Outcome and survival analysis of surgical repair of post-infarction ventricular septal rupture

The reported incidence of acute ventricular septal rupture (VSR) following acute myocardial infarction (AMI) is 1% to 3% in the era prior to widespread reperfusion therapy[1]. The incidence has since declined to around 0.3% following the advent of thrombolysis[2].

Despite advances in critical care and prompt surgical intervention, the incidence of post-infarction VSR has not changed during the past 2 decades. The mortality rate of post-infarction VSR remains high and relatively constant[3]. Reported 30-day mortality rates range from 19 to 54%[3–11].

This study aims to investigate the survival outcome and prognostic factors associated with surgical repair of post-infarction VSR at a tertiary referral centre over a 13 year period.

A retrospective case-note and database review was performed on consecutive patients who had undergone surgical repair of post-infarction VSR between January 1999 and December 2011 at our tertiary referral centre. This study was approved by the Singhealth Centralised Institutional Review Board (reference: 2011/881/C).

Patients

During the period January 1999 to December 2011, 42 consecutive patients were referred to our tertiary cardiac centre following diagnosis of VSR. Four patients with anterior VSRs underwent medical therapy as 2 patients declined surgery and the other 2 were not offered surgery in view of their advanced age (both 91 years old), multiple medical comorbidities and poor premorbid functional status. All 4 patients died within 5 days of diagnosis of VSR. Thirty eight patients underwent surgical repair of VSR. No patients were treated with percutaneous closure devices.

Data was retrospectively collected from patient case notes and electronic records. Patient demographics and comorbidities are shown in Table 1. All patients were followed-up at our institution, with 1 patient lost to follow-up. Male patients accounted for 52.6%. Mean age was 65.7 ± 9.4 years (range 45 – 83). The mean age of females was significantly higher than males (70.0 ± 6.0 vs 61.8 ± 10.3 years, P=0.005). In all patients, a new cardiac murmur was present in the setting of an ST-segment elevation myocardial infarction (STEMI). Transthoracic echocardiography (TTE) was performed for all patients to confirm the diagnosis of VSR and exclude the differential diagnosis of acute mitral regurgitation from papillary muscle rupture. Evidence of right ventricular (RV) dysfunction such as RV dilatation or wall motion abnormality was also diagnosed with pre-operative echocardiography. RV dysfunction was present in 10 patients (26.3%), of which 7 had presented with inferior MI and 3 with anterior MI. RV dysfunction was present in 8 of 10 (80%) with posterior VSRs and 2 of 28 (7.1%) with anterior VSRs, (P<0.0005). The VSR location was anterior in 28 patients (73.7%) and posterior in 10 patients (26.3%). One patient had a concomitant ventricular free wall rupture. The median time interval between myocardial infarction and septal rupture was 1 day (1, 4). The interval between myocardial infarction and septal rupture was 1 day in 20 patients (52.5%), 2 to 7 days in 13 patients (37.0%) and 8 to 13 days in 4 patients (10.5%).

Variable	Survivors n = 23 (%)	Non-survivors n = 15 (%)	P-value
Demographics
Age (years)	64.7 ± 9.2	67.3 ± 9.9	0.406
Gender (Male)	11 (47.8)	9 (60.0)	0.463
BMI (kg/m2)	25.1 ± 3.6	23.7 ± 2.0	0.186
LVEF (%)	40.3 ± 11.0	39.1 ± 9.4	0.746
Comorbidities
Smoking	8 (34.8)	8 (53.3)	0.258
Diabetes mellitus	8 (34.8)	4 (26.7)	0.599
Renal failure	9 (39.1)	6 (40.0)	0.957
Hypertension	16 (69.6)	10 (66.7)	0.851
Hyperlipidaemia	9 (39.1)	5 (33.3)	0.717
Previous stroke	2 (8.7)	2 (13.3)	0.649

BMI: Body Mass Index.

LVEF: Left Ventricular Ejection Fraction.

Eleven patients (28.9%) had undergone thrombolysis with either streptokinase or recombinant tissue plasminogen activator. Coronary angiography was performed in 36 patients (94.7%), which showed single vessel disease (SVD) in 38.8%, double vessel disease (DVD) in 30.6% and triple vessel disease (TVD) in 30.6% of patients. Within the SVD group, 10 of 14 patients (71.4%) had total occlusion of the infarcted artery. Figure1 shows the distribution of coronary disease according to VSR location.

The availability of early reperfusion therapy for myocardial infarction has led to the declining incidence of post-infarction VSRs. Despite this trend, VSR remains one of the most challenging conditions encountered by cardiac surgeons today.

The 30-day operative mortality in this series was 39.5%, similar to the rates of 34% to 37% reported in other series[4–7]. Lower operative mortality rates ranging from 19% to 23% have been reported[8–11]. Univariate analysis identified emergency operation as a predictor of early mortality in this series, which was similar to the findings of other authors[4, 6, 9, 10]. The same authors also identified the duration of aortic cross clamp as a predictor of early mortality but this association was not demonstrated in this study. Increased cardiopulmonary bypass time has also been reported as a predictor of operative mortality but this association was not evident in this series[11–14].

The optimal timing of surgical repair for VSRs is critical. A longer interval before surgery has been reported to be associated with improved survival[15]. Despite this, only 2 of our 38 patients remained sufficiently stable in NYHA class II to permit delayed surgery. Although both these patients survived with good outcomes and returned to NYHA class I after recovery, we reserve delayed VSR repair only for a carefully selected group of patients who remain haemodynamically stable.

We adopted the single patch technique in 35 cases (92.1%). The incidence of post-operative residual shunt was 23.7%, comparable to the incidence of 24 to 26% described in other series in which a variety of techniques were used[6, 16]. An alternative double patch technique has been reported to decrease the incidence of residual shunt from 11% to 0%[17].

We routinely perform coronary angiography for all haemodynamically stable patients diagnosed with VSR at our centre. This provides an opportunity for the insertion of IABP under fluoroscopic guidance. In this series, coronary angiography was performed in all except for 2 patients who presented in extremis. SVD was significantly more common in anterior VSRs compared to posterior VSRs in this series, similar to the findings of Davies et al.[18]. In the group with SVD, 71.4% had total occlusion of the infarcted artery. Reported figures range from 57% to 82%[2, 19].

Lundblad et al.[20] found that concomitant CABG during VSR repair reduces both early and late mortality when compared with patients with unbypassed coronary artery disease. Although there was no significant difference in early mortality within other series, mid to long-term survival benefit has been reported[16, 20–22].

In a review of recent literature, Perotta et al.[23] reported an improvement of mortality rates from 26.3% in those without CABG to 21.2% in those who had undergone CABG. These results applied to patients with multi-vessel disease where complete myocardial revascularization was achieved by bypassing all stenotic coronary arteries supplying non-infarcted areas. Actuarial survival at five years from this series ranged from 29% to 72%. In this study, there was no statistically significant early or long-term survival advantage for patients who had undergone concomitant CABG. Other authors have reported similar findings[4, 6, 7, 10, 11, 15, 17, 24]. Despite the lack of significant survival benefit associated with concomitant CABG in their studies, some authors advocate concomitant CABG during VSR repair as long as it can be performed safely[4, 5, 15, 17]. The aim is to reduce further ischaemic risk associated with multi-vessel coronary artery disease by improving collateral flow to the myocardium.

Similar to the findings of Moore et al.[25], our data shows that RV dysfunction has a negative impact on early and late survival. As RV failure ensues, the left-sided cardiac chambers are unable to fill, leading to biventricular failure. This exacerbates the low cardiac output state, contributing towards a higher mortality rate.

Percutaneous closure has been increasingly used in patients with postinfarction VSRs, initially in patients with recurrent ventricular septal defects (VSDs) after primary surgical repair, but more recently as primary therapy in patients with acute VSR and high surgical risk, or as a temporizing bridge to surgery[26]. Ventricular septal occluder devices were initially developed for percutaneous closure of congenital VSDs. Percutaneous deployment of an occluding device may be precluded by the geometry and location of the VSR, or interference from valvular apparatus, especially in posterior VSRs. The highest technical success has been reported with the Amplatzer device (AGA Medical Corp, Plymouth, MN), with procedural success rates ranging from 86% to 89%[27, 28]. Lee et al.[29] recently described the use of a hybrid approach with the advantage of allowing direct manipulation of the Amplatzer device. This approach prevents interference of the valve apparatus, avoiding the need to perform ventriculotomy in an already infarcted ventricle.

Large (≥15 mm) VSRs should undergo immediate surgery as these are prone to device embolization or residual VSR. Amplatzer closure can be used as definitive therapy for small or medium VSRs. It can also be used to stabilize patients and allow myocardial fibrosis, thus facilitating delayed surgical correction[27]. Despite a less invasive technique, procedural mortality and morbidity are high, especially in patients with cardiogenic shock (mortality in cardiogenic shock 88% vs. 38% non-shock P<0.001)[26]. Overall 30-day mortality rates for percutaneous closure range from 28% to 65%[26–28]. Other procedure-related major complications include residual shunting, left ventricular rupture and device embolization. The incidence of residual shunts ranges from 8.3% to 13.8%, much lower than 73% reported in an earlier series[26–28].

Ventricular assist devices (VADs) are useful adjuncts for the treatment of VSRs in the setting of univentricular or biventricular failure, either preoperatively as a bridge to surgery or postoperatively following VSR repair. Mechanical circulatory support with VADs allows restoration of peripheral organ perfusion and provides an opportunity for recovery and maturation of the infarcted myocardium before secondary VSR repair. Applying this concept, a staged approach of initial left ventricular or biventricular mechanical support with an implantable left ventricular assist device (LVAD) or biventricular assist device (BiVAD) followed by secondary VSR repair has been described[30, 31]. Previously, BiVAD support without VSR repair has been described as a bridge-to-transplantation approach[32].

La Torre et al.[33] reported a similar bridge to surgery strategy in their series of patients with posterior VSRs using the Impella Recover (ABIOMED, Inc, Danvers, Mass), a percutaneous left ventricular assist device (LVAD). Utilizing an intravascular microaxial blood pump inserted in the femoral artery, successful reduction in left-to-right shunting and increase in cardiac output was achieved while awaiting VSR repair about 3 weeks later. Despite successful surgical repair of VSR, some patients with extensive infarction and persistent ventricular failure remain in cardiogenic shock and require VAD support to maintain physiologic haemodynamics. Implantable LVADs have been used successfully to bridge these patients to cardiac transplantation[34].

This is a retrospective descriptive study with inherent biases in data collection. Due to the relatively rare occurrence of post-infarction VSR, the small sample size underpowered the statistical analysis and could have limited the number of statistically significant variables. A prospective multi-centre study incorporating a larger sample size would be useful to assess the prognostic value of the risk factors identified.

Post-infarction VSR remains a serious and challenging complication of acute myocardial infarction in the modern surgical era. Surgical repair is associated with an operative mortality of 39.5%. Overall survival at 10 years is 44.4% ± 8.4%. The NYHA class at presentation and the need for post-operative RRT are independent predictors of early mortality. Right ventricular dysfunction, lower LVEF and NYHA class IV symptoms at presentation are predictors of poor long-term survival. Concomitant CABG during VSR repair does not confer a significant survival advantage.