CRCM Journal

Postcardiotomy shock (PCS) remains a critical complication after cardiac surgery and portends high mortality. It may occur in up to 6% of patients after cardiac operations. Despite aggressive treatment including high-dose vasopressors and temporary mechanical circulatory support (MCS), mortality remains as high as 75% [1]. Venoarterial extracorporeal membrane oxygenation (VA-ECMO) is a mainstay treatment for PCS, used in up to nearly 4% of patients after cardiac surgery, but associated complications such as bleeding, limb ischemia, or stroke make it high risk [2]. Despite increased use of VA-ECMO over the last couple decades, mortality remains high [3]. It is a technology that requires many resources and experience and results in many patients being transferred to tertiary centers for either initiation, or management after initiation at outside facilities.

There are many studies investigating VA-ECMO for PCS including timing of initiation, cannulation techniques, and risk factors [1,4– 6]. Several studies have reported experiences transferring critically ill patients to tertiary centers, with an increasing number of reports on patients on VA-ECMO support [7–13]. No studies have focused on the implications of transferring patients with PCS who have already been cannulated for VA-ECMO at outside facilities. The objective of this study was to compare outcomes for patients with PCS treated with VA-ECMO who were either cannulated in a tertiary center or cannulated at an outside hospital (OSH) and subsequently transferred to a tertiary center.

Methods

Study design

This study was approved by the Institutional Review Board of the University of Minnesota Medical Center (protocol #1703M11301). Requirement for consent was waived due to the retrospective nature of the study. Adult patients (age ≥ 18 years old) who underwent cardiac surgery and subsequently developed PCS and were treated with VA-ECMO between January 2017 through June 2022 were identified in an internally and prospectively kept institutional ECMO registry. This included patients who were cannulated at an OSH and transferred to our tertiary center for further management. Cannulation techniques included percutaneous peripheral cannulation in the cardiac catheterization laboratory (CCL) or intensive care unit (ICU), or central cannulation or peripheral surgical cutdown in the operating room (OR). The decision to initiate VA-ECMO support was at the discretion of the surgeon; there were no standardized protocols for VA-ECMO initiation. Postcardiotomy shock was defined as shock after cardiac intervention by median sternotomy or thoracotomy without evidence of alternative causes such as sepsis or hemorrhage. We excluded lung transplant and left ventricular assist device (LVAD) recipients but included heart transplant recipients. The electronic medical record was retrospectively reviewed to confirm the index cardiac operation and determine cardiac surgery end-time, use of additional MCS prior to VA-ECMO initiation, cardiopulmonary bypass (CPB) time, aortic cross clamp time, circulatory arrest time, laboratory values at time of ECMO initiation, discharge location, and 30-day readmission. Calculated variables included total time on VA-ECMO, and time from cardiac surgery to VAECMO initiation. Delay in VA-ECMO initiation was defined as initiation of support after the index operation was done and the patient had left the OR. Patients were divided into two cohorts based on where cannulation occurred: in-hospital (INH) or OSH.

Statistical methods

All statistical analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc., Cary, NC, USA). A p-value of <0.05 was considered statistically significant. Baseline characteristics were reported by a mean ± standard deviation for continuous variables and percentage for categorical variables. Univariate analysis comparing INH vs OSH cannulation were done using Chi-square test or Fisher’s Exact tests for any cell count less than 5 for categorical variables, and two-sample Wilcoxon tests for continuous variables as they were not normally distributed. Primary outcome was 30-day mortality. Multivariate logistic regression models were used to identify predictors of 30-day mortality in those treated with VA-ECMO for PCS. Kaplan-Meier plots were generated to estimate 30-day and 1-year mortality for the INH vs OSH cannulation cohorts and were assessed using the log-rank test.

Results

A total of 105 patients on VA-ECMO for PCS were treated at our center. Fifty-five (52%) patients were cannulated at our center (INH), and 50 (48%) were cannulated at an OSH and subsequently transferred to our center. All patients except for one had their cardiac operation and VA-ECMO initiated at the same facility. (Figure 1) depicts the locations of the referring hospitals. Overall median age was 62 years [IQR 52-69]; there were 24 patients (22.9%) ages 70-87 years old. (Table 1) shows the univariate analysis comparing baseline characteristics between the two cohorts. Patients cannulated at OSH were older (p=0.004) and had lower rates of chronic kidney disease (CKD) (p<0.001), peripheral arterial disease (PAD) (p=0.005), and prior venous thromboembolism (VTE) (p=0.027).

Figure 1: Map of Minnesota and surrounding states depicting the locations of referring centers who transferred patients with PCS on VA-ECMO. There were an additional six referring hospitals in the Minneapolis/St. Paul metropolitan area.

Table 1: Baseline characteristics of patients with postcardiotomy shock treated with VA-ECMO

Table 2 shows differences in perioperative characteristics. Overall, 71/105 (68%) patients were centrally cannulated and 34/105 (32%) had peripheral cannulation; this was similar between the two cohorts (central cannulation at INH 65.5% vs OSH 68.0%, p=0.782). There were similar trends in the facility location of where VA-ECMO cannulation occurred (p=0.162); most patients were cannulated in the OR (72.7% INH vs 88% OSH), with slightly higher percentage being cannulated in the ICU or CCL at INH. There were differences in additional temporary MCS utilized between the two groups (p=0.033); those cannulated at OSH had higher use of intra-aortic balloon pump (IABP), (p=0.048). Index operation differed between the two groups (p<0.001); most patients cannulated at OSH had undergone coronary artery bypass graft (CABG) (40.0%) and multi-procedure operations (22.0%), while those cannulated INH had undergone multi-procedure operations (25.5%), OHT (23.6%), or aortic operations (20.0%). More patients from OSH had redo sternotomy, approaching significance (12% vs 1.8%, p=0.052). Postoperative cardiac arrest events were similar between the two groups (INH 12.7% vs 18.0% OSH, p=0.453). There were no differences in CPB, cross clamp, or circulatory arrest times between the two groups. There were no differences in delay of VA-ECMO initiation, or median lactate, creatinine, aspartate transaminase (AST), or alanine transaminase (ALT) at initiation between the two cohorts. Median time on VAECMO was 102.1 hours [IQR 61.3-183.8]; there was no difference between the two groups (p=0.312).

Table 2: Perioperative characteristics for patients with postcardiotomy shock treated with VA-ECMO

For the entire cohort, 30-day mortality was 59% (62/105); this was equal to in-hospital mortality (Table 3). Fifty-seven percent (60/105) survived to decannulation, and 41.0% (43/105) survived to discharge. For those age ≥ 70 years, 6/24 (25%) survived to discharge. There was no statistical difference between patients cannulated at OSH versus INH with regards to 30-day mortality (66% vs 53%, p=0.167), survival to decannulation (52% vs 62%, p=0.310), or survival to discharge (34% vs 47%, p=0.167). Postoperative length of stay was longer for those cannulated INH compared to OSH (17 days vs 12 days, p=0.021). For those who survived to discharge, overall the mean postoperative length of stay was significantly longer (41.8 days vs 8.7 days, p<0.001). There were no differences in postoperative complications including stroke, acute limb ischemia, tracheostomy, or dialysis, and no differences in discharge location. Multivariate analysis identified age (OR 1.04, CI 1.00-1.08, p=0.040), BMI (1.20, CI 1.09-1.32, p<0.001), and post-operative dialysis (OR 4.26, CI 1.59-11.45, p=0.004) as risk factors for increased 30-day mortality (Table 4). Kaplan-Meier estimates for 30-day survival between INH vs OSH were 49% vs 36% (p=0.126) (Figure 2), and for 1-year survival, 38% vs 30% (p=0.163), respectively (Figure 3).

Table 3: Postoperative characteristics for patients with postcardiotomy shock treated with VA-ECMO

Table 4: Multivariate logistic regression identifying predictors of 30-day mortality for patients treated with VA-ECMO for PCS.

Figure 2: 30-day survival for those cannulated at OSH vs INH. 30-day survival estimates:  INH: 49%, OSH: 36%.

Figure 3: 1-year survival for those cannulated at OSH vs INH. 1-year survival estimates:  INH: 38%, OSH: 30%.