Results
Between 2011 and 2016, 576 patients underwent mitral valve surgery for severe MR due to degenerative valve disease. Majority or 77.2% of the operations (n= 444) were performed by 3 experienced surgeons, and 6 surgeons with less experience performed 22.8 % of the operations (n=132). No patient left the OR with an unsatisfactory repair (moderate or higher MR on post-pump ECHO). Patients operated on by less experienced surgeons were more likely to be non-Caucasian, to have higher body mass index, chronic lung disease, CHF, lower ejection fraction, and to be undergoing an urgent or emergent operation. Other characteristics, including age, sex, smoking status, other comorbidities including DM, hypertension, dyslipidemia, dialysis, peripheral vascular disease, cerebrovascular disease, and previous myocardial infarction, New York Heart Association (NYHA) class and atrial fibrillation were not different between the two groups (Table 1).
Procedural details by surgeon experience are presented in Table 2. Experienced surgeons were more likely to attempt repair (P=0.024) and more likely to succeed in repair (P=0.001). Experienced surgeons were more likely to attempt repair of both anterior leaflet prolapse and bi-leaflet prolapse (P=0.005). There was no difference in attempted repair of posterior leaflet prolapse (P=0.871). Fewer patients had residual MR in the experienced group (P=0.03). Experienced surgeons had shorter mean cross clamp times (P<0.001). Less experienced surgeons used the trans-septal approach more often (P<0.001) and were more likely to describe leaflet restriction (P<0.001). Valve pathology represented by leaflets affected, calcification, and annular dilatation were not statistically different between both groups.
The technical score (number of techniques used in the repair) was not different between both groups, but the techniques used differed according to surgeon experience. Experienced surgeons were more likely to use neochordae (P<0.001), and less experienced surgeons used chordae transfer more often (P<0.001). Rates for other techniques were not significantly different between the two groups. (Table 3).
Repair rate was higher in the experienced group (81.3% vs 69.7%, P=0.005), and rationale for valve replacement differed by surgeon experience (P=0.001). Extensive calcification was the primary reason for replacement in the experienced group, whereas failure of attempted repair was the most common in the less experienced group (Table 4).
The overall rate of recurrence was 13% (n=61) over the study period. Most (69%) recurrences happened in the first 2 years after surgery (Figure 1). On multivariable logistic regression analysis; in the first two post-operative years, surgeon experience was not a risk factor for recurrence. Annular calcification was the only independent factor for higher risk of recurrence (OR = 8.98 CI 3.19-25.28). Patient’s age, male sex, DM, hypertension, urgent/emergent surgery and anterior/bi-leaflet prolapse were not independent risk factors for recurrence (Figure 2). Of the patients with recurrent MR, 23% (n=14) underwent mitral reoperation, and the other 77% (n=38) were either asymptomatic or high risk for reoperation. Reasons for recurrence according to surgeon description in the operative reports included dehisced ring in 35.7% (n=5), new lesions in 35.7% (n=5), endocarditis in 14.3% (n=2), and torn neochordae in 14.3% (n=2). Of the patients who underwent reoperation, re-repair was performed in 35.7% (n=5), all of which were performed by experienced surgeons, and the remainder underwent valve replacement.
The overall mortality (throughout 2011-2020) was 11.1% (n=64), with a rate of 8% (n=37) in the experienced group and 21% (n=27) in the less experienced group. By KM method, adjusted survival was higher in patients treated by more experienced surgeons (log rank P<0.0001) (Figure 3). 5-years survival in the patients operated on by experienced surgeons was 93.9% (n=417) and 80.2% (n=105) in the patients operated on by less experienced surgeons. Independent risk factors for mortality on Cox model were: less experienced surgeon (HR= 2.64, P=0.002), age (HR=1.03, P=0.012), valve replacement (HR=1.75, P=0.04), CHF (HR=2.01, P=0.029) and chronic lung disease (HR=2.25, P=0.005) . DM, dyslipidemia, HTN, Afib and urgent/emergent surgery were not independent factors for mortality (Table 5).