Plasma proteome analysis in patients with pulmonary arterial hypertension: an observational cohort study

Abstract

Background: Idiopathic and heritable pulmonary arterial hypertension form a rare but molecularly heterogeneous disease group. We aimed to measure and validate differences in plasma concentrations of proteins that are associated with survival in patients with idiopathic or heritable pulmonary arterial hypertension to improve risk stratification.

Methods: In this observational cohort study, we enrolled patients with idiopathic or heritable pulmonary arterial hypertension from London (UK; cohorts 1 and 2), Giessen (Germany; cohort 3), and Paris (France; cohort 4). Blood samples were collected at routine clinical appointment visits, clinical data were collected within 30 days of blood sampling, and biochemical data were collected within 7 days of blood sampling. We used an aptamer-based assay of 1129 plasma proteins, and patient clinical details were concealed to the technicians. We identified a panel of prognostic proteins, confirmed with alternative targeted assays, which we evaluated against the established prognostic risk equation for pulmonary arterial hypertension derived from the REVEAL registry. All-cause mortality was the primary endpoint.

Findings: 20 proteins differentiated survivors and non-survivors in 143 consecutive patients with idiopathic or heritable pulmonary arterial hypertension with 2 years' follow-up (cohort 1) and in a further 75 patients with 2·5 years' follow-up (cohort 2). Nine proteins were both prognostic independent of plasma NT-proBNP concentrations and confirmed by targeted assays. The functions of these proteins relate to myocardial stress, inflammation, pulmonary vascular cellular dysfunction and structural dysregulation, iron status, and coagulation. A cutoff-based score using the panel of nine proteins provided prognostic information independent of the REVEAL equation, improving the C statistic from area under the curve 0·83 (for REVEAL risk score, 95% CI 0·77-0·89; p<0·0001) to 0·91 (for panel and REVEAL 0·87-0·96; p<0·0001) and improving reclassification indices without detriment to calibration. Poor survival was preceded by an adverse change in panel score in paired samples from 43 incident patients with pulmonary arterial hypertension in cohort 3 (p=0·0133). The protein panel was validated in 93 patients with idiopathic or heritable pulmonary arterial hypertension in cohort 4, with 4·4 years' follow-up and improved risk estimates, providing complementary information to the clinical risk equation.

Interpretation: A combination of nine circulating proteins identifies patients with pulmonary arterial hypertension with a high risk of mortality, independent of existing clinical assessments, and might have a use in clinical management and the evaluation of new therapies.

Funding: National Institute for Health Research, Wellcome Trust, British Heart Foundation, Assistance Publique-Hôpitaux de Paris, Inserm, Université Paris-Sud, and Agence Nationale de la Recherche.

Figure 1

Study design

Figure 2

Prognostic protein panel analysis (A) Volcano plot illustrating differences in protein expression between survivors and non-survivors. (B) ROC analysis of 20 selected proteins showing sensitivity and 1 – specificity at cutoffs. (C) Kaplan–Meier survival analysis of patients with idiopathic pulmonary arterial hypertension in cohort 2 divided by TIMP-2 cutoff derived from ROC analysis of cohort 1. (D) Hazard ratios and 95% CI from Cox regression analysis comparing 20 prognostic proteins with established prognostic marker, NT-proBNP. (E) Commercially available ELISA or Luminex assays targeting the 14 independently prognostic proteins used to validate SomaScan measurements in a subset of 80 plasma samples selected from cohort 1 (n=55) and healthy controls (n=25), with samples with high and low concentrations of the analytes chosen. Nine proteins were validated and further studied in cohort 3 (serial samples) and cohort 4 (validation cohort). This scatter-plot illustrates TIMP-1 measurements by SomaScan and Luminex assays in idiopathic pulmonary arterial hypertension cohort 4. Cutoffs for SomaScan and Luminex values derived by percentile equalling ROC-derived cutoff in cohort 1 are indicated by dashed lines. Statistics indicate Spearman's rank correlation. ROC=receiver operating characteristic. RFU=relative fluorescence unit.

Figure 3

Survival analysis of panel score and established prognostic factors Kaplan–Meier survival estimates in patients with different panel scores, in all patients with idiopathic pulmonary arterial hypertension from (A) cohorts 1 and 2 and (B) cohort 4. ROC analysis of Cox models before and after addition of the prognostic protein panel to the established equation, in all patients with idiopathic pulmonary arterial hypertension from (C) cohorts 1 and 2 and (D) cohort 4. ROC=receiver operating characteristic. AUC=area under the curve.

Figure 4

Prognostic value of changes in the protein panel score from diagnosis to after initiation of therapy (A) Cox proportional hazard estimates associated with changes in individual proteins and the overall panel score, showing only the combination of proteins into the score is significantly associated with outcomes. (B) Kaplan–Meier survival estimates in patients with serial panel score measurements (cohort 3), showing an increase in the panel score from diagnosis to after initiation of therapy is associated with poor outcomes.