P28 A MULTIDIMENSIONAL COMPOSITE SCORE USING NON-INVASIVE BASELINE VARIABLES TO PREDICT MORTALITY IN PATIENTS WITH PULMONARY ARTERIAL HYPERTENSION

W. N. Lee, S. Kerridge, A. J. Peacock, M. K. Johnson. Scottish Pulmonary Vascular Unit, Glasgow, UK

Introduction In order to facilitate decision making regarding treatment escalation and referral for lung transplantation in pulmonary arterial hypertension (PAH), an accurate way to predict prognosis is desirable.

Aim The aim of the study was to develop a composite score incorporating a number of prognostic non-invasive outcome variables measured at diagnosis to predict mortality in patients with PAH.

Methods Data from 331 patients diagnosed with inoperable chronic thromboembolic disease and PAH (excluding congenital heart disease) in the Scottish Pulmonary Vascular Unit from October 1996 to May 2009 were retrospectively reviewed. Univariate and multivariate logistic regression analysis was used to determine the association between age, gender, pulmonary function, pulmonary haemodynamics at right heart catheterisation, six-minute walk distance (6MWD), WHO functional class (FC), NTproBNP and CAMPHOR scores (on symptoms, activities and quality of life (QOL)) and mortality at 6 months, 1 year and 2 years. A composite score was derived including baseline values of 6MWD, WHO FC, NTproBNP and CAMPHOR QOL score. The weighting of each variable was adjusted using the coefficients of the logistic regression model. The predictive value of the composite score was assessed using ROC analysis.

Results The composite score is outlined in table 1. Patients with a composite score of ⩾7 had a higher mortality with a median survival of 468 days versus 1477 days in patients with a composite score of <7 (log rank, p<0.005). The area under the curve of the composite score was 0.88 (p<0.005) for 6-month mortality, 0.75 (p<0.05) for 1-year mortality and 0.73 (p<0.01) for 2-year mortality.

Conclusions Using non-invasive baseline variables, we have derived a multidimensional composite score which appeared to predict mortality in our cohort of patients. Its …