The following data were collected in all patients at diagnosis: demographic characteristics (age and sex), clinical features (New York Heart Association [NYHA] functional class, 6-minute walk test, syncope), hemodynamic variables (right atrial pressure, mean pulmonary artery pressure, cardiac index and arterial pulmonary vascular resistance, acute vasodilator test response), and pulmonary function parameters (forced expiratory flow in 1 s, forced vital capacity, and carbon monoxide diffusing capacity). Clinical and radiological screening for small patella syndrome was performed in carriers of the TBX4 mutation.

Sanger sequencing was performed of all exon and exon-intron boundaries of BMPR2, KCNK3, and TBX4. Screening for structural variants and chromosome rearrangements was performed using MLPA (multiplex ligation-dependent probe amplification) (Appendix 1 of the supplementary material).

The variants identified were classified according to their frequency in the general population. Variants absent from public databases (Exome Variant Server, 1000G, and ExAc1S-3S of the supplementary material) or with very low allele frequencies (< 0.01%) were considered mutations. The pathogenicity of variants was assessed by taking into account their effect on the protein level, their presence in previous reports in the literature, the evidence of cosegregation with the disease, and the in silico predicted impact on the protein using widely used software (Polyphen2, MutationTaster4S,5S of the supplementary material, SIFT [Sorting Intolerant from Tolerant], MutPred, SNPs&GO). The variants identified were classified as pathogenic, possibly pathogenic, variants of unknown significance, and nonpathogenic variants (Appendix 2 of the supplementary material). Once the proband study was completed, a genetic study of the first-degree relatives of the mutation carriers was performed and the genealogical tree was constructed.

The statistical software Stata (version 13, Stata Corp.; College Station, Texas, United States) and IBM SPSS 22 (SPSS, Inc.; Chicago, Illinois, United States) were used for the statistical analysis and the data are presented as mean ± standard deviation with a level of significance of P ≤ .05. Kaplan-Meier univariate survival analysis used the time of diagnosis as the start of follow-up and the 15th June 2015 as the end of follow-up, with a maximum follow-up of 15 years (generalized type I censoring). The log rank test was used to compare the survival rates of the IPAH and HPAH patients and the forms associated with the different genes; outcome was defined as death or placement on the lung transplant waiting list. The hazard ratio and 95% confidence interval were calculated for classical prognostic factors (right atrial pressure, cardiac index, 6-minute walk test, NYHA functional class at diagnosis, sex, mean pulmonary artery pressure, and pulmonary vascular resistance), as well as hereditary disease. Also evaluated was the interaction of the classical prognostic factors with hereditary disease. Those variables showing a significance level of P < .10 in the univariate analysis were included in the multivariate Cox regression model, using forward and backward stepwise selection to confirm the results.

The study was performed in accordance with the principles of the Declaration of Helsinki and Spanish law on data protection and was approved by the ethics committees of the participating centers. All patients provided informed consent.

A positive genetic test was observed in 31 probands (18.78%). Specifically, in 16 patients with IPAH (11.10%) and in 15 patients with HPAH (68.18%). Nineteen BMPR2 variants (Table 1S of the supplementary material) were identified in 24 probands, as well as 3 TBX4 variants in 4 probands and 2 KCNK3 variants in 3 probands (Table 2S of the supplementary material). The baseline clinical characteristics according to the affected gene are shown in Table 2.

Prognostic Impact of the Genetic Alterations (Kaplan-Meier Survival Analysis)

The mean follow-up was 7.74 ± 4.46 years. Survival analysis failed to show significant differences between IPAH and HPAH (Figure A) and only showed differences among the HPAH forms associated with TBX4, KCNK3, and BMPR2 (Figure B).

Figure.

Kaplan-Meier survival curves. A: IPAH vs HPAH. B: HPAH with a BMPR2, KCNK3, or TBX4 mutation. HPAH, hereditary pulmonary arterial hypertension; IPAH, idiopathic pulmonary arterial hypertension.

(0.11MB).

The results of the univariate and multivariate analyses are shown in Table 4 and Table 5 (log likelihood function = –103.29; log-rank chi-square = 13.03; P = .005).

The genetic study was refused by 2 of 19 families and accepted by 72 of the 90 family members (80%) belonging to the remaining 17 families. The study identified 21 healthy carriers (29.1%) and 6 affected carriers (8.3%). The characteristics of the carriers are shown in Tables 4S and 5S of the supplementary material. The estimated penetrance was 23.5% for BMPR2 mutations, 25.0% for KCNK3 mutations, and 33.0% for TBX4 mutations.

In our series, the clinical profile of hereditary forms as a whole was similar to that associated with BMPR2.4,7,9,12,19 Notably, in 31.81% of patients with HPAH, the genetic study failed to identify the causative mutation; this percentage is higher than in previous reports.2,3

Of the 19 mutations identified in BMPR2, 42% were missense mutations, a higher rate than previously reported.4 Of these, 11 were novel, showing the complexity of this gene, which continues to show new alterations5,20 despite having been widely studied. Although these changes were distributed throughout the gene and affected various protein domains, 8 of them (42%) were located in exon 2, which was a much higher frequency than previously described.4 This finding could indicate the presence of a mutation “hot spot” in this exon, at least in the Spanish population.

The identification of the c.653_676dupATAAAGGCTCCTTGGATGAGCG mutation, a duplication of 22 base pairs, is notable because this type of change has not previously been linked to PAH development. This mutation was characterized by a higher penetrance (50.0% vs 23.5%), a novel finding with possible future implications.

The previous findings on the phenotype associated with BMPR2 were confirmed in our series (Table 2),4–9,20 except that there was a lower proportion of women with the hereditary form of the disease, another novel finding. The “respiratory advantage” of the forms associated with BMPR2 described by Girerd23,24 was confirmed, absent from the idiopathic forms and the other hereditary forms. This finding could reflect a pathophysiological process distinct at the pulmonary vascular level related to the BMPR2/Smad pathway. There were no significant differences in the phenotype according to the type of BMPR2 mutation, in contrast to the findings of other studies.8

So far, 9 cases of PAH have been linked to KCNK3: 6 in the series of Ma et al.13 and 3 in our series, with 2 novel pathogenic variants. The identification of further mutations in KCNK3 is expected, given that the gene has only relatively recently been linked to PAH. Our group has described for the first time a homozygous mutation in KCNK3 (p.Gly106Arg) in a patient with consanguineous parents and an aggressive form of the disease; the mother is a heterozygous carrier of this mutation. Although further confirmatory studies are required, the presence of both mutated alleles in the index case could explain the aggressive symptoms.

Similar to the work of Ma et al.,13 our study suggests the presence of a more aggressive phenotype of the forms associated with KCNK3, even if they share certain characteristics with the forms associated with BMPR2, such as autosomal dominant inheritance, incomplete penetrance, and predominantly adult onset. Various experimental approaches are being attempted to reestablish the functioning of this channel,25,26 possibly opening a new line of treatment for these patients.

Besides our current series with 4 patients with pathogenic mutations in TBX4, the only other known patients with TBX4 mutations are the 7 patients reported by Kerstjens-Frederikse et al.14 Notably, all of our patients with childhood-onset PAH showed changes in exon 8, whereas the patients with adult-onset PAH had mutations in exon 3. This phenomenon was not seen in the series of Kerstjens-Frederikse et al.,14 whose patients with childhood-onset PAH showed variations in either of the 2 exons.

The phenotype of this subgroup of patients has been poorly described, although it has been suggested that this form shows predominantly childhood onset and lower aggressiveness and is often associated with small patella syndrome. However, none of our patients met the diagnostic criteria for this skeletal dysplasia. Our results support these findings, with no death or transplant event after a mean follow-up of 21.75 years, despite the baseline hemodynamic severity. Thus, PAH associated with mutations in TBX4 could represent the most benign form of this disease, although further studies are required to confirm this hypothesis.

In the population with IPAH, univariate analysis showed a prognostic impact of the classical prognostic factors derived from population registries, such as male sex, 6-minute walking distance, and arterial pulmonary vascular resistance. These results are similar to those of previous reports,27,28 even if the impact of hereditary disease and the baseline NYHA functional class were not significant, in contrast to previous findings (Table 4)12,27,28,9S,14S of the supplementary material. Only the 6-minute walking distance and male sex were still significant in the multivariate analysis. In HPAH patients, male sex and the classical prognostic parameters, related to advanced stages of the disease (right atrial pressure, elevated vascular resistance, and advanced functional class), also had a negative impact in the univariate analysis, although only the functional class was still significant in the multivariate analysis of this subgroup. Thus, NYHA functional class III or IV in the HPAH subgroup was related to a 3.5 times higher risk of placement on the lung transplant list or death (Table 5), a novel finding indicating the importance of an early diagnosis.

We obtained a significant percentage of negative results in the genetic and echocardiographic screening of the family members, as published by other groups.28 However, due to the prognostic impact of early diagnosis, particularly in the hereditary forms, appropriate screening of family members is vital to detect asymptomatic carriers at an early stage. Thus, of particular interest is the development of follow-up programs of healthy carriers, a field being intensively investigated.28–30

Regarding the penetrance of the alterations described, incomplete penetrance with considerable interfamily variability was observed in the 3 genes studied in our population. This finding highlights the marked differences according to the type of mutation in the forms associated with BMPR2, even if the global penetrance of BMPR2 in our series is almost identical to that of previous findings.6,19 Our group is the first to describe the penetrance associated with mutations in TBX4 and further studies are required to confirm our results.

Although the present study has analyzed the largest Spanish cohort of patients with PAH, our findings should be interpreted with caution, largely because both prevalent and incident cases were included, which might have caused a significant dispersion in the date of the diagnosis with a possible underlying survival bias. In addition, just 3 of the genes related to the development of the disease were studied, including BMPR2, the main gene linked to the disease, and the many other genes linked to the disease should be included in future studies.13,15,16 The categorization of the pathogenicity of the mutations found was an estimate based on diverse clinical and genetic/molecular parameters. Thus, for their confirmation, further work is required that demonstrates cosegregation of the mutation with the phenotype and/or involves functional studies.