Surgical outcomes of post‐infarct ventricular septal defect repair: Insights from the UK national adult cardiac surgery audit database

Ventricular septal defect (VSD) is becoming a progressively less frequent mechanical complication of myocardial infarction (MI). However, this event is still associated with high operative mortality. We aimed to describe the trends and the risk factors associated with surgical VSD repair outcomes and to provide a clinical benchmark for percutaneous VSD closure strategies.


| INTRODUCTION
Ventricular septal defect (VSD) due to ischemic septal rupture represents a rare event complicating ST-elevation myocardial infarction (MI). There has been a significant decrease in the incidence of this dreaded complication, from 1% to 3% in the prethrombolysis eras to 0.2%-0.34% in the latest years. 1 However, this event is still associated with very poor clinical outcomes. The mortality rate is as high as 94% after medical treatment whereas surgical repair of VSD has been reported to have an operative mortality higher than 30% and it varies largely according to the timing of surgery. Currently, there is no consensus on which is the optimal timing for the repair of VSD as patients undergoing surgery early after the MI show worse outcomes while patients in which surgical repair is attempted after 2-3 weeks after the culprit MI have the lowest operative mortality. 2,3 More recently, the use of percutaneous devices to close the septal defect has emerged as a potential alternative to surgical repair.
Although interventional VSD closure is associated with a more rapid correction of the interventricular shunt, it still remains impacted by high mortality and morbidity and therefore, surgical repair remains in the majority of cases the gold standard for these patients. 4 This study aimed to provide a clinical benchmark for the treatment of post-MI VSD by reporting on the outcomes of VSD repair over a 23-year period and identify risk factors associated with worse short-term outcomes after surgical repair.

| METHODS
The study is part of a research project approved by the Health Research

| Outcomes
The primary outcome was in-hospital mortality. Secondary outcomes investigated were postoperative cerebrovascular events, need for postoperative dialysis, deep sternal wound infection, reexploration for bleeding, insertion of intra-aortic balloon pump (IABP) or ventricular assist devices and length of stay.

| Statistical analysis
Categorical variables are presented as count and proportion and were compared using the Pearson's Chi-squared test or Fisher's exact test, as appropriate, whereas continuous variables were reported as median and interquartile range and were compared using Wilcoxon rank sum test.
Patients characteristics were reported in the overall sample and were stratified by patient status at discharge (survivor vs. nonsurvivors). Predictors of operative mortality were investigated using a multivariable, generalized linear mixed model, which included the patient characteristics found significant at the univariable analysis as fixed effect covariates. A clustering effect was anticipated for patients operated in the same hospital or by the same surgeon and therefore these two variables were included in the model as random intercepts. As surgical outcomes may have improved in the last decade, year of surgery was also included as random intercept. Effect estimates for fixed effect covariates were reported as odds ratio (OR) and 95% confidence interval (CI). The marginal R-squared considers only the variance of the fixed effects, while the conditional R-squared takes both the fixed and random effects into account. 6 The primary outcome was also investigated according to the time interval from MI to the surgery, and according to the procedure performed (isolated VSD repair, concomitant CABG or CABG plus mitral valve surgery).
p value <.05 was considered significant in all the analysis. Statistical analysis was performed using R version 4.0.0 using the packages sjplot, lme4, lmertest, gtsummary. The overall median number of VSD repairs per surgeon was 2 (1-3) and the overall median hospital volume 24   Table 2).
Patients who died were more likely to be older, have an impaired left ventricular and renal function, and present with unstable angina. Also, nonsurvivors tended to have undergone a previous PCI and be more critically ill presenting with cardiogenic shock, inotropic support and preoperative ventilation. VSD repair was more likely to be nonelective and take place within the first 24 h after MI in patients who died (Table 1).
Over the years, the operative mortality ranged from 27.8% to 48.8% and did not change significantly ( Figure 2). When the primary outcome was analysed according to the interval between MI and the surgical repair, the operative mortality was significantly higher in patients undergoing surgery within the first 24 h (75% and 61.3% if operated within 6 h or between 6 and 24 h, respectively) compared to patients with a delayed repair ( Figure 3). Moreover, operative mortality did not differ significantly when isolated VSD repair was compared to VSD repaired with concomitant CABG and mitral valve surgery ( Figure 4). The incidence of secondary outcomes is presented in Table 3. The

| DISCUSSION
This study is one of the largest reported cohort of VSD surgical repair and the first from the national United Kingdom database over a 23-year period.
We found that the number of VSD surgical repairs did not change over this long period of time suggesting that the introduction and wider adoption of interventional strategies for the closure of VSD did not impact the overall surgical management of this dreaded event, for which we found an overall operative mortality of 38.9%. We also found that the closer the operation to the onset of the culprit MI the worse the op-

CONFLICT OF INTERESTS
The authors declare that there are no conflict of interests.