Ventricular septal defect (VSD) is a rare complication of myocardial infarction (MI), with an estimated incidence of 0.2% of all MI patients in the current era of widely available reperfusion techniques.

Despite the low reported incidence, post-MI VSD carries a high mortality risk for patients, whether treated medically (exceeding 90%) or surgically (ranging from 19% to 60%). Although there has been considerable improvement in the treatment of acute coronary disease in the past two decades, VSD patients remain a challenging niche.1,2 In this regard, several issues involving the surgical treatment of VSD patients remain controversial, including the timing of surgery, the surgical technique, extracorporeal membrane oxygenation (ECMO) as a bridge to closure, concomitant procedures, and the role of percutaneous closure.

Most of the evidence regarding post-MI VSD is derived from small single-center studies with small sample sizes. Three national registries have been reported in the literature.3–5 These multicenter studies provide the best available evidence regarding this rare entity.

Our objective was to establish the first dedicated post-MI VSD national registry in Portugal. We hope that the collective analysis of data on VSD patients on a national scale will contribute to the evidence available regarding risk factors, surgical treatment, perioperative management and long-term follow-up of post-MI VSD patients.

Patient and procedural characteristics are summarized in Table 1. Fifty-one percent of patients were female. The most common comorbidities in the cohort were hypertension (68%) and diabetes (34.7%). The majority of patients presented no significant hemodynamic lesions in either aortic or mitral valves. Left ventricular ejection fraction was normal in 53.6% of patients. Pulmonary hypertension (pulmonary arterial systolic pressure >60 mmHg) was present in 14.7% of patients. Qp/Qs was 1–1.9 in 10.5% and 2–2.9 in 19.7% of patients, however, Qp/Qs measurement was not available for the majority of patients. Regarding coronary anatomy, the most commonly affected vessels were the left anterior descending and right coronary arteries (Figure 1). Seventy-three percent of patients received preoperative inotropic support. Mechanical support by intra-aortic balloon pump (IABP) as a bridge to closure was used in 30 (39.5%) and ECMO in five (6.6%) patients. The most common closure technique (94.5%) was based on the classic technique described by Daggett.6 Associated percutaneous closure was necessary in four patients (5.3%), all for residual VSD after the index closure. Forty-six percent of patients had a concomitant procedure at the same time as VSD closure, most commonly coronary artery bypass grafting (CABG) (36.8%). The most common postoperative complications were infection (non-specified location) (25%) and renal failure requiring dialysis (18.4%).

Figure 1.

Coronary anatomy of the cohort. Cx/OM: circumflex or obtuse marginal artery; LAD: left anterior descending artery; LM, left main; RCA: right coronary artery.

(0,09MB).

Cumulative survival

Median survival of the cohort was 72 months (95% confidence interval [CI] 4–144 months) (Figure 2). Two-, six- and 10-year survival rates were 54.3% (95% CI 44.1–67.0), 49.0% (95% CI 38.6–62.3), and 46.9% (95% CI 36.3–60.5), respectively. Survival at discharge was 61.8% and 30-day survival was 69.7%. Patients who underwent VSD closure early rather than later after onset of symptoms seem to have performed worse (first 10 days vs. after 10 days: log-rank p=0.036) (Figure 3). This difference was evident even when only patients alive at the 30-day mark were considered in the analysis (Figure 4).

Figure 2.

Kaplan-Meier curve of the entire cohort. Mean survival was 72 months.

(0,16MB).

Figure 3.

Kaplan-Meier curve of the early (first 10 days) vs. delayed intervention groups (log-rank p=0.036).

(0,17MB).

Figure 4.

Kaplan-Meier curve of the early (first 10 days) vs. delayed intervention groups in patients alive at the 30-day mark (log-rank p=0.048).

(0,18MB).

Kaplan-Meier analysis showed no difference in cumulative survival in patients undergoing concomitant CABG (log-rank p=0.4), different closure techniques (log-rank p=0.5), ECMO as a bridge to closure (log-rank p=0.9) or IABP as a bridge to closure (log-rank p=0.6), or with different locations of the VSD (log-rank p=0.9) (Figure 5).

Figure 5.

Kaplan-Meier curve discriminating: (A) concomitant coronary artery bypass graft procedure (log-rank p=0.41); (B) location of ventricular septal defect (log-rank p=0.86).

(0,35MB).

Regarding the timing of post-MI VSD surgical closure, most authors agree that early surgery is indicated, as it avoids prolonged exposure to congestive heart failure and cardiogenic shock leading to multiorgan dysfunction. Advocates of deferred closure argue that a delayed strategy provides time to stabilize the patient, permitting VSD closure in optimal clinical and tissue conditions.7 This is an almost impossible dilemma to resolve in an observational setting, as critically ill patients tend to be operated early and patients in whom surgery is delayed are usually less ill. Adding to this problem is the fact that patients considered inoperable or who died waiting for surgery are not usually included in surgical cohorts. According to the 2013 American College of Cardiology Foundation/American Heart Association guideline for the management of ST-elevation MI,8 surgical treatment should be initiated without delay after diagnosis, whereas the 2017 European Society of Cardiology guidelines on ST-elevation MI9 recommend that delaying intervention may be considered in certain patients.

In our analysis, cumulative survival of delayed intervention was superior both for the overall cohort (Figure 3) and for patients alive at 30 days (Figure 4). This, however, was not apparent in the multivariate model for 30-day mortality. This may be related to underpowering of our model to detect a significant difference. More importantly, however, as previously described, observational studies in the setting of VSD cannot provide a clear picture of the best surgical timing due to the fact that critically ill patients tend to be operated sooner and hemodynamically stable patients tend to be operated later.

One reasonable conclusion is that these patients benefit from being operated as late as their hemodynamic profile permits.

ECMO has been advocated as a bridge to VSD closure. Theoretically ECMO limits the patient's exposure to the detrimental hemodynamic consequences of the VSD, reducing end-organ malperfusion and unloading the heart, while permitting a more opportune timing of closure. However, ECMO is not without its own inherent complications. Morimura et al. reported mechanical circulatory support, including ECMO and IABP, as a bridge to VSD closure in a small cohort, with positive results.10

In our cohort ECMO was used in five patients as a bridge to closure. This strategy was not associated with different cumulative survival (log-rank p=0.9); however, all ECMO patients had a follow-up of less than five years, and three (60%) were alive at discharge.

Model simulations suggest that the Impella device may present a superior hemodynamic benefit in the post-MI VSD setting compared to ECMO, thus permitting a delayed closure strategy with fewer associated complications.11

There is further controversy relating to the benefit of concomitant CABG. Although the majority of studies show no benefit of concomitant CABG regarding early and long-term mortality, results with both improved and reduced survival have been reported. The aforementioned registries3–5 showed no significant difference in survival for patients undergoing concomitant CABG, except for mid-term survival after exclusion of patients who died in the first 30 days. Perrotta et al. reviewed the best available evidence regarding concomitant revascularization in this setting and found that overall, it suggests that revascularization of non-infarcted areas of myocardium and consequent improvement of collateral flow may contribute to better long-term survival.12 Our findings are consistent with the results reported from previous registries, with no difference in cumulative survival (Figure 4), although a protective effect for 30-day mortality was found (OR 0.23; 95% CI 0.06–0.90; p=0.034).

Various techniques of VSD closure have been described in the literature.2,6,13 It is unclear whether any one technique has definite advantages over any other. Furthermore, extent of infarcted area, time of reperfusion and center/surgeon case volume may be more significant factors affecting residual VSD and overall outcome. Our cohort has a clear dominance of the classic technique described by Daggett, with a small number of alternate techniques, which precludes interpretation regarding the relation between closure technique and outcome.

Since the first percutaneous closure of a post-MI VSD in the 1980s, percutaneous intervention has evolved and is viewed with renewed interest. Technical success rates of over 75% with associated in-hospital/30-day mortality of 32% have been reported.14 However, patient selection is an inherent characteristic of such interventional reports, with anatomical feasibility as a limiting factor and most patients not being considered candidates for surgery.

Percutaneous closure in post-MI VSD has been proposed both as a bridge to surgical treatment in patients who are not expected to survive an operation and as a definitive treatment in simple <15 mm defects in the sub-acute or chronic phase.15 Additionally, percutaneous closure is an alternative for reintervention for residual VSD after index closure, which itself consists of a particular subset of high-risk patients. In our cohort we identified four patients in which this approach was used. Three out of these four (75%) patients were discharged alive.

Operative mortality was high in our cohort, albeit in agreement with that described in previous registries: 33.0% in the Japanese registry,5 42.9% in the Society of Thoracic Surgeons Adult Cardiac Surgery Database3 and 41% in the Swedish registry4 (30-day mortality) and 47% in the Global Utilization of Streptokinase and T-PA for Occluded Coronary Arteries (GUSTO) trial.1

Several risk factors for early and late death have been proposed in previous studies.3–5,16,17

Multivariate analysis showed the significant impact of concomitant CABG (OR 0.23; 95% CI 0.06–0.90; p=0.034) and age (OR 1.08; 95% CI 1.01–1.15; p=0.27) on 30-day mortality, as well as a trend regarding concomitant mitral surgery (OR 9.20; 95% CI 0.90–93.96; p=0.061) (Table 2). Other authors have reported similar findings regarding age3,5 and significant mitral regurgitation.3 These findings suggest that surgical complexity and frailty may play a role in early outcomes. A beneficial effect of concomitant CABG has been proposed but remains controversial.15 In our study, due to the outcome being 30-day mortality, long hospitalizations (>30 days) may have skewed the influence of other important baseline characteristics.

As a side note, a uniform approach regarding timing of intervention was not discernible in the three centers included in our cohort, with both epoch- and surgeon-related variation found. A common notion, however, is the need for a pragmatic rather than a dogmatic approach, in the sense that the timing of intervention depends on close monitoring and patient-tailored strategies rather than a predefined timeframe.

VSD is a dreadful complication of MI. Registries may provide important evidence regarding an entity for which randomized controlled trials are impractical. Percutaneous closure and mechanical support are emerging techniques that may prove useful in selected patients. In our study, age negatively influenced and concomitant CABG positively influenced 30-day survival.

Timing of surgery remains a controversial issue. Later closure seems to be advantageous but there is significant bias. Overall, it seems that these patients benefit from being operated as late as their hemodynamic profile permits.

VSD is a dreadful complication of MI.

Registries may provide important evidence regarding this entity.

CABG positively influenced 30-day survival.

Age negatively influenced 30-day survival.

Later closure seems to be advantageous but an important bias exists.

No funding was received for this project.

The authors have no conflicts of interest to declare.