This was an ambispective observational cohort study of patients diagnosed with group 1 PAH (idiopathic, familial, or hereditary PAH, forms associated with connective tissue disease, human immunodeficiency virus [HIV], portal hypertension, rapeseed oil, congenital heart disease, pulmonary veno-occlusive disease, and other less common causes [Osler-Weber-Rendu and hemolytic anemia]) from January 1984 to December 2014, who were treated at the Pulmonary Hypertension Multidisciplinary Unit of the Hospital Univeristario 12 de Octubre. The Hospital Universitario 12 de Octubre Ethics Committee approved the study. Patients seen from the year 2000 onwards were prospectively included in the database; those seen prior to the year 2000 were retrospectively included. Date of diagnosis was defined as the first right heart catheterization performed. We excluded congenital heart diseases with Eisenmenger syndrome and chronic thromboembolic pulmonary hypertension (Figure 1).

Figure 1.

Flow diagram of patients and clinical deterioration. A: distribution of patients. B: distribution of clinical deterioration. CHD, congenital heart disease; CTD, connective tissue disease; CTEPH, chronic thrombo-embolic pulmonary hypertension; ES, Eisenmenger syndrome; HIV, human immunodeficiency virus; HPAH, hereditary pulmonary artery hypertension; IPAH, idiopathic pulmonary artery hypertension; PAH, pulmonary arterial hypertension; PH, pulmonary hypertension; POPH, portopulmonary hypertension; PVOD, pulmonary veno-occlusive disease; TOS, toxic oil syndrome.

(0.26MB).

Patients were divided according to 3 time intervals, to coincide with the guidelines available for each period (European Society of Cardiology/European Respiratory Society guidelines) as follows:

• •

  Diagnosis prior to 1 January 2004.

• •

  From 1 January 2004 to 31 December 2009 (2004 guidelines).4

• •

  From 1 January 2010 to 31 December 2014 (2009 guidelines).5

Diagnosis of PAH was based on a diagnostic algorithm and the hemodynamic criteria recommended in the guidelines from each period.

The following variables were analyzed at diagnosis: demographic data, PAH etiology, functional class (FC), 6-minute walk test (6MWT), right atrial pressure, cardiac output, mean pulmonary artery pressure, pulmonary vascular resistance, pericardial effusion, and initial treatment. The event “clinical deterioration” was defined by the first event to occur: death, inclusion on the LT waiting list, or atrial septostomy.

Descriptive results are presented as frequency and percentage for qualitative variables and mean ± standard deviation or median [interquartile range] for quantitative variables. To calculate the statistical significance (P < .05) of comparisons for qualitative variables between groups, the Fisher exact test was used. For quantitative variables with normal distribution, we used the Student t test, and analysis of variance for independent groups when 2 or more groups were being compared. Nonparametric tests were used when the variable did not follow a normal distribution.

To identify the parameters that were “predictors of clinical deterioration”, bivariate analysis was performed for each variable regarding the time until deterioration; Kaplan-Meier curves and the log-rank test were used to compare the curves. Subsequently, those variables that were statistically significant (P < .05) in the bivariate analyses were selected and included in the multivariate Cox regression model for time until deterioration.

To determine if etiology was a prognostic factor, we grouped etiologies according to the median survival obtained on the Kaplan-Meier curves such that there were no significant differences between the etiologies forming each group. The first group comprised idiopathic PAH/hereditary PAH/rapeseed oil-associated PAH/congenital heart disease-associated PAH and HIV-associated PAH (median survival, 11.9 years, 95% confidence interval [95%CI], 8.4-18.1 years); the second group comprised PAH associated with connective tissue disease/portopulmonary hypertension/other (median survival, 5.9 years; 95%CI, 4.4-8.1 years); and group 3 comprised those with PVOD (median survival, 2.5 years; 95%CI, 1.2-7.1 years). The Kaplan-Meier graphs of this analysis are available in the Figure of the supplementary material.

Taking into account the most recent clinical practice guidelines,6 we established composite criteria as predictors of clinical deterioration, which combined FC and 6MWT distance or right atrial pressure. The criteria were defined as follows: criterion 1, FC III or IV with 6MWT < 475 m; criterion 2, FC III or IV with right atrial pressure > 8 mmHg. Finally, variables with a P-value < .05 in the multivariate model were selected as prognostic factors. The final result shows the relative risk (hazard ratio [HR]) and 95%CI. The C-index was used to calculate the concordance rate between the multivariate model prediction and the observed result for the whole sample.

Median follow-up was 4.5 years [2.2-8.2 years] and 14 patients (4%) were lost to follow-up. Figure 3 contains the variables that showed a significant difference in clinical deterioration in the bivariate analysis and were later included in the multivariate analysis. Survival free from clinical deterioration was higher in women, patients younger than 56 years, patients with right atrial pressure < 8 mmHg, with 6MWT > 475 m, and those in FC I-II (Figure 3). In terms of PAH etiology, group 3 had the worst survival (Figure 3C). The median survival free from transplant or death was 9 years (95%CI, 7.532-11.959 years). Figure 4A shows the data on survival free from clinical deterioration (91.9%, 80.1%, and 67.6% at the first, third, and fifth year, respectively) with a median survival of 8.2 years (95%CI, 7.0-11.6 years). Figure 4B shows the survival free from death or transplant in the first, third, and fifth year, which were 92.2%, 80.6%, and 68.5%. Cause of death was PAH-related in 52.4% (79% of these heart failure and 21% sudden death), in 25.4% it was unrelated to PAH (18% infections and 12% neoplasia), and in 22.2% cause of death was undetermined. Table 3 shows the multivariate analysis of prognostic factors for clinical deterioration. The C-statistic for the multivariate model was 0.8 (95%CI, 0.66-0.91).

Figure 3.

Bivariate analysis of clinical deterioration A: sex. B: age. C: grouped etiologies. D: distance achieved on 6MWT. E: functional class. F: right atrial pressure. 6MWT, 6-minute walk test; PVOD, pulmonary veno-occlusive disease; RAP, right atrial pressure.

(0.49MB).

Figure 4.

Survival. A: survival free from clinical deterioration (SFCD). B: survival free from death or transplant. OS, overall survival.

(0.2MB).

This was the most common etiology. The median age at diagnosis of 44 years remained unchanged over time. In contrast, the COMPERA registry8 (2007-2012) and the United Kingdom registry9 (2001-2009) showed a median age at diagnosis of 71 and 62 years, respectively. This could be explained by fewer referrals of older patients to our hospital in recent years, related to the greater number of PAH units in Spain.

The median survival free from death or transplant was 11 years, longer than the 7 years reported by the REVEAL registry.6 Advanced age, FC III or IV, and the distance achieved in the 6MWT have been reported as established prognostic factors in patients with idiopathic PAH.6 In our study, the patients were younger, had better FC, and achieved a longer distance in the 6MWT, and prostacyclin use was greater; these factors could have contributed to the observed survival differences.

In our study, 50% of patients older than 56 years had connective tissue disease-associated PAH; 92% of these had scleroderma. In the REVEAL registry,10 as in our series, patients with scleroderma had a mean age of 62 years and were the most prevalent group among patients older than 50 years. In patients with scleroderma-associated PAH in our series, survival in the first year was 81%, comparable to that described in the REVEAL registry.10 This one-year mortality close to 20%—which should act as a red flag to the medical community—led to the proposal of new screening11 and treatment protocols; our unit implemented these last year. Early diagnosis is key to improve survival.12

The toxic oil syndrome epidemic related to the ingestion of denatured rapeseed oil occurred in 1981 and affected around 20 000 people. Approximately 20% of those affected developed PAH, with apparent complete spontaneous remission in many of them; a small percentage developed a chronic severe form.1 Our hospital centralized the follow-up of patients with this type of PAH; 74% of the cases were diagnosed prior to 2004. However, new diagnoses continue to be made, some of which have intervals between exposure and PAH onset of more than 30 years. Such latency periods of variable duration are also described in the development of PAH after ingestion of anorexigens: in a series from the French registry, 44% of patients13 were diagnosed more than 5 years after exposure to the drug. Therefore, a history of ingesting rapeseed oil will be present in our setting and must be identified as a risk factor particular to Spain, despite the long latency period.

Patients with Eisenmenger syndrome were excluded because this has a different natural history and treatment. More than half of the rest of the patients had forms associated with systemic-pulmonary shunt repairs done in childhood, and the diagnosis of PAH had been made in the fourth decade of life. The REHAP registry14 describes that 25% of the 240 patients with congenital heart disease-associated PAH presented pulmonary hypertension following repair and, as in our series, survival was comparable to that of idiopathic PAH.

The second most common clinical group was that associated with restrictive defects (mostly interatrial communication), with a survival comparable to idiopathic PAH. In patients who received a LT for simple congenital heart disease-associated PAH with pulmonary-systemic shunt, the defect was closed during LT. Our unit is a national pioneer in the repair of congenital heart disease at the time of transplant, as recommended in the current transplantation guidelines.15

Pulmonary venooclusive disease is characterized by rapid progression and a lack of treatment options. Mutations in the eukaryotic initiation factor-2 kinase A4 (EIF2KA4) gene have been associated with PVOD development, with an autosomal recessive inheritance pattern. Our group recently discovered a founder mutation (c.3344C>T [p.P1115L]) in said EIF2AK4 gene in 5 families of Romani ethnicity with PVOD characterized by disease onset at < 35 years old, consanguinity, low carbon monoxide diffusion capacity, and poor survival.16

In our series, 75% of patients with PVOD arrived at our hospital in a critical status, and 50% of them required support with an extracorporeal membrane oxygenator in the first 3 months after diagnosis, with no deaths while on the transplant waiting list. Other groups have described a high waitlist mortality rate in these patients, up to 22% at 6 months.17 The fast progression from time of diagnosis and lack of specific treatment mean that immediate referral is essential when PVOD is suspected, so that assessment by the transplant team may begin. In hereditary forms, a genetic study and screening of first degree relatives should be performed3 and heterozygous carriers should be identified. Medical practitioners in Spain should be particularly alert to this possibility and the diagnosis should not be delayed if there is clinical suspicion in patients from this ethnic group.

Prostacyclins were used in 205 patients (56%). Seventy percent of those who died and 75% of patients in FC IV were receiving prostacyclin treatment. In the REVEAL registry, prostacyclin use was drastically lower and only 43% of those who died were receiving this treatment.18 The REHAP registry showed that from a national perspective, prostacyclins were underused, with only 50% of patients in FC IV receiving prostacyclins.2 This difference between data from national registries and an expert center can be explained by the need for a complex care organization that supports the use of these drugs and guarantees their treatment safety and excellence. Trained nursing staff is essential, as is a structured health education program and a 24-hour immediate response system (both in-person and telephone).

Inclusion on the LT waiting list depends on multiple factors, essentially the speed of progression of the disease and the scarcity of organs that leads to an ever-increasing waiting list time. Therefore, the latest clinical guidelines3 differentiate between 2 aspects: a) referral for LT assessment, which should be done early before all therapeutic possibilities have been exhausted, and b) active inclusion in the LT program. It is recommended to develop strategies as a bridge to transplant to minimize waiting list mortality. In our unit, after 2011, 15% of LTs for PAH were performed as an emergency and half needed a bridge to transplant strategy, essentially extracorporeal membrane support in 4 patients. In 5 patients, atrial septostomy allowed clinical stabilization of the patient and successful elective LT.19

The nonmodifiable demographic variables with a prognostic effect highlight the need to establish therapeutic strategies from the outset for patients with a high risk profile. In our series, the modifiable prognostic variables related to disease progression, such as FC III-IV, 6MWT < 475 m, and right atrial pressure > 8 mmHg (included in the composite criteria), were comparable to most other registries1,6 and reinforce the recommendation from the clinical guidelines to perform targeted assessments.5

An analysis was also performed with composite variables that distinguished patients in FC III-IV from those with higher risk who may need more aggressive treatment. Kane et al.20 explored this strategy in a cohort of 484 patients; the multifactorial prognostic estimation obtained a model with good predictive capacity, with a C-index of 0.84. Our multivariate model, with the inclusion of composite criteria (FC+6MWT and FC+right atrial pressure) had a C-index of 0.80, which highlights the importance of multifactorial evaluation in the prognostic estimation, as recommended in the current clinical practice guidelines.5

This was an ambispective observational study (12% of the patients were included retrospectively) of the experience of one center with a cohort of patients studied over a long period of time, with marked differences in each period in the etiology of PAH and the therapeutic options available. The analysis of risk factors and survival must be interpreted with caution, as there is a temporal bias. However, we think that a description of the clinical reality observed in a referral center over 3 decades gives a comprehensive view of the disease over time, and can help establish strategies aimed at improving the management of this disease. As this was a 30-year series, we did not use the newer tools for prognostic stratification available from 2010, namely N-terminal-pro-brain-natriuretic peptide, ergospirometric parameters, and the newer imaging techniques.