Little is known about the effects of “second-generation drugs” (prostanoids, endothelin receptor antagonists, 5-phosphodiesterase inhibitors) in children with pulmonary arterial hypertension (PAH). This study describes the outcome of a national cohort of children with PAH in an era when these drugs became available. From 1993 to 2008, 52 consecutive children with idiopathic PAH (n = 29) or systemic-to-pulmonary shunt-associated PAH (n = 23) underwent baseline and follow-up assessments. Treatment was initiated depending on functional class, acute pulmonary vasoreactivity response, and drug availability. Observed survival was evaluated depending on time of diagnosis in relation to second-generation drug availability and subsequently compared to calculated predicted survival. Children for whom second-generation drugs were available had improved survival compared to their predicted survival (1-, 3-, and 5-year survival rates 93%, 83%, and 66% vs 79%, 61%, and 50%, respectively). However, this improved survival was observed only in patients for whom second-generation drugs became available during their disease course. No improved survival was observed in patients for whom drugs were available already at diagnosis. Baseline variables associated with decreased survival included higher functional class, higher pulmonary-to-systemic arterial pressure ratio, lower cardiac index, and higher serum levels of N-terminal pro–brain natriuretic peptide and uric acid. After start of second-generation drugs, functional class, 6-minute walking distance, and N-terminal pro–brain natriuretic peptide improved but gradually decreased after longer follow-up. In conclusion, survival of pediatric PAH seemed improved since the introduction of second-generation drugs only in selected patients for whom these drugs became available during their disease course. Start of second-generation drugs initially induced clinical improvements, but these effects decreased after longer follow-up.
Pulmonary arterial hypertension (PAH) is a progressive, detrimental pulmonary vascular disease, characterized by advanced remodeling of the small pulmonary arteries leading to increased pulmonary vascular resistance and right ventricular failure. The disease can occur at all ages as an idiopathic entity (iPAH) or associated with various underlying conditions, such as congenital heart defects with systemic-to-pulmonary shunt (PAH-CHD). Without treatment, estimated median survival after diagnosis of patients with iPAH is reported to be dismal, i.e., 0.8 year in children and 2.8 years in adults. In the treatment of PAH, calcium channel blockers (“first-generation drugs”) are prescribed to responders to acute pulmonary vasodilator testing. In recent years, 3 classes of “second-generation drugs” have been introduced, namely prostanoids (epoprostenol, treprostinil, iloprost, beraprost), endothelin receptor antagonists (bosentan, sitaxsentan, ambrisentan), and 5-phosphodiesterase inhibitors (sildenafil, tadalafil). These drugs are presumed not only to have pulmonary vasodilator activity but also to improve remodeling of the pulmonary arteries. With these new drugs, functional and hemodynamic improvements and increased survival have been demonstrated in adults with PAH. Although no randomized clinical trials have been published in children, uncontrolled studies have suggested beneficial effects in pediatric PAH. The objective of this study was to report the outcome of a national, consecutive cohort of children with PAH treated in a period when second-generation drugs became available. Observed survival of these children is assessed in relation to availability of these drugs. In addition, potential baseline predictors for survival and the effect of second-generation drugs on follow-up assessments are evaluated.
Methods
From 1993 to 2008, 52 consecutive children with PAH (iPAH, n = 29; PAH-CHD, n = 23) were seen within the Dutch network for diagnosis and treatment of pediatric PAH. Patients with persistent pulmonary hypertension of the newborn were excluded. This network includes all 8 Dutch pediatric cardiology centers, of which 1 serves as an expert center to which patients with PAH are referred for diagnostic workup, initiation of therapy, and serial follow-up. Patient data are entered in a database registry with informed consent from parents/caregivers and with institutional review board approval.
At presentation at the referral center, all but 3 children underwent cardiac catheterization. Diagnosis of PAH was confirmed by measurement of a mean pulmonary arterial pressure ≥25 mm Hg, pulmonary capillary wedge pressure ≤15 mm Hg, and pulmonary vascular resistance >3 Woods units/m 2 . Hemodynamic evaluation including acute pulmonary vasodilator testing with oxygen and/or nitric oxide was used to exclude operability in patients with PAH-CHD. Acute responders were identified according to criteria as defined by Barst et al. In 3 patients, cardiac catheterization was not performed due to clinical instability. In these patients, diagnosis of PAH-CHD and Eisenmenger syndrome physiology was established echocardiographically by measurement of a right-to-left shunt through the cardiac defect. Obstructive anatomic lesions in the right ventricular outflow tract and proximal pulmonary arteries and signs of pulmonary venous congestion were excluded.
Diagnosis of PAH-CHD and iPAH was determined after extensive diagnostic workup and evaluation of associated conditions, as described in detail previously. In short, patients with PAH-CHD had congenital heart defects consisting of large unrestrictive post-tricuspid shunts (e.g., ventricular septal defects). Five of these patients had undergone previous shunt closure before developing persistent PAH for ≥1 year.
Baseline and follow-up assessments further included World Health Organization functional class (WHO class), transcutaneous oxygen saturation at rest, body mass index, blood pressure, 6-minute walking distance (in children ≥5 years old), hematocrit, N-terminal pro–brain natriuretic peptide (NT–pro-BNP), uric acid, and creatinine. Follow-up assessments were obtained at least every 6 months. In addition, survival was analyzed.
Because walking distance in childhood is correlated with age, we corrected 6-minute walking distance for age by calculating SD scores for 6-minute walking distance using reference values for healthy Caucasian children 4 to 11 years old. Reference values for children 12 to 18 years old were derived by extrapolating 6-minute walking distance values from those obtained in healthy Caucasian adults. In addition, SD scores for body mass index were calculated using reference values for healthy Dutch children.
Supportive drugs, including anticoagulants and diuretics, and calcium channel blockers were available throughout the study period. Second-generation drugs, although not yet registered for pediatric use, were available for children seen at the expert center from 2000: epoprostenol from 2000, bosentan and sildenafil from 2002. Because evidence-based treatment guidelines for children with PAH are lacking, treatment during the period of second-generation drug availability was given according to an adapted version of the established treatment algorithm for adults with PAH. This adapted treatment algorithm stratifies patients by diagnosis, responder status to acute pulmonary vasodilator testing, and WHO class. In iPAH, responders to acute pulmonary vasodilator testing who were in WHO class II or III were started on calcium channel blockers. In iPAH and PAH-CHD, nonresponders in WHO class III were started on bosentan. In case of clinical worsening, these patients received additional sildenafil. If clinical worsening persisted despite additional sildenafil, additional intravenous epoprostenol was initiated. Responders and nonresponders in WHO class IV were started on epoprostenol. However, in patients with Eisenmenger syndrome, intravenous epoprostenol administration was used only in highly selected patients. Evaluation of clinical worsening was based on the physician’s judgment during follow-up.
Drugs were dosed according to body weight. Calcium channel blockers (diltiazem or nifedipine) were administered in a dose of 1 to 8 mg/kg/day. Bosentan target dose was 31.25 mg (10- to 20-kg weight), 62.5 mg (20 to 40 kg) or 125 mg (weight >40 kg) 2 times/day. Before reaching this dose of bosentan, patients received 1/2 the target dose 1 time/day for 4 weeks. Sildenafil was prescribed at a target dose of 1 to 4 mg/kg/day. Epoprostenol was delivered continuously by an implantable intravenous delivery system. The initial target dose was 17 to 20 ng/kg/min and subsequently uptitrated until a satisfactory clinical response was obtained (range 20–55 ng/kg/min). Supportive drugs included diuretics and anticoagulants (coumarin derivatives or acetylsalicylic acid). All patients were prescribed anticoagulants, unless contraindications such as hemoptysis were present.
For analysis purposes, we categorized patients into 3 cohorts based on their time of diagnosis in relation to the availability of second-generation drugs: (1) patients who were diagnosed before and died before second-generation drugs were available to the patient (cohort 1, n = 7), (2) patients for whom second-generation drugs were not available at time of diagnosis but became available during the disease course (cohort 2, n = 21), and (3) patients for whom second-generation drugs were available at time of diagnosis (cohort 3, n = 24). Cohorts 1 and 2 included patients diagnosed before the introduction of these drugs (n = 26) and patients in whom these drugs were recommended but not initiated because of late referral to the expert center (n = 2).
Data are presented as mean ± SD or median and range, where appropriate. Log transformation was used to normalize the distribution of variables. To analyze differences in baseline assessments among the 3 cohorts, 1-way analysis of variance with Bonferroni post hoc testing (continuous variables) or Kruskal-Wallis followed by Mann-Whitney post hoc testing with Bonferroni correction was performed (WHO class, age, weight). To analyze differences in baseline assessments between iPAH and PAH-CHD and between cohort 2 and cohort 3, independent-samples t tests (continuous variables) and Mann-Whitney U tests (WHO class, age, weight) were used. Survival for all patients and for the study cohorts is depicted using Kaplan-Meier curves.
Predicted survival of cohorts 2 and 3 was calculated using the equation for probability of survival formulated by the landmark National Institutes of Health registry of patients with iPAH and validated in other cohorts of adults and children. This equation uses mean pulmonary arterial pressure (mPAP), mean right atrial pressure (mRAP), and cardiac index (CI) to calculate predicted survival (P) at different follow-up time points (t): P(t) = (H[t]) exp e exp(0.007325 × mPAP + 0.0526 × mRAP − 0.3275 × CI), where H(t) = (0.88 − 0.14t + 0.01t 2 ).
To identify baseline predictors for survival in all patients, clinical, laboratory, and hemodynamic baseline assessments, diagnosis (iPAH vs PAH-CHD) and cohort (cohorts 2 and 3 vs cohort 1) were analyzed by univariate Cox proportional hazards model. Baseline variables found to be predictive in univariate analysis (p <0.05) were subsequently corrected for age, sex, diagnosis, and cohort using multivariate Cox regression analysis.
To explore time-specific differences in the effect of second-generation drugs on different clinical and laboratory follow-up assessments from start of second-generation drugs, a random intercept and random coefficient mixed model for repeated measures was used. This approach accommodated imbalance in the data due to variable numbers of repeated measures and unequal intervals between measurements. In addition, subject-specific effects such as age and sex were considered in the individual random effects component. All p values were 2-tailed and those ≤0.05 were considered statistically significant. Statistical analyses were performed with SPSS 16.0 (SPSS, Inc., Chicago, Illinois) and STATA 10 (STATA Corp., College Station, Texas) for Windows.
Results
Baseline characteristics of all patients at presentation at referral center are presented in Table 1 . There were significantly more patients with PAH-CHD in cohort 2 than cohort 3. In cohort 2, median time from diagnosis to second-generation drug availability was 4.0 years (range 0.8 to 13.3). This period was significantly longer in patients with PAH-CHD than in those with iPAH (median 5.2 years, range 1.6 to 13.3, and median 2.7 years, range 0.8 to 4.9, respectively, p = 0.03). Cohort 2 had significantly lower respiratory rate, transcutaneous oxygen saturation at rest, hematocrit, and NT–pro-BNP. No statistically significant differences between iPAH and PAH-CHD were found, except for transcutaneous oxygen saturation at rest, which was significantly lower in PAH-CHD (p = 0.02; Table 2 ).
| Variable | All Patients (n = 52) | Cohort 1 (n = 7) | Cohort 2 (n = 21) | Cohort 3 (n = 24) |
|---|---|---|---|---|
| Diagnosis | ||||
| Idiopathic pulmonary arterial hypertension | 29 (56%) | 4 (57%) | 8 (38%) ⁎ | 17 (71%) ⁎ |
| Pulmonary arterial hypertension associated with systemic-to-pulmonary shunts | 23 (44%) | 3 (43%) | 13 (62%) ⁎ | 7 (29%) ⁎ |
| Age at presentation (years) | 6.1 (0.04–17.4) | 1.2 (0.04–14.9) † | 7.2 (1.4–17.4) † | 4.7 (0.1–15.8) |
| Age at first diagnosis (years) | 3.1 (0.01–15.8) | 1.1 (0.01–6.9) † | 3.1 (0.04–15.8) † | 4.4 (0.1–15.4) † |
| Female | 33 (63%) | 4 (57%) | 12 (57%) | 17 (71%) |
| Weight (kg) | 18.0 (2.5–89.0) | 7.7 (3.2–60.9) | 20.0 (8.2–62.0) | 16.4 (2.5–89.0) |
| Body mass index (kg/m 2 ) | 16.7 ± 4.7 | 14.7 ± 1.9 | 17.8 ± 4.9 | 16.4 ± 4.9 |
| Body mass index SD score | −0.8 ± 1.9 | −1.2 ± 1.3 | −0.3 ± 1.6 | −1.2 ± 2.2 |
| Blood pressure (mm Hg) | ||||
| Systolic | 97 ± 18 | 88 ± 17 | 101 ± 12 | 97 ± 21 |
| Diastolic | 60 ± 13 | 53 ± 16 | 62 ± 8 | 60 ± 15 |
| Heart rate (beats/min) | 99 ± 23 | 113 ± 38 | 92 ± 10 | 103 ± 27 |
| Respiratory rate (breaths/min) | 31 ± 10 | 35 ± 12 ‡ | 27 ± 6 ⁎ ‡ | 33 ± 11 ⁎ ‡ |
| Transcutaneous oxygen saturation (%) | 92 ± 7 | 90 ± 6 ‡ | 89 ± 7 ⁎ ‡ | 95 ± 6 ⁎ ‡ |
| World Health Organization functional class | ||||
| I | 1 (2%) | 0 | 0 | 1 (1%) |
| II | 13 (25%) | 1 (14%) | 6 (29%) | 6 (25%) |
| III | 26 (50%) | 4 (57%) | 11 (52%) | 11 (49%) |
| IV | 12 (23%) | 2 (29%) | 4 (19%) | 6 (25%) |
| 6-minute walking distance (m) | 349 ± 100 | 327 | 312 ± 112 | 381 ± 86 |
| 6-minute walking distance SD score | −3.3 ± 2.1 | −4.0 | −4.2 ± 2.1 | −3.2 ± 1.9 |
| Hematocrit | 0.44 ± 0.10 | 0.50 ± 0.14 | 0.48 ± 0.11 ⁎ | 0.39 ± 0.05 ⁎ |
| Uric acid (mmol/L) | 0.30 ± 0.10 | 0.42 | 0.31 ± 0.09 | 0.30 ± 0.11 |
| N-terminal pro–brain natriuretic peptide (log value) | 2.7 ± 0.6 | 3.3 ‡ | 2.3 ± 0.6 ⁎ ‡ | 2.8 ± 0.7 ⁎ ‡ |
| Creatinine (μmol/L) | 51 ± 19 | 30 | 52 ± 16 | 53 ± 20 |
| Responder to acute pulmonary vasodilator testing (%) | 8 (15%) | 0 | 4 (19%) | 4 (17%) |
| Aortic saturation (%) | 92 ± 9 | 96 ± 6 | 91 ± 8 | 91 ± 10 |
| Mixed venous saturation (%) | 62 ± 10 | 58 ± 6 | 65 ± 8 | 60 ± 12 |
| Mean right atrial pressure (mm Hg) | 7 ± 4 | 5 ± 2 | 8 ± 4 | 7 ± 4 |
| Mean pulmonary arterial pressure (mm Hg) | 55 ± 18 | 68 ± 22 | 53 ± 16 | 53 ± 18 |
| Mean pulmonary arterial pressure/mean systemic arterial pressure | 0.9 ± 0.3 | 1.1 ± 0.3 | 0.9 ± 0.3 | 0.9 ± 0.3 |
| Cardiac index (L/min/m 2 ) | 2.8 ± 1.2 | 2.4 ± 0.8 | 3.1 ± 1.5 | 2.6 ± 0.7 |
| Pulmonary/systemic blood flow index | 1.0 ± 0.4 | 1.0 ± 0.3 | 1.0 ± 0.4 | 1.1 ± 0.5 |
| Pulmonary vascular resistance index (Woods units/m 2 ) | 20.5 ± 13.7 | 28.2 ± 16.4 | 18.5 ± 12.9 | 20.2 ± 13.6 |
| Pulmonary vascular resistance index/systemic vascular resistance | 1.0 ± 0.6 | 1.4 ± 0.8 | 0.9 ± 0.7 | 0.9 ± 0.5 |
⁎ Cohort 2 versus 3 for respiratory rate (p = 0.04), transcutaneous oxygen saturation at rest (p = 0.01), hematocrit (p = 0.001), N-terminal pro–brain natriuretic peptide (p = 0.02), and diagnosis (p = 0.03).
† Cohort 1 versus 2 for age at presentation and at diagnosis (p = 0.03); cohort 1 versus 3 for age at diagnosis (p = 0.01).
‡ Difference among 3 cohorts for transcutaneous oxygen saturation at rest (p = 0.01) and hematocrit (p = 0.001).
| Variable | iPAH (n = 29) | PAH-CHD (n = 23) |
|---|---|---|
| Cohort | ||
| 1 | 4 (14%) | 3 (13%) |
| 2 | 8 (28%) | 13 (57%) |
| 3 | 17 (58%) | 7 (30%) |
| Age at presentation (years) | 5.0 (0.04–15.8) | 6.9 (0.05–17.4) |
| Age at first diagnosis (years) | 4.4 (0.01–15.8) | 2.8 (0.04–15.2) |
| Female | 16 (55%) | 17 (74%) |
| Weight (kg) | 17.8 (3.4–89.0) | 19.3 (2.5–60.9) |
| Body mass index (kg/m 2 ) | 17.0 ± 4.7 | 16.4 ± 4.7 |
| Body mass index SD score | −0.7 ± 1.9 | −0.9 ± 1.9 |
| Blood pressure (mm Hg) | ||
| Systolic | 101 ± 18 | 92 ± 16 |
| Diastolic | 63 ± 13 | 57 ± 12 |
| Heart rate (beats/min) | 100 ± 22 | 98 ± 24 |
| Respiratory rate (breaths/min) | 30 ± 9 | 32 ± 10 |
| Transcutaneous oxygen saturation (%) | 94 ± 5 ⁎ | 88 ± 7 ⁎ |
| World Health Organization functional class | ||
| I | 1 (3%) | 0 |
| II | 8 (28%) | 6 (26%) |
| III | 13 (45%) | 12 (52%) |
| IV | 7 (24%) | 5 (22%) |
| 6-minute walking distance (m) | 374 ± 104 | 312 ± 83 |
| 6-minute walking distance SD score | −2.8 ± 2.2 | −4.1 ± 1.7 |
| Hematocrit | 0.39 ± 0.05 | 0.50 ± 0.11 |
| Uric acid (mmol/L) | 0.30 ± 0.10 | 0.31 ± 0.10 |
| N-terminal pro–brain natriuretic peptide (log value) | 2.6 ± 0.7 | 2.6 ± 0.6 |
| Creatinine (μmol/L) | 53 ± 20 | 49 ± 18 |
| Responder to acute pulmonary vasodilator testing | 7 (24%) | 1 (4%) |
| Aortic saturation (%) | 94 ± 8 | 89 ± 8 |
| Mixed venous saturation (%) | 61 ± 11 | 62 ± 9 |
| Mean right atrial pressure (mm Hg) | 7 ± 4 | 7 ± 4 |
| Mean pulmonary arterial pressure (mm Hg) | 55 ± 17 | 54 ± 20 |
| Mean pulmonary arterial pressure/mean systemic arterial pressure | 0.9 ± 0.3 | 1.0 ± 0.1 |
| Cardiac index (L/min/m 2 ) | 2.8 ± 0.8 | 2.8 ± 1.5 |
| Pulmonary/systemic blood flow index | 1.0 ± 0.2 | 1.1 ± 0.6 |
| Pulmonary vascular resistance index (Woods units/m 2 ) | 19.9 ± 12.3 | 20.1 ± 15.1 |
| Pulmonary/systemic vascular resistance index | 1.0 ± 0.6 | 1.0 ± 0.7 |
⁎ Idiopathic pulmonary arterial hypertension versus pulmonary arterial hypertension associated with congenital heart defects with systemic-to-pulmonary shunt for transcutaneous oxygen saturation at rest (p = 0.02).
Calcium channel blockers were started in 8 patients. Second-generation drugs were administered to 37 patients, of whom 5 had previously received calcium channel blockers. In these 5 patients, second-generation drugs were started due to clinical deterioration and disappearance of acute pulmonary vasodilator response after start of calcium channel blocker therapy (median 2.0 years, range 0.6 to 6.3). Patients received monotherapy (n = 22) or combination therapy with 2 (n = 13) or 3 (n = 2) second-generation drugs. Second-generation drugs included bosentan (n = 33), beraprost (n = 1), sildenafil (n = 11), and epoprostenol (n = 9). Twelve patients received no specific anti-PAH drugs due to good clinical condition (WHO class I or II, n = 3), refusal to receive drugs (n = 1), rapid death within a few days due to end-stage disease (n = 1), and death before second-generation drug availability (n = 7).
During a follow-up of 0.02 to 14.3 years (median 3.3), 18 patients died: 10 of 29 (34%) with iPAH and 8 of 23 (35%) with PAH-CHD. Causes of death included progressive right ventricular failure (n = 9), hemoptysis leading to acute circulatory failure (n = 5), acute pulmonary hypertensive crisis during anesthesia for heart catheterization (n = 1) or renewal of intravenous epoprostenol delivery system (n = 1), sepsis due to infection of the intravenous epoprostenol delivery system (n = 1), and sudden death (n = 1).
Survival from time of diagnosis for all patients is depicted in Figure 1 . Survival did not differ significantly between iPAH and PAH-CHD. In contrast, survival rates for cohort 2 were higher than those for cohort 3 (p = 0.05) and cohort 1 (p = 0.002; Figure 1 ). Observed survival from time of second-generation drug availability for cohorts 2 plus 3 was longer compared to their calculated predicted survival ( Figure 2 ). When analyzed separately, survival for cohort 2 was significantly better than predicted survival, whereas survival for cohort 3 was not improved ( Figure 2 ).
