The safety and benefit of minimally invasive aortic valve replacement in patients with left ventricular dysfunction has not been well investigated. We conducted a retrospective review of 140 patients with ejection fraction <or= 40% who underwent isolated aortic valve replacement between July 1996 and March 2005. Aortic valve replacement was performed via an upper hemisternotomy in 73 patients and via a full sternotomy in 67. Two matched cohorts of 41 patients each were constructed using propensity score analysis, and the outcomes were compared. There was no significant difference in operative mortality (hemisternotomy, 2.4% vs 4.8% for full sternotomy), incidence of postoperative complications, blood transfusion requirement, length of hospital stay, or discharge to home rates. Aortic valve replacement via an upper hemisternotomy can be performed safely, even in patients with left ventricular dysfunction, with morbidity and mortality outcomes similar to those of a full sternotomy.