Minimally invasive versus conventional mitral valve surgery: a propensity-matched comparison.

OBJECTIVE Less invasive approaches to mitral valve surgery are increasingly used for improved cosmesis; however, few studies have investigated their effect on outcome. We sought to compare these minimally invasive approaches fairly with conventional full sternotomy by using propensity-matching methods. METHODS From January 1995 to January 2004, 2124 patients underwent isolated mitral valve surgery through a minimally invasive approach, and 1047 underwent isolated mitral valve surgery through a conventional sternotomy. Because there were important differences in patient characteristics, a propensity score based on 42 factors was used to obtain 590 well-matched patient pairs (56% of cases). RESULTS In-hospital mortality was similar for propensity-matched patients: 0.17% (1/590) for those undergoing minimally invasive surgery and 0.85% (5/590) for those undergoing conventional surgery (P = .2). Occurrences of stroke (P = .8), renal failure (P > .9), myocardial infarction (P = .7), and infection (P = .8) were also similar. However, 24-hour mediastinal drainage was less after minimally invasive surgery (median, 250 vs 350 mL; P < .0001), and fewer patients received transfusions (30% vs 37%, P = .01). More patients undergoing minimally invasive surgery were extubated in the operating room (18% vs 5.7%, P < .0001), and postoperative forced expiratory volume in 1 second was higher. Early after operation, pain scores were lower (P < .0001) after minimally invasive surgery. CONCLUSION Within that portion of the spectrum of mitral valve surgery in which propensity matching was possible, minimally invasive mitral valve surgery had cosmetic, blood product use, respiratory, and pain advantages over conventional surgery, and no apparent detriments. Mortality and morbidity for robotic and percutaneous procedures should be compared with these minimally invasive outcomes.