In 1875, Latham et al.1 first described ischemic ventricular septal rupture (VSR) and since then it has been considered one of the most serious complications of acute myocardial infarction (AMI). Currently, the epidemiology of VSR has changed with a prevalence of 0.2% to 0.3%, in comparison with the estimated 3% rates before widespread reperfusion therapies.2–4 However, survival in affected patients does not appear to have improved. Contemporary registries show extraordinarily high 30-day mortality rates, between 38% and 88%, due to the development of cardiogenic shock and multiorgan failure.3,4 Widespread application of stent-based reperfusion strategies has probably helped to reduce the rates of AMI-related mechanical complications, but despite continuous technical improvements in interventional cardiology, in-hospital mortality does not appear to be decreasing.5–7

The early adoption of mechanical circulatory support could represent a turning point in the management of mechanical complications, such as VSR or papillary muscle rupture. The hemodynamic improvement can lead to adequate stabilization, avoiding the systemic consequences of low tissue perfusion and multiorgan failure before corrective surgery.8–14 This strategy has shown positive results even in patients with a high-risk profile assessed by the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) classification and with established multiorgan failure.14 Nonetheless, although early identification and support of cardiogenic shock seems to change the poor prognosis of post-AMI VSR, the evidence remains limited. Moreover, current data on VSR in the era of optimal reperfusion with primary percutaneous coronary intervention (PCI) is scarce.

In this study, we aimed to analyze developments in the management and prognosis of post-AMI VSR patients in a multicenter registry involving tertiary centers with well-developed reperfusion networks and locally available or rapid access to cardiac surgery facilities.

The main results of our study were the following: a) VSR is still a relatively rare complication of AMI with a high mortality rate; b) VSR patients have been more frequently managed with VA-ECMO support in the last few years, with a trend to more frequent successful achievement surgical repair or heart transplant; c) there is a temporal trend of increased 1-year survival, although in-hospital mortality rates are still higher than 50%.

Despite all current advancements in AMI treatment, mortality rates of postinfarction VSR are still extremely high, at around 60% in our series and in other registries.2 Our data are similar to previously published data, with a prevalence of VSR of less than 1% of all AMI.3 We observed no temporal changes in patient characteristics, except for younger age in the last 5-year period. We have no clear explanation for the increase in the number of patients in the second half and decrease in the age of presentation of the patients in this cohort. It might be related to less thorough investigation with bedside echocardiogram during the first years of this retrospective analysis or missing diagnosis of this AMI-related mechanical complication in our database, which was composed of prospectively collected data from several hospitals. This finding is intriguing and could be related to the implementation of highly active primary PCI networks in some of the included centers, beginning in 2013 to 2014 (table 1 the supplementary data) with an increase in the total number of ST-segment elevation myocardial infarction patients treated in larger institutions. However, we believe this aspect should be explored in a larger, preferably nationwide study.

Analysis of clinical aspects of the ischemic event showed that apical location of VSR was more frequent, as previously described, with a similar distribution between right and left coronary artery infarctions. The delay in the diagnosis is also important, this issue being a reiterated and classic problem for mechanical complications of AMI. Like previous literature, our study demonstrates that post-AMI VSR is more frequent in patients who delay receiving medical attention during symptom onset, and in a substantial number of patients, the VSR occurs once the patients are in the hospital.

We believe that despite initial symptoms at the time of diagnosis, rapid decline in the hemodynamic status of these patients should be anticipated and elicit a rapid response based on application of mechanical circulatory support therapies. VA-ECMO support, with or without intra-aortic balloon pump, is the preferred option in most centers.17,18 Stabilization with VA-ECMO could potentially favor best overall conditions toward a timely definitive surgical repair. In our study, the increased trend to VA-ECMO support was associated with a slightly better survival in the univariate analysis.

Better repair techniques of the septal defect are still needed. Percutaneous closure has been described with different devices, but the high residual or recurrence of the left-to-right shunting are both associated with very poor overall results, as presented in our study.19–23 Surgical repair is usually the preferred option but is limited by the need for large patches to overcome the fragility of VSR borders, leading to small and/or restrictive left ventricular cavities and refractory cardiogenic shock in some cases. Given the overall poor results of percutaneous and surgical options, it is noteworthy that heart transplant was used as a first-line strategy in some of our centers with good results (survival in 5 out of 6 patients). Heart transplant is not available in all included centers but transferring patients to specific advanced heart failure units is relatively common and feasible within the Spanish health system, even for sick patients such as those that are considered possible candidates. Despite the well-known limitations of the shortage of donors, patient suitability and logistic complexity, heart transplant may be the best option in specific subgroups, such as those with severe and irreversible right/left ventricular failure or extensive defects.

Another important question is the timing between the use of VA-ECMO and definitive surgical repair. From one perspective, allowing time for definitive scarring of VSR borders could theoretically facilitate surgical repair sutures.14 However, prolonged support is associated with more vascular and bleeding complications.17,18,24 Some studies observed that the real benefit of a delayed surgical repair appears 7 days after diagnosis.25 Ariza-Solé et al.14 suggest that surgery should be delayed until the recovery of multiorgan failure, 3 to 4 days after placement of ECMO, and the device should be withdrawn 2 to 3 days after surgical repair when the patient is completely stabilized. Malhotra et al.26 have proposed a prognostic scoring system to establish the optimal timing of surgical repair. In the absence of randomized clinical trials, large observational data such as the present study suggest a benefit of an VA-ECMO-based strategy as a bridge to final reparative surgery or heart transplant. We observed a nonsignificant trend in delayed repair in patients with VA-ECMO, increasing the mean time to repair from 2 to 5 days. In patients treated with VA-ECMO, the optimal timing to surgery appears from the fourth day, with an in-hospital mortality of 36%. This information is congruent with the above-mentioned previous studies.14

One noteworthy finding is that we identified 3 patients in which the cardiac catheterization showed nonobstructive coronary arteries. MINOCA had been previously identified as an etiology of VSR in several clinical cases.27–30 The appearance of this complication in MINOCA seems to be exceedingly rare, and its pathophysiologic features are poorly understood.

This study has several limitations. The observational and retrospective character of our registry, which is supported with historical data from the collaborating centers, is a potential source of selection bias and is the main limitation. The original study design does not allow strong causal associations to be established. This design can favor the loss of data in patients with VSR and early cardiogenic shock, who died in the first few hours of receiving medical attention and could possibly have been excluded from the databases of individual units. In the case of the time to PCI, we do not have information about the exact time until the procedure in hours, which could better explain the incidence of VSR related to a delay in treatment. We did not use an external quality assurance resource for the analysis of these data. However, all selected centers have a prospectively collected database, which should have helped minimize loss of relevant information. The incidence of VSR is an estimated value taken from reports of voluntary participating centers, and therefore should be interpreted with caution. Finally, the contribution of only large centers in this database could limit extrapolation of the prevalence or clinical manifestations of VSR to other settings, although we believe it had a limited impact in the analysis of our primary endpoint and the present data should be viewed in the context of similar tertiary centers.

Post-AMI VSR still carries high in-hospital and 1-year mortality. There is a temporal trend toward a decrease in the mortality of post-AMI VSR that appears to be multifactorial. Management differences among 2 time periods included an increase in use of VA-ECMO support and in heart transplant as a possible alternative in anecdotal cases. The contribution of VA-ECMO to overall survival should be addressed in appropriate clinical trials.

The authors received no financial support for the research, authorship, and/or publication of this article.

M. Sanmartín Fernández developed the original project and the hypothesis along with the design and objectives and revised the final manuscript. J.D. Sánchez Vega, G.L. Alonso Salinas, J.M. Viéitez Flórez and J.L. Zamorano Gómez contributed to the analyses of the data and writing of the manuscript. The rest of the authors, including those already listed, contributed to the recruitment process. All authors critically revised the article for important intellectual content and have contributed to the work and agreed on the final version.

The authors declare no competing interests.

Collaboration in data collection: Nagore Horrillo Alonso and Jorge Díaz Calvo from Hospital Universitario de Cruces (Baracaldo), Sandra Rosillo from Hospital Universitario la Paz, IDIPAZ, Madrid.

The authors would like to thank Sara Rosenstone Calvo, M.D. for the help provided in editing this manuscript.