OBJECTIVES: To compare the performance of pulmonary arterial hypertension (PAH) drug classes based on the risk of mortality reported in clinical trials, and inform a revised PAH treatment algorithm. METHODS: The study used Bayesian Analysis Using Gibbs Sampling in Windows (WinBUGS) and Monte Carlo Simulations to conduct a multiple treatment comparison of placebo and three PAH drug classes; prostanoids (Treprostinil & Iloprost), endothelin-receptor antagonists (Bosentan, Sitaxsentan, & Ambrisentan), and phosphodiesterase (sildenafil).  Direct and indirect pairwise odds ratios (OR) were obtained. Published PAH studies through 2010 were identified from MEDLINE (PubMed) database and an extended manual search was also conducted based on references from identified studies. Inclusion was restricted to randomized controlled trials lasting at least 12 weeks and at most 16 weeks, with subjects having either idiopathic or associated PAH, and studies that reported mortality as an endpoint. Studies that compared combinations of drugs were excluded from the analysis. Results are reported in OR with 95% credible intervals. RESULTS: In total 8 studies (20 treatment arms, 2,015 subjects enrolled) were included in the analysis. 2 studies were 3-arm trials and 1 was a 4-arm trial. With placebo as the reference class, the OR of mortality was 1.00 (0.06, 4.20) for prostanoids; 1.14 (0.15, 4.69) for endothelin-receptor antagonists and 0.62 (0.05, 2.83) for phosphodiesterase. Using prostanoids as the reference class, the OR of mortality was 4.28 (0.10, 23.10) for endothelin-receptor antagonists and 2.40 (0.04, 12.66) for phosphodiesterase. Using endothelin-receptor antagonists as the reference class, the OR of mortality was 1.11 (0.04, 5.55) for phosphodiesterase.  None of the ORs were statistically significant. CONCLUSIONS: Based on the risk of mortality reported in clinical trials, there is no statistically significant difference among PAH drug classes. Including more studies/drugs and the use of different PAH outcomes could inform a more detailed comparison.