MV repair with CABG is more effective at reducing moderate-severe ischemic functional MR as compared to CABG alone. In a group of 107 consecutive patients with moderate or severe ischemic MR, 50 patients underwent MV repair with concomitant CABG and 57 patients, CABG alone. Among patients with severe MR, a 1-year improvement of MR severity was observed in all patients in the repair group and only in 67 % in the CABG group (p < 0.001) [6]. In a larger cohort including patients with moderate-severe ischemic MR, persistence of moderate-severe MR at 1 year was more frequent in patients with isolated CABG (i.e. without MV repair) as compared to those with CABG and MV repair (48 % versus 12 %, p < 0.0001) [7]. In this regard, it seems that CABG alone is not sufficient to adequately address MR.

However, no study has demonstrated improvement in survival rates or long-term functional status with MV repair combined to CABG compared to CABG alone. Indeed, despite significant better improvement in MR severity in the MV repair + CABG as compared to CABG alone, both 5-year survival rate (88 % versus 87 %, p = 0.48) and long-term postoperative NYHA functional class were similar in the two groups [6]. These results were confirmed by the study from Mihaljevic et al. where survival rates at 1, 5 and 10-year were 88, 75 and 47 % in CABG alone and 92, 74 and 39 % in MV repair combined to CABG. In addition, NYHA functional class improvement was also similar in both groups at 5 years, as well as the persistence of severe symptoms (NYHA class ≥ III) [7]. Nevertheless, those data come from mono-centric retrospective studies without standardized surgical approach of the MV repair, allowing not reaching definitive conclusion. In addition, the surgeon could have influenced the decision to perform or not MV repair. Indeed, decision to not perform MV repair could be the result of downgrading of MR severity by intraoperative trans esophageal echocardiography (TEE) or of a long and difficult CABG leading to hemodynamic instability and longer ischemic time [8]. Ischemic MR represents a heterogeneous group of patients with a wide range of left ventricular (LV) dysfunction, remolding and extent of revascularization. However, chronic ischemic MR depicts a graded relationship between MR severity and reduced survival or cardiovascular event-free survival [9]. Even if survival benefit of MV repair has not been demonstrated, the association of chronic ischemic MR with worse prognosis and the fact that severe ischemic MR is not adequately corrected by revascularization alone could justify MV surgery at the time of revascularisation (Fig. 7.1).

As for moderate-severe ischemic MR, CABG alone appeared to be not efficient to significantly reduce MR in patients with moderate ischemic MR. In a cohort of 89 patients with moderate ischemic MR receiving CABG alone, only 62 % of patients exhibited markedly reduced MR severity, 36 % had no changes and 2 % increased MR [10]. Similarly, Aklog et al. reported a postoperative remaining rate of 40 % of moderate MR following isolated CABG, whereas 51 % of patients improved to mild and only 9 % had no MR [11]. Those data are consistent with the randomized trial published in 2009 and showing, in 102 patients with moderate MR, that CABG alone was less efficient to improve MR, LV size and function, left atrial dimension, systolic pulmonary arterial pressure and NYHA functional class than CABG + MV repair [12]. However, some specific patients could significantly improve ischemic MR after isolated CABG. Presence of large extend (≥5 segments) of viable myocardium, especially in the areas adjacent to papillary muscles, and absence of dyssynchrony (<60 ms) between papillary muscles have been identified as independent predictors of long-term reduction of MR severity after revascularization alone, whereas 93 % of patients with both characteristics showed an improvement to no or mild MR at 12-month follow-up. Preoperative dyssynchrony was the best reliable parameter to identify MR improvement after CABG alone with a positive predictive value of 89 % and a negative value of 85 % [4]. Then, assessment of myocardial viability and dyssynchrony could help in clinical decision-making of adding or not a MV repair to a revascularization procedure in patients with ischemic moderate MR [4, 5].

Controversy exists regarding impacts of residual moderate MR and benefit of MV repair on survival. In the study from Ryden et al., the improvement of functional status and survival at 1 and 3-year did not differ significantly according to residual MR [10]. Wong et al. reported similar results with no statistical difference in outcome between patients with postoperative MR and those without MR. Even if the MV repair subgroup had less postoperative MR, MV repair was not independently associated with improvement of long-term survival [13]. In the randomized trial from Fattouch et al., considering that it was not powered to detect difference in survival, 5–year mortality was similar in both CABG alone and CABG + MV repair groups (89 % vs. 94 %, p = NS) [12]. Conversely, Lam et al. have demonstrated a reduced survival in patients with residual MR after CABG alone. In patients with moderate ischemic MR who underwent CABG alone, the5-year survival was worse in patients with moderate residual MR than those without MR (73 % vs. 85 % p = 0.003) [14]. Fattouch et al. have also shown that 5-year freedom from deaths and cardiac events among patients with postoperative recurrent moderate ischemic MR versus no MR were, respectively, 91 % versus 74 % and 88 % versus 62 % (p < 0.0001). However, moderate ischemic MR affected significantly the late survival only in patients with LVEF of 40 % or less, as cardiac-related events were independent of the value of the LVEF. After robust adjustment, postoperative moderate MR was associated with a 2.7-fold increase in risk of death. The others independent predictors of late mortality in this cohort were age, LVEF <40 %, NYHA functional class ≥ III and renal failure [15]. Of note, it should be emphasized that only postoperative moderate MR is a risk factor that may potentially be avoid and controlled using a more aggressive strategy such as the use of MV repair. In this regard, although the lack of available randomized studies contributes to prolong the controversy, it seems reasonable to correct moderate ischemic MR during CABG, especially in patients with LVEF less than 40 %, dyssynchrony and absence of significant myocardial viability, in order to avoid residual MR and its harmful consequences (Fig. 7.1).

Moreover, an important thing to consider is that chronic ischemic MR is a dynamic condition very dependent on preload and after load. In many studies, there was a considerable variability in the degree of MR severity on the intraoperative TEE despite the fact that all patients had preoperative moderate MR [11, 13, 14]. In a cohort of patients with moderate ischemic MR who underwent revascularization and concomitant intraoperative TEE, 90 % of patients had MR downgraded to mild or less and one third had no detectable MR [11]. Similarly, in another cohort of 95 patients, the degree of MR was downgraded in 57 % of patients and even more frequently in patients who underwent isolated CABG (71.2 %) [13]. Generally, the unloading effect of anaesthesia is the causal mechanism with the induction of arterial and venous vasodilatation leading to the decreasing of after load and preload. Hence, the decision to treat ischemic MR should be taken before surgery in patients undergoing revascularization.

In patients with chronic ischemic MR, important incremental information can be obtained from exercise Doppler echocardiography, as resting MR severity cannot predict the extent of exercise-induced changes in MR severity [16]. A large exercise-induced increase (≥13 mm²) in effective regurgitant orifice area is indicative of significant dynamic component of functional MR and predicts mortality and hospitalization for heart failure, regardless resting MR severity [17, 18]. Hence, exercise Doppler echocardiography can unmask significant dynamic MR in patients with ischemic left ventricular dysfunction and mild to moderate MR at rest and is recommended in those patients before surgical revascularization. However, the prognosis importance of dynamic MR is not necessary applicable to non-ischemic cardiomyopathy.

Mitral valve repair using an undersized (one or two sizes), rigid ring annuloplasty with revascularization is considered as the best approach to surgically treat ischemic MR. However, late results are controversial according to patient’s heterogeneous preoperative characteristics. Restrictive mitral valve annuloplasty has been shown to improve NYHA functional class with minimal residual MR and to significantly reduce LV and left atrial dimension at 2 years follow-up in patients with ischemic LV dysfunction and severe MR who underwent CABG [19]. Even if ischemic MR is a ventricular and not a valvular problem and that MV repair targets the consequence rather than the cause, this technique provides good results in terms of MR correction and improvement of symptoms in well-selected patients. Indeed, it seems that the best outcome results achieved by restrictive annuloplasty is obtained in patients with a preoperative LV end-diastolic diameter less than 65 mm [20].

Combined CABG with mitral procedure is associated with significant mortality. Actually, the necessity to add revascularization to a mitral procedure means that you address a sicker patient with more commodities leading to a higher rate of morbidity and mortality. Moreover, MV repair and MV replacement results are difficult to compare, as population was very different between studies [21, 22]. A recently published meta-analysis comparing mortality according to the procedure performed to correct ischemic MR demonstrated a significantly increase likelihood of both short-term and long-term mortality for MV replacement compared to repair. However, inherent limitations of meta-analysis including retrospective studies should be acknowledged, especially regarding the heterogeneous characteristics of patients studied [23]. In fact, patients who undergo MV repair are often different from those who undergo replacement being generally younger with a better NYHA functional class and less LV remolding [24]. According to this, three large matched cohorts of patients have shown no evidence that MV repair provide better short or long-term survival compared to MV replacement. Gillinov et al. compared 482 propensity-score matched patients with ischemic MR, both early (30 days) and late (1–5 years) survivals were not statistically different after MV repair compared to replacement. To note, within the propensity-matched better-risk group, a trend favoring MV repair compared to replacement was observed, as within the poor-risk group, survivals were strictly similar. Also, early and late survivals were not related to the type of procedure, but to the preoperative characteristics of the patients [25]. In a study by Magne et al., 370 patients with ischemic MR underwent MV repair (186) or replacement (184). Compared to patients with MV repair, those with MV replacement had higher prevalence of preoperative renal failure, pulmonary hypertension, NYHA functional class ≥ III and severe MR, as well as lower LVEF and bigger LV end-diastolic diameter. Although operative mortality was significantly lower after MV repair compared to replacement (9.7 % versus 17.4 %, p = 0.03), 6-year survival was not statistically different between procedures (73 % versus 67 %, p = 0.17). In addition, the type of procedure was not an independent predictor of operative or 6-year mortality in multivariate analysis. Performing identical analysis in a propensity score matched cohort resulted in similar findings [26]. More recently, Lorusso et al. studied 244 pairs of propensity score matched patients who underwent either MV repair or replacement in association with CABG for chronic ischemic MR and LV ejection fraction less than 40 % from a large multicenter study. Early (30-day) mortality (3.3 % versus 5.3 %; p = 0.32) and 8-year survival (81.6 % versus 79.6 %; p = 0.42) were similar in patients with MV repair compared to replacement and, again, the type of procedure was not an independent predictor of early or late mortality. However, freedom from valve-related reoperations was significantly higher in MV replacement compared to MV repair (85.5 % versus 71.3 %; p < 0.001) and freedom from valve-related complications was similar (69.9 % versus 68.3 %; p = 0.78) Moreover, MV repair was a strong predictor of reoperations, mainly for repair failure [27]. Consequently, these studies emphasized that the outcome of patients with ischemic MR is more related to the patient’s clinical status and comorbidities rather than the type of procedure. Those findings underline the importance of tailoring the choice of procedure according to the preoperative characteristics of the patients (Table 7.1).

Both procedures present some advantages and inconvenient. First, even if MV repair protects against postoperative MR in the vast majority of patients, substantial rate of persistent and recurrent MR are currently reported in the literature. Consistently, MV replacement provides better freedom from both persistent and recurrent MR without a higher incidence of valve-related complications [34]. Concern has been raised about the relatively high incidence of persistent and recurrent MR found in some patients after MV repair. In fact, MV repair is known to be associated with 15–25 % rates of persistent MR early after operation and with 50–70 % rates of recurrent MR at 5 years [35]. Moreover, the presence of persistent and recurrent MR is associated with higher incidence of cardiac events and reduced survival [28, 36]. Patients with persistent MR had a significantly lower 3-year event-free survival compared to patients without residual MR (26 % versus 75 %; p = 0.01) [28]. Persistent MR is likely related to the alteration of mitral valve geometry with the persistence or worsening of the tethering of MV leaflets, especially the posterior (Fig. 7.2), as recurrent MR is more associated to progressive LV remodeling with increasing of LV dimension and sphericity (Fig. 7.3) [28, 30, 36]. However, all recurrent MR cannot be explained by adverse global LV remolding. Localized LV remolding, more particularly in the vicinity of papillary muscle insertion site leading to increase tenting on the anterior leaflet, has been shown to be an alternative mechanism in patients without LV dilatation (Fig. 7.2) [37]. According to this, the evaluation of the risk of persistent/recurrent MR from preoperative echocardiography is primordial (Table 7.1). In patients at high risk of MV repair failure, MV replacement with preservation of the subvalvular apparatus could be recommended (Fig. 7.1) [35, 38, 39]. Despite the relatively high rates of residual MR, the under sizing of more than two sizes is not recommended as small annuloplasty size did not influence post-operative MR [40]. In addition, restrictive annuloplasty may induce some degree of functional mitral stenosis [41] associated with higher pulmonary arterial pressure, reduced functional capacity and heart failure symptoms [42, 43]. Finally, the outcome of patients undergoing MV replacement may be affected by prosthetic valve thrombogenicity, durability and hemodynamic. The risk of thromboembolic events is higher in mitral than in aortic position, in mechanical valve and in the early postoperative phases (<3 months). In addition, the risk of anticoagulant-related bleeding should also be considered. Risk factors associated with accelerated structural valve deterioration include younger age and bioprosthetic valves in mitral position and expose patients to the risk of reoperation. Moderate or severe prosthesis-patient mismatch following MV replacement could be present in around one third of patients and lead to reduce long-term postoperative survival [44, 45]. Even if MV repair may avoid the complications related to MV replacement, this may be counterbalanced by the rate of persistent/recurrent MR requiring surgical re-intervention and by the induced functional mitral stenosis which could explain, at least in part, the lack of overt evidence-based benefit of MV repair compared to replacement in patients with ischemic functional MR.