Conference: 2013 PH Professional Network Symposium
Release Date: 09.28.2013
Presentation Type: Abstracts
Download Adobe Acrobat
N. Kim, MD1; F. Rahaghi, MD2; J. Elwing, MD3; J. Sager, MD4; K. Chin, MD5; Z. Safdar, MD6; D. Ross, MD7; S. Short, MPH8; D. Villaneuva,MD9; W. Benton, PharmD9; C. Daniels, MD10
1University of California, San Diego, Calif.; 2Cleveland Clinic, Weston Fla.; 3University of Cincinnati, Cincinnati, Ohio ; 4Santa Barbara PulmonaryConsultants, Santa Barbara Calif.; 5Univeristy of Texas Southwestern; 6Baylor College of Medicine, Houston, Texas; 7Ronald Reagan - UCLAMedical Center, Los Angeles, Calif.; 8ICON Late Phase & Outcomes Research, San Francisco, Calif.; 9Actelion Pharmaceuticals US, Inc., South SanFrancisco, Calif.; 10Nationwide Children’s Hospital and Ohio State University, Columbus, Ohio
PURPOSE: To identify predictors of persistence on iloprost therapy in patients with pulmonary arterial hypertension (PAH) using data from RESPIRE (Registry To Prospectively Evaluate Use Of Ventavis® in Patients with Pulmonary Arterial Hypertension [NCT00902603]).
BACKGROUND/SIGNIFICANCE: PAH is a chronic progressive disease characterized by increased pulmonary artery pressure and vascularresistance, leading to right heart failure and death. Treatments include the prostacyclin analogue iloprost (Ventavis®, Actelion, San Francisco, CA) –an inhalation solution indicated for PAH (WHO group 1) NYHA functional class (FC) III-IV.
METHODS: RESPIRE is an ongoing, observational, 32-center, US study, initiated 30 March 2009, following outpatients with PAH receivinginhaled iloprost for ≥3 months at enrollment. Demographic and clinical characteristics at enrollment and treatment persistence data collectedover 24 months’ follow-up were analyzed. Candidate factors identified a priori were entered into a multivariate model as potential predictors of discontinuation. An α level of 0.1 was used in the stepwise selection process to determine which factors entered/remained in the model.Discontinuations were the sum of discontinuations from: (1) study discontinuation forms, (2) PAH medications log, and (3) >28-day gap betweentreatments from post-enrollment records. The latter had almost no overlap with the first two. Persistence was defined as freedom from iloprostdiscontinuation. In the case of highly corr0elated candidate predictors, variables considered to be most clinically meaningful were included. NYHAFC could not be evaluated as a predictor because missing FC was associated with persistence. Nine subjects missing post-enrollment treatmentdata were excluded from the analysis.
FINDINGS: Enrollment data for 139 patients are presented (data lock February 19, 2013) with a mean post-enrollment persistence on iloprostof 1.02 years. One-year persistence was 54.3 ± 4.4% and 61.2% of patients discontinued iloprost during follow-up. The mean age was 62.3years, 80% were women, 16.5% used a nursing service, mean time from diagnosis to enrollment was 4 years, and mean time on iloprost pre-enrollment was 1.18 years. Other candidate predictors included in the analysis were PAH etiology, employment status, living situation, iloprost concentration at enrollment, mean inhalation time in the 3 months pre-enrollment (minutes/complete treatment), combination PAH therapy, and concomitant medications (diuretics, oxygen, warfarin, aspirin, beta blocker, and statin). Pre-treatment hemodynamics were not included in themodel. Proportional hazards regression modeling identified ‘Using a nursing service’ was a significant predictor of iloprost discontinuation (HR1.646 (95% CI 0.977–2.773); P = 0.061 vs. not using). Increased ‘Time from diagnosis to iloprost start’ (HR 0.917; P = 0.021 per year) predictedgreater likelihood of iloprost persistence. ‘Time from diagnosis to enrollment’ was considered for inclusion in the model but was highly correlatedwith ‘Time from diagnosis to iloprost start’ so only the latter was entered as a candidate variable.
IMPLICATIONS: Type of PAH, iloprost concentration, inhalation time, and common concomitant medications, among others, were not associatedwith treatment persistence. Longer illness duration and not using a nursing service at enrollment were independent predictors of longer persistence. Reasons for needing a nursing service (cognitive and dexterity issues) may contribute to increased likelihood of discontinuation.
*This study was sponsored by Actelion Pharmaceuticals US, Inc.