ABSTRACT
BACKGROUND: Severe prosthesis-patient mismatch (PPM) is considered to further decrease survival compared to moderate PPM. This study aimed to assess the impact of severe PPM on survival after aortic valve replacement (AVR). METHODS: We retrospectively studied 2404 consecutive patients with PPM who underwent first-time AVR for pure stenosis between January 2003 and December 2014. Mismatch was moderate for indexed effective valve orifice >0.65 to <0.85cm2/m2 and severe for indexed effective valve orifice ≤0.65cm2/m2. Moderate mismatch occurred in 2165 patients (89%), and severe in 239 (11%) patients. Logistic multiple regression with bootstrapping and propensity score analyses were performed using 29 clinical and demographic data to assess the risk-adjusted impact of severe mismatch on mortality. The Cox proportional hazards model was constructed to process the long-term outcome. RESULTS: Early mortality was 2.3% (51/2165) in moderate mismatch group and 3.7% (9/239) in severe mismatch group (p=0.2). Mortality at 5 and 10 years, was 218/1470 (14.8%) and 252/585 (43.1%) for moderate mismatch and 43/198 (21.7%) and 61/105 (58.1%) for severe mismatch (p=0.02 and p=0.006). Multivariable predictors of late mortality were as follows: age ≥70 years, left ventricular ejection fraction ≤40%, indexed left ventricular mass >220g/m2 and concomitant coronary artery revascularization. After propensity score matching, conditional logistic regression analysis demonstrated no relationship between severe mismatch and increased mortality at 5 postoperative years (HR, 0.9; 95% CI, 0.7-1.6; p=0.06), whereas it was significant at 10 postoperative years (HR, 1.9; 95% CI, 1.2-2.5; p=0.03). During the follow-up, severe mismatch was associated with more frequent hospital readmissions for cardiac events (0.12 vs. 0.08 events/patient/year, p=0.007). CONCLUSIONS: Patients with severe mismatch had lower long-term survival and higher incidence of hospital readmissions for cardiac events. However, the effect of severe mismatch on outcome appeared mainly related to the preoperative risk profile of each patient.
