52 Mitral valve surgery in heart failure

Functional, or secondary, mitral regurgitation (MR) develops as part of the process of remodelling which in turn occurs as a result of left ventricular failure. Although this complication can result whatever the primary pathology causing the left ventricular failure, the pathophysiological process varies to a certain extent depending on the aetiology. There is some confusion in the literature with respect to ischaemic MR, with some authors calling all ischaemic MR functional. For the sake of clarity, this chapter deals with functional MR resulting from left ventricular failure which for the sake of argument I will define as a left ventricular ejection fraction (LVEF) of less than 35%.

Functional MR is a common consequence of left ventricular systolic dysfunction (LVSD) and it has an important bearing on both symptoms and outcome in patients suffering from heart failure (HF). In a study of 2057 patients with symptomatic HF and with a LVEF of less than 40% but without structural valve disease, 56.2% had some degree of MR and in 29.8% of these it was moderate or severe. Furthermore, the presence of moderate or severe MR was associated with a significantly higher mortality at 1, 3, and 5 years (Fig. 52.1). In a multivariate analysis, MR of any degree was an independent predictor of death.¹

The success of mitral valve surgery, and in particular the flourishing of mitral valve repair techniques, along with the ever-increasing burden of patients with HF, has led to the adaptation of repair techniques, originally developed to correct structural abnormalities of the mitral valve, to attempt to halt or even reverse the remodelling process.

As much as 40 years ago it was appreciated that an adequate mitral valve repair might have advantages over valve replacement so long as ‘such repair would hold up over a long period of time’.² During that era of cardiac surgery most patients coming to valve surgery had either congenital abnormalities or, more commonly, rheumatic heart disease. Subsequently, through the 1970s Carpentier and his colleagues in Paris were developing more complex techniques to repair mitral valves with degenerative as well as rheumatic pathology.³

The first published series of patients undergoing repair for functional MR in 1995⁴ heralded a rapid expansion in the number of patients with a combination of left ventricular failure and MR being put forward for surgery. Subsequently, in the absence of appropriate randomized clinical trials it has become less clear which patients should undergo this procedure and even what procedure should be offered.

The surgical literature, although replete with studies of mitral valve surgery for structural abnormalities, is not so well furnished with respect to correction of functional regurgitation. Much of this subjective literature is also devalued by primitive assessment of the degree of MR. From the earliest years of open heart surgery onwards there have been nonrandomized comparisons of mitral repair versus replacement⁵ and although there are several ongoing randomized trials there are no trials yet published which measure the effect of mitral valve surgery in HF. Thus, as is so often the case in heart surgery, we must fall back on evidence based on cohort studies and, to some extent, inadequately tested conjecture.

The reader should also bear in mind that mitral valve surgery is only one weapon in the surgeon’s armamentarium against HF and that it can be employed in combination with other procedures such as revascularization, ventricular remodelling, and mechanical ventricular assist.⁶

The three principle components of left ventricular remodelling which contribute to the development of MR are annular dilatation, papillary muscle displacement, and discoordination of ventricular contraction resulting from dyssynchrony. One effect of the latter two processes is to cause tethering of the mitral leaflets. Once MR has developed, the consequent increase in volume loading of the ventricle leads to an increase in ventricular cavity volumes, greater transmural tension, and yet more ventricular dilatation.

Carpentier’s original pathophysiological classification of MR consisted of three types (figure) according to leaflet motion: normal (type I), increased (type II), or restricted (type III) (Fig. 52.2).⁷ This classification was designed to help in the decision-making process before and during valve repair. Thus, type I MR can be resolved by an annuloplasty of some kind, type II by elimination of prolapse, and type III by relieving whatever is causing the restrictive motion. In both dilated and ischaemic cardiomyopathy, the posterior leaflet can be restricted by an outward, paradoxical motion of the left ventricular wall at the base of a papillary muscle. This has come to be known as type IIIb leaflet motion. In many patients more than one mechanism may be at work at the same time.

To some extent the mechanism of MR is dependent on the underlying pathology. Functional regurgitation is either type I, as a result of annular dilatation, or type IIIb, as a result of abnormal ventricular wall motion leading to leaflet tethering. In idiopathic dilated cardiomyopathy, the process can be considered to be passive although there can still be both annular dilatation and leaflet tethering as result of ventricular dilatation. In ischaemia there is the possibility of an active component to leaflet tethering with the potential for reversing this by revascularization. There is no direct evidence for this and experimental models of ischaemia of the papillary muscles, in the absence of ventricular dilatation, failed to produce MR.⁸

It might seem logical that the presence of MR is associated with a reduction in left ventricular afterload. Indeed, it is not uncommon to hear the view that restoration of mitral competence in the presence of left ventricular dysfunction might result in worsening HF because the mitral valve is no longer acting as a safety valve for the ventricle. In fact, there is evidence that the opposite is more likely to be the case. Physiological studies, albeit in patients with chronic MR resulting from structural valve abnormalities, have demonstrated that there is an increase in mean, peak, and end-systolic stress⁹ when compared to controls without MR, all three being measures of afterload. Furthermore, the increase in afterload is greater in the presence of reduced ventricular function. At face value this would suggest that restoration of mitral competence, providing that it can be achieved safely, is always going to be to the patient’s advantage. There is also evidence that once functional MR is corrected there is an arrest and reversal of remodelling. This has been demonstrated in the ACORN trial in which elimination of MR with repair or replacement, in patients with a LVEF of less than 35%, was associated with progressive reduction in left ventricular end-diastolic volume and left ventricular end-systolic volume and an increase in LVEF compared with the preoperative baseline up to 24 months after surgery.¹⁰

Preservation of left ventricular function in the face of mitral valve surgery is particularly important when dealing with a ventricle whose function is impaired in the first place. Nonrandomized studies suggested, from an early stage in the evolution of mitral valve repair techniques, that repair was associated with better preservation of left ventricular function than replacement and that this was linked to lower operative mortality and morbidity.

There is evidence that the advantage which repair holds over replacement is the result of the preservation of the subvalvar apparatus and the consequent maintenance of left ventricular geometry. Nearly 50 years ago Dr C. Walton Lillehei, one of the greatest of the founding fathers of modern cardiac surgery, published experimental and clinical studies which demonstrated that mitral valve replacement with preservation of the subvalvar apparatus significantly reduced the ill-effects of mitral valve replacement on left ventricular function.¹¹ More recent publications have born this out to the extent that mitral valve replacement with preservation of the subvalvar apparatus may be just as good as mitral valve repair with respect to ventricular function and geometry.^12,13 There is one small randomized controlled trial of mitral valve replacement comparing preservation of posterior leaflet attachments with preservation of all chordal structures,¹⁴ the findings of which support the contention that mitral valve replacement with preservation of all chordal attachments is as efficacious as valve repair in preserving left ventricular function.

There are only two fundamental questions in surgery: ‘what to do?’ and ‘when to do it?’ With respect to functional MR, the first can be simply expressed as ‘repair or replace?’

Before the advent of valve prostheses there was a profusion of techniques described for repair of MR. Most of them were varying forms of suture annuloplasty.^15,–18 Reed and his colleagues brought a greater application of pathophysiological and haemodynamic principles to the subject.¹⁹ They had become aware that annular dilatation predominantly affected the posterior part of the annulus, that part to which the posterior leaflet is attached. Their suture repair technique left the unaffected anterior annulus unaltered but reduced the posterior annulus to a measured length corresponding to a reduced but adequate mitral valve orifice area. In due course these same principles governed the development of the rigid and semirigid annuloplasty rings developed by Carpentier and others.

The initial generation of valve prostheses at the start of the 1960s brought with them a host of new complications, so that for many surgeons valve repair was maintained while they awaited the development of ‘safe’ prostheses. When that happened, through the 1970s and 1980s, the torch of valve repair was kept alight by a group of vocal advocates. They subsequently provided evidence that convinced the rest of the surgical community that repair was safer than replacement both initially and in the long term after surgery, so that repair became the gold standard.²⁰ Unfortunately, throughout these changing times, there were no randomized controlled trials to confirm or refute the evidence of cohort studies and, as we shall see, it is still unclear whether repair or replacement is the most appropriate procedure for functional MR.²¹

The principal repair technique used to correct both type I and type IIIb components in functional MR is that of reduction annuloplasty. The surgical technique consists of sizing for an annuloplasty ring in the same way as in a patient with structural valve disease. When carrying out a repair in the face of functional MR, an annuloplasty two sizes smaller than the obturator or sizer is chosen. The prosthesis is then sutured in place in the usual way. This was first reported by Bolling and colleagues in a series of 16 patients, 12 with idiopathic dilated cardiomyopathy and 4 with irreversible ischaemic cardiomyopathy and all with a LVEF of less than 25% assessed by left ventricular angiography or radionuclide ventriculography. The early results in terms of symptomatology and left ventricular reverse remodelling were remarkably good.⁴ There were no operative deaths and 1-year actuarial survival was 75%. The patients showed marked symptomatic improvement and an improvement in left ventricular function. Subsequent authors have mirrored these findings.^22,23

However, in larger and later studies the benefits were not so clear cut. In a cohort of 126 patients, including both ischaemic and dilated cardiomyopathy, undergoing reduction mitral annuloplasty the 30-day mortality was 4.8%. Furthermore, surgery in this group of patients did not confer a survival advantage compared with a matched group who did not undergo surgery and in whom confounding clinical variables were controlled by propensity scoring. This finding held true whatever the aetiology of the HF.²⁴

One of the trade-offs between repair and replacement rests on the incidence of recurrence of regurgitation after repair. After valve replacement the incidence of regurgitation should approach zero. Since the life expectancy of patients with severe left ventricular function is relatively poor no matter what, they are unlikely to suffer the consequences of bioprosthetic valve degeneration so that the hazards of mechanical valve replacement can be avoided.²⁵ Valve repairs for structural valve disease are robust and reoperation rates for subsequent failure are low. The situation after reduction mitral annuloplasty for functional MR is not so clear. Certainly high rates of recurrent regurgitation have been found in patients with an ischaemic aetiology following annuloplasty,²⁶ with moderate or severe regurgitation present in 57% of patients at an average follow-up of 28 months after surgery despite satisfactory initial echocardiographic outcomes. Providing the initial annuloplasty is sufficiently restrictive and the prosthetic ring does not dehisce, the recurrence must be due to a type IIIb mechanism.

This is borne out in a study of a subgroup of 117 patients from the ACORN trial, all with idiopathic dilated cardiomyopathy. Of these patients 25 had at least moderate MR within 6 months of surgery.²⁷ Those patients with recurrent MR failed to show the improvement in left ventricular dimensions and function compared with the patients with competent mitral valves following surgery. Preoperative anterior leaflet tethering strongly correlated with recurrence of MR after surgical repair. After annuloplasty the posterior leaflet is fixed and passive and in effect is maximally restricted, so that it is the type IIIb motion of the anterior leaflet which becomes the predominant lesion in these patients. When that lesion exists prior to surgery then it comes to prominence postoperatively. There is little doubt that further developments in valve repair will be required to overcome this difficulty.

There exist a whole range of different designs of annuloplasty prostheses, including complete and incomplete rings of all degress of flexibility. While there is echocardiographic evidence that flexible rings result in better ventricular function following repair for structural valve disease,²⁸ there is no good evidence that this is reflected in clinical outcome.²⁹ Indeed, there is limited evidence that, at least in ischaemic aetiology, a complete rigid ring gives the best results for functional MR.³⁰

As to when to operate, there really is no evidence on which to base surgical practice. At present it would seem reasonable to exhaust all other therapies before advising any patient to go forward to what is a relatively high-risk procedure with a limited outcome. We need to see the results of the ongoing randomized clinical trials in the hope that these will provide some clues as to which patients might benefit most from correction of functional regurgitation.

Functional MR occurs commonly as part of the remodelling process resulting from left ventricular failure. Surgical correction of the regurgitation either by repair or replacement can be effective, although there are no randomized clinical trial results to substantiate this claim, and little evidence to guide timing of mitral surgery for functional regurgitation. Reduction annuloplasty is the most favoured means of surgical correction, although mitral valve replacement with preservation of all the subvalvular structure should not be discounted, especially if restrictive motion of the anterior mitral leaflet is present preoperatively.