Background

Methods

Results

Conclusions

• chronic obstructive pulmonary disease
• pulmonary hypertension
• lung transplantation

The development of pulmonary hypertension (PH) during the course of chronic obstructive pulmonary disease (COPD) is a well-known phenomenon with a prevalence depending on the severity of airway obstruction. When present, PH, defined as group 3 PH according to the classification of sixth World Symposium on Pulmonary Hypertension,1 is usually of moderate severity; the mean pulmonary artery pressure (mPAP) level at rest ranges from 25 to 35 mm Hg, with preserved cardiac output.²,³ However, a subset of COPD patients present a much higher mPAP.^{2, 3, 4, 5, 6, 7, 8, 9, 10, 11} In the most characteristic cases, the level mPAP is disproportionate to the degree of bronchopulmonary involvement. For these patients who seem to have a particular involvement of pulmonary circulation and who could be potential candidates for vasoactive therapy, 2^, 3  the term “out-of-proportion” PH has been replaced with “severe PH-COPD” defined by mPAP ≥35 mm Hg or mPAP ≥25 mm Hg with low cardiac index (<2l/min/m²).^{2, 3} However, little is known about the characteristics of COPD patients with severe PH. We designed a prospective multicenter study including COPD patients with severe PH followed over several years to provide a more complete description of this entity. Our results focus on the characteristics of patients at inclusion and at 6- and 12-month visits as well as survival.

Materials and methods

In 2000, the French clinical research network led by the French Reference Center for Pulmonary Hypertension (Université Paris-Sud, Hôpital Kremlin-Bicêtre, Le Kremlin-Bicêtre, France) initiated a national prospective registry to collect data on pulmonary arterial hypertension (PAH) and other forms of PH from 17 university hospitals that followed at least five newly diagnosed PAH patients per year. The centers contributing to this national prospective registry complete an inclusion form and follow-up forms (every 6 months) for each new patient included. Demographic, clinical, pulmonary function tests, biological, and hemodynamic data (right heart catheterization [RHC] and echocardiography) are collected in the registry at inclusion and follow-up. The registry, which has already provided important information on PAH,¹² already includes patients with severe PH-COPD, but the information relative to the broncho-pulmonary disease is not well detailed, and the registry is a mix of incident and prevalent cases. To provide more precise information on this subset of PH patients, an additional prospective registry (severe PH-COPD registry), connected to the national PH registry and dealing only with incident patients, was set up in 2012. The primary end-point was survival. We anticipated a 30% 1-year mortality rate. We decided to enroll 100 patients with incident disease to obtain a 20% to 40% 95% confidence interval (CI). This sample size was considered a pragmatic objective given the low prevalence of severe PH-COPD and an inclusion period of 4 years. The planned follow-up was 3 years. To be included in the severe PH-COPD registry by any of the centers of the French clinical research network, the patients had to fulfill the following criteria: (1) give informed consent to be included in the severe PH-COPD registry; (2) age > 18 years; (3) diagnosis of COPD with a pulmonary function test showing forced expiratory volume in 1 sec (FEV₁)/forced vital capacity (FVC) ratio <70% and FEV₁ ≤80% of predicted; (4) severe PH with mPAP ≥35 mm Hg measured on RHC, pulmonary artery occlusion pressure (PAOP) ≤15 mm Hg at rest, well after an exacerbation (more than 6 weeks); (5) diagnosis of PH by RHC <1 year before inclusion (incident case); (6) absence of evolving disease (other than PH and COPD) yielding a life expectancy <6 months; and (7) possibility to ensure a regular medical follow-up. In addition, patients had to be classified as having group 3 PH related to COPD with lack of an alternative overt cause of PH. However, the presence of additional factors that might affect pulmonary circulation (history of pulmonary embolism, previous thoracic surgery, obstructive sleep apnea, intravenous drug use, etc.) was authorized and recorded. We restricted the definition of severe PH-COPD to mPAP ≥35 mm Hg because the study was designed before the 2013 edition of the world symposium on PH, which proposed to include the association of mPAP ≥25 mm Hg and cardiac index <2 L/min /m² in the definition of severe group 3 PH. 2 Before inclusion in the severe PH-COPD registry, thoracic CT scans were checked to exclude cases of combined pulmonary fibrosis and emphysema, given the known association between this syndrome and severe PH. The patients included in the PH-COPD registry were simultaneously included in the national PAH registry, so the information collected in the severe PH-COPD registry was limited to items not already collected in the national registry. In particular, the information collected at inclusion and follow-up in the severe PH-COPD registry is listed in the online data supplement (Table E1, Table E2). After inclusion, patients were followed for 3 years, with a visit every 6 months. The PH-COPD registry was set up in agreement with the Commission Nationale de l’Informatique et des Libertés, dedicated to information technology and civil rights in France (December 2012). The registry was approved by the ethics committee of hospital Bichat, Paris, France (CEERB Paris Nord, November 2012, no. 12-071).

Statistical analysis

Continuous variables are reported with median (interquartile range [IQR]) and categorical variables with number (percentage). Spearman correlation analysis was used to explore the correlation of mPAP and PaO₂ with the number of exacerbations, FEV_1, and lung diffusing capacity for carbon monoxide (DLCO) at inclusion. The overall survival for the whole population and survival stratified by New York Heart Association (NYHA) class was analyzed by the Kaplan-Meier method and NYHA class groups were compared by log-rank test. Patients were censored after 18 months of follow-up or at their transplantation date. As the Kaplan-Meier curve did not drop below 50%, the mean survival time restricted to 18 months was used to describe survival. Univariate and multivariable Cox proportional-hazards models were used to estimate hazard ratios (HRs) with 95% CIs for survival. Univariate Cox models were used to evaluate the association between each variable and survival. A multivariable Cox regression model was then used with variables selected by clinical relevance and missing data rates.

Results

From December 2012 to December 2016, 100 patients from 13 French centers were prospectively included in the study. One patient was excluded from the analysis because he had prevalent disease. Thus, the remaining 99 patients with incident disease form the basis of the study. The flow chart of the study is given in Figure 1.

Characteristics at inclusion

The full characteristics of the 99 patients at inclusion are reported in Table 1. Almost 60% were included in 3 centers and the others in the 10 remaining centers. The patients included in each center are described in Table E3 (supplementary material). The diagnosis of severe PH was obtained at a median of 4 years after the diagnosis of COPD. The median age of patients at inclusion was 66.0 [62.0-72.0] years; 82 (82.8%) were male. The median body mass index (BMI) was 24.1 kg/m² (IQR 21.1-26.6; range 16-38); BMI was >30 kg/m² for 7% of cases. Patients were current smokers (10.2%), former smokers (85.7%), or never smokers (4.1%). Of the four never-smokers, two had been clearly exposed to occupational dust, one was a farmer with presumed professional exposure, and the last one had no identified occupational or domestic exposure but had a history of chronic asthma. The proportion of patients with NYHA class I, II, III and IV was 2%, 20.2%, 55.6%, and 22.2%, respectively. During 12 months before inclusion, 42 (42.9%) had an exacerbation requiring hospitalization (severe exacerbation). The median number of such exacerbations during the previous year was 1.0 [1.0-3.0]. The median COPD assessment Test (CAT), performed in 78 of the 99 patients of the cohort was 20.0 [15.0-23.0], which indicates symptomatic impact. The prevalence of associated comorbidities was 37% (obesity n = 7, diabetes n = 12, coronary artery disease n = 10, dysthyroidism n = 4, chronic kidney failure n = 4, hematologic disease n = 3, atrial fibrillation n = 1, other n = 4; Table 1). Medical history, including factors that may have an impact on the pulmonary circulation, was characterized by obstructive sleep apnea (n = 21), history of thrombo-embolic disease (n = 10), history of thoracic surgery (n = 9) including lobectomy (n = 6), history of intravenous drug use (n = 1), benfluorex intake (n = 1), depression n = 8, and systemic hypertension (n = 41; Table 1).

Table 1Patient Characteristics at Baseline