Feasibility of microRNA signatures for detection early stage pulmonary hypertension

Introduction: Increasing evidence suggests that borderline and exercise pulmonary hypertension (PH), bPH and ePH respectively, are associated with increased mortality. bPH and ePH may represent early stages of PH. Micro-RNAs (miRNAs) have been shown to play a key role in the pathophysiology of PH. We sought to identify miRNA signatures for early detection of PH, bPH and ePH. Methods: 245 plasma samples were collected at rest from PH (mPAP >25 mmHg, n=75), bPH (mPAP 21-24 mmHg, n=27), ePH (mPAP 30mmHg with exercise, n=39) and non-PH dyspnea patients (DC, n=104)) and assayed for N-terminal pro b-type natriuretic peptide (NTproBNP) and 600 miRNAs. Samples were randomly partitioned into training and validation sets. MiRNA signatures were generated with or without NTproBNP from the training set for distinguishing DC from PH, bPH, “bPH+ePH”, and “PH+bPH+ePH” and tested on a naive (split-sample) validation set. Results: The average Area Under the Curve (AUC) for miRNA vs. NTproBNP was 0.71 vs. 0.59, 0.60 vs. 0.56, 0.75 vs. 0.70, and 0.74 vs. 0.81 for distinguishing “DC” from “bPH+ePH”, “bPH”, “PH+bPH+ePH”, and “PH”, respectively. Conclusion: miRNA signatures can discriminate early stage disease (bPH and ePH) from DC better than NTproBNP, suggesting that miRNA signatures are capable of identifying the pulmonary vascular disease phenotype before disease progression to PH with significant RV dysfunction.