Impact of valve prosthesis-patient mismatch on intermediate-term outcome and regression of left ventricular mass following aortic valve replacement with mechanical prosthesis.

BACKGROUND: The impact of aortic valve replacement (AVR) with prosthesis-patient mismatch (PPM) on intermediate-term outcome and left ventricular mass (LVM) regression in patients with aortic stenosis (AS) was investigated.

METHODS: One hundred fifty patients with AS (87 pure stenosis and 63 combined stenosis and regurgitation) were classified into a PPM group (n = 34, indexed effective orifice area (EOAI) >0.65 cm(2)/m(2) and < or =0.85 cm(2)/m(2); moderate PPM) and a non-PPM group (n = 116, EOAI > 0.85). Mean age, mean and peak aortic pressure gradient (PG) were not different between the groups (PPM, 99.7 +/- 37.2 and 54.9 +/- 23.2 mmHg; non-PPM, 95.9 +/- 29.2 and 54.4 +/- 16.0 mmHg). The absolute and relative regression in indexed left ventricular mass (LVMI) was estimated by preoperative and postoperative echocardiography (n = 98).

RESULTS: Twelve patients died (valve-related death in 7) during 5 years of follow-up. Comparing the PPM and non-PPM groups, overall survival (78.7% vs. 87.8%) and survival free from valve-related death (96.8% vs. 92.1%) were not significantly different. New York Heart Association (NYHA) functional class improved in all patients and there were no patients in class III or IV. The postoperative mean PG was 14.6 +/- 6.1 mmHg in the PPM group and 9.4 +/- 3.8 mmHg in the non-PPM group (p = 0.0005), with an inverse correlation (r = -0.48, p < 0.0001) between EOAI and the postoperative mean PG. However, there was no significant difference in the absolute and relative LVMI regression between the two groups. Multiple linear regression analysis was performed and higher preoperative LVMI and mean aortic PG were independent predictors of greater LVMI regression after AVR.

CONCLUSIONS: Moderate PPM does not appear to alter LVMI regression, NYHA class, or intermediate-term outcome in AS patients undergoing AVR with mechanical prostheses. In multivariate analysis, preoperative LVMI and mean aortic PG were important independent predictors of LVMI regression.