Influence of patient age on procedural selection in mitral valve surgery

Abstract

Background: Previous studies suggest that mitral valve replacement is comparable to repair in the elderly, and a national trend exists toward tissue valves. However, few direct comparison data are available, and this study evaluated the effects of patient age on risk-adjusted survival after mitral procedures.

Methods: From 1986 to 2006, 2,064 patients underwent isolated primary mitral operations (±CABG). Maximal follow-up was 20 years with a median of 5 years. Valve disease etiology was the following: degenerative, 864; ischemic, 450; rheumatic, 416; endocarditis, 98; and "other," 236. Overall, 58% had repair and 39% had concomitant coronary artery bypass grafting. Survival differences were evaluated with a Cox proportional hazards model that included baseline characteristics, valve disease etiology, and choice of repair versus replacement with tissue or mechanical valves.

Results: Baseline risk profiles generally were better for mechanical valves, and age was the most significant multivariable predictor of late mortality [hazard ratio = 1.4 per 10-year increment, Wald χ(2) = 32.7, p < 0.0001]. As compared with repair, risk-adjusted survival was inferior with either tissue valves [1.8, 27.6, <0.0001] or mechanical valves [1.3, 8.1, 0.0044], and no treatment interaction was observed with age (p = 0.18). At no patient age did tissue valves achieve equivalent survival to either repair or mechanical valves.

Conclusions: Mitral repair is associated with better survival than valve replacement across the spectrum of patient age. If replacement is required, mechanical valves achieve better outcomes, even in the elderly. These data suggest that tissue valves should be reserved only for patients with absolute contraindications to anticoagulation who are not amenable to repair.

Fig 1

Incidence of mitral procedures over time. (— = repair; ··· = replacement-tissue; --- = replacement-mechanical.)

Fig 2

Unadjusted Kaplan-Meier survival analysis. Log-rank p value less than 0.0001 for tissue valves versus either mechanical valves or repair.

Fig 3

Unadjusted Kaplan-Meier survival analysis for patients greater than 65 years of age. Log-rank p value = 0.001 for tissue valves versus either mechanical valves or repair.

Fig 4

Long-term survival after adjusting for differences in baseline characteristics. Cox model p values. (Repair versus mechanical replacement = 0.0044; repair versus tissue replacement = <0.0001; mechanical versus tissue replacement = 0.0017.)

Fig 5

Adjusted survival estimates conditional on 90-day survival. Cox model p values. (Repair versus mechanical replacement = 0.0493; repair versus tissue replacement <0.0001; mechanical versus tissue replacement <0.0001.

Fig 6

On the y axis is risk-adjusted survival at 10 years after valve surgery, and on the x axis is patient age at the original surgical procedure. Adjusted 10-year survival probability was best for patients receiving mitral valve repair, followed by mechanical valve replacement for all ages. Tissue mitral valve replacement was associated with decreased adjusted 10-year survival at all ages, even in the elderly. Thus, outcome differences for the 3 procedures were fairly constant across all patient ages. (— = repair; ··· = tissue replacement; --- = mechanical replacement.)