Heart Failure Reviews

, Volume 18, Issue 1, pp 15–25

Surgical approaches to left ventricular reconstruction: a matter of perspective


    • Department of Cardiothoracic SurgeryJena University Hospital – Friedrich Schiller University Jena

DOI: 10.1007/s10741-011-9296-5

Cite this article as:
Doenst, T. Heart Fail Rev (2013) 18: 15. doi:10.1007/s10741-011-9296-5


Surgical reconstruction of physiological shape and size of a postischemically remodeled left ventricle has been advocated to improve ventricular function and improve patient long-term outcome. What initially started as linear aneurysm resection surgery developed over the years into the endoventricular repair techniques (surgical ventricular reconstruction, SVR) that have also been applied in patients with postischemically dilated ventricles and mainly anterior akinesia. SVR improved function as measured by the ejection fraction. Whether it affects survival was finally tested in the largest surgical trial ever conducted, the STICH trial (Surgical Treatment for IsChemic Heart failure). The trial, however, presented rather sobering information with its Hypothesis 2 outcome by demonstrating identical 5-year survival rates between SVR plus bypass grafting (CABG) and CABG alone. SVR also did not improve quality of life. This neutral finding spawned a series of critical responses with respect to trial design and conduct accompanied by appropriate responses by the trial’s leadership. At the end of this dispute, it appears that SVR has been accepted as not very useful for most patients and is less and less performed in daily practice. What remains is a series of different perspectives that will be discussed in this review. The conclusion will be that SVR may be of low value for the patient with dilated and massively remodeled ventricles, but the technique bears therapeutic potential for some patients for different reasons so that the surgeon’s ability to perform this operation should not get lost.


Cardiac surgeryDor plastyAneurysm resectionIschemic heart failure


Surgical correction of postischemically dilated and functionally impaired left ventricles has been performed for more than 50 years [1, 2]. The need to develop surgical treatment options was based on the dismal natural course of patients developing postinfarct left ventricular (mainly anterior) aneurysms [3, 4]. Over the years, modifications of the initial, Cooley-type, linear closure technique [1] have been described and the indication was extended to patients with dilated ventricles and akinetic rather than dyskinetic anterior walls [5]. This extension in indication was mainly induced by the work of Dor and colleagues [68]. The “Dor procedure” as infarct exclusion, possibly endocardial resection and endoventricular patch implantation, has gained the greatest acceptance [9, 10].

Outcomes of the different techniques have been abundantly reported, but since the patient population undergoing these procedures also changed over the years, results are difficult to compare. An attempt of comparison is made in Fig. 1. The rationale behind this figure was based on the thought that improvements in technique should result in improved short- and/or long-term outcomes. Figure 1a shows the perioperative mortality of all studies reporting outcomes after ventricular reconstruction surgery as a feature of time. There was no trend toward lower mortality that would allow suggesting that operative technique had become safer. However, over time, patients undergoing this procedure may have undergone surgery with more comorbidities, and any improvement in surgical safety may have been overcome by this increase in “risky” comorbidites. Figure 1b shows the reported 5-year mortality rates. It is interesting to note that they all range around the 70% mark irrespective of the time when the study was performed or the technique that was used. It is then striking to realize that, with the advances in medical therapy over the years, treatment for heart failure patients with beta-blockers and ACE inhibitors can achieve the same long-term outcome [1114].
Fig. 1

Perioperative mortality (a) and 5-year survival (b) of ventricular reconstruction surgery (irrespective of surgical technique) as reported in the literature. The number on the graph reflects the reference

It is clear that the comparisons made in Figure 1 have countless confounding factors as has been the case with all reports on this surgical technique. The desperately needed evidence was finally generated by the STICH trial [15]. Here, 2,136 patients with ischemic heart failure and an ejection fraction below 35% were randomized to test two hypotheses. The first one assessed whether in patients with systolic, ischemic heart failure (EF < 35%), CABG provides a survival advantage over medical therapy alone. The results of this hypothesis were recently presented [16] and are not subject of this review. The second hypothesis tested whether SVR added to CABG provided a survival advantage free of re-hospitalization compared to CABG alone [15]. One thousand of the 2,136 patients of the trial were randomized for this comparison. The outcome was sobering to cardiac surgeons. The 5-year survival was identical in the two groups. The survival curves are practically superimposed on each other (Fig. 2). In addition, length of ICU stay was longer, hospital cost was higher, and there was no difference in long-term quality of life [17].
Fig. 2

Kaplan–Meier estimates of STICH Hypothesis 2 outcomes. a The probability of the primary outcome (death from any cause or hospitalization for cardiac causes), which did not differ significantly between the two groups. The primary outcome occurred in 292 patients (59%) assigned to undergo coronary artery bypass grafting (CABG) alone and in 289 patients (58%) assigned to undergo CABG with surgical ventricular reconstruction (SVR) (hazard ratio 0.99; 95% CI 0.84–1.17). b The probability of death from any cause, which occurred in 141 patients (28%) assigned to undergo CABG and in 138 patients (28%) assigned to undergo CABG with SVR (hazard ratio 1.00; 95% CI 0.79–1.26). (Reproduced with permission from Jones et al. NEJM 2009)

This neutral outcome led to heavy criticism [18] regarding trial design and conduct, with appropriate responses of the trial’s leadership to follow [19]. Details are described in the perspectives below. Irrespective of the criticisms, the results of this large prospective randomized trial are there and they signal the end of SVR as we know it. However, surgeons today are left with a certain bad aftertaste because SVR may still be a valuable treatment option for some patients, specifically those that would not have fulfilled the STICH trial’s inclusion criteria (imagine a patient with a large anterior aneurysm but an ejection fraction of still 45%). For these and other reasons, it is important to point to these specific conditions and describe the reasons why it may be useful to keep SVR in the surgical armamentarium of the future.

In the three perspectives below, I will therefore address the different conditions, reasons, and criticisms in this area and hope to provide the physician caring for these patients (may it be surgical or non-surgical) with important information enabling them to judge whether SVR may still be a reasonable option for their patients after the STICH trial. In order to understand the perspectives and the controversy regarding therapeutic potential and differences in the surgical techniques, a short review of some basic principles of the underlying disease, i.e., ischemic remodeling is in order.

Terminology, pathomechanisms, and surgical technique

The term “remodeling” was introduced first by Pfeffer et al. in 1985 when the investigators noted a significant morphological response in terms of hypertrophy and dilatation of the remote muscle to regional ischemia in the rat (LAD occlusion) [20, 21]. They defined “remodeling” as global changes to myocardial shape, size, and geometry in response to local ischemia. The underlying mechanisms have since been intensely investigated, but the true mechanism is still at large. In general, inflammatory responses, neurohormonal activation, changes in the extracellular matrix, and hemodynamic load have been discussed [2224].

With the observation that beta-blockers and ACE inhibitors can result in the reduction in ventricular volume and the reversal of some of these remodeling responses, the term “reverse remodeling” was coined [1114].

In the surgical field, efforts to reshape the ventricle have been made for long and are in the surgical literature sometimes referred to as “surgical remodeling” [25]. However, surgical remodeling should in theory be more closely associated with “reverse remodeling” (as defined from the mechanistic basic science perspective) than with the originally described, rather detrimental “postischemic remodeling.” For this reason, I suggest to use the term “ventricular reconstruction” for the surgical activities to reshape the left ventricle, giving a possibly clearer, but certainly goal-guided terminology for these surgical procedures.

Figure 3 shows a simplified scheme of the most commonly propagated technique for surgical ventricular reconstruction [6]. The a- or dyskinetic anterior wall is incised laterally to the left anterior descending artery. The ischemic border zone is to be identified either by inspection or by palpation (if the procedure is performed on the beating heart). And a purse string suture is placed along the border zone. It is then recommended to tie the suture over a shaping device to assure proper shape and size. The remaining opening is then closed by a Dacron patch implanted with a running suture. Depending on the amount of excluded myocardium, the remaining tissue is closed over the circular or oval-shaped patch in linear fashion.
Fig. 3

Schematic illustration of surgical ventricular reconstruction (SVR). The a- or dyskinetic anterior wall is incised laterally to the left anterior descending artery. The ischemic border zone is to be identified either by inspection or by palpation (if the procedure is performed on the beating heart). A purse string suture is placed along the border zone. It is then recommended to tie the suture over a shaping device to assure proper shape and size. The remaining opening is then closed by a Dacron patch implanted with a running suture. Depending on the amount of excluded myocardium, the remaining tissue is closed over the circular or oval-shaped patch in linear fashion

With the terminology and the basic surgical technique clarified, we can now examine the different perspectives.

SVR: a basic science perspective

A basic science perspective always contains the desire to understand the underlying mechanism of disease. So, why does a heart remodel and what makes us believe that SVR should stop or reverse this process?

As addressed above, remodeling is a process that affects the cellular level and the extracellular matrix. Surgical reshaping would then have to interrupt pathways that support the remodeling process and/or stimulate those that counteract it. Based on this rationale and the principle basic mechanisms described above, there are two ways SVR may work: first, by a reduction in wall stress and second, by affecting ventricular geometry in a “beneficial manner.” The latter, rather unspecific aspect, includes the procedure’s effect on mitral valve function, specifically mitral regurgitation.

With respect to wall stress, it is readily accepted that the reduction in radius will result in less ventricular wall tension, which should result in more efficient pump function since wall tension correlates with oxygen consumption [26]. However, the results from the Batista procedure may argue against this line of reasoning since outcome was not improved by the Batista-type volume reduction [27]. However, the principle wall tension-reducing effect of this procedure as assessed by pressure volume loops was demonstrated by the Batista group [28] in a selected set of patients. However, any such improvements may have been overcome by significant diastolic dysfunction brought about by the massive scar lining the ventricle after this extensive muscle resection [29]. After all, the Batista procedure has almost completely disappeared from clinical practice. In any case, a different technique may still utilize the potential of reducing wall tension by volume reduction. However, the question whether reducing wall tension actually stops remodeling has still not been answered.

The second aspect in this context is ventricular geometry. A recent review demonstrated superior outcomes of circular reconstruction techniques compared to linear closures [30]. While it is not clear why these results are different, they may relate to the suggestion that surgical ventricular reconstruction or restoration may be able to reestablish “physiological” myofiber orientation. Buckberg et al. [18] have promoted the Torrent-Guasp concept [31] that myocardial dilatation is accompanied by changes of myofiber alignment in the heart. The concept suggests that the heart is made from one tube of musculature that is oriented in a helical way in healthy hearts and becomes more spherical in dilated hearts. SVR procedures are then to reestablish helical myofiber orientation during ventricular reshaping [18]. As convincing as this concept has been promoted, as poor is our actual ability assessing its relevance. Thus far, it is impossible to image myofiber orientation in vivo. In addition, one may pose the question how a procedure, such as SVR that is applied to the apical part of a dilated heart is supposed to reorient fiber alignment at the basal area and at the cellular level.

The same argument applies to the suggestion that SVR increases synchronicity of contraction. The CRT device literature provides an overwhelming amount of data on the life-extending effect of cardiac resynchronization [32, 33]. If increasing synchronicity by SVR would be a relevant mechanism, the effect should be seen clinically.

More evidence-oriented investigations on the impact of SVR procedures on cardiac functional parameters have been performed in the animal laboratory or by cardiac magnetic resonance imaging. Nicolosi et al. [29] simulated an anterior aneurysm in pigs by implanting a 6–8 cm Dacron patch into the anterior wall. They measured regional work and diastolic function and compared linear and circular repair techniques with each other when they removed the Dacron patch again. The authors found no differences in contractile function between the two repair techniques after patch removal. However, they demonstrated significant impairment of diastolic function, which they attributed to the generated scar. We demonstrated in a small series of patients having received cardiac magnetic resonance imaging before and after SVR no impact of any of several geometric parameters on functional recovery [34]. Thus, the impact of surgically reconstructing proper geometry remains unclear as does the ability of any of these techniques to stop remodeling.

Another argument that was made for the establishment of physiological size and shape by SVR was made by Klein et al. [30] stating that proper volume reduction and shape should stop remodeling. However, both Menicanti and Dor, two of the biggest protagonists of this type of operation, show significant enlargement after adequate volume reduction. The Menicanti group shows significant re-enlargement of patients having undergone SVR with aggressive ventricular volume reduction (roughly 40%) [35, 36]. Dor et al. [5] present a series of patients having received SVR for either a- or dyskinetic myocardium. Both patient groups (appox. 50 patients each) presented with a roughly 20% re-increase in ventricular enddiastolic volume 1 year after SVR that initially achieved a 40–50% volume reduction.

Finally, mitral regurgitation is often present in these patients, affects ventricular geometry and may be altered by SVR. The mere presence of mitral regurgitation is an indicator for poorer prognosis [3739]. Unfortunately, current evidence does not clearly support an improvement of survival by mitral valve repair [40]. From a geometric perspective, the presence of mitral regurgitation creates a stimulus for ventricular dilatation. The suggestion to repair the valve either by SVR or as added procedure to SVR is therefore only logical. The reduction in MR may therefore reduce ventricular volume in the long run and also acutely abort a dilatatory stimulus. Thus, correction of mitral regurgitation may support beneficial effects of SVR, but also potentially camouflages the true effect of SVR on ventricular function and remodeling.

Thus, the issue of being able to stop the remodeling process surgically has not been resolved and the controversy is likely to continue for several reasons. First, not every patient after infarction remodels [41]. Second, the remodeling process is not continuous in all patients. Third, some patients develop massive anterior aneurysms without a tendency of ventricular enlargement for years to come, and fourth, the role of mitral regurgitation in this context is not fully clear. From an experience-based standpoint, taking the discussed basic mechanisms of remodeling into account, it appears reasonable to assume that patients undergoing SVR, while they experience remodeling of their ventricle will continue to do so after surgery (those may be the ones that redilate) and patients undergoing SVR in the absence of active remodeling also keep their new volume and shape—but this is just a speculation at this point.

Which parameter assesses the true impact of SVR?

One key difficulty in assessing the efficacy of any SVR technique is, except for survival, the lack of proper parameters to measure the impact of this surgical procedure. Thus far, almost all studies have assessed the efficacy of SVR based on the changes in ejection fraction and volume. Survival curves are often demonstrated (see also Fig. 1), but except for the STICH trial, a proper comparative group is usually missing. Thus, an increase in ejection fraction and a reduction in volume are considered a marker for a proper functioning of the procedure. However, the value of the EF to assess SVR efficacy may be called into question. Consider a patient with an ejection fraction of 30% and an end-diastolic LV volume of 200 ml and an anterior wall akinesia or dyskinesia of more than 35% as advocated by Dor and Menicanti et al. [5, 36]. If we assume that the SVR procedure reduces volume to 150 ml by only excluding non-contractile myocardium, and also assuming that contractility of the remaining myocardium is also not changed, stroke volume should remain the same or should even increase. Before reduction, stroke volume in our patient was 60 ml at an EF of 30%. The same stroke volume would then result in an ejection fraction of 40% after SVR. Thus, the improvement in EF may be considered a mathematical necessity. In other words, if a patient does not demonstrate an increase in ejection fraction that is directly related to the reduction in LV volume, one may argue that contractility has been negatively impacted by the procedure. This exemplary assumption finds support from clinical investigations where these exact numbers were generated by Batista-type volume reductions and assessment of EF by echo and pressure volume loops [42].

It is interesting to note in this context that most prognostic predictions in heart failure and SVR patient populations have thus far depended on preoperative EF and/or volume [43]. The RESTORE group published their results indicating that the larger ventricles have worse outcome [6]. If the volume-reducing effect improves survival, the postoperative volume should reflect the new prognostic perspective. However, these analyses have not been performed.

In summary, from a basic science experimental perspective, SVR may provide clinical benefit by reducing wall tension, increasing synchronicity, reducing mitral regurgitation, and improving efficiency of cardiac work. Some thoughts and ideas on SVR mechanisms are not fully plausible. In addition, proper functional assessment requires volume-independent parameters that are difficult to measure in patients.

SVR: the clinical perspective

From a clinical perspective, epidemiologic outcome studies are needed because real life shows that clinical practice often speaks a different language than one may expect from a basic science point of view or from the animal laboratory. Many of the studies addressing SVR in the clinical setting have already been mentioned above. With the conflicting results from the different observational and retrospective database studies on SVR, a prospective randomized trial was needed. It was performed as Hypothesis 2 of the STICH trial over the last decade and finally presented its primary outcome in 2009 [15].

Figure 2 displays the survival curves of the two groups in hypothesis 2 as shown in the original publication of the STICH trial results [15]. As already addressed in the introduction, 5-year survival was identical in the two groups. The survival curves are practically superimposed on each other. In addition, perioperative mortality tended to be slightly higher (n.s.) in the group receiving CABG + SVR and more patients were in postoperative low output syndrome receiving intra-aortic balloon counterpulsation. Furthermore, length of ICU stay was longer, hospital cost was higher, and there was no difference in postoperative quality of life [17]. This outcome was sobering to many surgeons, and in the aftermath of the primary outcome presentation, heavy criticism was issued [18].

The critics argued that initially clearly defined inclusion criteria were later loosened so that inappropriate patients were included [18]. The volume data from the STICH trial demonstrate that patients with large ventricles and significant anterior mainly a- but also some dyskinesia were included [19]. According to the classification set fourth by the Milano group around DiDonato and Menicanti, these patients seem to belong mainly to the shape classified as type 3 [35]. The classification describes type 1 as the classic anterior aneurysm with maintained contractility in the posterior and basal portion of the ventricle. Type 2 is the intermediate type between 1 and 3 and is characterized by less contractility basally and posteriorly. Type 3 finally describes a largely dilated ventricle with anterior scar including the apex and septum with limited contractility in the remaining myocardium (see [35] for details). Di Donato et al. report in their series of patients a 30-day mortality of 5.3% (type 1), 12.7% (type 2), and 5.9% (type 3) [35]. Survival at 5-years survival was 69, 60, and 51% in type 1, 2, and 3, respectively. The RESTORE database reports operative mortality of 5.3% and a 5-year survival of 69% [6]. Considering that the STICH trial included mainly type 3 and the Restore database contained mainly type 1, these numbers are practically identical to those of the two STICH arms with perioperative mortalities of 4–6% and 5-year survival of roughly 60% [15].

The next criticism, which is intimately linked to the previous one, was that the volume reduction achieved by the STICH SVR surgeons was not enough [18]. While it may be true that a 19% lower volume at 4 months in light of a 7% lower volume in the CABG group without SVR appears small, the criticism may change if data are put into perspective. First, if SVR was not performed appropriately, the impact of SVR on survival would still be questionable because 5-year survival between the STICH trial and the database studies ([6, 35], Fig. 2) is identical. In addition, operative mortality was low and again identical to the other studies, so it is safe to argue that the surgery did at least cause no additional harm. Second, the time point of volume assessment was 4 months after surgery in the STICH trial and mainly before discharge in the other studies. Thus, it may be possible that the patients after SVR had already redilated underestimating the initial volume-reducing effect. Menicanti et al. [36] support the latter argument. They report in their series of 1161 patients with SVR an overall volume-reducing effect of 32 percent assessed within the first 10 days after surgery. A late follow-up between 6 and 24 months on two-thirds of the patients (the ones that died during follow-up are likely the ones with the larger ventricles, [36]) shows an increase in volume again by 18%. If one assessed the volume-reducing effect of SVR by comparing the preop value with the late postop one, the total reduction is only 21%, again the same order of magnitude as the STICH trial [15].

Finally, it was criticized that viability studies were not performed [18]. We have performed our own analysis on a small series of patients having received CMR before and after SVR to look for volume-independent parameters of SVR efficacy [34]. We were able to demonstrate that the amount of scar tissue in the basal portion of the ventricle may be a predictor for poor recovery of post-SVR function. However, we also assessed function by EF, and the study was too small to assess a survival impact. In addition, very recent evidence from the hypothesis 1 part of the STICH trial demonstrates that viability testing in ischemic heart failure patients is able to predict prognosis but is not useful for decision making if the question of revascularization is to be answered [44]. It would appear unlikely that decision making for SVR based on viability would change all that much.

If one, despite these convincing arguments against the STICH criticisms, still wants to entertain the thought that inappropriate patients were included in the trial we, as surgeons, need to look for the blame among ourselves. The trial was designed by surgeons and led by surgeons. The fraction of patients having been randomized in the trial with pure aneurysms was low (Oh et al.; late-breaking sessions of the American Heart Association, 2009). Those were the ones having the best long-term survival in the paper by DiDonato et al. [35]. Those would also be the ones where surgeons would still expect a survival benefit even after STICH. It is one interpretation to conclude that these patients do no longer exist for faster revascularization during myocardial infarction and less incidence of true transmural infarction. It is another interpretation to suggest that these patients were not randomized by us because of our conviction for the “need” of SVR independent of a randomized trial, so that randomization was just not performed. If the latter interpretation is true, the finger should be turned at ourselves and the question arises whether this “pre-selection of patients” for the STICH trial resulted in a masking of a potential survival benefit for certain subgroups of patients. Thus, a certain bad after taste remains after STICH. So where does STICH and the revolving discussion take us with SVR?

The future perspective

It is important to realize that, based on the STICH outcome, it is still justifiable to perform SVR because no significant detrimental effect to the patient has been demonstrated. Even more so, it should be seriously considered to perform SVR in appropriate patients when they present with conditions that have not been tested in the STICH trial. Table 1 shows a list of conditions where SVR may still be a valid treatment option for the appropriate patient, for instance, a young patient with a circumscribed large anterior aneurysm and good contractility in the remaining myocardium (the classic type 1 aneurysm). Here, SVR is well tolerated in addition to CABG and wall tension will be reduced for many years to come. In patients with significant remodeling and concomitant mitral regurgitation, the SVR procedure may result in reduction in mitral regurgitation without the need for additional mitral valve reconstruction. In addition, SVR is a formidable operation for patients presenting with ischemic ventricular septal defects, specifically those affecting the mid- to apical septum. Here, the placement of a large patch over the VSD anchoring it to non-infarcted septal muscle and excluding the infarcted area from the left ventricle allows to operatively address these patients at almost all time points and does not require to wait for weeks until the infarcted tissue has healed. Another possible indication is the treatment for ventricular arrhythmias in a patient presenting as possible SVR candidate. Dor et al. report an over 90% cure rate of both spontaneous and inducible ventricular tachycardia [45]. However, the report by O’Neill et al. [46] showing increased rates of ventricular arrhythmias within the first 3 months after surgery needs to be mentioned in this context. The difference between the two studies may lie within the endocardial resection that may have been more aggressively performed by Dor et al. [45]. Finally, it appears important to maintain the experience of ventricular surgery in our field. Closure of ventricular septal defects or contained ventricular ruptures presents a formidable surgical challenge. Experience with ventricular reconstruction procedures is an undisputed basis on which the necessary extension of technique can be built. In these patients, although low in numbers, the presence or absence of the required surgical expertise results in immediately measurable differences in patient survival.
Table 1

Clinical conditions where SVR may be considered even after the STICH trial

Circumscribed anterior LV aneurysm with normal contractility in the remaining myocardium

If mild to moderate MR can be treated by SVR

Ischemic ventricular septal defects (specifically mid-septal to apical location)

A-or dyskinetic anterior wall with ventricular tachycardia focus in borderzone

Aneurysms and contained ventricular ruptures in other areas of the left ventricle (modified SVR needed)


SVR is an attractive surgical procedure that failed to provide a survival advantage in patients with postischemic globally remodeled left ventricles. SVR may provide functional benefit by reducing wall tension, increasing synchronicity, reducing mitral regurgitation, and improving efficiency of cardiac work. However, proper functional assessment requires volume-independent parameters that have thus far not been applied in general practice. Despite a neutral outcome in the STICH trial, the procedure still bears therapeutic potential for some patients for different reasons so that the surgeon’s ability to perform this operation should not get lost.


TD was supported by grant Do602/9-1 from the Deutsche Forschungsgemeinschaft (DFG). I wish to thank Andrea Schrepper, Dr. Kristin Rödiger, Dr. Michael Schwarzer, and Dr. T. Dung Nguyen for editorial assistance.

Copyright information

© Springer Science+Business Media, LLC 2012