50 Revascularization and remodelling surgery

There is a wide spectrum of clinical presentations of coronary heart disease ranging from mild asymptomatic single-vessel disease, through multivessel disease with mild or moderate left ventricular dysfunction, to severe endstage ischaemic cardiomyopathy. As the severity of left ventricular disease increases, so does the potential benefit of surgical intervention to the patient’s outcome. Unfortunately the clarity of the specific indications for coronary artery bypass grafting decreases. For example, in a patient with multivessel coronary artery disease and moderate left ventricular dysfunction, the indications for coronary artery bypass grafting (CABG) are clear and well-defined by multiple large-scale prospective and retrospective studies. In contrast, for a patient with severe ischaemic cardiomyopathy and potentially graftable coronary arteries who is being evaluated for a variety of treatment options including medical treatment, resynchronization therapy, percutaneous coronary intervention, coronary bypass grafting, valvular reconstruction, ventricular remodelling, mechanical assistance, and transplantation, the exact role of CABG is not so clear.

Nevertheless, there is persuasive experimental and clinical evidence that revascularization and perfusion of ischaemic or injured myocardium enhances cardiomyocyte integrity and contractility, thereby augmenting ventricular function. Thus the drive to revascularize is strong. In this high-risk population, the challenge is to identify those patients who will derive the most long-term benefit while incurring the least perioperative risk.

The pathophysiology of advanced heart failure (HF) has been extensively reviewed.^1,2 The common pathology in all cases of systolic HF is that of impaired myocyte contractility, activation of the renin–angiotensin–aldosterone and sympathoadrenal systems, leading subsequently to ventricular remodelling: myocyte hypertrophy, fibroblast proliferation, and fibrosis, and eventually to dilatation and increased sphericity of the ventricle. Specific pathological processes are present in ischaemic HF which are relevant to surgical treatment.

Ischaemic HF is brought about by two mechanisms: (1) impaired contractility of poorly perfused regions of myocardium supplied by stenotic coronary arteries, and (2) noncontractile or poorly contractile regions of scarred myocardium following myocardial infarction (MI). Impaired contractility of poorly perfused myocardium due to coronary artery disease was first described by Rahimtoola³ who coined the term ‘hibernating myocardium’. It was suggested that this occurred as a result of cardiomyocytes down-regulating their function and metabolism in the presence of hypoxia, as a means of surviving. More recently, it has been suggested that hibernating myocardium may be a result of repeated episodes of ischaemia causing myocardial stunning.⁴ Such ischaemic, hibernating myocardium has been shown to improve its function upon revascularization.

Ventricular dilatation occurs in all cases of HF. The ventricle dilates in an attempt to maintain stroke volume in the setting of decreased myocardial contractility. In the case of ischaemic HF, expansion of the infarct area in the early days of recovery following a moderate- to large-sized MI may be responsible for the initial stages of ventricular dilatation. In the longer term, adverse ventricular remodelling is responsible for continued ventricular dilatation.

Following MI, loss of contractile function in the infarct-related region of the ventricle increases wall stress in the remote myocardium. Myocytes in regions immediately adjacent to the infarct area also become dysfunctional with decreased contractility. Eccentric hypertrophy of myocytes occurs with extra intracellular sarcomeres (contractile proteins) laid down in series.⁵ This is thought to be the main mechanism of ventricular dilatation. Slippage of myocytes also occurs and contributes further to ventricular dilatation. This may be due to activation of matrix metalloproteinases (MMPs) which dissolve the intermyofibrillar collagen struts in the extracellular matrix. MMP activation may also be responsible for thinning of the infarcted segment.⁵ The result is that the ventricular chamber diameter and volume increases and the ventricle assumes a more spherical shape.⁵

The increase in ventricular chamber diameter results in a raised tension or stress at the ventricular wall to support any given intraventricular pressure. This occurs in accordance with Laplace’s law which states that T = P × R/2h (where T is circumferential ventricular wall stress, P is intraventricular pressure, R is radius of curvature, and h is wall thickness). The increase in ventricular wall stress has several adverse effects: (1) oxygen consumption is increased as the increased wall stress results in the ventricle facing an increased workload at the beginning of systole, (2) subendocardial perfusion is compromised, and (3) myocyte systolic shortening is impaired.⁶

The increase in ventricular diameter further compromises ventricular function. First, the increased ventricular size means that the ventricle operates on a flat portion of the Frank–Starling curve.⁵ Secondly, the more spherical shape of the dilated ventricle alters the orientation of myofibrils. Normally most myofibrils lie obliquely to the axis of the heart, such that a 15% shortening during systole achieves a 60% ejection fraction. In the dilated, spherical ventricle, Ingel and colleagues⁷ have shown that most of the myofibrils lie in a transverse direction to the axis of the heart and a 15% shortening during systole only achieves an ejection fraction of 30%. The problem is further compounded in areas of infarcted myocardium where nonfunctional myocytes and scar tissue exist. Such areas of akinesia or dyskinesia further compromise the efficiency of ventricular systole.

The increased ventricular wall stress is a potent stimulus for further myocyte hypertrophy. The elongation of myocytes by eccentric hypertrophy is out of proportion to an increase in its thickness by concentric hypertrophy so that ventricular wall stress continues to be significantly raised. Progressive adverse ventricular remodelling occurs as long as ventricular wall stress is raised, leading eventually to endstage HF.⁵

The aim of myocardial revascularization, whether by CABG or percutaneous coronary intervention (PCI), is: (1) to correct myocardial ischaemia and hence prevent further adverse ventricular remodelling and MIs, and (2) to improve myocardial contractility in regions of ischaemic hibernating myocardium that have been shown to be viable.

The determination of hibernating myocardium and viability can be made by stress echocardiography or by various nuclear imaging techniques such as PET, SPECT, or thallium or by cardiovascular magnetic resonance imaging (CMR) with gadolinium enhancement.⁸ It has been reported that the improvement in left ventricular function after CABG is related to the number of viable segments present; at least eight viable segments should be present to ensure an absolute improvement in ejection fraction of at least 5%.⁹ The early randomized trials of CABG versus medical treatment excluded patients with HF symptoms (NYHA class 〉II) and those with severe impairment of left ventricular function (ejection fraction 〈35% in the CASS study and 〈50% in the ECSS study).¹⁰ Subgroup analysis of 160 trial patients who had an ejection fraction of less than 50% and three-vessel coronary artery disease or proximal left main stem or proximal left anterior descending artery stenosis showed that the 10-year survival in those who had CABG was better compared to those who were treated with medication (79% vs 61%, p = 0.01). The survival advantage for CABG was present regardless of the severity of impairment of left ventricular function.

This survival advantage with CABG in patients with impaired left ventricular function has been consistently reported in the other early randomized trials.^11,12 No trial included patients with left ventricular ejection fraction (LVEF) less than 35%, but registry data from CASS¹³ and the Duke University Cardiovascular Database¹⁴ indicated that coronary bypass surgery carried a distinct survival advantage in patients with the worst ventricular function, most extensive coronary artery disease, and most severe angina. It should be emphasized that most of these patients presented with angina and not HF. Only 4% of the trial patients had HF symptoms and only 7.2% had an ejection fraction less than 40%. In addition, advances in the medical treatment of advanced HF in the last decade have improved survival significantly. The results of these early randomized studies may therefore not be applicable at the present day in a patient with advanced ischaemic HF, i.e. a patient with NYHA class III and IV HF symptoms and an ejection fraction less than 30%.

Several more recent nonrandomized studies have reported that CABG in patients with advanced ischaemic HF can be performed with an acceptable risk (operative mortality of 1.7–5.3%) and improves ejection fraction by up to 40% above the baseline value. The reported 5-year survival is 60–75%.^15,–19 The larger studies are summarized in Table 50.1. Unfortunately, complete data is not reported for many of these studies and the patient population is not uniform across the studies. Hibernation studies were performed preoperatively in some studies^17,18 but not in others.¹⁶ In addition, only 23–43% of patients had HF symptoms preoperatively. Accurate reporting of ventricular function, HF symptoms, and NYHA functional class at late follow-up is absent from most studies. This is important, as many of the benefits noted in the studies may not be sustained at late follow-up. Recurrence of HF symptoms is reported in 53% of patients at 5 years by Luciani.¹⁷ However, only 48% of patients in this study had hibernation studies preoperatively. More favourable results are reported by Lorusso’s group¹⁸ which performed hibernation studies in all patients (18% recurrence of HF symptoms at 4 years, 40% at 8 years). Interestingly, the same study reports that left ventricular function, although improving significantly immediately postoperatively (ejection fraction 40 ± 2% compared to 28 ± 9%, p 〈 0.01), subsequently fell at late follow-up and was only marginally better compared to preoperatively (ejection fraction 30 ± 9% compared with 28 ± 9%). This decline in left ventricular function at late follow-up is also reflected in the NYHA functional class (35% in NYHA class III and IV at 8 years compared to 24% immediately postoperatively).

There is some evidence that recurrence of HF symptoms after CABG for ischaemic HF may be related to the severity of ventricular dilatation. Yamaguchi et al.²⁰ reported a recurrence of HF following CABG in 69% of patients when the left ventricular end-systolic volume index (LVESVI) was greater than 100 mL/m², compared to only 15% when the LVESVI was less than 100 mL/m² (p 〈 0.01). Five-year survival was also worse when LVESVI was greater than 100 mL/m² (53.5% vs 85%; p 〈 0.01). Similarly, Louie et al.²¹ reported a failure of CABG in 27% of patients with ischaemic cardiomyopathy undergoing CABG for HF symptoms, all of whom had a left ventricle that was significantly more dilated compared with those in whom CABG was successful (left ventricular end-diastolic diameter of 81 mm vs 68 mm). In these patients, it may be necessary to perform some form of ventricular restoration surgery in addition to CABG.

Diabetes mellitus in patients with coronary artery disease is associated with a poor outcome. Evidence from the SAVE trial²² showed that diabetic patients were more prone to HF after MI. But those with diabetes exhibited less left ventricular cavity dilation than control patients without diabetes.²³ In a prospective study of 129 patients (31 diabetic, 98 nondiabetic), Rizello and colleagues²⁴ assessed myocardial viability before and after myocardial revascularization. LVEF increased in 44% of diabetic and in 40% of nondiabetic patients. LVEF only improved in patients with viable myocardium. Indeed, viability was the only predictor of both early (30 days) and late (5 years) survival.

There has been a widespread reluctance among HF physicians to refer patients with HF for higher-risk coronary surgery, not only because of the lack of a high level of evidence but because of the difficulty in predicting outcome in an individual patient. A recent prospective study from Japan by Mizuno and coworkers²⁵ studied 31 diabetic and 33 nondiabetic patients with ischaemic cardiomyopathy before and after surgical revascularization. At 6 months after revascularization, subepicardial perfusion was markedly improved in both populations. In contrast, subendocardial perfusion markedly improved only in the nondiabetic patients and was little changed in the diabetic patients. Improvement in left ventricular function was greater in nondiabetics and persistent HF was found more often in the diabetic patients. Diabetes appears to be an important clinical modifier of the remodelling process. This is one of the first studies to demonstrate the intramural heterogeneity of recovery of myocardial perfusion and its relation to persistent HF after surgical revascularization.

Nonrandomized studies have reported that CABG in patients with advanced ischaemic HF can be performed with an acceptable risk and that it improves ejection fraction and NYHA functional class. There is concern that these improvements may not be sustained in the long term and many of these patients have a recurrence of HF symptoms with deterioration in ventricular function within 5 years. Patient selection is crucial and those with severely dilated ventricles may do less well with CABG alone. Of interest, is not only the influence of CABG on long-term survival in patients with advanced ischaemic HF, but also its impact on functional capacity, quality of life, heart function, and whether any benefits are sustained in the long term. Other factors in patient selection may also be important such as the presence of good target coronary vessels for revascularization, complete revascularization, absence of right HF, or raised pulmonary artery pressures.^17,26 In cases where the left ventricle is significantly dilated, e.g. above a LVESVI of 100 mL/m², some sort of ventricular restoration surgery may be necessary in addition to myocardial revascularization.

The aim of ventricular restoration is to restore the size, shape, and geometry of the dilated left ventricle towards normal. Restoration of ventricular size reverses many of the pathophysiological processes described earlier by decreasing ventricular wall stress.²⁷ This in turn enhances myocardial perfusion, decreases oxygen consumption, and enables improved contractility of myocytes. Restoration of ventricular shape and geometry towards a more elliptical structure also leads to greater efficiency of ventricular systole as previously described. Ventricular restoration is achieved by resection of myocardium and reconstructing the remaining ventricle into a more elliptical shape. Several different techniques for ventricular restoration can be used depending on the underlying cause of the cardiomyopathy.

As Buckberg²⁸ has pointed out, there are certain specific questions that need to be asked before surgical remodelling occurs:

Ventricular restoration in ischaemic cardiomyopathy, also referred to as ventricular restoration surgery (VRS), is most commonly performed using the Dor procedure.²⁹ The modified linear closure technique described by Mickleborough³⁰ is also sometimes used. The Dor procedure was initially described for the resection of left ventricular aneurysms and was later modified for use in ischaemic cardiomyopathy. Both techniques involve resection of the akinetic or dyskinetic anterior free wall of the left ventricle (Fig. 50.1). Typically, these patients have had an anterior MI with scarring and akinesia or dyskinesia of the left ventricle anterior free wall which may extend on to the septum. This segment of nonfunctional myocardium is resected. In order to reshape the ventricle an endoventricular suture is placed to reduce the size of the defect. This judgement, which will determine the final stroke volume of the left ventricle, can be guided by the use of a balloon of known volume. In the Dor procedure, an oval Dacron patch is then placed which excludes the infarcted part of the septum from the rest of the ventricle. The size of the patch is tailored to the required size of the ventricle, and the shape of the patch is tailored such that it helps restore the geometry of the left ventricle towards a more elliptical configuration (Fig. 50.1b, c). VRS adds about 20 minutes to the duration of an operation for CABG and has not been found to increase the operative risk.³¹ CABG is always performed at the same time. The aim is to (1) recruit hibernating myocardium and hence enhance myocardial contractility; (2) resect nonfunctional akinetic or dyskinetic myocardium and hence improve the efficiency of ventricular contraction; and (3) restore the left ventricle to its normal size, shape, and geometry with the benefits discussed previously. In a nonrandomized study involving 814 patients, Dor³² reported an increase in LVEF from 22% to 38%, and an increase in cardiac index from 1.7 to 2.5 L/min/m² following surgery. The operative mortality was 6.6%. Survival at 10 years was 80% in those with an ejection fraction of 30–40% preoperatively and 60% in those with an ejection fraction less than 30%.

Numerous nonrandomized studies have been reported.^33,–36 The three larger studies with more than 200 patients each are summarized in Table 50.2. These studies report a hospital mortality of 2.8%-8.1%, an absolute improvement in ejection fraction of 10–13% above baseline (or a relative improvement of up to 40% above baseline), and an improvement in NYHA functional class. The benefits of VRS appear to be sustained. In the RESTORE trial Athanasuleas and coworkers³³ examined how VRS affected early and late survival in postanterior infarction CHF patients. The investigators applied VRS to 1198 postinfarction patients between 1998 and 2008. Anteroseptal, apical, and anterolateral left ventricular scarred segments were identified and excluded by an intracardiac patch. Concomitant procedures included CABG in 95% of patients, mitral valve repair in 22%, and mitral valve replacement in 1%. Overall 30-day mortality after SVR was 5.3% (8.7% with mitral repair vs 4.0% without repair, p 〈 0.001). Perioperative mechanical support was uncommon (〈9%). Global systolic function improved postoperatively. Ejection fraction increased from 29.6% ± 11.0% preoperatively to 39.55 ± 12.3% postoperatively (p 〈 0.001). The LVESVI decreased from 80.4 ± 51.4 preoperatively to 56.6 ± 34.3 mL/min postoperatively (p 〈 0.001). Overall 5-year survival was 68.6% ± 2.8%. Logistic regression analysis identified LVEF of 〈30%, LVESVI of at least 80 mL/min², advanced NYHA functional class, and age of at least 75 years as risk factors for death. In this study, 85% of patients were free of congestive HF symptoms at 18 months. Five-year freedom from hospital readmission for CHF was 78%. Similarly, Di Donato³⁴ reported an event-free survival of 82.1% at 3 years. Actuarial survival was 89.2% at 18 months in the RESTORE study and 74% at 3 years in Di Donato’s study.

The results of these nonrandomized studies are encouraging. They suggest that VRS in these very sick patients can be performed with acceptable hospital mortality, and improves ejection fraction and functional capacity. The early results of up to 3 years suggest that the improvement in congestive HF symptoms is sustained. The 3-year actuarial survival also appears impressive considering the patient population. Clearly, patient selection is important. The ideal patient may be one who has had an anterior MI with akinesia or dyskinesia of the left ventricle anterior free wall, dilatation of the left ventricle, good target coronary vessels which can be grafted, and viable hibernating myocardium. The results of VRS may not be as good in patients who have pathology outside the left ventricle such as right ventricular failure or raised pulmonary artery pressures.

The need for a randomized trial has been widely recognized and the design much discussed.³⁷ The Hypothesis 2 substudy of the Surgical Treatment for Ischemic Heart Failure (STICH) has recently been reported by Jones et al.³⁸ This substudy compared CABG alone with the combined procedure of CABG with surgical ventricular reconstruction. Eligible patients were required to have coronary artery disease amenable to CABG, a LVEF of 35% or less, and a dominant anterior region of myocardial akinesia or dyskinesia that was amenable to surgical ventricular reconstruction. All patients received standard medical and device treatment for HF. In total, 1000 patients were recruited from 96 medical centres in 23 countries. The patients in the two study groups were closely matched for demographic characteristics, comorbidity, the proportion who were on HF drugs, CCS angina class, NYHA class, coronary anatomy, and the extent of anterior myocardial akinesia or dyskinesia. Both groups of patients were equally successful in improving the postoperative CC angina and NYHA class. There was similar improvement in the six-minute walk test and similar reductions in symptoms. As one would expect, there was a greater reduction in the end-systolic volume index with the combined procedure (16 mL/m² of body surface area), as compared with CABG alone (5 mL/m²). Unfortunately these data were obtained from only 373 patients at baseline and at 4 months.

The primary outcome of the trial was a composite of death from any cause or hospitalization for cardiac causes. There was no difference in the occurrence of the primary outcome between the CABG group (59%) and the combined procedure group (58%). The 30-day surgical rates of death for CABG alone (5%) and for the combined procedure (6%) were similar and low overall, and no difference in the rate of death from any cause was observed in a mean follow-up period of 48 months.

On the basis of this trial, Eisen³⁹ stated in an editorial that the routine use of surgical ventricular reconstruction in addition to CABG cannot be justified. There may be specific subgroups of patients who might benefit from the combined procedure, but such an effect is not apparent so far in the results of the STICH trial and may be difficult to detect, given the heterogeneity of the study population.

There were several major problems with the conduct of this trial.⁴⁰ Myocardial viability was assessed in only 20% of the patients, so it is not clear whether the study was examining the treatment of scar tissue or hibernating myocardium. The conduct of the surgery also raises questions: 501 SVR procedures were performed in 127 sites over 5 years, an average of 0.7 SVR operations per site per year. Were steps taken to assure the eligibility of the surgeons and effectiveness of the units? The ESVI was reduced by only 19% in STICH, which compares unfavourably with a 36% reduction in the RESTORE study. In 41% of patients undergoing the Dor procedure in STICH a Dacron patch was not used, whereas in the RESTORE studies all patients received a patch. This raises the question of whether operative procedures were inadequate in a large proportion of the patients randomized to the Dor procedure. Despite the caution of Eisen,³⁹ it will be important to examine the subgroups in detail to identify those patients who may benefit from the Dor procedure.

Effective protection in these operations can be challenging but is essential to prevent the vulnerable left ventricular endocardium from undergoing subendocardial necrosis. A review of protection during the Dor procedure in the 1198-patient RESTORE study⁴¹ showed a fairly even distribution between cold blood cardioplegia and continuous perfusion in the beating heart on cardiopulmonary bypass. The beating heart method was used more frequently in older patients and those with the lowest ejection fractions, larger left ventricular volumes, and more advanced HF. There was a tendency, which did not achieve statistical significance, towards a higher 5-year survival in this subset.

The safety of the beating heart method was shown in acute studies. Preferential subendocardial blood flow occurred during perfusion in chronically dilated hearts, whereas cardioplegia caused diminished subendocardial perfusion. Higher perfusion pressures are required during blood cardioplegia or beating heart methods because of the vascular remodelling which occurs in the coronary bed of failing versus normal hearts.

Flexibility is a critical factor in planning protection strategies in failing hearts.

New treatments such as the Acorn CoCap and Paracor cardiac support device (CSD) have been evaluated with respect to their usefulness in limiting adverse ventricular remodelling. Dynamic cardiomyoplasty was the precursor of passive prosthetic ventricular support. Unfortunately the results from animal and clinical studies were inconsistent and limited, despite frequently observed clinical benefit. In a canine model of chronic dilated cardiomyopathy, Patel and colleagues⁴² suggested that the haemodynamic benefit of cardiomyoplasty was due to the passive effect of the skeletal muscle wrap around the heart. The relief of wall stress produced by girdling of the conditioned muscle wrap was shown to stabilize the remodelling process of HF, preventing progressive deterioration of systolic and diastolic function. These studies led to the development of a device that would relieve wall stress, similar to the skeletal muscle wrap.

The device that has been studied most extensively is the CorCap CSD (Acorn Cardiovascular Inc., St. Paul, Minnesota, USA) It is a multifilament polyester mesh implant which is placed around both ventricles to decrease diastolic wall stress without resultant constriction (Fig. 50.2). Mann and colleagues⁴³ assessed the safety and efficacy of the CSD in 300 patients with HF. Of the 300 patients enrolled, 193 were randomized to mitral surgery alone or mitral surgery plus CSD. The 107 patients who did not need mitral surgery were randomized to medical treatment or medical treatment plus CSD. The primary endpoint was a composite based on changes in clinical status, the need for major cardiac procedures for worsening HF, and a change in NYHA class. All patients had an LVEF of less than 35%, a LVEDD of 60 mm or greater, and a six-minute walk test of less than 450 m. The proportional odds ratio for the primary endpoint favoured treatment with the CSD (1.73; 95% CI 1.07–2.79; p = 0.024). When compared with the baseline, LVEF increased significantly at 12 months (p = 0.0009) in the CSD-treated group compared with controls (p = 0.65), but the changes in LVEF between groups were not significant (p = 0.45). Therefore the CorCap CSD may have a role in preventing adverse remodelling after MI. It requires an operation for its insertion but this could be through a small anterior thoracotomy.

A more recent development, the Paracor device, is currently being assessed in clinical trials in Europe and the United States. This is an elastic nitinol mesh that is designed to mechanically reinforce the heart to retard or hopefully halt the remodelling process. It can be deployed in a minimally invasive fashion. Klodell and coworkers have reported their early results.⁴⁴ Fifty patients in NYHA class II or III underwent the procedure, which was well tolerated. At 6 months there was a significant improvement in the six-minute walk (+65.7 m, p = 0.002) and Minnesota Living with Heart Failure scores (–15.7, p = 0.002). Long-term functional results are not yet available.

Recent advances in medical treatment have greatly improved symptom control and survival in HF, but morbidity and mortality continue to remain significant especially in the advanced stages of HF. This may be because medical treatment alone neither corrects the cause, nor reverses all of the pathophysiological changes that occur in advanced HF, especially those related to ventricular dilatation and alteration in the geometry of the ventricle. It is likely that significant numbers of patients with advanced HF, especially that due to ischaemic heart disease, will benefit from surgery. A combination of surgical treatments may be necessary depending on the pathological changes present.

Clinical studies on revascularization of ischaemic, hibernating myocardium are promising but more randomized studies are needed to confirm its long-term efficacy. Similarly, large non randomized studies have shown that VRS combined with CABG in ischaemic cardiomyopathy improves cardiac function and functional capacity and may improve survival. The results of the STICH trial were disappointing, but the design of the trial has been heavily criticized.

Advanced HF is a complex disease leading to many different pathologies and the ideal treatment for each patient may need to be tailored individually. Some patients may need a combination of different treatments. The armamentarium is enlarging and the future appears promising for these very sick patients.