BACKGROUND: Eisenmenger's syndrome (ES) is the most advanced form of pulmonary arterial hypertension related to congenital heart disease. Evolution of pulmonary vascular disease differs markedly between patients with atrial septal defects (ASD) versus ventricular septal defects (VSD), potentially affecting response to treatment. We compared the effects of bosentan and placebo in patients with isolated ASD (ASD subgroup) versus patients with isolated VSD or both defects (VSD subgroup). METHODS: Post-hoc analysis of a 16-week, multicenter, randomized, double-blind, placebo-controlled trial was performed. Fifty-four patients (13: ASDs, 36: VSDs, 5: VSD+ASD) were randomized to bosentan 62.5 mg bid for four weeks (uptitrated to 125 mg bid thereafter) or placebo. Main outcome measures were: indexed pulmonary vascular resistance (PVRi), exercise capacity, mean pulmonary artery pressure (mPAP), pulmonary blood flow index (Qpi), and changes in oxygen saturation (SpO). RESULTS: Placebo-corrected median (95% CI) treatment effects on PVRi were -544.0 dyn.s.cm (-1593.8, 344.7) and -436.4 dyn.s.cm (-960.0, 167.0) in the ASD and VSD subgroups, respectively. Effects of bosentan on exercise capacity and mPAP were similar in both subgroups. No changes in SpO or Qpi were observed in either bosentan or placebo subgroups. CONCLUSIONS: Improvements in exercise capacity and cardiopulmonary hemodynamics, without desaturation, were observed in ES patients with both ASDs and VSDs. Although not reaching statistical significance, improvements were similar to those in the BREATHE-5 analyses, suggesting that the location of septal defects is not a key determinant of treatment response. These data further support the use of bosentan for the treatment of ES, independent of shunt location.