Impact of untreated mild-to-moderate mitral regurgitation at the time of isolated aortic valve replacement on late adverse outcomes^☆

Abstract

Objective: The impact of untreated mild-to-moderate mitral regurgitation (MR) on patients undergoing isolated aortic valve replacement (AVR) is uncertain. The aim of this study is to investigate its long-term effects on outcomes. Method: We retrospectively reviewed 193 consecutive patients undergoing isolated AVR between 1993 and 2007. The mean age of the study group was 64 ± 12 years, 59% were male and the mean preoperative ejection fraction was 59 ± 12%. The pathologic aetiology and degree of MR was determined on preoperative echocardiogram. Patients were stratified into preoperative no/trivial MR (group I; n = 134) versus mild-to-moderate MR (group II; n = 59). The aetiology of MR in group II was either organic (n = 35, 60%) or functional (n = 24, 41%). Survival and functional outcome were compared between the two groups and analyses for predictors of adverse events were performed by the Cox proportional hazard model. Results: Operative mortality was 2.6% (n = 5). In group II, mean degree of MR significantly decreased from 2.1 ± 0.3 to 1.6 ± 0.8 during the late period (p = 0.003). The improvement in MR grade was more obvious in patients with functional aetiology. Although the actuarial survival was not significantly different between groups, freedom from re-admission for heart failure at 10 years was significantly lower in group II than in group I (23% vs 83%; p = 0.002). Multivariate analysis demonstrated that independent predictors of heart failure were presence of mild-to-moderate MR (p = 0.012, odds ratio (OR) 3.8) and left ventricular ejection fraction (p = 0.004, OR 0.95). Conclusion: Despite the significant reduction after isolated AVR, preoperative mild-to-moderate MR is an independent risk factor impacting long-term functional outcome. Our results suggested that the concomitant mitral valve surgery for mild-to-moderate MR is warranted, especially in patients with reduced left ventricular function.

1 Introduction

Mitral regurgitation (MR) is a common finding in patients undergoing aortic valve replacement (AVR) with an incidence of more than half of patients being evaluated [1,2]. When the degree of MR is severe, a double-valve operation, either repair or replacement, is indicated. However, when MR is mild to moderate (2+ or 3+), concerns are raised whether a second valve should be simultaneously operated upon at the time of AVR, because concomitant surgery of the aortic and mitral valve is associated with an increased morbidity and mortality [3]. Although some reports have shown improvement in the degree of MR in up to half of patients after isolated AVR, there are conflicting reports showing unchanged or even worsened MR in suboptimal percentage of patients [1,2,4–14]. Moreover, relatively few studies to date have examined the clinical impact of preoperative mild-to-moderate MR on long-term outcome [2,10,11].

In this study, we sought to determine the degree of improvement in MR after AVR alone and examine whether patients with concomitant mild-to-moderate MR experienced an increase on long-term adverse outcomes compared with those without significant MR at the time of AVR.

2 Patients and methods

We retrospectively reviewed 377 patients, who underwent AVR between January 1993 and December 2007, after obtaining the ethical committee approval at the Osaka University Hospital. Of those, we excluded patients with combined procedures including coronary artery bypass grafting, treatment of endocarditis and previous aortic or mitral valve surgery at the time of AVR. Thirty-two patients who underwent mitral valve operation for moderate-to-severe MR were also excluded. After excluding those, 193 patients were included in the study cohort.

2.1 Definition and grading of mitral regurgitation

All patients had a preoperative transthoracic echocardiogram before AVR. Patients were defined as having predominant aortic stenosis (AS) if they had aortic stenosis and 1+ or 2+ aortic regurgitation (AR), as predominant AR if they had 3+ or 4+ aortic insufficiency and as mixed aortic stenosis and regurgitation (ASR) if they had both aortic stenosis and 3+ or 4+ aortic insufficiency.

The severity of MR was graded as follows: none (0), trivial (1+), mild (2+), moderate (3+) and severe (4+), based on the distance of regurgitation extending into the left atrium beyond the mitral valve. Then the patients were stratified into group I (none/trivial MR) versus group II (mild/moderate MR). In group II patients, the pathologic aetiology of MR was also determined from the preoperative echocardiogram reports and categorised as: (1) functional, if occurring without valvular morphologic abnormality with or without wall motion abnormalities or LV dysfunction; or (2) organic, if associated with evidence of leaflet, annular, chordal or papillary muscle pathology [7].

In group II patients, postoperative echocardiogram reports at the time of discharge and at the late period (1–3 years after the operation) were collected and reviewed to determine the change in MR. A change in MR was considered as an improvement when there was at least a one score decrease from the preoperative degree of MR.

2.2 Surgical procedures

Aortic valve or root replacement with or without other minor procedures was performed with conventional cardiopulmonary bypass at mild hypothermia. In the majority of patients, both antegrade and retrograde cold blood cardioplegia were used for myocardial protection. In patients over the age of 65, a biologic valve had been preferably implanted. Regarding the implanted valve size used, particular effort including annular enlargement to insert a larger prosthesis was made to avoid the prosthesis–patient mismatch, especially in patients with small aortic annulus.

2.3 Patient characteristics

The clinical characteristics of the two groups are shown in Table 1 . There were 134 patients (69%) in group I and 59 patients (31%) in group II. The patients in group II were significantly older; therefore, they were more likely to receive a bioprosthesis than were those in group I (p = 0.03). Left ventricular (LV) systolic and diastolic dimensions on preoperative echocardiogram were significantly larger in group II, while left ventricular ejection fraction (LVEF) was similar between the groups. There were no differences between the two groups in any other preoperative co-morbidity and aetiology of aortic valve disease. The degree of MR was grade 2+ in the majority of patients (88%) in group II. None of patients had grade 4+ MR. This is likely a reflection of our collective principle to intervene on severe MR at the time of AVR.

2.4 Follow-up

The primary end point was to evaluate the impact of mild-to-moderate MR on long-term survival and freedom from re-admission for congestive heart failure. Follow-up data after the operation were obtained and completed in 183 patients (95%) through a review of patient medical records and telephone or postal mail interviews. Ten patients were lost to follow-up after discharge. Mean follow-up period was 3.3 ± 0.5 years. All of the cardiac and non-cardiac events including death were recorded. Early mortality was defined as in-hospital mortality and late mortality was death occurring beyond that period.

2.5 Statistical analysis

SPSS (version 11.0, SPSS Inc., Chicago, IL, USA) software was used for statistical analysis. The quantitative data are presented as mean ± standard deviation. Values obtained from pre- and postoperative data were compared by a paired t test. The unpaired t test and chi-square (χ²) test were used to compare the variables. Predictors for improvement of MR were identified using logistic regression analysis. Pre- and intra-operative variables with statistical significance by univariate comparisons were entered into the subsequent multivariate analysis. Kaplan–Meier analyses were used to assess the rate of freedom from all-cause death, and re-admission for congestive heart failure. Logrank testing was performed to compare groups. To clarify the independent predictor of adverse events, for pre- and intra-operative variables with statistical significance by univariate comparisons, multivariate analysis by using Cox hazard proportional analysis was performed. Statistical significance was defined as p ≪ 0.05.

3 Results

3.1 Change of mitral regurgitation in group II patients

The pathologic aetiology of MR in group II as determined by preoperative echocardiogram was either a functional one in 24 patients (41%) or an organic one in 35 patients (59%). Of the early survivors, 58 patients (98%) underwent early examination and 36 patients (61%) underwent late examination. Fig. 1 shows a serial change in the degree of MR in the early and late postoperative periods. At the time of discharge, MR was 1+ or less in 35 patients (60%), 2+ in 21 patients (36%) and 3+ in two patients (4%). In the late period, MR was 1+ or less in 17 patients (47%), 2+ in 16 patients (44%), 3+ in two patients (6%) and 4+ in one patient (3%).

Improvement of MR severity in patients with functional MR was more obvious than that in patients with organic MR (Table 2 ). A total of 16 patients out of 23 (70%) showed improvement in MR at the time of discharge and 11 of 15 (73%) during the late period in the functional group compared with 19 of 35 (54%) at the time of discharge and 8 of 21 (38%) during the late period in the organic group. Significant predictors of improvement in MR during the late period included male sex (p = 0.008), larger aortic prosthesis (p = 0.03) and functional aetiology (p = 0.03) by univariate comparisons. Conversely, in group II female patients, small aortic prosthesis used and in organic aetiology, MR tended to remain unchanged or worsen in the late period. The aortic valve aetiology had no significant impact on outcome. Subsequent multivariate analysis, however, failed to elucidate any significant predictors. Two patients in group II required mitral valve replacement for slow progression of organic mitral valve disease 9 and 11 years after the initial operation.

3.2 Mortality and morbidity

In general, in-hospital mortality was 2.6% (5 of 193), which was associated with congestive heart failure (n = 1), severe respiratory failure (n = 1), cerebral bleeding (n = 1) and multiple organ failure (n = 2). The in-hospital mortality of group II (1.7%) was not significantly different from that of group I (2.9%) (p = 0.6). Of the early survivors, late mortality was observed in 11 patients (5.9%) during the follow-up period. Survival of the whole cohort was 91.2% at 5 years and 89.4% at 10 years. Univariate pre- and intra-operative predictors of overall mortality included hypertension, preoperative use of catecholamine and the use of a biologic valve (p ≪ 0.05 for each). By multivariate analysis, however, only preoperative use of catecholamine was an independent predictor of overall mortality (p = 0.01). Preoperative significant MR (group II) showed no significant impact on long-term survival.

The overall mortality and late postoperative complications in each group are summarised in Table 3 . The incidence of postoperative complications including cerebrovascular accidents, infective endocarditis and re-do AVR remained low and was similar between groups. There was no re-operation for progression of mitral valve disease in group I. There was also no difference in late mortality between the two groups (4.6% vs 5.2%). Actuarial survival rates at 1, 5 and 10 years were similar between groups; 94.6%, 92.3% and 90.3% for group I versus 94.7%, 88.0% and 88.0% for group II (p = 0.49). However, the rates of freedom from re-admission for heart failure at 1, 5 and 10 years were significantly different between two groups; 98.4%, 91.9% and 88.8% for group I versus 96.2%, 76.5% and 23.5% for group II (Fig. 2 ). Among group II patients, a comparison of freedom from recurrent heart failure demonstrated a better trend in patients with improvement in their MR versus those that did not improve at the time of discharge (Fig. 3 ). However, this difference was not statistically significant (p = 0.15). By the Cox proportional hazard model, preoperative LVEF and group II were independent risk factors predicting re-admission for heart failure (Table 4 ).

Because preoperative LV function and presence of MR are likely to be key factors affecting long-term functional outcomes in patients undergoing isolated AVR, we stratified all the patients according to both preoperative LVEF (less than 50% or not) and degree of MR (less than 2+ or not). Kaplan–Meier curves depicting freedom from re-admission for heart failure of these four groups demonstrated that patients with both LVEF ≪50% and more than MR 2+ had significantly worse functional prognosis than the other group did (p ≪ 0.0001, Fig. 4 ).

4 Discussion

The major findings of our study are: (1) mild-to-moderate (2+ to 3+) MR improved immediately after AVR in 70% of patients with functional MR and in 50% of patients with organic MR and remained stable up to the late period; (2) preoperative mild-to-moderate MR is not an independent predictor of early and late mortality but a significant predictor for late adverse functional outcome, especially in patients with LV dysfunction. The role of concomitant mitral valve surgery at the time of AVR is still controversial and no clear guidelines exist. Our findings may suggest a selective approach to intervene on the mitral valve at the time of AVR when MR is mild to moderate.

There is evidence to suggest that relief of the aortic valve disease, particularly in patients with AS [4,5,7–9,12–14], results in some reduction in MR without mitral valve surgery. The regression of left ventricular hypertrophy and resolution of volume overload may promote further reduction. Several studies demonstrated that functional MR decreased in the majority of patients postoperatively [2,4–14]. In our cohort, similar to previous reports, more than 70% of patients with mild-to-moderate functional MR had benefited from isolated AVR and the incidence of worsened MR was low at the late period. With respect to organic MR, there are conflicting data, although it is generally known that the degree of benefit will be limited. Most recently, Unger et al. quantitatively assessed the change in MR after AVR and concluded that the improvement of MR was regardless of its aetiology [7]. Vanden Eynden et al. reported that the most of rheumatoid and myxomatous MR remained unchanged or even deteriorated after AVR [9]. Our result was consistent with the latter; 62% of patients with organic MR, in contrast to functional aetiology, were either unchanged or worsened. Moreover, the aetiology of MR was a possible predictor for improvement of MR. Therefore, preoperative echocardiographic inspection of the mitral valve morphology gives surgeons an important prognostic factor for the change in MR severity.

There were a few studies to examine the long-term survival in AVR patients who had concomitant MR. Among those with functional MR, two case-matched studies demonstrated no significant impact of moderate MR at the time of isolated AVR on long-term survival [2,12]. Among elderly patients over 70 years, Barreiro et al. found moderate MR, including both functional and organic aetiology, to be an independent predictor of late mortality [10]. In their study, however, patients with moderate MR had significant co-morbidities, including previous myocardial infarction and lower LVEF, with a mean of 47.6%. In our study, the groups were relatively homogeneous in their co-morbidities despite no matching. The AVR operation for group II patients could be safely performed and their long-term survival was also comparable to that in group I patients, suggesting that surgical correction of mild-to-moderate MR at the time of AVR is a parsimonious indicator in terms of survival benefit. However, the present study warned of another deleterious aspect of MR on long-term functional outcome. Less data are available regarding whether co-existing mild-to-moderate MR has an impact on long-term symptomatology. A study observed that patients with preoperative moderate MR had increased composite end points of congestive heart failure and re-intervention [11]. These authors also identified several important risk factors, including left atrial size, degree of aortic gradient, chronic atrial fibrillation in AS patients and LV end-systolic dimension in AR patients. These risk factors may reflect the underlying existence of more advanced cardiac disease in patients with MR. As a fact, in their series, half of the patients with moderate MR had LV systolic dysfunction with LVEF less than 50%. In our cohort, preoperative mild-to-moderate MR was an independent predictor for congestive heart failure, particularly when combined with LV dysfunction (LVEF ≪50%). It is consistent with other reports in which the burden of MR, even mild, is an independent prognostic factor for worse long-term outcome as seen in the setting of LV dysfunction [15,16]. Moreover, in ischaemic mitral regurgitation, acute pulmonary oedema inducing congestive heart failure symptoms is associated with exercise-induced change of mitral regurgitant volume in patients with less than mild MR at rest [17]. The same scenario could take place after isolated AVR for patients with co-existent MR. Although we could not elucidate statistically significant impact of persistent or worsened MR in group II patients after AVR on long-term functional outcome, possible exacerbation of MR after some loading might be responsible for higher incidence of late congestive heart failure in patients with residual 2+ or 3+ MR. The present study does not establish the indication for concurrent mitral valve surgery at the time of AVR, but our results suggested that mild-to-moderate MR might impose the substantial risk of congestive heart failure, especially in patients with organic aetiology and LV dysfunction, if left untreated. We acknowledge the increased operative mortality and morbidity by a double-valve surgery [3]. However, a later study also demonstrated the acceptable operative mortality of less than 10% in patients undergoing AVR with either concomitant mitral valve repair or replacement [18]. Therefore, we think that the possible increase in risk by simultaneous mitral valve surgery might be lower than the possible advantage with regard to late functional outcomes in these patients.

There are several limitations to this study. The study was not of randomised design and, therefore, selection bias or unidentified confounding bias such as postoperative medication may have influenced the results. Patients with mild-to-moderate MR were older and a greater number of them had remodelled LV than those without MR. Although LV systolic function was comparable between groups, those factors may suggest more advanced LV diastolic dysfunction, which is the important pathogenesis of heart failure symptom [19], in group II patients. Further investigations are needed to examine these limitations. The size of the valve implanted is another concern because prosthesis–patient mismatch is reported to be associated with increased mortality and morbidity after AVR [20,21]. In our cohort, 32 patients received a small valve (17 mm, n = 5 and 19 mm, n = 27). However, because we have paid much attention to avoid prosthesis–patient mismatch, the effective orifice area index calculated from the supplementary data provided by the valve manufacturer was not clinically significant (>0.85 cm² m⁻²) in these patients [20,21]. Finally, as a characteristic of a retrospective study, follow-up echocardiographic measurements were incomplete and opportunistic and there is potential for sampling bias in the assessment of MR at the late period.

5 Conclusion

Improvement in mild-to-moderate MR, especially with functional aetiology, can be expected in the majority of patients with significant aortic valve disease after isolated AVR. Although the presence of mild-to-moderate MR at the time of AVR does not constitute a risk factor for early and late mortality, it is an independent risk factor impacting long-term functional outcome in relation to LV systolic dysfunction. Our results suggested that the concomitant mitral valve surgery is warranted in patients with reduced LV function.

Acknowledgement

We wish to express our appreciation for the excellent technical assistance of Chikako Matsuo in the data acquisition.

Appendix A

Conference discussion

Dr H. Vanermen (Aalst, Belgium): I’ve read your paper with great interest. There aren’t that many papers, about the natural history of mitral valve regurg when you do an aortic valve replacement, so your paper is very interesting, although obviously there are some shortcomings.

If we look at the clinical status of the patient with organic MR, I think your conclusion is pretty good. We should be more aggressive in this group and especially in patients with bad ejection fraction. But in view of the fact that, first of all, a patient with aortic stenosis and very often a diastolic dysfunction and a hypertrophic left ventricle, their contractile reserve when there is myocardial regurg may be lost quicker; in view of the fact that it is an easy and a quick fix in the functional group, and in view of the fact that if you have to go back later on, it becomes a terribly difficult operation, don’t you think actually that we should be way more aggressive in the group, the functional group, where the MR is grade 3, grade 2, and where there is deterioration of the left ventricular function?

Dr Takeda: I think that there are still concerns about the treatment of mild-to-moderate functional MR in patients with aortic valve disease because, as we all know, the degree of functional MR decreases after isolated aortic valve replacement in most patients. In our results, functional MR decreased in 70% of patients. But unfortunately, we could not predict which patients will benefit from only isolated aortic valve replacement. At least in our small series, patients with heart failure symptoms and mild-to-moderate functional MR had significantly reduced left ventricular function compared with those without heart failure symptoms. So I think patients with symptomatic mild-to-moderate functional MR, had already more advanced cardiac disease. Therefore we should not underestimate the degree of mitral regurgitation in such patients, because many papers show that the existence of mild-to-moderate functional MR imposes a dismal prognosis on patients with left ventricular dysfunction.

So from our results we should more aggressively perform annuloplasty procedure, because many papers, including Dr Dion's, show that annuloplasty itself can be performed with a low operative mortality, even in high risk patients.

Dr K. Fukamachi (Cleveland, Ohio, USA): I have a question on your last conclusion that concomitant mitral valve surgery is warranted. Do you have the data from the patients with left ventricular dysfunction with concomitant mitral valve surgery that shows better results than this group?

Dr Takeda: I’m sorry that I have actually no available data about that. But, in the near future, technology advancement including percutaneous valve procedures enables us to easily perform double valve procedures in very high risk patients with left ventricular dysfunction.

Dr G. Mani (New Delhi, India): I’m just wondering whether perioperative transesophageal echocardiography done after aortic valve replacement and then reassessing the MR, especially in functional MR, would be a better strategy?

Dr Takeda: We reassess just after the operation the degree of mitral regurgitation. We analyse the degree of residual MR. But in functional MR groups, the degree of improvement is stable in the late period.

Dr Mani: We have observed that in functional MR after aortic valve replacement especially for aortic stenosis. And it does regress, and then you can do away with doing anything to the MR, after aortic valve replacement.

Dr Takeda: In group 1 patients?

Dr Mani: Yes.

Dr Takeda: In group 1 patients, no patients required reoperation due to progression of MR in the late period. In group 2 patients, however, about 7% of patients developed moderate to severe MR and two patients required mitral valve surgery in the late period.

References