Introduction
Extracorporeal life support (ECLS) is initiated in patients with acute refractory congestive heart failure as the initial mechanical circulatory support. Patients who show sufficient functional recovery under ECLS can wean from mechanical support, whereas those failing to show functional recovery require upgrade of mechanical support to a durable left ventricular assist device (LVAD) [
1,
2]. We recently reported the feasibility, safety, and therapeutic efficacy of central ECLS on salvage and/or bridge to recovery or to a durable LVAD for acute refractory congestive heart failure caused by myocardial infarction, myocarditis, or cardiomyopathy [
3]. However, this bridge from central ECLS to a durable LVAD strategy is associated with several positive and negative factors [
4‐
6].
In the bridge strategy, hemodynamic stabilization and left ventricular unloading are completed under central ECLS before durable LVAD implantation. Therefore, hemodynamic and structural changes by durable LVAD implantation are minimal, potentially enhancing recovery from this implant surgery [
6]. In contrast, insufficient general recovery from ECLS surgery at the time of durable LVAD implant surgery may be associated with postoperative complications, such as poor morbidity and/or infectious diseases that prolong the in-hospital stay. Additionally, intrapericardial adhesion induced by central ECLS surgery may potentiate blood product use and exacerbate coagulopathic change, which lead to a suboptimal anticoagulation state after durable LVAD implantation. Our research group recently reported that this bridge strategy was more strongly associated with 1-year composite events compared with primary durable LVAD implantation [
3]. Imamura et al. also recently reported that this bridge strategy was associated with lower 3-year mortality compared with primary durable LVAD implantation [
7]. However, the underlying mechanisms that determine the difference between primary and bridge implantation are not fully understood.
We hypothesized that adverse events associated with the bridge strategy might be predicted by background/characteristics or intraoperative variables at durable LVAD implantation surgery. To test this hypothesis, we aimed to review in-hospital and mid-term outcomes of a bridge from central ECLS to the durable LVAD strategy compared with those of primary LVAD implant surgery, and to examine predictive factors associated with adverse events after durable LVAD implantation.
Materials and methods
Study cohort and data collection
The institutional surgical database contained information on a consecutive series of 158 patients who underwent either the HeartMate II (Abbott, Chicago, IL, USA) or HeartMate 3 (Abbott) device for bridge-to-transplantation in the Department of Cardiac Surgery between April 2013 and September 2020. The patients’ medical charts, surgical reports, and referral letters were reviewed to collect study data. Clinical follow-up was completed at the end of the study in all patients. Data collection was performed in November 2020. Either patients or their legal representatives provided written informed consent for surgery and the use of their data for diagnostic and research purposes preoperatively. The study was conducted in compliance with the Declaration of Helsinki and the International Conference on Harmonization/Good Clinical Practice. The study was approved by the National Cerebral and Cardiovascular Center Institutional Review Board for Clinical Research (approval number: M30-026).
Study endpoints
The primary study endpoint was all-cause mortality and cerebrovascular accidents (CVAs) after durable LVAD implantation, and the secondary endpoints were adverse events defined in the Interagency Registry for Mechanically Assisted Circulatory Support report [
8], apart from CVAs. CVAs consisted of any stroke event, transient ischemic attack, or seizures. Stroke was further categorized by subtype (ischemic and hemorrhagic) and by severity with the modified Rankin Scale (mRS). The mRS was divided into the following severity classifications: nondisabling (mRS score 0–3) or disabling (mRS score ≥ 4) strokes. Additionally, driveline infection was defined as a positive culture from the skin and/or tissue surrounding the driveline, coupled with the requirement to treat with antimicrobial therapy when there was clinical evidence of infection, such as pain, fever, drainage, or leukocytosis [
9]. Pump failure was defined as cessation of any component of the device system to operate to its designed performance specifications or otherwise failure to perform as intended [
10]. Re-sternotomy was performed in patients with sustained chest drain output, pericardial tamponade, or deep sternal wound infection.
Indication and selection of a durable LVAD
All patients were pathologically diagnosed on the basis of biopsy specimens of the right ventricle (RV) before LVAD implantation. LVAD was implanted as bridge-to-transplantation in 157 patients who were listed in the Organ Transplantation Network, Japan before surgery, whereas one patient underwent durable LVAD implantation as the destination therapy. The primary selection of the durable LVAD device was the HeartMate II (n = 120) between April 2013 and January 2019 and the HeartMate3 (n = 38) between August 2019 and September 2020.
Cohort grouping and surgical procedure
A durable LVAD was implanted as the primary mechanical support device in 115 patients (primary group), whereas the remaining 43 underwent durable LVAD implantation as the bridge from central ECLS (bridge group). In the bridge group, the mean duration of central ECLS was 52.4 ± 44.4 days (range 3–173 days). Initial central ECLS was an extracorporeal LVAD in 30 (69.8%) patients, extracorporeal bilateral ventricular assist device in 7 (16.3%), and extracorporeal membrane oxygenation by central cannulations in 6 (1%).
The durable LVAD was implanted by the median sternotomy approach in all patients. In the bridge group, the left ventricular apex cuff of the central ECLS was completely removed to replace the cuff of the HeartMate II or HeartMate 3. Tricuspid annuloplasty was performed by using a prosthetic ring in patients with moderate or more tricuspid regurgitation preoperatively (
n = 45). Aortic valve plasty was performed by the Park’s stich [
11] in patients with mild or more aortic valve regurgitation preoperatively (
n = 19). In patients who failed to wean off cardiopulmonary bypass under durable LVAD support owing to poor RV function, mechanical RV support was added by cannulations into the main pulmonary artery and right atrium via the femoral vein, which was connected to an extracorporeal membrane oxygenation circuit (
n = 4).
Perioperative medical treatments and laboratory examinations
In patients who were on a vitamin K antagonist before surgery, prothrombin complex was administered before the skin incision to a target international normalized ratio (INR) of < 1.5. The chest was not left open in any patients at entry to the intensive care unit. Considering that early postoperative heparinization reduces hemolysis [
12], heparin infusion was started within 24–48 h after implantation, depending upon the chest drain output. This infusion was titrated by targeting an activated prothrombin time of 50–70 s until the INR reached 2.0. A vitamin K antagonist was started at days 1–2 postoperatively by targeting an INR of 2.0–3.0. Aspirin (100 mg daily) was started when the platelet count in the blood was > 100,000/mm
3.
All patients were examined by standard transthoracic echocardiography and a right heart catheter study within 1 month preoperatively and postoperatively. A blood test was performed 1 day before surgery and daily for 1 week postoperatively.
Statistical analysis
All statistical analysis was performed by using JMP software, ver. 15.2 (SAS Institute Inc., Cary, NC, USA). For comparison of preoperative background and characteristics of patients between the primary and bridge groups, continuous variables were compared by using the unpaired t test and categorical variables were compared using the chi-square test. Continuous variables are shown as mean ± standard error and categorical variables are shown as numbers and percentages. Overall survival and CVA-free survival were estimated using Kaplan–Meier curves and compared across groups using the log-rank test. Multivariable logistic regression models were used to identify the risk factors for CVAs during the follow-up period. Statistical significance was defined as a P value < 0.05.
Discussion
This study examined patients who were bridged from central ECLS and were more frequently complicated by CVAs compared with those who underwent primary implantation of HeartMate II or HeartMate 3. Survival outcome was not significantly different between the two groups. The postoperative survival rate and cumulative incidence of CVAs in this cohort are comparable with the outcomes of previous studies [
13‐
17]. Notably, this bridge strategy was the only independent predictive factor of CVAs post-implantation of HeartMate II or HeartMate 3.
Various factors could be associated with our finding that a bridge from central ECLS was the only independent predictive factor of CVAs as follows. First, ischemic and myocarditis etiologies, which predominated in the bridge group, potentiate regional motion abnormalities in the LV wall and result in formation of LV thrombus. However, idiopathic dilated cardiomyopathy, which predominated in the primary group, causes a reduction in LV wall motion globally and homogeneously. Second, LV size at the durable LVAD implantation was significantly smaller in the bridge group compared with the primary group. A smaller LV under durable LVAD is more likely to lead to formation of wedge thrombus [
18]. Third, a poor nutritional state at durable LVAD implantation, represented by serum albumin levels, in the bridge group may have been associated with CVAs post-durable LVAD implantation [
19]. Fourth, refractory cardiogenic shock before admission causes severe multiorgan dysfunction, which might not have fully recovered before durable LVAD implantation. Finally, a prolonged operation time with predominant use of blood products in the bridge group could have been the cause of CVAs. Adhesiolysis of the pericardial space was additionally required before durable LVAD implantation in all patients in the bridge group, whereas only 20 patients required intrapericardial adhesiolysis owing to previous cardiac procedures in the primary group. Complicated and long operation requiring a large amount of blood products could result in postoperative delayed heparinization, which showed a strong tendency to evoke CVAs. Our institute recently showed that hemolysis in HeartMate II was correlated with delayed heparinization. Thus, we consider that optimum coagulation profile should be established as soon as the surgery is completed to prevent thrombogenic complications, such as CVAs or hemolysis.
The finding that CVAs in the bridge strategy could be caused by etiology, LV size, nutritional state, or complicated surgery can be explained by the timing and severity of CVAs in our study. All CVAs in the bridge group occurred within 1 year after LVAD implantation. Additionally, hemorrhagic and disabling stroke predominated in the bridge group, representing a coagulopathic state potentially associated with a poor nutritional state. However, once the general condition became stabilized under durable LVAD support, outcomes were not different between the bridge and primary groups. Therefore, physicians need to pay attention to these factors for patients under the bridge strategy. This includes prompt establishment of anticoagulant therapy for patients with regional LV wall motion abnormalities and/or a small LV size, a durable LVAD implant under a normalized nutritional state, or a meticulous adhesiolysis procedure.
This study may have a limitation of its design in which durable LVADs, apart from HeartMate II or HeartMate 3, were not included. A variety of durable LVADs have been implanted in our institute depending on the feasibility or availability of the devices in the last 20 years. Among them, HeartMate II and HeartMate 3 were the most safely used primary choice of devices throughout the study period. Inclusion of other devices in this study would have substantially increased the bias in surgical indications, the implantation procedure, or the postoperative management. Additionally, our study may have been limited by the treatment strategy in Japan, in which a durable LVAD is implanted for patients who are listed as transplant candidates before surgery. Therefore, ECLS is established first for patients with acute refractory heart failure to assess candidacy for transplant followed by durable LVAD implantation. However, this bridge to candidacy strategy is practiced worldwide. Additionally, mechanisms underlying CVAs in the bridge strategy, which were described in this study, could have important implications in use of durable LVADs. As a statistical limitation, in this cohort, CVAs occurred only in 22 cases, limiting up to 4 or 5 factors which can be included in the multivariate analysis. We selected factors that were included in the multivariate analysis, based on P values in the univariate analysis and on cofounding tendency among the factors.
In conclusion, a bridge from central ECLS was a risk factor of CVAs after durable LVAD implantation, and these CVAs could have been associated with a small LV size, poor nutritional state, or intraoperative blood product use. Attention to these factors would improve outcomes of patients who have a bridge from central ECLS.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.