International Journal of Cardiology 181 (2015) 270–276

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

New predictors of mortality in adults with congenital heart disease and pulmonary hypertension: Midterm outcome of a prospective study Mark J. Schuuring a,b, Annelieke C.M.J. van Riel a,b, Jeroen C. Vis a, Marielle G. Duffels a, Arie P.J. van Dijk c, Rianne H.A.C.M. de Bruin-Bon a, Aeilko H. Zwinderman d, Barbara J.M. Mulder a,b,⁎, Berto J. Bouma a a

Department of Cardiology, Academic Medical Center, Amsterdam, The Netherlands Interuniversity Cardiology Institute of the Netherlands, Utrecht, The Netherlands Department of Cardiology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands d Department of Clinical Epidemiology and Biostatistics, Academic Medical Center, Amsterdam, The Netherlands b c

a r t i c l e

i n f o

Article history: Received 2 May 2014 Received in revised form 12 November 2014 Accepted 26 November 2014 Available online 13 December 2014 Keywords: Mortality Predictors Pulmonary hypertension Congenital heart disease

a b s t r a c t Background: Patients with CHD-PAH have a limited prognosis. In daily practice, combination therapy is often initiated after a clinical event. Although clinical events have been associated with a poor prognosis in idiopathic PAH, data on this association are limited in CHD-PAH. The aim of this study was to determine whether baseline characteristics and clinical events associate with mortality in patients with pulmonary hypertension (PAH) due to congenital heart disease (CHD). Methods: In total 91 consecutive adults (42 ± 14 year) with CHD-PAH were referred for therapy between January 2005 and June 2013. Cox proportional hazard analysis was performed to identify determinants of mortality, including clinical events as time dependent covariates. Results: Twenty-four patients (nine with Down) died during the median follow-up of 4.7 (range 0.1–7.9) years. The one and eight year mortality rates were 7.3% and 37.3%, respectively. Clinical events included admission for heart failure (n = 9), arrhythmias (n = 9), haemoptysis (n = 5), change to a worse NYHA class (n = 16), vascular events (n = 1), syncope (n = 1) and need for red blood cell depletion (n = 4). In univariate analysis, both baseline characteristics and clinical events were associated with mortality. In multivariate analysis, only baseline NT-pro-BNP serum level ≥ 500 ng/L and TAPSE b 15 mm at echocardiography were significant determinants of mortality. None of the clinical events remained significant. Patients with both a NT-pro-BNP serum level ≥ 500 ng/L and TAPSE b 15 mm at echocardiography have a nine fold higher mortality rate than patients without both risk factors. Conclusion: Prognosis is still poor in contemporary patients with CHD-PAH. Both baseline NT-pro-BNP serum level and right ventricular function are superior to clinical events in prognostication. These two baseline characteristics should have a major impact on therapeutic management in patients with CHD-PAH, such as initiation of combination therapy. © 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Pulmonary arterial hypertension (PAH) is a common complication in adults with congenital heart disease (CHD) [1,2]. Life expectancy and quality of life are markedly reduced in patients with CHD-PAH [3–5]. The guidelines of the ESC and ACC/AHA on PAH recommend that monotherapy should be considered in symptomatic patients with PAH and that combination therapy may be considered in these patients [6–8]. Benefit of combination therapy may outweigh the risk of side effects in patients with a poor prognosis. In daily practice, patients with CHD-PAH often are initiated on monotherapy and they are treated ⁎ Corresponding author at: Department of Cardiology, Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. E-mail address: [email protected] (B.J.M. Mulder).

http://dx.doi.org/10.1016/j.ijcard.2014.11.222 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.

with combination therapy after the occurrence of a clinical event, as a clinical event is believed to herald a poor prognosis in patients with CHD-PAH. However, this expectation is mainly based on data obtained in patients with idiopathic PAH [9,10]. In patients with CHD-PAH, studies on prognosis are limited. A few studies on prognostic value of baseline characteristics reported several determinants associated with poor outcome in CHD-PAH, including increased baseline NT-pro-BNP serum level [11], right ventricular (RV) dysfunction [12], six minute walking distance [13] and severely impaired renal function [14]. No association was found between clinical events and a poor prognosis in three studies on adults with CHD-PAH. However, these three studies were hampered by a retrospective design [15–17]. In patients with CHD-PAH, there is an urgent need for reliable determinants of a poor prognosis to guide therapeutic management. Using prospectively collected data, we aimed to determine whether baseline

M.J. Schuuring et al. / International Journal of Cardiology 181 (2015) 270–276

characteristics and clinical events are associated with mortality in patients with CHD-PAH. These findings may be of help to select patients at highest risk of mortality who might benefit of therapeutic adaptations. 2. Methods 2.1. Study population and design Consecutive adults with CHD-PAH referred between January 2005 and May 2013 were studied prospectively using a standardized protocol. To establish PAH diagnosis, minimal systolic pulmonary arterial pressure had to be 40 mm Hg at rest on a transthoracic echocardiogram with a tricuspid regurgitation velocity of more than 2.9 m/s, according to the guidelines [6,18]. Patients with and without Down syndrome were included [19,20]. Patients with moderate to severe liver disease, simultaneous use of cyclosporine A or pregnancy were not included [21]. Bosentan monotherapy 62.5 mg twice daily was started in all patients, increasing the dose to 125 mg twice a day after four weeks, as tolerated. Decisions on type of drug and timing to start combination or triple therapy were at the discretion of the treating physician, considering patient preferences and existing contraindications. Sildenafil was commonly started on a dose of 20 mg three times daily. Iloprost inhalation as a third therapy was started three times daily. If three times daily was well tolerated, iloprost was extended to six or nine times a day. Approval of the research protocol by the local ethics committee was obtained. Informed consent was not required, as all investigations were performed for routine clinical care.

2.3. Data collection Patients visited the outward clinic for an exercise test, an echocardiogram and laboratory measurement. All tests were performed on the same day. Exercise capacity was assessed using the six minute walking test, according to the guidelines of the American Thoracic Society with continuous pulse oximetry monitoring [22]. During exercise the oxygen saturation and heart rate were recorded.

271

Baseline echocardiography was performed with a Vivid 7 ultrasound system (General Electric). Pulmonary stenosis was ruled out in all patients [23]. Tricuspide annulus plane systolic excursion (TAPSE) was measured in the lateral tricuspid valve annulus with Mmode imaging in the lateral tricuspid valve annulus in the apical 4-chamber view. Qp and Qs were calculated with both left- and right ventricular outflow tract diameters and both left- and right VTI measurements. Systemic to pulmonary shunts include moderate to large defects, systemic-to-pulmonary shunting is still prevalent, whereas cyanosis is not a feature [24]. Those patients with small defects who develop PAH have a defect which does not account for the development of elevated pulmonary vascular resistance [24]. All echocardiographic images were acquired and recorded digitally. NT-pro-BNP levels were measured at baseline, which correlates with the start of bosentan. NT-proBNP serum levels were determined by electrochemiluminescence immunoassay on an Elecsys 2010 analyzer (Roche Diagnostics, Almere, The Netherlands). 2.4. Definition of clinical events The decision on type of events that were collected was based on the determinants of mortality in patients with idiopathic PAH [6,10]. Heart failure was defined as an increase in non-study treatment requirements or hospital admission for worsening symptoms of heart failure [25]. Arrhythmias were defined as any episode of documented atrial- or ventricular brady- or tachyarrhythmia that required electrocardioversion, pacemaker implantation or permanent change of medication. Vascular events were determined as a myocardial infarction or ischemic stroke. An episode of syncope was defined as a transient loss of consciousness, with a short onset and spontaneous recovery. Haemoptysis was defined as expectoration of blood ranging from blood-streaking of sputum to the presence of gross blood in the absence of any accompanying sputum. Red blood cell depletion was achieved with phlebotomy, and decisions on phlebotomy were at the discretion of the treating physician based on complaints of dizziness and elevated hematocrit. 2.5. Statistical analysis Statistical analysis was performed with SPSS 20.0 (IBM). Continuous variables were expressed as mean ± standard deviation when normally distributed, and median

Table 1 Baseline characteristics.

Median follow-up, years (range) Clinical assessment Age, years (sd) Male gender, n (%) Down syndrome, n (%) Clinical subgroup Eisenmenger syndrome, n (%) Systemic to pulmonary shunt, n (%) Small defect, n (%) Closed defect, n (%) NYHA class II, n (%) III, n (%) IV, n (%) Medication Diuretics, n (%) Beta blockers, n (%) ACE/angiotensin 2 inhibitors, n (%) Calcium antagonists, n (%) Resting saturation, percentage (sd) Systolic blood pressure, mm Hg (sd) Electrocardiogram Heart frequency, beats per minute (sd) QRS duration, ms (sd) QT dispersion, ms (sd) Exercise testing Six minute walk distance, m (sd) Heart rate maximum, beats per minute (sd) Saturation maximal exercise, % (sd) Echocardiography TAPSE, mm (sd) TAPSE b 15 mm Right atrial area, cm2 (sd) Right ventricular systolic pressure, mm Hg (sd) Moderate or severely impaired LV function Laboratory NT-pro-BNP serum level, ng/L (range) Glomerular filtration rate, mL/min (sd)

All patients

Non-survivors

Survivors

n = 91

n = 24

n = 67

p

4.7 (0.1–7.9)

3.4 (0.2–6.2)

6.0 (0.1–7.9)

0.001

41 (14) 36 (40) 40 (44)

48 (16) 10 (42) 9 (38)

39 (12) 26 (39) 31 (46)

0.006 0.806 0.458

69 (76) 14 (15) 4 (5) 4 (4)

20 (83) 3 (13) 0 1 (4)

49 (73) 11 (16) 4 (6) 3 (4)

0.603

25 (28) 62 (68) 4 (4)

5 (21) 16 (67) 3 (13)

20 (30) 40 (69) 1 (1)

0.067

23 (25) 8 (9) 8 (9) 7 (8) 85 (8) 118 (20)

12 (50) 2 (8) 4 (17) 5 (24) 85 (7) 117 (26)

11 (16) 6 (9) 4 (6) 2 (3) 85 (9) 119 (18)

0.001 0.926 0.112 0.005 0.943 0.691

79 (14) 109 (25) 75 (24)

86 (16) 108 (23) 74 (15)

77 (12) 114 (26) 75 (26)

0.007 0.347 0.907

356 (127) 112 (18) 70 (13)

322 (132) 108 (18) 71 (12)

367 (125) 113 (18) 70 (13)

0.142 0.339 0.794

20 (5) 8 (12) 22 (7) 87 (23) 3 (4)

0.072 0.014 0.466 0.689 0.078

20 (5) 16 (18) 22 (7) 87 (22) 6 (7) 556 (35–8753) 88 (27)

18 (5) 8 (33) 23 (8) 90 (17) 3 (13) 1441 (69–6586) 77 (28)

400 (35–8753) 91 (25)

0.003 0.047

sd: standard deviation; n = number; ns: not significant; NYHA: New York Heart Association; ACE: angiotensin converting enzyme; TAPSE: tricuspid annular plane systolic excursion. Bold p levels mean p b 0.05.

272

M.J. Schuuring et al. / International Journal of Cardiology 181 (2015) 270–276

3. Results 3.1. Baseline characteristics Ninety-one adults (42 ± 14 year) with CHD-PAH were referred for therapy. None of the patients were excluded from the analysis. Every single patient referred was included in the study. Table 1 summarizes patients' baseline characteristics. 3.2. Follow-up One patient refused treatment initiation and reimbursement of bosentan by the health insurance was rejected in 5 patients. In total 85 patients were initiated on bosentan therapy. Ten patients (11%) were classified as lost to follow-up after a median follow-up of 3.8 (range 0.3–5.3) years, because of a change of treating physician. One of these patients died. Twenty-four patients (nine with Down) died, see Fig. 1. Table 2 shows detailed baseline information about the deceased patients. The one and eight year mortality rates were 7.3% and 37.3% respectively. Clinical events were common during a median follow-up of 4.7 (range 0.1– 7.9) years (Table 3 and Fig. 1). The composite event rates were high at one (20.9%), two (32.5%) and eight (79.6%) year follow-up.

Fig. 1. Kaplan–Meier curves for mortality and events combined with mortality.

(range) if otherwise. Categorical variables were expressed as number (percentage) and differences were analyzed using a Chi-square test. Log rank test was performed to determine significant differences in mortality rate between two groups. Relevant cut-offs for age, heart rate and NT-pro-BNP serum level were obtained using receiver operating characteristic (ROC) curves in order to make the data easy to interpret for clinical use. Relevant cut-off for TAPSE (15 mm) was reported before [12]. Missing data were handled by multiple imputations using SPSS. The relation between determinants and clinical events was assessed using univariate and multivariate Cox proportional hazard analyses. Events during follow-up were modeled as time dependent covariates. Date of bosentan initiation was used as start date. In patients who had a clinical event during follow-up, each individual's person-years from the date of inclusion was calculated to the date of the event or the end of follow-up, to use clinical events as a time dependent covariate. Combination therapy was in general initiated after occurrence of an event. Therefore, the initial events were modeled and we were unable to evaluate whether combination therapy was associated with mortality. The significant univariate determinants were entered in a multivariate model using five imputed datasets. Determinants were missing in a range of 0–15%. The multivariate Cox model was analyzed with a forward conditional algorithm, again including time dependent covariates for clinical events. In case of multiple events, the first event was used. A value of p b 0.05 was considered to be significant.

3.3. Determinants of mortality Baseline characteristics univariately associated with mortality are shown in Table 4. Fig. 2 illustrates Kaplan–Meier graphs on NT-proBNP serum level and TAPSE in relation to mortality. During follow-up arrhythmias and a composite of all clinical events were associated with mortality, see Table 4. In multivariate regression, baseline NT-pro-BNP serum level ≥ 500 ng/L and TAPSE b 15 mm only remained significant determinants of mortality. None of the clinical events during follow-up were significant in multivariate analysis (see Table 4). Some variables were interdependent in uni- and multivariate analyses. However, the interaction terms were not significant in the multivariate analysis. Results of multivariate analysis were not affected by adjustment for type of congenital heart disease and age. Baseline NT-pro-BNP serum level ≥ 500 ng/L and TAPSE b 15 mm remained the strongest determinants of mortality.

Table 2 Detailed baseline information of deceased patients.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Defect

Subgroup

Down

Gender

BL NTproBNP N 500 ng/L

BL TAPSE b 15 mm

Last TAPSE b 15 mm

Age death

Cause of death

VSD VSD VSD VSD VSD VSD TGA pAVSD Monoventricle Closed ASD AVSD AVSD AVSD AVSD AVSD ASD SV ASD II ASD I ASD ASD ASD II, VSD VSD VSD VSD

Eisenmenger Eisenmenger Eisenmenger Eisenmenger Eisenmenger Eisenmenger Eisenmenger Eisenmenger Eisenmenger Closed Eisenmenger Eisenmenger Eisenmenger Eisenmenger Eisenmenger Eisenmenger Syst-to-pulm Eisenmenger Syst-to-pulm Eisenmenger Eisenmenger Syst-to-pulm Eisenmenger Eisenmenger

0 0 1 1 0 1 0 1 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0

Female Female Male Male Male Female Female Female Female Female Female Male Male Female Male Female Male Male Female Female Male Male Female Female

1 1 1 0 1 1 1 1 1 1 N/A 1 1 1 0 0 N/A N/A 1 1 N/A 0 N/A 1

1 1 1 0 0 1 1 0 0 1 0 0 1 0 0 1 0 0 0 0 0 N/A N/A 0

1 1 1 1 0 1 1 0 0 Died within 3 months 1 1 1 Died 6 months later 0 1 0 0 Died within 3 months 0 Died within 3 months 1 N/A N/A

64 53 48 58 39 56 38 51 41 76 38 54 34 28 30 59 52 75 79 66 45 26 65 49

Right sided heart failure Right sided heart failure Sudden Right sided heart failure Unknown Right sided heart failure Heart failure, MOF Heart failure due to pneumonia Ventricular arrhythmia Right sided heart failure Right sided heart failure Right sided heart failure Right sided heart failure Respiratory failure Cerebral infection Abdominal sepsis Unknown Respiratory failure Unknown Unknown Sudden Unknown Right sided heart failure Sepsis

BL: baseline; VSD: ventricular septal defect; AVSD: atrioventricular septal defect; ASD = atrial septal defect; TGA: transposition great arteries; 6MWD: six minute walking distance; yr: year; syst-to-pulm: systemic to pulmonary shunt; RV: right ventricular; MOF: multi-organ failure; N/A: not available.

M.J. Schuuring et al. / International Journal of Cardiology 181 (2015) 270–276

273

Table 3 Clinical events during follow-up.

Admission for heart failure Arrhythmias Haemoptysis Change to a worse NYHA class Vascular event Syncope Need for red blood cell depletion Transplantation Any clinical event during follow-up (composite of all above)

All patients n = 93

Survivors n = 69

Non-survivors n = 24

p

9 9 5 16 1 1 4

5 4 4 12 0 0 3

4 5 1 4 1 1 1

0.276 0.058 0.662 0.725 0.525 0.112 0.930

0 40

0 16

0 24

n/a b0.01

3.4. Impact of baseline NT-pro-BNP serum level and TAPSE on mortality A scatterplot on baseline NT-pro-BNP serum level and TAPSE is shown in Fig. 3. Patients with both a baseline NT-pro-BNP serum level ≥ 500 ng/L and TAPSE b 15 mm had a mortality rate of 83% versus Table 4 Cox proportional hazard analysis on determinants of mortality. Univariate

Clinical characteristics at baseline Clinical assessment Age N 50 years Down syndrome Eisenmenger syndrome NYHA class III or worse Resting saturation (per 5% decrease) Electrocardiogram Heart rate N 80 beats per minute QRS duration N 120 ms QT dispersion (per 10 ms increase) Exercise testing Six minute walking distance (per 50 m decrease) Six minute walking distance b 300 m Heart rate maximum b 120 beats per minute Saturation maximal exercise (per 5% decrease) Echocardiography TAPSE b 15 mm Right- to left atrial ratio N 1.5 Right atrial area N 25 cm2 Impaired left ventricular function Cardiac output (per 0.5 liter decrease) Laboratory NT-pro-BNP serum level N 500 ng/L Glomerular filtration rate b 60 mL/min Clinical events during follow-up Admission for heart failure Arrhythmias Haemoptysis Change to a worse NYHA class Need for red blood cell depletion Any event during follow-up (composite of all)

Forward multivariate

HR 95% CI

p

HR 95% CI

2.4 0.7 1.1 1.0 1.0

1.0–5.4 0.3–1.5 0.4–3.1 0.4–2.6 0.8–1.3

0.040 0.329 0.900 0.960 0.884

2.3 1.0–5.4

0.059

1.7 0.7–4.0 0.9 0.8–1.1

0.260 0.324

1.1 0.9–1.4

0.125

2.2 1.0–5.0

0.065

1.8 0.6–5.0

0.265

1.0 0.8–1.2

0.737

3.5 3.4 1.1 3.2

0.008 4.0 1.4–11.0 0.009 0.087 0.909 0.070

Fig. 2. Kaplan–Meier graphs on NT-pro-BNP serum level and TAPSE in relation to mortality.

p

20% in patients without both risk factors (log rank 12.9; p b 0.001). A predicted survival model on baseline NT-pro-BNP serum level and TAPSE is shown in Fig. 4. A ROC curve for mortality demonstrated superiority of baseline NT-pro-BNP serum level and TAPSE (AUC of 0.77) to clinical events (AUC of 0.56), see Fig. 5. 3.5. Combination therapy

1.4–8.6 0.8–13.4 0.4–3.2 0.9–11.2

1.0 0.9–1.1

In eleven patients who changed to a worse NYHA class and in one patient with haemoptysis combination therapy was started. Sildenafil was prescribed eleven times. One patient preferred iloprost over sildenafil. In that patient iloprost was added to bosentan. None of the treating physicians initiated combination therapy aside of an event. In six

0.750

4.8 1.6–14.6 0.005 4.3 1.2–15.4 0.019 2.6 0.9–7.1

0.069

2.9 2.7 1.0 3.0 1.2

0.9–8.7 1.0–7.6 0.1–7.7 1.0–9.1 0.2–9.1

0.063 0.050 0.988 0.054 0.856

3.5 1.5–8.8

0.005

NYHA: New York Heart Association; TAPSE: tricuspid annular systolic excursion; BL: baseline. Bold p levels mean p b 0.05.

Fig. 3. Scatterplot on baseline NT-pro-BNP serum level and TAPSE. Caption mm: millimeter; ng/L: nanogram per liter.

274

M.J. Schuuring et al. / International Journal of Cardiology 181 (2015) 270–276

Fig. 4. Predicted survival model based on NT-pro-BNP serum level and TAPSE. Caption mm: millimeter; ng/L: nanogram per liter; HR: hazard ratio.

patients initiation of combination therapy was performed more than six months prior to the final study date. Therefore, follow-up data on combination therapy only was available in these patients. No significant change in NT-pro-BNP serum level and TAPSE was found after initiation of combination therapy in these six patients. In our population combination therapy was not started more often in patients with a NT-proBNP serum level ≥ 500 ng/L and TAPSE b 15 mm as compared to patients without both risk factors (p = 0.45). 3.6. Determinants of clinical events Baseline NT-pro-BNP serum level and right ventricular function were tested as determinants of clinical events (see Table 5). Both baseline characteristics were significantly associated with arrhythmias and admission for heart failure. 4. Discussion We are the first to show that RV baseline characteristics are superior to clinical events in prognostication, in a prospectively followed cohort with long-term follow-up of contemporary CHD-PAH patients. Patients

with both a NT-pro-BNP serum level ≥ 500 ng/L and TAPSE b 15 mm at echocardiography have a nine fold higher mortality rate than patients without both risk factors. 4.1. Interpretation Current therapeutic management is hampered by the lack of reliable determinants of prognosis in patients with CHD-PAH. Especially the prognostic value of baseline characteristics and clinical events remains unclear. Patients with CHD-PAH often are subject to an increased pulmonary vascular resistance for a long time. As a consequence, RV overload causes RV hypertrophy and dilatation [26]. This leads to RV dysfunction and congestion [14,27,28]. Ultimately, the vast majority of patients with CHD-PAH die because of right-sided heart failure [29,30]. In adults with CHD-PAH, for the first time, RV baseline characteristics are shown to dominate prognosis. Both baseline NT-pro-BNP serum level and TAPSE at echocardiography are early markers of RV dysfunction and fatal right-sided heart failure [31,32]. In our study one patient with baseline TAPSE b 15 mm died suddenly and one patient with baseline TAPSE b 15 mm died due to abdominal sepsis. Interestingly, three patients with baseline TAPSE N 15 mm died because of right-

Fig. 5. Receiver operating characteristic curves on baseline NT-pro-BNP serum level and TAPSE versus clinical events. Subtitle A: ROC curve for mortality of baseline NT-pro-BNP serum level and TAPSE combined. Subtitle B: ROC curve for mortality of clinical events.

M.J. Schuuring et al. / International Journal of Cardiology 181 (2015) 270–276

275

Table 5 Cox proportional hazard analysis on determinants of clinical events. Admission for heart failure

TAPSE b 15 mm NT-pro-BNP N 500 ng/L

Arrhythmias

Haemoptysis

Change to a worse NYHA

Need for red blood cell depletion

HR

95% CI

p

HR

95% CI

p

HR

95% CI

p

HR

95% CI

p

HR

95% CI

p

3.1 9.9

0.8–11.4 1.2–79.1

0.095 0.031

5.1 8.9

1.3–18.9 1.1–71.5

0.015 0.040

1.2 0.280

0.1–10.9 0.0–2.5

0.875 0.258

1.4 2.5

0.4–5.5 0.8–7.3

0.564 0.084

8.1 0.5

0.7–89.1 0–5.9

0.088 0.609

Bold p levels mean p b 0.05.

sided heart failure. Two of these patients developed heart failure over time, reflected by TAPSE b 15 mm on their last echocardiogram. One patient with Down syndrome and a preserved RV function died because of heart failure secondary to pneumonia. However, in general our finding that baseline NT-pro-BNP serum level and RV function outperformed other baseline characteristics and clinical events in prognostication is new and important. Interestingly, the degree of NYHA class was not significantly related to outcome (p = 0.067). The majority of patients (68%) were in the same NYHA class (class III) and only 4% were in NYHA class IV. In a larger study population or different distribution NYHA class might be related to outcome. Before our results were available, a clinical event was believed to herald a poor prognosis in patients with CHD-PAH. This expectation was based on studies in idiopathic PAH [9]. Our data, however, restrict the prognostic value of clinical events in CHD-PAH patients. Between patients with CHD-PAH and idiopathic PAH large differences exist in cause and rate of progression of the disease [33,34]. Mechanistic differences (for instance in blood viscosity and associated shear stress) between both patient groups might have different effects on early NTpro-BNP serum levels and RV function [35]. Moreover, patients with CHD-PAH often are able to maintain or increase their systemic cardiac output during acute conditions, such as heart failure and haemoptysis, by shunting over the defect [36]. In patients with idiopathic PAH, however, pulmonary hypertension per se limits pulmonary as well as systemic blood flow during acute conditions [37]. So, applicability of clinical events found in studies on patients with idiopathic PAH to determine prognosis seems to be limited in adults with CHD-PAH.

4.2. Clinical impact Our data suggest that initiation of combination therapy should be considered as soon as signs of RV failure (NT pro-BNP N 500 U/L and TAPSE b 15 mm) are found. Up till now, combination therapy often is started after a clinical event in patients with CHD-PAH. This strategy is mainly based on a study in patients with predominantly idiopathic PAH [9]. Patients with idiopathic PAH on combination therapy after an event had a significantly better survival than patients in a historical control group (83.9% versus 60.2%; p = 0.01) [9]. Other data on combination therapy are scarce. In our study, no significant changes in NT-proBNP serum level and TAPSE were found during follow-up after initiation of combination therapy in six patients. In 79 patients with CHD-PAH one retrospective analysis demonstrated an afresh improvement in six minute walking distance after initiation of combination therapy at symptomatic deterioration [38]. However, in 21 clinically stable patients with CHD-PAH one randomized controlled cross-over trial evaluated addition of sildenafil or placebo to ongoing bosentan monotherapy. This small randomized controlled cross-over trial did not demonstrate an improvement in six minute walking distance or pulmonary vascular resistance [39]. The study, however, might have been limited by the short follow-up period of three months. Value of clinical events in therapeutic management seems to be limited in adults with CHD-PAH. Therefore, we propose to initiate combination therapy in patients with CHD-PAH with NT-pro-BNP serum

level ≥ 500 ng/L and TAPSE b 15 mm at echocardiography. A prospective trial is warranted to evaluate this strategy. 4.3. Study limitations A potential limitation was the relatively small study size, as in most studies in patients with CHD. However, in patients with CHD-PAH this study is one of the largest cohort studies with a median follow-up of six years to date. Many of the deaths occurred in adults with Down syndrome, a condition with a reduced life expectancy [40]. Furthermore, data on time from PAH diagnosis to study inclusion was lacking. It should be noted that Holter recordings and invasive hemodynamic data were unavailable. A small proportion of missing data were imputed. Analyzing datasets with imputed data seems more efficient than analyzing only complete cases [41,42]. A recent study focusing on the development and validation of a prediction model where missing predictor values had been multiply imputed, showed a good performance of that prediction model when validated externally, even with the inclusion of predictors with a high proportion of missing values [43]. Invasive hemodynamics are recommended in the guidelines for the diagnoses of patients with PAH. However, echocardiography is an adequate noninvasive modality in patients with an evident diagnoses of PAH [44] in patients with CHD. The vast majority of patients (76%) had Eisenmenger syndrome, a cyanotic condition with a systolic pulmonary artery pressure up to 100 mm Hg. Due to the higher complication risk in patients with CHD-PAH cardiac catheterization was not performed in these patients. Patients with CHD-PAH often have abnormal hemostasis, including thrombocytopenia, making them at risk for both bleeding and thrombosis [6]. In particular, parietal thrombosis of enlarged proximal pulmonary arteries can be found in up to 20% of patients, it may cause peripheral embolization and pulmonary infarctions, and is associated with biventricular dysfunction and reduced pulmonary flow velocity [45]. Right heart catheterization only was performed at baseline in case diagnosis of PAH was not clearly evident at echocardiography. 5. Conclusion Prognosis is still poor in contemporary patients with CHD-PAH. Both baseline NT-pro-BNP serum level and right ventricular function are superior to clinical events in prognostication. Patients with both a NT-proBNP serum level ≥ 500 ng/L and TAPSE b 15 mm at echocardiography have a nine fold higher mortality rate than patients without both risk factors. These two baseline parameters should have a major impact on therapeutic management in patients with CHD-PAH, such as initiation of combination therapy. Conflict of interest None. Acknowledgments The work described in this study was carried out in the context of the Parelsnoer Institute (PSI). PSI is part of and funded by the Dutch Federation of University Medical Centers.

276

M.J. Schuuring et al. / International Journal of Cardiology 181 (2015) 270–276

References [1] G.P. Diller, M.A. Gatzoulis, Pulmonary vascular disease in adults with congenital heart disease, Circulation 115 (2007) 1039–1050. [2] M.J. Schuuring, B.J. Bouma, R. Cordina, M.A. Gatzoulis, W. Budts, M.P. Mullen, et al., Treatment of segmental pulmonary artery hypertension in adults with congenital heart disease, Int. J. Cardiol. 164 (2013) 106–110. [3] J.C. Vis, M.G. Duffels, P. Mulder, R.H.A.C.M. de Bruin-Bon, B.J. Bouma, R.M.F. Berger, et al., Prolonged beneficial effect of bosentan treatment and 4-year survival rates in adult patients with pulmonary arterial hypertension associated with congenital heart disease, Int. J. Cardiol. 164 (2013) 64–69. [4] G. Savarese, S. Paolillo, P. Costanzo, C. D'Amore, M. Cecere, T. Losco, et al., Do changes of 6-minute walk distance predict clinical events in patients with pulmonary arterial hypertension? A meta-analysis of 22 randomized trials, J. Am. Coll. Cardiol. 60 (2012) 1192–1201. [5] Z. Koyak, L. Harris, J.R. de Groot, C.K. Silversides, E.N. Oechslin, B.J. Bouma, et al., Sudden cardiac death in adult congenital heart disease, Circulation 126 (2012) 1944–1954. [6] N. Galiè, M.M. Hoeper, M. Humbert, A. Torbicki, J.-L. Vachiery, J.A. Barbera, et al., Guidelines for the diagnosis and treatment of pulmonary hypertension: the Task Force for the Diagnosis and Treatment of Pulmonary Hypertension of the European Society of Cardiology (ESC) and the European Respiratory Society (ERS), endorsed by the International Society of Heart and Lung Transplantation (ISHLT), Eur. Heart J. 30 (2009) 2493–2537. [7] N. Galiè, M. Beghetti, M. Gatzoulis, J. Granton, R. Berger, A. Lauer, et al., Bosentan Randomized Trial of Endothelin Antagonist Therapy-5 (BREATHE-5) Investigator. Bosentan therapy in patients with Eisenmenger syndrome: a multicenter, doubleblind, randomized, placebo-controlled study, Circulation 114 (2006) 48–54. [8] M.M. Hoeper, H.J. Bogaard, R. Condliffe, R. Frantz, D. Khanna, M. Kurzyna, et al., Definitions and diagnosis of pulmonary hypertension, J. Am. Coll. Cardiol. 62 (2013) D42–D50. [9] M.M. Hoeper, I. Markevych, E. Spiekerkoetter, T. Welte, J. Niedermeyer, Goaloriented treatment and combination therapy for pulmonary arterial hypertension, Eur. Respir. J. 26 (2005) 858–863. [10] V.V. McLaughlin, M.D. McGoon, Pulmonary arterial hypertension, Circulation 114 (2006) 1417–1431. [11] G.P. Diller, R. Alonso-Gonzalez, A. Kempny, K. Dimopoulos, R. Inuzuka, G. Giannakoulas, et al., B-type natriuretic peptide concentrations in contemporary Eisenmenger syndrome patients: predictive value and response to disease targeting therapy, Heart 98 (2012) 736–742. [12] P. Moceri, K. Dimopoulos, E. Liodakis, I. Germanakis, A. Kempny, G.P. Diller, et al., Echocardiographic predictors of outcome in Eisenmenger syndrome, Circulation 126 (12) (2012;Sep 18) 1461–1468. [13] A. Kempny, K. Dimopoulos, R. Alonso-Gonzalez, M. Alvarez-Barredo, O. Tutarel, A. Uebing, et al., Six-minute walk test distance and resting oxygen saturations but not functional class predict outcome in adult patients with Eisenmenger syndrome, Int. J. Cardiol. 168 (5) (2013) 4784–4789. [14] K. Dimopoulos, G.P. Diller, E. Koltsida, A. Pijuan-Domenech, S.A. Papadopoulou, S.V. Babu-Narayan, et al., Prevalence, predictors, and prognostic value of renal dysfunction in adults with congenital heart disease, Circulation 117 (2008) 2320–2328. [15] L. Daliento, J. Somerville, P. Presbitero, L. Menti, S. Brach-Prever, G. Rizzoli, et al., Eisenmenger syndrome. Factors relating to deterioration and death, Eur. Heart J. 19 (1998) 1845–1855. [16] G.P. Diller, K. Dimopoulos, C.S. Broberg, M.G. Kaya, U.S. Naghotra, A. Uebing, et al., Presentation, survival prospects, and predictors of death in Eisenmenger syndrome: a combined retrospective and case–control study, Eur. Heart J. 27 (2006) 1737–1742. [17] W.J. Cantor, D.A. Harrison, J.S. Moussadji, M.S. Connelly, G.D. Webb, P. Liu, et al., Determinants of survival and length of survival in adults with Eisenmenger syndrome, Am. J. Cardiol. 84 (1999) 677–681. [18] H. Baumgartner, P. Bonhoeffer, N.M.S. De Groot, F. de Haan, J.E. Deanfield, N. Galie, et al., ESC Guidelines for the management of grown-up congenital heart disease (new version 2010), Eur. Heart J. 31 (2010) 2915–2957. [19] J.C. Vis, R.H. de Bruin-Bon, B.J. Bouma, S.A. Huisman, L. Imschoot, K. van den Brink, et al., Congenital heart defects are under-recognised in adult patients with Down's syndrome, Heart 96 (2010) 1480–1484. [20] M.G.J. Duffels, J.C. Vis, R.L.E. van Loon, R.M.F. Berger, E.S. Hoendermis, A.P.J. van Dijk, et al., Down patients with Eisenmenger syndrome: is bosentan treatment an option? Int. J. Cardiol. 134 (2009) 378–383. [21] M.G.J. Duffels, J.C. Vis, R.L.E. van Loon, P.T. Nieuwkerk, A.P.J. van Dijk, E.S. Hoendermis, et al., Effect of bosentan on exercise capacity and quality of life in adults with pulmonary arterial hypertension associated with congenital heart disease with and without Down's syndrome, Am. J. Cardiol. 103 (2009) 1309–1315. [22] ATS statement: guidelines for the six-minute walk test, Am. J. Respir. Crit. Care Med. 166 (2002) 111–117. [23] L.G. Rudski, W.W. Lai, J. Afilalo, L. Hua, M.D. Handschumacher, K. Chandrasekaran, et al., Guidelines for the echocardiographic assessment of the right heart in adults:

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37] [38]

[39]

[40] [41]

[42] [43]

[44]

[45]

a report from the American Society of Echocardiography endorsed by the European Association of Echocardiography, a registered branch of the European Society of Cardiology, and the Canadian Society of Echocardiography, J. Am. Soc. Echocardiogr. Off. Publ. Am. Soc. Echocardiogr. 23 (2010) 685–713 (quiz 786–788). G. Simonneau, I.M. Robbins, M. Beghetti, R.N. Channick, M. Delcroix, C.P. Denton, et al., Updated clinical classification of pulmonary hypertension, J. Am. Coll. Cardiol. 54 (2009) S43–S54. Randomised, placebo-controlled trial of carvedilol in patients with congestive heart failure due to ischaemic heart disease. Australia/New Zealand Heart Failure Research Collaborative Group, Lancet 349 (1997) 375–380. M.J. Schuuring, S.M. Boekholdt, A. Windhausen, B.J. Bouma, M. Groenink, M. Keijzers, et al., Advanced therapy for pulmonary arterial hypertension due to congenital heart disease: a clinical perspective in a new therapeutic era, Neth. Heart J. 19 (2011) 509–513. P. Engelfriet, E. Boersma, E. Oechslin, J. Tijssen, M.A. Gatzoulis, U. Thilén, et al., The spectrum of adult congenital heart disease in Europe: morbidity and mortality in a 5 year follow-up period. The Euro Heart Survey on adult congenital heart disease, Eur. Heart J. 26 (2005) 2325–2333. P.M. Engelfriet, M.G.J. Duffels, T. Möller, E. Boersma, J.G.P. Tijssen, E. Thaulow, et al., Pulmonary arterial hypertension in adults born with a heart septal defect: the Euro Heart Survey on adult congenital heart disease, Heart 93 (2007) 682–687. A. Van De Bruaene, P. De Meester, J.-U. Voigt, M. Delcroix, A. Pasquet, J. De Backer, et al., Right ventricular function in patients with Eisenmenger syndrome, Am. J. Cardiol. 109 (2012) 1206–1211. A. Manes, M. Palazzini, E. Leci, M.L. Bacchi Reggiani, A. Branzi, N. Galiè, Current era survival of patients with pulmonary arterial hypertension associated with congenital heart disease: a comparison between clinical subgroups, Eur. Heart J. 35 (11) (2013) 716–724. I.I. Tulevski, H. Romkes, A. Dodge-Khatami, E.E. van der Wall, M. Groenink, D.J. van Veldhuisen, et al., Quantitative assessment of the pressure and volume overloaded right ventricle: imaging is a real challenge, Int. J. Cardiovasc. Imaging 18 (2002) 41–51. I.I. Tulevski, A. Dodge-Khatami, M. Groenink, E.E. van der Wall, H. Romkes, B.J.M. Mulder, Right ventricular function in congenital cardiac disease: noninvasive quantitative parameters for clinical follow-up, Cardiol. Young 13 (2003) 397–403. M.G.J. Duffels, M.N. van der Plas, S. Surie, M.M. Winter, B. Bouma, M. Groenink, et al., Bosentan in pulmonary arterial hypertension: a comparison between congenital heart disease and chronic pulmonary embolism, Neth. Heart J. 17 (2009) 334–338. M.G.J. Duffels, M. Hardziyenka, S. Surie, R.H.A.C.M. de Bruin-Bon, E.S. Hoendermis, A.P.J. van Dijk, et al., Duration of right ventricular contraction predicts the efficacy of bosentan treatment in patients with pulmonary hypertension, Eur. J. Echocardiogr. J. Work. Group Echocardiogr. Eur. Soc. Cardiol. 10 (2009) 433–438. T. Oosterhof, I.I. Tulevski, H.W. Vliegen, A.M. Spijkerboer, B.J.M. Mulder, Effects of volume and/or pressure overload secondary to congenital heart disease (tetralogy of fallot or pulmonary stenosis) on right ventricular function using cardiovascular magnetic resonance and B-type natriuretic peptide levels, Am. J. Cardiol. 97 (2006) 1051–1055. M.G.J. Duffels, P.M. Engelfriet, R.M.F. Berger, R.L.E. van Loon, E. Hoendermis, J.W.J. Vriend, et al., Pulmonary arterial hypertension in congenital heart disease: an epidemiologic perspective from a Dutch registry, Int. J. Cardiol. 120 (2007) 198–204. A. Rajdev, H. Garan, A. Biviano, Arrhythmias in pulmonary arterial hypertension, Prog. Cardiovasc. Dis. 55 (2012) 180–186. G.-P. Diller, R. Alonso-Gonzalez, K. Dimopoulos, M. Alvarez-Barredo, C. Koo, A. Kempny, et al., Disease targeting therapies in patients with Eisenmenger syndrome: response to treatment and long-term efficiency, Int. J. Cardiol. 167 (2013) 840–847. K. Iversen, A.S. Jensen, T.V. Jensen, N.G. Vejlstrup, L. Søndergaard, Combination therapy with bosentan and sildenafil in Eisenmenger syndrome: a randomized, placebocontrolled, double-blinded trial, Eur. Heart J. 31 (2010) 1124–1131. C. Hayes, Z. Johnson, L. Thornton, J. Fogarty, R. Lyons, M. O'Connor, et al., Ten-year survival of Down syndrome births, Int. J. Epidemiol. 26 (1997) 822–829. I.R. White, J.B. Carlin, Bias and efficiency of multiple imputation compared with complete-case analysis for missing covariate values, Stat. Med. 29 (2010) 2920–2931. D.B. Rubin, Mult Imput Non-Response Surv, first ed. John Wiley & Sons, New York, 1987. (n.d.). Y. Vergouwe, P. Royston, K.G.M. Moons, D.G. Altman, Development and validation of a prediction model with missing predictor data: a practical approach, J. Clin. Epidemiol. 63 (2010) 205–214. S.B. Eysmann, H.I. Palevsky, N. Reichek, K. Hackney, P.S. Douglas, Two-dimensional and Doppler-echocardiographic and cardiac catheterization correlates of survival in primary pulmonary hypertension, Circulation 80 (1989) 353–360. C.S. Broberg, M. Ujita, S. Prasad, W. Li, M. Rubens, B.E. Bax, et al., Pulmonary arterial thrombosis in Eisenmenger syndrome is associated with biventricular dysfunction and decreased pulmonary flow velocity, J. Am. Coll. Cardiol. 50 (2007) 634–642.