Pulmonary arterial hypertension (PAH) is a rare, incurable and often fatal disorder. Current management includes conventional therapy, prostacyclin therapy for more advanced patients, and newer agents such as bosentan, an endothelin antagonist. To date, no systematic review has assessed these interventions. OBJECTIVE: To perform a systematic review of the medical interventions used in the management of PAH. METHODS: A literature search of EMBASE and MEDLINE was performed in April 2002 and in November 2002. Data for three outcome measures; survival, exercise capacity, and right atrial pressure were extracted. Where possible, data were segregated according to intervention, the type of PAH and the NYHA/WHO functional class of patients evaluated in the studies. RESULTS: The literature search yielded 1503 abstracts. 54 papers were included after first pass assessment. Following second pass assessment 28 papers were included, and were extracted. An additional four studies were identified after running the second literature search. Three studies met second pass inclusion criteria and were included in the analysis. Three-year survival rates were 47% to 76% across all interventions. Mean changes in right atrial pressure from baseline ranged from a reduction of 6 mmHg with epoprostenol to an increase of 2 mmHg with aerosolised iloprost. Mean changes in exercise capacity from baseline varied from an increase of 10 m with trepostinil in combination with conventional therapy to an increase of 142 m with epoprostenol. The endothelin-receptor antagonist, bosentan, was found to have significant beneficial effects on exercise capacity, right atrial pressure superior to iloprost and in certain instances comparable to epoprostenol. CONCLUSION: Quantitative comparisons of the interventions are difficult within the context of this review, but the homogeneity of the data may make meta-analysis of the evidence possible in certain circumstances. To make the most valuable comparison, long-term data is needed, in particular assessing mortality.