Keywords
end-stage heart failure - bosentan - pulmonary hypertension - cardiac transplantation
- mortality - survival - endothelin-1
Introduction
Pulmonary hypertension (PH) is a pathophysiological condition defined as an increase
in mean pulmonary arterial pressure (PAP) ≤25 mmHg at rest.[1] PH due to left heart failure is associated with increased pulmonary vascular resistance
(PVR) and elevated transpulmonary pressure gradient (TPG).[2] In addition, it carries a poor prognosis. Mortality rates of 57% in patients with
moderate PH compared with 17% in patients without PH have been reported after 28 months
of follow-up.[3] Survival rates are lower in patients with pulmonary hypertension and poor right
ventricular (RV) function compared to patients with elevated PAP but preserved RV
function.[4]
In end-stage heart failure patients, preoperatively existing PH is associated with
a high risk of right heart failure and mortality after cardiac transplantation.[5]
[6]
[7] According to the International Society for Heart and Lung Transplantation and the
German Standing Committee on Organ Transplantation, cardiac transplantation is contraindicated
as long as values of PAP, PVR, and/or TPG are above 40 mmHg, 240 dyn × s × cm,−5 and 15, respectively.[5]
[8] At present, no specific therapy for PH due to left heart disease is available.[1]
The endothelin A and B receptor antagonist bosentan has been shown to improve pulmonary
and systemic hemodynamics in patients with heart failure.[7]
[9]
[10] The plasma concentration of endothelin-1 (ET-1) is elevated in chronic heart failure
patients and is inversely related to left ventricular ejection fraction (EF) and cardiac
index (CI).[9]
[11] ET-1 seems to be a key mediator in the pathogenesis of PH in chronic heart failure.[9]
[11]
[12]
The REACH-1 (Research on Endothelin Antagonism in Chronic Heart Failure) trial has
used bosentan doses of 500 mg daily in patients with advanced heart failure. The duration
of the trial was planned for 24 weeks. Unfortunately, however, the REACH-1 trial has
been terminated prematurely due to concerns about elevations in hepatic transaminases.[13] Nevertheless, this study also demonstrated a significantly greater clinical improvement
in patients who received bosentan therapy for 6 months compared to controls. We therefore
administered bosentan in several of our patients with end-stage heart failure and
PH as an off-label use during recent years. But we used a considerably lower daily
bosentan dose than that used in the REACH-1 trial. We provide the results of a retrospective
data analysis based on our medical records. We performed this analysis to assess whether
low dose bosentan treatment is associated with improved pulmonary hemodynamics and
clinical outcome in cardiac transplant candidates with PH.
Materials and Methods
Patients and study design
Since January 2006, 183 patients with end-stage heart failure and longstanding PH
were eligible for cardiac transplantation at our institution ([Fig. 1]). All patients fulfilled the functional, echocardiographic, and hemodynamic indications
for cardiac transplantation provided by the German Standing Committee on Organ Transplantation.[8] However, all patients also had values of PAP >35 mmHg, PVR >240 dyn × s × cm−5, and/or TPG >15 mmHg despite optimal treatment with angiotensin-converting enzyme-inhibitors,
β-blockers, anticoagulants, digitalis derivatives or/and diuretics,[1] whereas PAP and PVR decreased with the administration of intravenous phosphodiesterase
III inhibitor. As mentioned before, elevated PAP, PVR, and/or TPG values are considered
a contraindication for cardiac transplantation.[8] To achieve transplantability, we decided to treat the aforementioned group of patients
with oral bosentan as off-label use. Generally, all 183 patients with PH and end-stage
heart failure were eligible for the off-label use of oral bosentan. However, initial
bosentan administration was done under intensive care unit monitoring, including Swan-Ganz
catheter measurement. Since intensive care unit capacity was limited, only 115 patients
received bosentan. Sixty-eight other patients did not receive bosentan. The decision
for bosentan administration was made by the head of the transplant unit, based on
inclusion criteria and according to bed capacity. All patients gave their written
informed consent to receive bosentan. The majority of patients started with an oral
bosentan dose of 62.5 mg bid (n = 3 with 32.5 mg bid, n = 8 with 125 mg bid) and were up-titrated to 125 mg bid. We performed a retrospective
data analysis of clinical outcome parameters in patients who received bosentan or
did not receive bosentan. For data analysis, we checked only those patients who had
at least three echocardiographic and hemodynamic measurements within a time interval
of 15 months. Patients with simultaneous sildenafil or iloprost treatment were excluded
from data analysis. We finally included only those patients who received bosentan
treatment or did not receive bosentan treatment during the entire follow-up period.
Thus, we could retrospectively analyze the data of 54 patients receiving bosentan
(BOS group) and of 28 patients not receiving bosentan (CON group). The latter group
served as the control group. Patients were maintained on 125 mg bid bosentan until
the end of the study, cardiac transplantation, or death. The CON group received standard
conventional medication. In both groups, echocardiographic parameters and hemodynamics
were assessed at baseline, e.g., before bosentan treatment (t0), after 4 months (t1; range 2–6 months), and after 12 months (t2; range 9–15 months). All measurements were performed as routine examinations at regular
time intervals. Hemodynamics were measured using a Swan-Ganz pulmonary artery thermodilution
catheter through the right internal jugular vein. Cardiac output (CO) was determined
in triplicate and calculated by the Stewart-Hamilton indicator dilution formula.[14] Cardiac index (CI) was calculated by CO, taking current body surface into account.
Central venous pressure (CVD), RV pressure, PAP, and PCWP were measured directly by
right heart catheterization. Systemic vascular resistance (SVR), PVR and TPG were
calculated according to standard formulas:
Figure 1 Study flow chart.
SVR = [(PAP–PCWP) × 80]/CI
PVR = [(PAP–PCWP) × 80]/CO
TPG = PAP-PCWP
Transthoracic echocardiography was performed to assess left ventricular end-diastolic
and end-systolic diameter, right ventricular shortening, and left ventricular ejection
fraction (Agilent Sonos 5500 device, Hewlett Packard, Andover, MA, USA). Since bosentan
shows liver toxicity,[15] biochemical safety parameters were also measured. In addition, we assessed clinical
symptoms such as nocturia and dyspnea by questionnaire and peripheral edema by physical
examination. We also registered clinical outcome parameters such as ventricular assist
device (VAD) implantation, cardiac transplantation, and death in both groups up until
June 2009.
Statistical analysis
We report categorical variables using the percentage of observations and express continuous
variables as mean values and standard deviation or median and interquartile range
as appropriate. We tested normal distribution of the data using the Kolmogorov-Smirnov
test. Normal distribution was considered present if p values were above 0.05. For comparative analyses between groups, we used Fisher's
exact test, McNemar test, unpaired t-test, and Mann-Whitney test as appropriate. We used ANOVA and the Friedman test for
evaluations of time-dependent effects. A p value <0.05 (two-tailed test) was considered significant.
We generated Kaplan-Meier estimates to investigate the association between bosentan
and survival probability during follow-up as a function of time after study inclusion.
Log-rank test was used to test for differences in survival rates between groups. We
then examined the associations between bosentan and mortality risk using Cox proportional
hazard analysis. Since treatment assignment was not based on random allocation and
the bosentan and control group were therefore not expected to be completely comparable
with regard to important covariates, we also analyzed whether propensity score adjustments
were necessary. We used 38 baseline demographic, anthropometric, clinical, and biochemical
variables for each patient, using logistic regression. The propensity score ranged
from a low of 0.13 to a high of 0.63 The discriminate power of the propensity score
was quantified by measurement of the receiver operating characteristics area and was
found to be 0.651 only (95% confidence interval: 0.465–0.836; p = 0.110), indicating that the two study groups were comparable with respect to baseline
and clinical characteristics. We used the statistical software package SPSS, version
18 (SPSS, Inc., Chicago, IL, USA) to perform the analysis.
Results
Baseline characteristics of the study groups are given in [Table 1]. Patients did not differ with regard to age, sex, and anthropometric data. Moreover,
the prevalence of smokers and concomitant diagnoses such as diabetes, hypertension,
elevated cholesterol levels, hyperuricemia, and dyspnea was similar between groups.
Likewise, concomitant medications did not differ between groups. The majority of patients
in each study group were in NYHA functional class III. The BOS group was, however,
significantly younger than the CON group and suffered less often from diagnoses other
than dilated cardiomyopathy or ischemic heart disease. More patients in the BOS group
had pacemaker implants than in the CON group. In addition, some hemodynamic parameters
such as right ventricular pressure, PAP, and PCWP were significantly higher in the
BOS group than in the CON group at baseline ([Table 2]). In the BOS group, the cut-off values for cardiac transplantation of PAP (>40 mmHg),
PVR (>240 dyn × s × cm−5), and TPG (>15) were exceeded in 75.9%, 67.9%, and 47.2% of patients, respectively.
The corresponding values for the CON group were 57.1%, 75.0%, and 25.0%, respectively.
Table 1
Baseline characteristics of the bosentan and control group.
|
Parameter
|
Bosentan group (n = 54)
|
Control group (n = 28)
|
p value
|
|
Age (years)
|
54 ± 11.0
|
59.2 ± 9.3
|
0.015
|
|
Height (cm)
|
176 ± 9
|
174 ± 9
|
0.269
|
|
Weight (kg)
|
82.0 ± 16.2
|
80.4 ± 14.0
|
0.665
|
|
Body mass index (kg/m2)
|
26.4 ± 4.2
|
25.8 ± 5.9
|
0.587
|
|
Gender (% males)
|
87.0
|
86.2
|
>0.999
|
|
Heart rate (per min)
|
77.4 ± 13.4
|
75.2 ± 11.0
|
0.472
|
|
Mean blood pressure (mmHg)
|
78.4 ± 12.1
|
80.2 ± 12.8
|
0.528
|
|
NYHA functional class ≥ III (%)
|
79.63
|
86.21
|
0.380
|
|
Diagnosis
|
|
Dilated cardiomyopathy (%)
|
46.3
|
34.5
|
0.351
|
|
Coronary heart disease (%)
|
44.4
|
27.6
|
0.240
|
|
Others[a] (%)
|
9.3
|
37.9
|
0.008
|
|
Concomitant diagnoses
|
|
Diabetes mellitus (%)
|
35.2
|
28.6
|
0.625
|
|
Hypercholesterolemia (%)
|
9.3
|
14.3
|
0.483
|
|
Hyperuricemia (%)
|
9.3
|
14.3
|
0.483
|
|
Chronic kidney disease (%)
|
11.1
|
28.6
|
0.064
|
|
Hypertension (%)
|
33.3
|
21.4
|
0.314
|
|
Nicotine abuse (%)
|
13.0
|
21.4
|
0.351
|
|
Pacemaker and defibrillator
|
|
Only pacemaker implantation (%)
|
57.4
|
32.1
|
<0.001
|
|
Only defibrillator implantation (%)
|
0
|
0
|
>0.999
|
|
Implantation of both (%)
|
7.4
|
7.1
|
>0.999
|
|
Medications
|
|
ACE inhibitors/AT blockers (%)
|
72.2
|
82.1
|
0.420
|
|
β-blockers (%)
|
79.6
|
71.4
|
0.420
|
|
Diuretics (%)
|
88.9
|
78.6
|
0.320
|
|
Digitalis derivatives (%)
|
42.6
|
25.0
|
0.150
|
|
Oral anticoagulants (%)
|
66.7
|
46.4
|
0.098
|
|
Aspirin (%)
|
39.3
|
24.0
|
0.201
|
a Amyloidosis, cor pulmonale, hypertrophic cardiomyopathy, condition after coronary
artery bypass grafting
Table 2
Hemodynamic and echocardiographic parameters of the bosentan (BOS) and control (CON)
group at baseline and during follow-up.
|
|
t0
|
t1
|
t2
|
p value
|
|
BOS group (n = 54)
|
BOS group (n = 54)
|
BOS group (n = 54)
|
|
CON group (n = 28)
|
CON group (n = 28)
|
CON group (n = 28)
|
|
Mean arterial pressure (mmHg)
|
|
▸BOS group
|
78.6 ± 12.1
|
77.7 ± 9.7
|
74.9 ± 9.0
|
0.007
|
|
▸CON group
|
80.2 ± 12.8
|
79.7 ± 12.7
|
73.4 ± 17.9
|
0.217
|
|
Hemodynamics
|
|
Central venous pressure (mmHg)
|
|
▸BOS group
|
13.5 ± 4.8
|
12.3 ± 5.7
|
11.2 ± 5.9
|
0.023
|
|
▸CON group
|
13.1 ± 5.2
|
13.3 ± 6.1
|
14.3 ± 4.7
|
0.611
|
|
Mean right ventricular pressure (mmHg)
|
|
▸BOS group
|
31.5 ± 8.6*
|
28.4 ± 9.5
|
25.8 ± 9.2
|
0.002
|
|
▸CON group
|
27.5 ± 6.9
|
27.6 ± 8.7
|
29.0 ± 8.9
|
0.944
|
|
Mean pulmonary arterial pressure (mmHg)
|
|
▸BOS group
|
45.2 ± 8.7**
|
42.2 ± 11.3
|
38.4 ± 12.5
|
0.001
|
|
▸CON group
|
39.6 ± 6.7
|
38.1 ± 8.5
|
37.4 ± 8.5
|
0.537
|
|
Pulmonary capillary wedge pressure (mmHg)
|
|
▸BOS group
|
30.6 ± 6.7*
|
29.0 ± 8.5
|
26.9 ± 9.3
|
0.049
|
|
▸CON group
|
26.6 ± 6.9
|
23.9 ± 9.2
|
25.6 ± 8.1
|
0.446
|
|
Cardiac output (l/min)
|
|
▸BOS group
|
3.4 ± 0.8
|
3.9 ± 1.2
|
4.1 ± 1.0
|
0.001
|
|
▸CON group
|
3.4 ± 0.9
|
3.6 ± 1.0
|
3.6 ± 1.0
|
0.154
|
|
Cardiac index (l/minm2)
|
|
▸BOS group
|
1.7 ± 0.5
|
2.0 ± 0.5
|
2.1 ± 0.4
|
0.003
|
|
▸CON group
|
1.6 ± 0.5
|
1.9 ± 0.7
|
1.9 ± 0.4
|
0.559
|
|
Systemic vascular resistance (dyn × s × cm−5)
|
|
|
|
|
|
▸BOS group
|
1600 ± 508
|
1410 ± 528
|
1236 ± 456
|
0.001
|
|
▸CON group
|
1717 ± 544
|
1644 ± 672
|
1432 ± 584
|
0.095
|
|
Pulmonary vascular resistance (dyn × s × cm−5)
|
|
|
|
|
|
▸BOS group
|
382 ± 208
|
301 ± 191
|
231 ± 127
|
0.001
|
|
▸CON group
|
336 ± 133
|
348 ± 193
|
293 ± 164
|
0.230
|
|
Transpulmonary pressure gradient (mmHg)
|
|
▸BOS group
|
14.7 ± 6.9
|
13.4 ± 6.7
|
11.2 ± 5.9
|
0.001
|
|
▸CON group
|
13.0 ± 6.1
|
14.2 ± 7.6
|
11.9 ± 4.6
|
0.209
|
|
Echocardiography
|
|
Left ventricular end-diastolic diameter (mm)
|
|
▸BOS group
|
68.4 ± 11.5
|
69.1 ± 11.3
|
66.2 ± 12.8
|
0.179
|
|
▸CON group
|
68.4 ± 12.3
|
68.2 ± 12.9
|
65.5 ± 14.1
|
0.334
|
|
Left ventricular end-systolic diameter (mm)
|
|
▸BOS group
|
60.6 ± 12.4
|
60.3 ± 13.4
|
57.8 ± 15.4
|
0.385
|
|
▸CON group
|
59.6 ± 13.3
|
60.3 ± 14.5
|
57.5 ± 15.2
|
0.405
|
|
Right ventricular fractional shortening (%)
|
|
▸BOS group
|
11.8 ± 6.9
|
12.0 ± 5.6
|
12.7 ± 7.4
|
0.987
|
|
▸CON group
|
12.8 ± 6.1
|
13.5 ± 8.0
|
13.1 ± 7.6
|
0.626
|
|
Left ventricular ejection fraction (%)
|
|
▸BOS group
|
28.9 ± 13.2
|
29.4 ± 11.8
|
29.6 ± 13.4
|
0.466
|
|
▸CON group
|
29.8 ± 12.8
|
31.1 ± 14.0
|
28.9 ± 10.4
|
0.067
|
t0: baseline; t1: 4 months after inclusion; t2: 12 months after inclusion; *, **: p < 0.05 and p < 0.01 vs. controls at the same time point
In the BOS group, the median daily bosentan dose was 125 mg (interquartile range:
125–125 mg; mean ± SD: 139 ± 50 mg/d) at the beginning and 250 mg (interquartile range:
250–250 mg; mean ± SD: 216 ± 77 mg/d) at the end of the follow-up period. Down titration
of bosentan was necessary in 6 patients due to marked decreases in PAP and PVR. In
2 of these 6 patients, a parallel increase in aspartate aminotransferase and alanine
aminotransferase occurred, whereas in 1 of the 6 patients only alanine aminotransferase
increased.
In the CON group, hemodynamic and echocardiographic parameters did not change significantly
during follow-up ([Table 2]). In the BOS group, echocardiographic parameters also remained unchanged. However,
mean blood pressure and all measured hemodynamic parameters showed significant improvements
in the BOS group during follow-up. The three determinants of PVR (i.e., PAP, PCWP,
and CO) improved significantly from t0 to t2 by 18.0%, 15.7%, and 23.5%, respectively. Consequently, mean PVR decreased significantly
from a mean of 382 dyn × s × cm−5 to 256 dyn × s × cm−5. In the BOS group, the percentage of patients with values below the cut-offs increased
from t0 to t2 for PAP from 24.1% to 44.4% (p = 0.007), for PVR from 32.1% to 56.6% (p = 0.013), and for TPG from 52.8% to 73.6% (p = 0.007). SVR values declined by 17.0%. None of the patients in the BOS group developed
hypotension (mean arterial pressure <50 mmHg).
In the CON group, the percentages increased nonsignificantly from 42.9% to 53.6% for
PAP (p = 0.581), from 25.0% to 39.3% for PVR (p = 0.388), and remained constant for TPG (75.0% vs. 71.4%, p >0.999). SVR values tended to decline to a similar extend compared with the BOS group.
NYHA functional class did not change significantly during follow-up, neither in the
BOS group nor in the CON group (data not shown).
Biochemical safety parameters are presented in [Table 3]. In the BOS group, mean concentrations of liver enzymes such as aspartate aminotransferase,
alanine aminotransferase, and gamma-glutamyltransferase remained constant. In the
BOS group, there was a decrease in hematocrit values, which also tended to occur in
the CON group. Sodium concentrations improved significantly in the BOS group from
t0 to t2.
Table 3
Biochemical parameters of the bosentan and control group at baseline and during
follow-up.
|
|
t0
|
t1
|
t2
|
p value
|
|
BOS group (n = 54)
|
BOS group (n = 54)
|
BOS group (n = 54)
|
|
CON group (n = 28)
|
CON group (n = 28)
|
CON group (n = 28)
|
|
Liver function
|
|
Aspartate aminotransferase (U/l)
|
|
▸BOS group
|
38.7 ± 37.0
|
28.22 ± 8.76
|
28.2 ± 8.8
|
0.319
|
|
▸CON group
|
28.0 ± 9.1
|
29.2 ± 10.4
|
31.4 ± 12.5
|
0.471
|
|
Alanine aminotransferase (U/l)
|
|
▸BOS group
|
54.4 ± 12.4
|
25.9 ± 14.0
|
26.4 ± 15.1
|
0.093
|
|
▸CON group
|
33.4 ± 18.2
|
28.3 ± 14.6
|
26.7 ± 36.9
|
0.042
|
|
gamma-glutamyltransferase (U/l)
|
|
|
|
|
|
▸BOS group
|
114 ± 88
|
103 ± 88
|
112 ± 99
|
0.048
|
|
▸CON group
|
150 ± 137
|
144 ± 124
|
190 ± 182
|
0.303
|
|
Total bilirubin (mg/dL)
|
|
▸BOS group
|
1.30 ± 0.86
|
0.92 ± 0.53
|
0.88 ± 0.52
|
0.163
|
|
▸CON group
|
0.97 ± 0.48
|
1.02 ± 0.66
|
1.15 ± 0.89
|
0.608
|
|
Kidney function
|
|
Creatinine (mg/dL)
|
|
▸BOS group
|
1.39 ± 0.4
|
1.45 ± 0.51
|
1.67 ± 0.80
|
0.080
|
|
▸CON group
|
1.39 ± 0.4
|
1.43 ± 0.44
|
1.51 ± 0.48
|
0.519
|
|
Blood urea nitrogen (mg/dL)
|
|
▸BOS group
|
68.2 ± 36.4
|
75 ± 47
|
85.0 ± 56.9
|
0.093
|
|
▸CON group
|
76.2 ± 43.6
|
82 ± 54
|
84.6 ± 50.0
|
0.639
|
|
Hematology
|
|
Leukocytes (109/l)
|
|
▸BOS group
|
8.97 ± 12.5
|
9.15 ± 12.4
|
8.92 ± 12.7
|
0.722
|
|
▸CON group
|
8.41 ± 3.16
|
8.44 ± 2.86
|
8.20 ± 2.29
|
0.857
|
|
Red blood cells (1012/l)
|
|
|
|
|
|
▸BOS group
|
4.33 ± 0.63
|
4.23 ± 0.66
|
4.17 ± 4.12
|
0.106
|
|
▸CON group
|
4.63 ± 1.9
|
4.24 ± 0.64
|
4.12 ± 0.83
|
0.435
|
|
Hemoglobin (mg/dL)
|
|
▸BOS group
|
12.5 ± 1.9
|
12.6 ± 2.1
|
12.5 ± 3.13
|
0.215
|
|
▸CON group
|
14.1 ± 5.7
|
13.0 ± 2.53
|
12.7 ± 4.1
|
0.025
|
|
Hematocrit (%)
|
|
▸BOS group
|
38.1 ± 5.2
|
37.2 ± 6.1
|
36.4 ± 5.18
|
0.005
|
|
▸CON group
|
40.4 ± 11.8
|
38.3 ± 7.6
|
37.8 ± 12.7
|
0.063
|
|
Platelets (109/l)
|
|
▸BOS group
|
197 ± 61
|
212 ± 72
|
212 ± 78
|
0.145
|
|
▸CON group
|
198 ± 81
|
204 ± 84
|
217 ± 107
|
0.476
|
|
Electrolytes
|
|
Sodium (mmol/l)
|
|
▸BOS group
|
137 ± 4
|
136 ± 4
|
138 ± 4
|
0.008
|
|
▸CON group
|
137 ± 6
|
136 ± 5
|
137 ± 5
|
0.513
|
|
Potassium (mmol/l)
|
|
▸BOS group
|
4.25 ± 0.50
|
4.08 ± 0.54
|
4.08 ± 0.54
|
0.611
|
|
▸CON group
|
4.05 ± 0.76
|
4.09 ± 0.80
|
4.00 ± 0.51
|
0.889
|
t0: baseline; t1: 4 months after inclusion; t2: 12 months after inclusion
In the BOS group, 35 patients (64.8%) were not hospitalized, whereas 15 patients (27.8%),
one patient (1.9%), and three patients (5.6%) were hospitalized once, twice, and three
times, respectively, during follow-up. The corresponding values for the CON group
were 20 (71.4%), 8 (28.6%), zero, and zero, respectively. In the BOS and CON group,
the percentage of patients with nocturia (22.2% vs. 28.6%, respectively), peripheral
edema (20.4% vs. 14.4%, respectively), and dyspnea (18.5% vs. 7.1%, respectively)
did not differ significantly (p = 0.205–0.592).
During follow-up, ten patients in the BOS group (18.5%) and one patient in the CON
group (3.6%) were transplanted (p = 0.088). All transplanted patients were alive at the end of the follow-up period.
One patient in the BOS group and 11 patients in the CON group died during follow-up.
Causes of death were acute lung edema in the BOS group, and multiple organ failure
(6 ×), low output syndrome (2 ×), intracranial bleeding (1 ×), intestinal ischemia
(1 ×), and pancytopenia (1 ×) in the CON group. [Fig. 2] illustrates the Kaplan-Meier survival curves on the waiting list after t2 according to study group. One-year survival on the waiting list was 93.3% in the
BOS group and 70.7% in the CON group (p = 0.020). The corresponding values for freedom from cardiac related deaths was 93.3%
and 100%, respectively (p = 0.407).
Figure 2 Probability of survival on the waiting list after t2 in the bosentan group and the control group.
In the univariate Cox regression analysis, bosentan treatment was significantly related
to reduced risk of mortality on the waiting list (RR = 0.109; 95%CI: 0.013–0.881;
p = 0.038). Bosentan treatment remained an independent predictor of reduced mortality
risk on the waiting list after propensity score adjustment (RR = 0.107; 95%CI: 0.013–0.869;
p = 0.036).
Discussion
Our data demonstrate that patients with end-stage heart failure and PH not receiving
bosentan have a poor prognosis. Results support earlier findings of high mortality
rates in heart failure patients with PH.[7]. However, this study could also demonstrate significantly higher survival rates
of end-stage heart failure patients with PH who were treated with bosentan compared
to patients who were not treated with bosentan while awaiting cardiac transplantation.
Moreover, compared with the patients not treated with bosentan a higher percentage
of bosentan-treated patients tended to be transplanted during follow-up.
Although the REACH-1 trial was stopped prematurely due to liver toxicity, this earlier
study already demonstrated a significantly greater clinical improvement in patients
with advanced heart failure who received bosentan therapy for 6 months.[13] Thus, our data support the assumption that bosentan may be a new treatment option
for end-stage heart failure patients. Compared to the REACH-1 trial, our retrospective
data analysis is based on a longer duration of follow-up and a sicker patient population.
In addition, our analysis includes invasive hemodynamic assessment, and defined hemodynamic
criteria including TPG for enrolment.
The development of new and more effective pharmacological treatment strategies for
PH in patients with end-stage heart failure is urgently needed. In this context, bosentan
seems to be a promising candidate. In our analysis, bosentan treatment was associated
with an improvement in various hemodynamic parameters. For example, PVR decreased
by 33% in the BOS group, whereas it declined only nonsignificantly by 12.8% in the
CON group ([Table 2]). Our results are in line with the fact that bosentan treatment is also able to
reduce PVR in patients with pulmonary arterial hypertension.[16]
[17] In addition, our data analysis demonstrates an improvement in RV pressure, which
is an important prognostic value for survival in patients with chronic heart failure.[4] The improvement in RV pressure and the partial revision in PH may therefore be an
important explanation for the fact that one year after t2 the survival rate on the waiting list was 22% higher (93% vs. 71%) in the BOS group
than in the CON group. Note that patients with end-stage heart failure are at an increased
risk for cardiac decompensation. This can lead to systemic immunosuppression and thus
to an increased risk of developing multiple organ failure.
In our analysis, bosentan treatment was associated with a more pronounced reduction
in PVR than in SVR. Similar results have been observed in patients with chronic heart
failure and high circulating ET-1 concentrations who received short-term intravenous
bosentan.[10] Likewise, our data of unchanged echocardiographic measurements in the BOS group
confirm an earlier multicenter randomized controlled trial (RCT) of bosentan in heart
failure patients with severe systolic dysfunction and PH.[18] In that earlier RCT, echocardiographic measurements were not influenced by bosentan
treatment.
Since bosentan is an ET-1 receptor antagonist, our data support the assumption that
ET-1 plays an important role in the etiology of PH in heart failure patients. This
assumption is in line with the fact that the plasma concentrations of ET-1 are increased
in patients with heart failure and correlate with the prognosis.[9]
[11] In animal models of heart failure, ET-1 antagonists are also able to improve hemodynamics
and survival.[19]
In addition to the improvement in the survival of patients on the waiting list, bosentan
seems to offer an important additional benefit. Due to the significant decrease in
hemodynamic parameters, a relatively high number of patients in the BOS group compared
to the CON group could be transplanted during follow-up (18.5% vs. 3.6%). Since all
patients had end-stage heart failure, cardiac transplantation was the last option
for these patients. None of the transplanted patients had died by the end of the follow-up
period, supporting the assumption that survival after cardiac transplantation seems
to be similar in patients without pre-existing PH and in patients where preoperatively
existing PH is reversible.[7]
Elevated liver aminotransferase levels seemed to be characteristic for bosentan treatment.[9]
[16] In the present data analysis, however, the BOS group in its entirety showed no pathological
elevations in serum alanine aminotransferase or aspartate aminotransferase concentrations
from t0 to t2. This is probably due to the fact that we prescribed only 250 mg bosentan daily,
whereas others have used daily doses of 500 mg.[4] In previous investigations, peripheral edema occurred more frequently in patients
treated with bosentan.[9] In our retrospective analysis, clinical complications such as edema and dyspnea
as well as the percentage of patients who reported nocturemia were not adversely affected
by bosentan treatment. In addition, the hospitalization rate was similar between the
BOS and CON groups.
In contrast to earlier results,[16] the BOS group showed no improvement in NYHA functional class. However, it should
be noted that NYHA functional class is a subjective parameter. Hemodynamic parameters
are much more reliable for the assessment of cardiac function. Initial CI measurements
demonstrate that our patients suffered from advanced heart failure, leaving only a
limited possibility of improvement in NYHA functional class.
It must be stated that our data analysis has several limitations due to its retrospective
nature. It lacks randomization, blinding, and a clear study protocol. In addition,
this retrospective data analysis is short of defined dose up- and down-titration,
follow-up visits, and adverse event reporting. Although we have performed propensity
score adjustments to assess mortality rates, this can not compensate for nonrandomization.
Finally, the study is a single center report without an independent core lab, adjudicated
event committee, data monitoring and safety committee. Since earlier randomized controlled
trials with bosentan focused on patients with pulmonary arterial hypertension,[16]
[17] a randomized controlled multicenter study with survival on the waiting list as the
primary endpoint and cardiac transplantation as a secondary endpoint is urgently needed.
In conclusion, our data demonstrate that bosentan treatment is associated with improvements
in hemodynamics and clinical outcome in end-stage heart failure patients with PH.
If these results are confirmed by randomized controlled trials bosentan may represent
a treatment option.
