Comparisons between groups were undertaken using χ2 for categorical data, independent sample t-tests for normally distributed continuous data and Mann-Whitney U test for non-normally distributed continuous data. A two-tailed p value of less than or equal to 0.05 was considered statistically significant.

A survival analysis, with being on AT as a time varying covariate, was performed for the composite endpoint of death or transplant from the study entry point of 1 January 2004, or date of presentation, whichever was later, to the date of last clinical review/death or transplant. This analysis was performed only on those 233 patients seen at least twice during the study period and in whom duration of AT was known. Proportional hazards regressions were undertaken with being on AT entered as a time varying covariate.

The multivariable Cox proportional hazards model was used to identify factors associated with survival, with on AT entered as a time varying covariate. The prespecified variables of interest were: gender, age at presentation, Down syndrome, the presence of a solitary pretricuspid lesion (ASD), a diagnosis of a ‘Complex’ lesion, WHO FC at diagnosis, resting oxygen saturation, exposure to anticoagulation and exposure to AT.


Subject characteristics

The study enrolled 253 adults with ES (60% female) with a mean age at presentation of 31 years (SD 12.2); 33% also had Down syndrome. The median number of follow-up visits for each patient was over 5.

The dominant underlying lesion is outlined in table 1 with VSD as the most common lesion (33%) and 21% having complex anatomical abnormalities. At the time of presentation, only 4.9% were WHO FC 1, with 31.1% WHO FC 2, 57.5% WHO FC 3% and 6.5% WHO FC 4. Those with Down syndrome presented with worse FC, with 75% WHO FC 3 or greater, as compared with 58% in the non-syndromic cohort (p=0.01). This finding should be interpreted cautiously, given the difficulty and subjectively of FC assessment in patients with Down syndrome. Baseline resting mean oxygen saturations, reported in 124 subjects, were 85.0% (SD 8.2) on room air (84.8% (SD8.2) in AT exposed and 87.4% (SD8.2) in no AT exposure; p=0.3). Only 135 subjects (53%) underwent right heart catheterisation.

Table 1

Baseline characteristics

Data on initial 6 min walk distance (6MWD) was available in 106 subjects (42%) and of those 94% (100 people) were exposed to AT during the study period. In this subgroup with available data, the mean 6MWD was 350.4 m and was significantly shorter in the WHO ≥3 group than in those with WHO <3 (332.5m (SD 124.7) vs 398.0 m (SD 104.6); p=0.01).

Advanced therapy and other medical therapies

At the time of last review, the majority of patients had no past or current exposure to anticoagulants (59%). Almost one-third of patients had taken diuretics at some point during the course of their illness (31.8%) with only a small percentage taking either digoxin (11.8%) or calcium channel blockers (1.9%).

AT was started for WHO FC 3 or 4 symptoms (or a select group of deteriorating class 2 patients in New Zealand). Changes in the specific therapies were made chiefly for adverse effects or treatment failure. During the course of the study, 72% of patients were exposed to AT with 69% remaining on therapy at the time of last review (mostly single agent, 49%) (table 2). The mean time on AT was 5.7 years (SD 3.6), median 5.4 years (IQR 3.1–8.1) with a total of 961 patient-years on AT in the study. A significantly lower proportion of those with Down syndrome were on AT (29% vs 71% without DS, p=0.005) despite their poorer WHO FC at diagnosis.

Table 2

Medications prescribed, since 2004 (entire cohort)

Bosentan was prescribed in 66% of subjects (168 people) with 26% of those ceasing that therapy during the study (44 people). In those who ceased bosentan, the most common reason was side effects (39%, 17 people), with other causes including poor clinical response (23%, 10 people) or transplantation (13%, 6 people). Sildenafil was commenced in 35% of ES subjects with 10% of those ceasing therapy (nine people) due to either side effects (three people) or transplantation (six people). Prostacyclins were commenced in 3% (seven people) and ceased in five of them due to a combination of cost (two people), transplantation (two people) and poor clinical response (one person). Fifty patients were on multiple therapies, as shown in table 2.

Patients on AT were more functionally impaired at presentation (69% on AT were WHO ≥3, 51% not on AT were WHO ≥3; p=0.007) and were more likely to have been prescribed anticoagulation (47% on AT ever exposed to anticoagulants, 27% not on AT; p=0.003) (figure 2).

Figure 2

Risk of being prescribed advanced therapy (in 233 subjects). Anticoagulant, past or current exposure to anticoagulants; complex, a diagnosis of complex congenital heart disease; pre-TV lesion, the presence of a solitary pretricuspid valve lesion (ASD); WHO ≥3, WHO functional class three or greater at presentation.

Advanced therapy and survival patterns in ES

At the time of last review, of the 253 patients seen since 2004, 66% were alive, 30% dead, 3.6% had received a transplant and 0.4% had been lost to follow-up (vital status unknown as they had not returned for routine follow-up). All-cause survival from 1 January 2004/date of presentation was 88% at 5 years and 66% at 10 years; median follow-up time 9.1 years (IQR 5.6–11.9) (figure 3).

Figure 3

Survival in entire cohort. Time, years from 1 January 2004 or presentation (if after 1 January 2004) to date of last review.

In those 233 subjects seen at least twice since 2004 and in whom exact duration of AT was known, those exposed to AT had a 4.8% rate of death or transplant per year (95% CI 3.5 to 6.4) versus 8.4% (95% CI 5.9 to 11.6) in those who received no specific PAH drugs; p<0.001 (figure 4).

Figure 4

Greater survival in those exposed to advanced therapy. The left panel depicts the rate of death/transplant per person year in % for those exposed to advanced therapy and those never exposed. Error bars represent 95% CI. The right panel depicts the survival HR for those exposed to advanced therapy based on Cox multivariable analysis. AT, exposure to advanced therapy during study; no AT, no exposure to advanced therapy.

Cause of death was known in 68% of the 76 patients who died (52 individuals). When looking at those with a known cause of death, ‘pulmonary hypertension’ or ‘cardiac- sudden or chronic’ was the cause in 25 people and ‘hypoxia/respiratory failure’ in four people. Infectious causes were common and accounted for 11 deaths (pneumonia in 6, sepsis 1, endocarditis 1, cerebral abscess 2, meningoencephalitis 1). Bleeding complications accounted for four deaths (brain haemorrhage/injury and gastrointestinal haemorrhage). The remaining causes were periprocedural deaths (two cases), renal failure (three cases), aortic dissection (one case), thyroid neoplasm (one case) and ‘found unresponsive’ (one case). There was no significant difference in the rates of cardiac death in those exposed to AT versus those unexposed (p=0.16).

Predictors of mortality—Cox univariable and multivariable models

On univariable analysis of the 233 participants, exposure to AT was associated with greater survival with a survival HR of 1.96 (95% CI 1.30 to 2.86, p=0.001). There was no difference in the proportion of individuals having an event between those on single therapy and those on double or triple therapy (p=0.26).

In a Cox multivariable model (since 1 January 2004 or date of presentation if later), including WHO status and age at presentation, AT remained significantly associated with greater survival with a 56% reduction in the risk of death or transplant (survival HR 2.27, 95% CI 1.49 to 3.45; p<0.001). WHO FC at diagnosis was also found to be independently related to mortality, with those WHO ≥3 having a mortality HR of 1.82 (95% CI 1.19 to 2.78), p=0.006, as was age at presentation (mortality HR for each 10 years 1.02, 95% CI 1.00 to 1.03; p=0.012). All other prespecified factors, including gender, oxygen saturations, the presence of Down syndrome, a solitary pretricuspid lesion, a diagnosis of ‘complex’ CHD and exposure to anticoagulation were not associated with survival (figure 4).


The salient finding in this current study is that in a multicentre, binational ES cohort followed for almost a decade (median follow-up time 9.1 years), those exposed to AT (72%) had significantly greater survival than those never exposed, despite being more functionally impaired than those not on therapy.

This current study represents the longest follow-up yet reported for ES patients treated with AT, with other contemporary studies having median follow-ups of 5 years or less.10–12 With over 900 patient treatment years, we were able to demonstrate that while those commenced on AT were sicker than those not on AT; they had significantly greater survival times. Contemporary data in this regard are mixed, with two recent cohort studies also finding a survival benefit in those treated with AT.11 13 Conversely, a recently published multicentre study of >1000 patients with ES (62.8% on AT) did not find an association between survival and AT on multivariable analysis (though an association was shown in univariable analysis).12 The reason for this discrepancy may be the relatively short follow-up time of approximately 3 years in the latter study.

We report overall survival rates of 88% at 5 years and 66% at 10 years, from study entry (1 January 2004 or date of presentation if later). This compares favourably with other survival analyses undertaken in the same patient group, including two recent studies in large cohorts.7 12 The relatively greater survival seen in this current study may be related to the large proportion of patients exposed to AT (72%) or due to the current era of more aggressive/early commencement of AT.

The frequency of ES in the adult CHD population is difficult to quantify but has been estimated to be approximately 4% in a European CHD population late last century.5 14 Ongoing advances in medical care in the current era have resulted in survival of most children born with CHD in the developed world into adulthood.4 6 With this improved medium-term survival comes a need to optimise medical management of this population with unique needs, who are known to be at risk of significant late complications related to their disease. Specifically, in the ES group, cyanosis and its related complications vastly increase their morbidity and mortality, with mortality rates greater than 12 times that of the general population.10 12 15 Thus, there is a great need to improve our understanding of long-term disease complications and the best evidence-based therapeutic options.2

Kempny et al 12 identified five factors independently associated with mortality in ES: age, underlying shunt, sinus rhythm, pericardial effusion and baseline oxygen saturations.12 In this current study, the only other factors, in addition to AT, associated with mortality were WHO FC and age at diagnosis, with WHO ≥3 having a mortality HR of 1.82 (95% CI 1.19 to 2.78; p=0.006) and age at presentation having a mortality HR for each 10 years of 1.02 (95% CI 1.00 to 1.03; p=0.012). This is an interesting finding given that AT was prescribed to those with a worse WHO FC and yet still (in our experience) was associated with a greater survival. There is variable agreement in the literature regarding an association between FC and survival, with questions regarding the subjectivity of WHO grading raised as a potential limitation to its use.12 16 17 We do not have data available in this study specifically recording rhythm and echocardiographic characteristics to allow us to assess for any potential survival association. Baseline oxygen saturations were available in just under 50% of patients, and in this small subset were not associated with survival; however, this finding must be interpreted with caution given the missing data. Similarly, initial 6MWD was only available in 42% of subjects, and thus it was not possible to look for an association with survival. Importantly, however, in the subgroup with available 6MWD information, those with WHO ≥3 had shorter 6MWD than those with WHO<3 (p=0.01). This implies that WHO FC is able to discern sicker or more functionally impaired individuals.

As stated, underlying shunt was not found to be related to mortality in this study, which is in disagreement with Kempny et al, who on multivariable analysis, quoted a mortality HR of 1.56 for those with pretricuspid shunts (p=0.041).12 This discrepancy between the two studies is difficult to reconcile given the low proportion of patients in the current study with pretricuspid lesions (10.7%) and the borderline significance of the association on multivariable analysis in the prior study. As such, this should be interrogated in future studies with larger cohorts.

Cause of death information was not available in all subjects in the current report; however, our limited information implies that cardiorespiratory causes predominated, with infection also common. This is in agreement with a recent retrospective study of over 1500 patients with ES that reported heart failure as the leading cause of death, followed closely by infection.18

The main strength of this study is that we report on a large multicentre cohort of patients over a long follow-up time. Furthermore, strict government mandated therapy guidelines have ensured a relative uniformity of treatment and follow-up strategy across the centres, and National Death Registries allowed us to know mortality data on all subjects followed.

The main potential limitation of this study is selection bias. We only report from ‘expert’ ACHD centres and only on those who survived up until the time of availability of AT (2004). We believe, however, that the majority of patients with ES in Australia and New Zealand are treated in specialist ACHD centres. Also, we lacked data on many important clinical variables for many patients, including ECG and echocardiography findings, exercise capacity measures, renal function and oxygen saturation results. The other limitation is the use of a time-varying covariate in AT exposure analysis. Care must be taken with interpretation of this analysis given different on versus off therapy follow-up times. Nonetheless, these limitations do not impact on the salient finding of the study, concerning the association between AT and greater survival.

In summary, in a contemporary ES cohort of over 250 patients with median follow-up of over 9 years, treatment with AT was independently associated with greater survival.

Key messages

What is already known on this subject?

  • Those with Eisenmenger syndrome (ES) have mortality rates greater than 12-fold the general population, in early adulthood. Specific factors associated with survival have been demonstrated in recent studies; however, the potential survival benefit of advanced pulmonary vasodilator therapy (AT) is contentious.

What might this study add?

  • In 253 patients with ES followed (median 9.1 years), overall survival from 1 January 2004 (or date first seen if after 1 January 2004) was 88% at 5 years and 66% at 10 years.

  • Prior to treatment, those receiving AT had been more functionally impaired than those untreated.

  • Despite this, those exposed to therapy had greater survival and fewer deaths/transplant per patient year (HR for survival of 2.27, 95% CI 1.49 to 3.45; p<0.001).

How might this impact on clinical practice?

  • The finding of greater long-term survival in those treated with AT as compared with those untreated may be instructive for both therapeutic algorithms in those with ES and in the design of future randomised trials to confirm this association.



  • Contributors All authors are responsible for the reported research and have approved the manuscript for submission. Specifically, CA and DSC were responsible for initial study concept, primary data collection and analysis and manuscript preparation and submission. They are responsible for overall content as guarantors. GS assisted with concept design, data collection and manuscript preparation. AB, CO, DJR and LG assisted with data collection, manuscript preparation and correction. ACK assisted with data analysis, manuscript preparation and correction.

  • Competing interests None declared.

  • Ethics approval Local ethics approval was obtained at each contributing site across Australia and New Zealand.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement There is no additional unpublished data from this study available to other parties.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles