Article Text
Abstract
Objective Pulmonary arterial hypertension is known to be associated with increased expression of endothelin (ET)-1 and its precursor big ET-1. Therefore, we hypothesised that in children with pulmonary hypertension (PH) altered levels of ET-1 and big ET-1 may have clinical and prognostic impact.
Methods Sixty-six children with different forms of PH (mean age 10.4±9.7 years) were included. Blood samples were taken from the pulmonary artery and a systemic artery. Levels of ET-1/big ET-1 were measured via ELISA method and compared with clinical and haemodynamic data. To assess prognostic relevance, Kaplan–Meier survival analysis was conducted with definition of end point as the composite of mortality, lung transplantation, use of intravenous prostanoids and Potts shunt creation.
Results ET-1 levels ranged between 0.09 and 11.64 (mean 1.48±2.34) fmol/mL, and big ET-1 levels between 0.05 and 2.92 (mean 0.84±0.58) fmol/mL. No significant relationships were found between ET-1/big ET-1 levels and functional class as well as haemodynamic indices of PH severity. Mean follow-up after catheterisation was 63.2±44.1 months. While 31 of the 66 (47%) patients with PH reached a predefined end point, there was no significant relation between levels of ET-1/big ET-1 and patient outcome.
Conclusions Although children with PH had alterations in ET-1/big ET-1 expression, which may reflect changes in net release or lung clearance, levels of ET-1/big ET-1 showed no correlation with clinical and haemodynamic parameters, and were not able to predict outcome.
Statistics from Altmetric.com
Introduction
Pulmonary arterial hypertension (PAH) is a rare disease with an incidence of 2–3 per million and a prevalence of 25–50 per million.1 Although recent developments of PAH-targeted therapies resulted in an improvement of prognosis, mortality is still high even in the paediatric population.2–6 The underlying pulmonary vascular disease (PVD) is characterised by pathologic remodelling of the pulmonary arterioles with intima thickening and media hypertrophy, resulting in increased pulmonary vascular resistance and pressure, and eventually right ventricular failure.1 ,7 ,8 The aetiology of PVD is multifactorial, but endothelin-1 (ET-1), a potent and long-lasting vasoconstrictive peptide, has been implicated in the pathogenesis of several models of pulmonary hypertension (PH) and appears to play a central role in the remodelling process,9 ,10 and furthermore, the endothelin pathway serves as an important target in PAH-specific drug therapy.6
ET-1 is produced and secreted into circulation by the endothelial cells of the pulmonary vessels after transformation from its precursor form big ET-1 by the ET-converting enzyme (ECE-1). Elevated plasma ET-1 levels have been detected in diverse forms of PAH and also in experimental disease models.7 Different stimuli, like hypoxia, ischaemia, shear stress, catecholamines and insulin were shown to increase synthesis of ET-1 in PAH. In addition, augmented ECE-1 levels and up-regulation of ET receptors have also been described,7 and in patients with severe PAH, all components of the endothelin system can be up-regulated within the pulmonary vascular endothelial cells.7 The lung usually acts as a clearance organ for ET-1, removing more than 50% of circulating ET-1 from the blood on each passage through the lungs. In patients with PAH however, the ratio of ET-1 levels in systemic arterial plasma (after lung passage) to that of mixed venous plasma (entering the lung) is frequently close to or greater than unity, suggesting reduced net clearance and, in some patients, net release.11 ,12
In view of these findings, ET-1 and big ET-1 might be potential candidates for biomarkers in patients with PH. Levels of plasma ET-1 and its precursor forms have been previously shown to correlate with disease severity in adults with PH and to be predictive in left heart disease associated with target-organ damage in systemic hypertension.13–15
The aim of this study was to assess alterations in the expression of ET-1 and big ET-1 in children and to evaluate its correlation with their severity of PH. In addition, we purposed to assess the predictive value of these biomarkers in paediatric PH.
Methods
The study population included all children referred for invasive clinical haemodynamic evaluation in the catheterisation laboratory to confirm PH (defined as mean pulmonary artery (PA) pressure >25 mm Hg) at our tertiary referral centre between March 2001 and February 2010. Before catheterisation, all patients underwent assessment of their functional status (age-dependent New York Heart Association (NYHA) class for older children and adolescents) or Ross class (for infants and younger children).16 Clinical data were retrospectively obtained from hospital medical records including date of birth, gender, anatomic diagnoses and medication. All patients or parents of the patients gave written informed consent to a specific protocol for patients with PH approved by the local ethics committee.
Haemodynamic assessment
Diagnostic cardiac catheterisation was performed under local anaesthesia through a percutaneous femoral approach using low dosages of midazolam, ketamine or propofol sedation. Measurements of baseline haemodynamics included venous and arterial saturations, blood gases, systemic and PA pressures, left atrial (or pulmonary capillary wedge) pressure and right atrial pressure in the standard manner with fluid-filled catheters. Oxygen saturations were measured by co-oximetry after sampling in the superior vena cava (SsvcO2), pulmonary vein (SpvO2), PA (SvO2) and systemic artery (SA; SaO2). Systemic and pulmonary blood flows were estimated from the Fick equation and systemic and pulmonary vascular resistances were calculated from standard equations (mean arterial pressure − mean atrial pressure divided by flow). Blood flow and vascular resistances were indexed to body surface area.
Blood sampling and ET-1/big ET-1 assay
The protocol included the assessment of ET-1 and its precursor big ET-1 levels during invasive cardiac catheterisation. Blood samples of 5 mL each were taken simultaneously from the PA and the femoral artery (SA) under baseline condition and spontaneous room-air respiration before vasoreactivity testing. The blood was subsequently transferred to potassium EDTA containing tubes which were placed on ice. After the blood was centrifuged for 20 min (1800 g at 4°C), the samples were frozen at −80°C in polypropylene tubes until further analysis. Levels of ET-1 and big ET-1 in plasma were measured by ELISA method with commercially available kits (BI-20052, Biomedica, Vienna, Austria).
Statistical analysis
Data were presented as mean±1 SD. Correlations were obtained by Pearson's correlation test. To compare values of ET-1 and big ET-1 of patients with different forms of PH, we used Mann–Whitney U test and Welch t test, as appropriate. To assess the prognostic relevance of ET-1 and big ET-1, Kaplan–Meier survival analysis was conducted with definition of end point as the composite of all-cause mortality, lung transplantation, use of intravenous prostanoids and creation of a Potts shunt. A log-rank test was conducted to determine the difference between the PH groups. The assumption of proportional hazards is not fulfilled in the comparison of the idiopathic PAH (IPAH) and PAH-congenital heart disease (CHD) group, accordingly no log-rank test was applied. Cox proportional hazards analysis was used to assess the influence of multiple variables, for example, ET-1 and big ET-1, on the end point. Number of variables were selected by a relaxed version of the 10 event per variable rule following the argumentation of Vittinghoff and McCulloch.17 Somers rank correlation, collinearity and the number of events were considered in the variable selection for each model. For model validity, we conducted an internal bootstrap validation with 200 replicates.
Statistical analysis was performed using IBM (IBM, Armonk, New York, USA), SPSS Statistics software V.19.0 (SPSS, Chicago, Illinois, USA) and R V.3.0.2 (R Foundation for Statistical Computing, Vienna, Austria) with R-package rms V.4.2-1. The null hypothesis was rejected when p<0.05.
Results
Study population
Sixty-six children with PH (defined as mean PA pressure >25 mm Hg) were included in the study (table 1). PH was related to IPAH in 26 patients (15 females, mean age 9.8±8.1 years), associated with CHD (PAH-CHD) in 29 patients (17 females, mean age 13.5±11.3 years) and associated with interstitial lung disease (PH-LD) in 11 patients (8 females, mean age 3.9±4.2 years). Additional 14 patients (11 females, mean age 7.3±4.8 years) with an atrial septal defect (ASD) without PH were also studied at the time of interventional ASD closure and served as a control group with increased pulmonary blood flow (IPF) and normal PA pressures.
Levels of ET-1, big ET-1 and their SA/PA ratio
Levels of ET-1 in the PA showed a wide range between 0.09 and 11.64 fmol/mL (mean 1.48±2.34 fmol/mL) and big ET-1 between 0.05 and 2.92 fmol/mL (mean 0.84±0.58 fmol/mL) in the entire PH group, as well as between 0.1 and 8.64 fmol/mL (mean 1.21±2.29 fmol/mL), respectively 0.02 and 4.38 fmol/mL (mean 0.87±0.37 fmol/mL) in patients with IPF (table 2 and figure 1).
Levels of ET-1 (PA) did not differ between the PH subgroups and compared with patients with IPF (p>0.05 in all) (figure 1). Patients with PAH-CHD and those with PH-LD had significant higher levels of big ET-1 than patients with IPAH (p=0.04 in both). There was no significant difference of big ET-1 levels between the entire PH group and patients with IPF.
The SA-PA ratio of ET-1 (normal <1) was unity (=1) or higher in 38 patients (58%) (IPAH 14 patients (54%), PAH-CHD 15 patients (52%), PH-LD 9 patients (82%) and in 10 of the patients with IPF (71%)). There was no significant difference of ET-1 SA/PA ratios between the PH subgroups and patients with IPF (p>0.05 in all).
Clinical and haemodynamic correlations
There were no significant relations of the ET-1/big ET-1 levels and their corresponding SA/PA ratios with NYHA/Ross functional class. Only in NYHA/Ross functional class IV, the corresponding ET-1 levels were higher than in patients of class II (p=0.03) (figure 2).
ET-1 and big ET-1 levels showed no significant correlations with haemodynamic indices of severity of PH such as cardiac index (ET-1: r=0.02, p=0.90; big ET-1: r=0.13, p=0.31), mean pulmonary art pressure (mPAP) (ET-1: r=0.39, p=0.70; big ET-1: r=0.05, p=0.69) and pulmonary vascular resistance index (ET-1: r=0.002, p=0.97; big ET-1: r=0.22, p=0.09). There was also no relationship between SA/PA ratio of ET-1 and big ET-1 with mPAP/mean systemic arterial pressure ratio (ET-1: r=0.09, p=0.49; big ET-1: r=0.19, p=0.14) and all remaining assessed haemodynamic parameters.
Overall survival
Over a mean follow-up of 63.2±44.1 months, 31 of the 66 (47%) patients with PH reached a predefined end point (IPAH: 14 patients, PAH-CHD: 10 patients, PH-LD: 7 patients): 25 patients died, in 2 patients the administration of continuous intravenous prostanoid infusion was required, 3 patients underwent surgical implantation of a Potts shunt and 1 patient underwent combined heart-lung transplantation (figure 3). Overall 2-year, 3-year and 5-year event-free survivals were 73.4% (63.4–85.1), 70.3% (59.9–82.4) and 62.6% (51.5–76.2), respectively. With a 5-year event-free survival of 40.0% (18.7–85.5), patients with PH-LD showed significantly worse outcome than patients with PAH-CHD (p=0.015) and a slightly worse outcome compared with patients with IPAH (p=0.056) (figure 3).
Multivariate analysis
On multivariate analysis (figure 4), no significant relation between levels of ET-1 and big ET-1 (and the corresponding SA/PA ratios) and patient outcome could be detected in the entire group of patients with PH. Only in the subgroup of PAH-CHD patients, higher big ET-1 levels (HR 7.71, 95% CI 2.09 to 28.39) and an increased ET-1 SA/PA ratio (HR 2.21, 95% CI 1.05 to 4.66) predicted worse outcome.
Discussion
This study is the first presenting levels of ET-1 and big ET-1 in a large paediatric PH patient cohort with corresponding clinical, haemodynamic and long-term outcome data in order to evaluate the predictive value of ET-1 and big ET-1 in this specific patient group. Our data demonstrate that although children with abnormal pulmonary haemodynamics may have some alterations in ET-1 and big ET-1 expression, which may reflect individual changes in net release or lung clearance, levels of ET-1 and big ET-1 showed no significant correlation with clinical as well as haemodynamic parameters, and were not able to predict outcome in paediatric PH. These results suggest that the role of ET-1 and big ET-1 as biomarkers in children and adolescents with PH is limited.
Our findings are in contrast to a previous report of a small study with only 16 adults with PAH, where the authors detected a considerable correlation between pulmonary vascular resistance, mean PA pressure, cardiac output, cardiac index, 6 min walk data and elevated plasma levels of big ET-1 as well as mature ET-1 plasma levels at all sites of blood sampling.18 Presumably, there might be many potential reasons for these different results (specific characteristics of children vs adolescents with PH; different PH aetiologies and mechanisms of the development of PVD; different assays used for the assessment of ET-1 levels), however, the reported correlations may also be a statistical phenomenon due to the small number of patients. Furthermore, in this previous study, 8 of the 16 patients with PAH were already receiving prostacyclin analogues at the time of ET-1 or big ET-1 measurement. Our results were confirmed by the study of Montani et al15 who also found in a similar study design that ET-1 plasma concentrations were not related to prognosis in patients with PAH.
In contrast to Stewart et al19 and Dupuis et al20 who found a tendency of higher ET-1 plasma levels in the arterial blood when compared with those in the venous blood that points to a transpulmonary gradient, our presented study shows similar big ET-1 and mature ET-1 plasma levels at all sites of blood sampling. Concerning mature ET-1, this observation may be caused by rapid pulmonary clearance of ET-1. In the case of big ET-1, a part of it released by pulmonary vasculature may be cleaved by endothelin converting enzyme (ECE-1), as suggested from findings of an increased immunoreactivity of lung ECE-1 in patients with PH, which equalises plasma levels at arterial and venous sampling sites.9–11
A central role in the pathogenesis of PAH has been proposed for ET-1 because it is overexpressed in patients with PAH and has been implicated in vascular remodelling. In addition to PAH, altered expression of ET-1 and big ET-1 have been reported also in other forms of abnormal pulmonary haemodynamics as in IPF due to left-to-right shunt, which was also confirmed by our results, thereby suggesting that increased expression of ET-1 and big ET-1 may also contribute to the vascular abnormalities associated with shunt disorders.19 ,21 Significant elevated ET-1 levels have been documented in various forms of PH, and consequently marked higher ratio of ET-1 levels in systemic arterial plasma to that of mixed venous plasma.19 The latter was also confirmed in our study. The majority of patients (61%) showed an ET-1 SA/PA ratio of unity or >1, which can be regarded as a typical finding in patients with PH. Furthermore, the PA ET-1/big ET-1 ratio was <1 in most of the patients (59%) as a consequence of increased ET-1 production, since a ratio >1 would rather suggest a pathologic ET-1 clearance or reduced degradation.
Correlation between levels of ET-1 as well as big ET-1 and haemodynamic indices of severity of PH
The elevation of ET-1 plasma levels in patients with IPAH as well as in patients with secondary PH, due to pulmonary diseases of known causes or heart diseases, was shown by Stewart et al19 and a strong correlation between PA pressure and ET-1 plasma levels was observed in patients with associated forms of PH.20 ,22 Nootens et al23 found correlations between ET-1 plasma levels and right atrial pressure as well as PA oxygen saturation in patients with IPAH. Giaid et al9 showed increased ET-1 expression in lung vascular endothelial cells in patients with IPAH and associated forms of PH, which indicated an association between tissue ET-1 expression and circulating ET-1. Remarkably, in our study we found no significant relationships between levels of ET-1 as well as big ET-1 and clinical as well as haemodynamic indices of severity of PH. As mentioned above, this may be due to differences in children compared with adults regarding aetiology of PH and pathobiology of the corresponding PVD. Future studies should be conducted to evaluate these differences in more detail.
Correlation between levels of ET-1/big ET-1 and patient outcome
Data about the predictive value of ET-1 and big ET-1 levels in children with PH are lacking so far. We presented data of 66 patients monitored during several years, giving us the unique opportunity to evaluate long-term prognosis according to baseline levels of ET-1 and big ET-1 (mean follow-up after catheterisation of 5.3±3.7 years). Thirty-one of the 66 (47%) patients with PH reached the predefined end point. Patients with associated LD showed the worst outcome. Remarkably, there was no significant relation between levels of ET-1 and patient outcome, whereas big ET-1 predicted outcome only in the subgroup of patients with PAH-CHD, but not in the whole PH group. Furthermore, in this context, it is of interest that patients with PAH-CHD showed significant higher levels of big ET-1 than patients with IPAH. This might be related to a different underlying pathophysiological development of PAH or a different PVD in patients with CHD compared with those with IPAH. Overall, however, our data suggest that the predictive value of ET-1 and big ET-1 for patient outcome in children and adolescents with PH is low.
Clinical relevance of the presented results
Beyond the information that ET-1 and big ET-1 have not shown to be reliable biomarkers in children with PH to predict clinical worsening or prognosis, it is also of clinical relevance that although there are some patients with PH who present with elevated levels of ET-1 and big ET-1, there are also some who do not show this increased expression of ET-1 and big ET-1. We have demonstrated that this does not seem to be related to disease severity in children with PH, however, it can be speculated that it might have impact on decision making of specific targeted therapy. Since endothelin-receptor antagonists obviously will show better effects in patients with markedly elevated levels of ET-1 and big ET-1, it might be reasonable to assess ET-1 and big ET-1 levels before initiating PAH-targeted therapy. The presented study was not designed to evaluate the interaction between ET-1/big ET-1 levels and effects of targeted therapy; clearly, further research of this phenomenon is required. Therefore, future studies should analyse whether PAH therapies alter ET-1 and big ET-1 levels and if this ET-1 and big ET-1 level reduction would correlate with better outcome.
Study limitations
ET-1 and big ET-1 are subject to rapid turnover and are rapidly cleared from the circulation and therefore requires a very systematic blood sampling and analysis protocol. Furthermore, its paracrine functions and the cross-reactivity of ET-1 measurement might be some limitations. Future studies will be required to confirm whether precursor forms of ET-1, that is, COOH-terminal pro-ET-1, which is derived from ET-1 propeptide and is released in equimolar amounts to ET-1 might be more stable and may serve as a more reliable in vivo biomarker than the mature form of ET-1.13
Since cardiac catheterisation is an invasive procedure, the assessment of ET-1 levels from the PA and the femoral artery is not applicable in a healthy population of children. Therefore, we used patients with IPF as a control group; however, these patients clearly have abnormal pulmonary haemodynamics as well. There are only few published data on levels of ET-1 in healthy children in the literature, however in these studies blood samples have been taken exclusively from peripheral vessels. In one study with 10 healthy children aged 4–12 years, the ET-1 levels ranged from 0.05 to 0.5 fmol/ml.24 Another small study with 13 healthy children aged 1.2–8 years found mean ET-1 levels of 0.34±0.07 fmol/ml and big ET-1 1.72±0.20 fmol/mL.25 Notably, the site of blood sampling is of great importance in the assessment of ET-1 levels. Unfortunately, previous studies have rarely differentiated between peripheral venous ET-1 levels and mixed venous ET-1 levels. However, the ET-l levels in blood from the antecubital vein may not be representative of mixed venous levels because production or clearance across the forearm bed may not parallel that of other systemic beds.19 This has to be recognised when comparing our study with previous results. In our study, we used exclusively blood sampling from the pulmonary and the SA to avoid this limitation.
The use of endothelin-receptor antagonists obviously may alter endothelin levels. Therefore, we preferably included patients before onset of PAH-specific therapy. Of the 66 patients included, only 7 (10%) were already treated with endothelin-receptor antagonists at the time of sample collection, which should have a negligible effect on our results.
Conclusions
In contrast to the results of previous studies, our data do not confirm that the activation of the ET-1 system in paediatric patients with PH is correlated with the severity of this disease. In addition, we did not find a significant relation to patient outcome suggesting no important role for ET-1 or big ET-1 as a biomarker for predicting the progression of the underlying PVD.
Key messages
What is already known on this subject?
In adults, pulmonary arterial hypertension (PAH) is known to be associated with increased expression of endothelin (ET)-1 and its precursor big ET-1. Furthermore, previous work in adult patients with PAH has demonstrated its close relation to disease severity, and its ability to serve as a potential marker to identify patients at risk and to predict outcome.
What might this study add?
Data about the predictive value of ET-1 and big ET-1 levels in children with pulmonary hypertension (PH) are lacking so far. We present data of 66 patients monitored during several years, giving us the unique opportunity to evaluate long-term prognosis according to baseline levels of ET-1 and big ET-1. Patients with associated lung disease showed the worst outcome. Remarkably, there was no significant relation between levels of ET-1/big ET-1 and patient outcome.
How might this impact on clinical practice?
The presented results demonstrate that ET-1 and big ET-1 have no important role as biomarker for predicting the severity of PH in children, therefore in the management of these patients.
References
Footnotes
Contributors HL, JK, DS and CA have contributed to the conception and design of the study. HL, GK, UB, KS, MK and JK have contributed to the acquisition of data, analysis and interpretation of the data. SM, JK, DS and CA have contributed to the drafting of the article and have revised it critically for important intellectual content.
Competing interests None declared.
Patient consent Obtained.
Ethics approval Local ethics committee, Justus-Liebig University Giessen, Germany.
Provenance and peer review Not commissioned; externally peer reviewed.