Surgical repair for postinfarction left ventricular free-wall rupture: 25-year single-centre experience^†

Abstract

OBJECTIVES

Postinfarction left ventricular free-wall rupture (LVFWR) is a rare, unpredictable and often fatal complication of acute myocardial infarction. We reviewed our surgical experience with postinfarction LVFWR over 25 years to identify risk factors for in-hospital mortality.

METHODS

Seventy-two consecutive patients with LVFWR who underwent surgical repair between 1994 and 2023 were retrospectively analysed. The primary end point was in-hospital mortality. The mean follow-up period was 3.9 ± 5.4 years (maximum 25 years, 283 patient-years).

RESULTS

Thirty-five patients (49%) were directly transported to our centre, whereas 37 (51%) were initially taken to a local hospital. Prior to surgery, 30 (42%) developed out-of-hospital (n = 8, 11%) or in-hospital (n = 22, 31%) cardiac arrest. Upon entry to the operating room, 45 (63%) presented hypotensive shock, while 27 (38%) suffered cardiac arrest. LVFWR was oozing type in 42 patients (58%) and blow-out type in 30 patients (42%), for which sutureless repair (n = 27, 38%) or sutured repair (n = 45, 62%) was performed. The in-hospital mortality rate was 35%, in association with haemodynamic condition before surgery: 19% in patients without cardiac arrest, and 55% with in-hospital and 63% with out-of-hospital cardiac arrest. Patient age [adjusted odds ratio (OR) 2.2 per 10-year period, 95% confidence interval (CI) 1.0–4.9, P = 0.044], out-of-hospital cardiac arrest (adjusted OR 8.6, 95% CI 1.4–66, P = 0.020), in-hospital cardiac arrest (adjusted OR 6.6, 95% CI 1.6–33, P = 0.010) and initial ambulance transport to a local hospital (adjusted OR 4.2, 95% CI 1.2–17, P = 0.020) were identified as independent risk factors of in-hospital mortality. The overall 5-year survival rate was 50%, while that for in-hospital survivors was 77%. When the present study period is split into 2 eras (1994–2010 and 2011–2023), the in-hospital mortality rate was comparable between them (41% vs 31%, P = 0.408). Notably, the in-hospital mortality of patients with a blow-out type left ventricular rupture was significantly lower in that latter study period (86% vs 39%, P = 0.025).

CONCLUSIONS

In surgically treated patients with LVFWR, the early mortality rate was high, while the long-term prognosis for in-hospital survivors was acceptable. These findings highlight the need for early diagnosis and prompt management of haemodynamic instability at a tertiary centre to improve outcomes.

Open in new tabDownload slide

INTRODUCTION

Mechanical complications following acute myocardial infarction (AMI) may involve the interventricular septum, ventricular free wall or papillary muscles. Left ventricular free-wall rupture (LVFWR) is an unpredictable and often life-threatening mechanical complication of AMI. With the advent of reperfusion strategies for AMI, including thrombolysis and percutaneous coronary intervention, LVFWR has become rare, with the incidence rate reported to range from 0.01 to 0.5% of AMI cases [1, 2]. Nevertheless, despite demonstrated improved clinical outcomes, LVFWR still accounts for up to 61% of in-hospital mortality cases [3–14]. Despite the challenges presented, prompt diagnosis is a key factor, and subsequent surgical repair is generally the treatment of choice. Because of its rarity, little is known regarding early clinical results of surgical LVFWR repair or late follow-up findings. The present retrospective study was performed to review our surgical experiences with postinfarction LVFWR over 25 years and identify determinants of in-hospital mortality.

MATERIALS AND METHODS

Data collection, analysis and reporting were approved by the Institutional Review Board of the National Cerebral and Cardiovascular Center (reference No.: M30-026; approval date: 25 March 2022). Data on patients’ clinical characteristics, surgical data and outcomes were obtained from the institutional surgical database. A definitive diagnosis of LVFWR was based on evidence of pericardial effusion or cardiac tamponade shown by transthoracic echocardiography, along with physical examination and electrocardiographic findings.

Surgical intervention

Surgery was performed under either a cardiopulmonary bypass condition, with or without cardioplegic arrest, or a beating heart condition (off-pump), based on haemodynamic status, type of rupture or need for an additional surgical procedure. Following removal of clots and blood from the pericardial space, the rupture site was carefully identified. For inactive bleeding cases, a sutureless repair was principally indicated. Prior to 2000, a Gore-Tex, bovine or autologous pericardium patch was fixed on the epicardium with fibrin glue to cover the area of haematoma and infarcted myocardium. Since 2000, TachoComb (CSL Behring, Tokyo, Japan) (2000–2011) or TachoSil (Nycomed, Zurich, Switzerland) (2011–present), types of ready-to-use haemostatic collagen sponge, have been utilized for sutureless repair. After removing clots from the ruptured site, the collagen sponge was trimmed to an appropriate size and secured to the surface of the infarct myocardium. Three- to 5-min manual compression was then performed until it became firmly adherent and haemostatic. For some cases, in order to prevent re-rupture, the surrounding infarcted myocardium was then widely covered with a bovine pericardial patch oversewn to the epicardium with a continuous running polypropylene suture, with meticulous attention given to avoid coronary involvement. Additionally, gelatin-resorcinol-formaldehyde glue (Cardial, Technopole, Sainte-Etienne, France) or Bioglue (CryoLife, Inc., Kennesaw, GA, USA) was injected beneath the patch to increase compression strength on the myocardium and prevent blood leakage (Video 1).

For active bleeding cases, the use of direct sutures reinforced with a polytetrafluoroethylene (PTFE) felt strip without a preceding infarct excision was primarily considered for closing the myocardial tear under a cardiopulmonary bypass condition, with or without cardiac arrest. Then, a glued patch repair using bovine pericardium and gelatin-resorcinol-formaldehyde (GRF) glue was further applied to the ischaemic region (Video 2). When the laceration was too fragile and difficult to repair with direct sutures, an infarct excision with closure of the defect using a tailored patch was performed to re-establish the geometry of the left ventricular (LV) chamber.

Outcome determination

The primary end point of this study was in-hospital mortality, defined as death from any cause occurring within 30 days after surgery or after 30 days during the same hospitalization related to the operation. The secondary outcome was identification of risk factors for early mortality after surgical repair of postinfarction LVFWR.

Statistical analysis

To elucidate factors related to incidence and predictors of in-hospital mortality, the enrolled patients were classified as hospital survivors and non-survivors. Continuous variables are presented as mean ± standard deviation, and categorical variables as frequency and proportion. For continuous variables, comparisons between hospital survivors and non-survivors were made using Student’s t-test or Mann–Whitney U-test, as appropriate. Likewise, categorical variables were compared using chi-square analysis or Fisher’s exact test.

Survival analysis was performed using the Kaplan–Meier method for estimation. Preoperative and surgical factors associated with in-hospital mortality were analysed using logistic regression analysis. Factors showing a P value <0.05 were then entered appropriately in a multivariable fashion. The results are summarized as odds ratio (OR) and 95% confidence interval (CI). Statistical analyses were performed using JMP Pro, version 15.1.0 (SAS Institute Inc., Cary, NC, USA).

RESULTS

Baseline clinical characteristics

Our institutional surgical database showed a consecutive series of 72 patients with postinfarction LVFWR who underwent surgical repair at the centre between 1994 and 2023 (Table 1). All patients (75 ± 9 years, 36 males) underwent surgical intervention on a salvage or emergency basis for postinfarction LVFWR. Thirty-five patients (49%) were brought directly to our centre, whereas the remaining 37 (51%) were initially taken to a local non-surgery hospital and then transferred to our centre for further evaluation and treatment. The majority of patients (n = 69, 96%) presented ST-elevation AMI at the initial evaluation, while only 3 (4.2%) suffered from non-ST-elevation AMI. Pericardial effusion was present in all, and a pericardiocentesis procedure was performed in 24 patients (33%). Prior to surgery, 45 patients (63%) underwent coronary angiography, which revealed single-vessel disease in 22 (43%), double-vessel disease in 10 (27%), triple-vessel disease in 8 (17%) and left main disease in 2 (3.3%), while 24 (33%) received percutaneous coronary intervention.

The time interval between symptom onset and diagnosis of LVFWR was 53 ± 75 h. Eight patients (11%) were presented with out-of-hospital cardiac arrest, either during ambulance transport (n = 6) or while at a local hospital (n = 2). In addition, 22 (31%) were presented with in-hospital cardiac arrest that occurred in an intensive care unit (n = 10), catheterization laboratory (n = 5), emergency room (n = 4) or operating room (n = 4). For the 30 patients with preoperative cardiac arrest, cardiopulmonary resuscitation with or without mechanical circulatory support (MCS) was necessary, and return of spontaneous circulation was successfully achieved in 3 before arrival at the surgical theatre. Consequently, upon entry to the operating room, 45 patients (63%) were suffering from hypotensive shock, while 27 (38%) presented cardiac arrest (Fig. 1). Preoperatively, 19 (26%) were supported with extracorporeal membrane oxygenation (ECMO) only, 10 (14%) with an intra-aortic balloon pump (IABP) only and 6 (8.3%) with both ECMO and an IABP.

As compared with the hospital survivors, non-survivors were older, initially transported to a local hospital more frequently and more likely to develop in-hospital cardiac arrest prior to surgery or upon entry to an operating room. On the other hand, there were no inter-group differences for coronary severity, interval time from onset to LVFWR diagnosis, rate of preoperative MCS or percutaneous coronary intervention before surgery.

Surgical repair and postoperative outcomes

The operative procedures used and postoperative outcomes are summarized in Table 2. The rupture site was the LV anterior wall in 35 (49%), posterior-lateral in 28 (39%) and inferior in 9 (13%) patients. Oozing type LVFWR was identified in 42 (58%), while blow-out type was seen in the remaining 30 (42%). Sutureless repair with (n = 4, 15%) or without (n = 23, 85%) on-pump support was performed for 27 patients with oozing type LVFWR (38%). Those 4 were presented with in- or out-of-hospital cardiac arrest; thus, ECMO support was initiated, and then on-pump supported (ECMO) repair was performed. Sutured repair was chosen for 15 oozing type and 30 blow-out type LVFWR cases (62%), either with the use of an off-pump LV repair technique (n = 5, 11%) or a cardiopulmonary bypass with (n = 23, 51%) or without (n = 17, 38%) cardioplegic arrest. Concomitant coronary artery bypass grafting was performed in 8 (11%) and ventricular septal perforation repair in 5 (6.9%) patients.

Following surgery, 9 patients (13%) were supported with ECMO only, 25 (35%) with an IABP only and 22 (31%) with both ECMO and an IABP. Related to poor preoperative haemodynamics in the non-survivors, the majority of those required postoperative MCS. LV re-rupture occurred in 2 patients at 4 and 5 days, respectively, after undergoing sutureless repair. One of them was rushed to a surgical theatre under ECMO support, received redo LV sutured repair and survived, whereas the other was presented with cardiac arrest, did not respond to medical treatment with ECMO support and died thereafter. In-hospital mortality occurred in 35% of the cases (n = 25). The most common cause of in-hospital mortality was heart failure (n = 10), followed by irreparable rupture (n = 5), brain death (n = 5), infection (n = 4) and AMI (n = 1).

While there was no inter-group difference regarding location of the LV rupture, non-survivors were more likely to be presented with a blow-out rupture and therefore underwent a sutured repair procedure under a cardiopulmonary bypass, along with longer operation time and higher incidence of postoperative MCS.

Preoperative and surgical factors associated with in-hospital mortality

Univariable analysis identified associations of in-hospital mortality with age at surgery, initial ambulance transport to a local non-surgery hospital, out-of-hospital cardiac arrest, in-hospital cardiac arrest and blow-out type LV rupture. Multivariable analysis identified age at surgery (adjusted OR 2.2 for each 10-year increase, 95% CI 1.0–4.9, P = 0.044), initial ambulance transport to a local hospital (adjusted OR 4.2, 95% CI 1.2–17, P = 0020), out-of-hospital cardiac arrest (adjusted OR 8.6, 95% CI 1.4–66, P = 0.020) and in-hospital cardiac arrest (adjusted OR 6.6, 95% CI 1.6–33, P = 0.010) as independent predictors of in-hospital mortality (Supplementary Material, Table S1).

Regarding the association with haemodynamic condition before surgery, in-hospital mortality was noted in 19% of patients without cardiac arrest, and 55% with in-hospital and 63% with out-of-hospital cardiac arrest (Fig. 2).

Follow-up

Follow-up examinations were completed in 100% of the patients who survived following surgery. The average follow-up period for the 47 survivors was 6.0 ± 5.6 years (0.1–24.8 years), while the overall average follow-up term was 3.9 ± 5.4 years with a cumulative 283 patient-years. Among the in-hospital survivors, 15 patients died, with the most common cause of death unknown (n = 6), followed by heart failure (n = 3), brain death (n = 2), malignancy (n = 2), infection (n = 1) and renal failure (n = 1). Consequently, a total of 40 patients (56%) died during follow-up, with the overall 5- and 10-year survival rates 50% and 40%, respectively, while those rates for the hospital survivors were 77% and 61%, respectively (Fig. 3A and B).

DISCUSSION

Because of the rarity of LVFWR, little is known regarding early clinical results for surgical LVFWR repair or late follow-up findings. Therefore, this retrospective study was designed to review surgical experience at our centre with postinfarction LVFWR over a 25-year period to identify determinants of in-hospital mortality. The major findings of this study can be summarized as follows. In surgically treated patients with postinfarction LVFWR, the early mortality rate was high, with age at surgery, out-of-hospital or in-hospital cardiac arrest before surgery and initial ambulance transport to a local non-surgery hospital shown to be associated factors. On the other hand, the long-term prognosis for in-hospital survivors was acceptable. These highlight the importance of prompt management of haemodynamic instability at a tertiary centre to improve the outcome of affected patients.

The in-hospital mortality rate noted in the present series of cases is consistent with previous studies, which have noted a range of 0–61%. This wide range is likely due to differences in patient demographics, including age at surgery, haemodynamic state at time of arrival in the operating room (i.e. hypotensive shock or cardiac arrest) and type of LV rupture (Supplementary Tables S2 and S3). In 2001, Iemura et al. [4] reported surgical outcomes of 17 patients with LVFWR (average age 65 ± 8 years, blow-out 18%) and noted an overall in-hospital survival of 12%, which was higher in those with blow-out as compared to oozing type (33% vs 7.1%). In the same year, McMullan et al. [5] presented findings showing a higher overall in-hospital survival rate of 61% for 18 surgically treated patients (average age 64 years, blow-out 78%) and also noted that the rate for blow-out type LVFWR cases was consistently higher as compared to oozing type (71% vs 25%). Unfortunately, despite improved understanding of this entity, as well as effective diagnostic and therapeutic strategies, this trend of higher mortality rate for blow-out type LVFWR has not substantially changed over time. Additionally, in accordance with the present findings, Formica et al. [12] in 2018 and Okamura et al. [13] in 2019 reported that the in-hospital mortality rate for patients with a blow-out rupture was as high as 50%, again substantially higher than that for those with an oozing rupture.

One of the major challenges faced when treating blow-out type LVFWR is that rapid pericardial effusion accumulation and consequent cardiogenic shock and/or cardiac arrest are frequently observed at the time of presentation, which results in only limited time for therapeutic intervention. Although few studies have investigated determinants of in-hospital mortality following surgical repair of LVFWR because of the limited number of cases available for analysis, the independent association of preoperative cardiac arrest with in-hospital mortality observed in the present study is consistent with findings noted in previous reports [12, 14, 15]. The results of our study provide further evidence that not only out-of-hospital cardiac arrest but also in-hospital cardiac arrest significantly increases the risk of in-hospital mortality, even when medical staff members perform a prompt response with early or bystander cardiopulmonary resuscitation. It is therefore of utmost importance to primarily consider possible LVFWR when a patient is presented with pericardial effusion and regional contraction abnormality shown by echocardiographic assessment and quickly transport them to an operating room before development of cardiac arrest. In consideration of the high mortality rate related to a cardiac arrest condition, we believe that immediate transport to a hybrid operating room, where coronary assessment and/or intervention are possible, is justified, so as to mitigate a prolonged state of end-organ hypoperfusion. We agree with statements in the report presented by McMullan et al. [5] that it seems wise to consider cardiac catheterization and coronary intervention at a later date if required, as it is difficult to know whether LVFWR will allow enough time to safely understand the coronary anatomy of the affected patient.

The association of initial ambulance transport to a local non-surgery hospital with a higher risk of in-hospital mortality after surgery is another novel finding of the present study. Notably, in-hospital mortality for patients who were transferred from a local hospital was 46% (17 of 37), significantly higher as compared to 23% (8 of 35) for those directly brought to our centre. On the other hand, there were no differences noted between patients transported first to a local hospital and those brought straight to our centre with regard to demographics, haemodynamics at presentation, LVFWR type or operative procedure used, except for the location in our centre when a median sternotomy was necessitated for the repair procedure. Patients who were initially transported to a local hospital were more likely to develop abrupt or unpredictable haemodynamic shock or cardiac arrest, which then required a salvage surgery procedure at a catheterization unit (local hospitals 43% vs our centre 11%) or intensive care unit (29% vs 0%), before being rushed to an operating room. These findings may support the need for direct transport to a tertiary centre for patients suspected of having LVFWR related to AMI, though mechanisms related to the better early outcome of patients initially treated at our centre remain to be determined.

Because of the rarity of LVFWR, few studies have investigated long-term survival of patients who underwent a related surgical procedure [8, 9, 12–15]. The 10-year survival rate of 50% noted in the present study is quite comparable to the rate of 53% reported by Formica et al. [12] and considered to be likely acceptable given that the mean age at time surgery for our patients was ∼7 years older as compared to their cohort (75 ± 9 vs 68 ± 9 years). The relatively high survival rate noted for hospital survivors indicates that surgical intervention is the treatment of choice for this catastrophic complication of AMI. Nevertheless, meticulous follow-up examinations and appropriate treatment for potential late complications, such as late LV pseudoaneurysm and residual myocardial ischaemia, are of paramount importance.

Clinical implications

Recent reports have shown increased use of MCS for patients who develop mechanical complications following AMI [1, 16, 17]. To evaluate the prognostic impact of preoperative and postoperative MCS use in the present patients with an LV rupture following AMI, long-term survival was compared between those with and without such support. While the results showed no significant difference in long-term mortality between patients with and without use of preoperative MCS (Fig. 4A), long-term mortality for those who required postoperative MCS was significantly lower as compared to those who did not (Fig. 4B). It is considered that MCS does not always stabilize the haemodynamics of a patient with an LV rupture because of continuous bleeding from the ruptured LV or the presence of pericardial effusion leading to compression of the heart, which would explain the insignificant impact of preoperative MCS. Thus, its introduction prior to surgery does not provide clinicians adequate time for performing diagnosis and definitive treatment. The lower survival rate noted for patients who required postoperative MCS was consistent, at least in part, with findings reported by Matteucci et al. [18]. We speculate that the lower rate might be attributable to a greater degree of ventricular dysfunction and low output syndrome following a large extent of myocardial infarction and does not necessarily negate the potential benefits of MCS following LV repair.

Over the previous decade, 2 significant changes have occurred in the strategy used for treating patients with an LV free-wall rupture. Based on the above-mentioned findings, a patient suspected to have an LV free-wall rupture will not be sent to a catheter laboratory or undergo a CT scan to obtain findings for accurate diagnosis but rather is rushed to an operating room. Another change in the strategy includes aggressive use of postoperative MCS. When the present study period is split into 2 eras (1994–2010 and 2011–2023), the in-hospital mortality rate was comparable between them (41% vs 31%, P = 0.408), despite a significantly greater incidence of blow-out type LV rupture cases in the latter era (26% vs 51%, P = 0.033) (Supplementary Material, Tables S4 and S5). Notably, the in-hospital mortality of patients with a blow-out type LV rupture was significantly lower in that latter study period (86% vs 39%, P = 0.025) (Fig. 5). The better outcomes achieved for blow-out LV rupture cases in the latter era might be attributable to not only technical improvements, resulting in shorter operation, cardiopulmonary bypass and aortic cross-clamp times, as well as aggressive use of postoperative MCS, but also to a decreased rate of cardiac arrest prior to surgery as a consequence of the policy change (100% vs 65%, P = 0.143).

Limitations

The main limitation of this study is its retrospective nature; thus, the presence of both selection bias and unmeasured confounding factors cannot be excluded. In addition, the number of patients enrolled may be considered relatively small despite being the largest sample size among previous single-centre reports. Also, the inclusion of patients who underwent concomitant coronary artery bypass grafting or ventricular septal perforation may have influenced the results. However, such concomitant procedures are usually required for very sick patients who are presented with LVFWR. Finally, data for patients managed conservatively or who died without surgery are lacking.

CONCLUSION

The early mortality rate for patients with postinfarction LVFWR and who received surgical treatment was high, with associations found with age at time of surgery, out-of-hospital or in-hospital cardiac arrest before surgery, and initial ambulance transport to a local non-surgery hospital. However, the long-term prognosis for in-hospital survivors was acceptable. These findings highlight the need for prompt management of patients with haemodynamic instability at a tertiary centre to improve outcomes.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

None declared.

Conflict of interest: none declared.

ACKNOWLEDGEMENTS

None declared.

DATA AVAILABILITY

The data underlying this article will be shared upon reasonable request to the corresponding author.

Author contributions

Satoshi Kainuma: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Resources; Writing—original draft. Naonori Kawamoto: Investigation. Kota Suzuki: Data curation. Naoki Tadokoro: Investigation. Takashi Kakuta: Investigation. Ayumi Ikuta: Methodology. Kohei Tonai: Data curation. Masaya Hirayama: Data curation. Hironobu Sakurai: Data curation. Yoshiyuki Tomishima: Data curation. Kota Murai: Resources. Kenichiro Sawada: Investigation. Takamasa Iwai: Methodology. Hideo Matama: Data curation. Hiroyuki Miura: Data curation. Satoshi Honda: Data curation; Validation. Shuichi Yoneda: Data curation. Masashi Fujino: Data curation. Kazuhiro Nakao: Data curation; Investigation. Kensuke Takagi: Investigation; Methodology. Fumiyuki Otsuka: Data curation. Yasuhide Asaumi: Data curation. Yu Kataoka: Data curation; Software. Yoshio Tahara: Investigation; Resources; Validation. Teruo Noguchi: Visualization. Tomoyuki Fujita: Supervision. Satsuki Fukushima: Conceptualization; Supervision; Validation.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Francesco Formica, Wade Dimitri and the other, anonymous reviewers for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• AMI

  Acute myocardial infarction

 
• CI

  Confidence interval

 
• ECMO

  Extracorporeal membrane oxygenation

 
• IABP

  Intra-aortic balloon pump

 
• LV

  Left ventricular

 
• LVFWR

  Left ventricular free-wall rupture

 
• MCS

  Mechanical circulatory support

 
• OR

  Odds ratio

Author notes