Initial experience with CytoSorb therapy in patients receiving left ventricular assist devices

2.1 Study population and design

This study conforms to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the local Ethical Committee (20-9283-BO). We retrospectively reviewed the data of 207 consecutive patients who underwent LVAD implantation at our institution between August 2010 and January 2020. All patients who received either HeartMate III™ (St. Jude Medical—Abbott, MN, USA) or a HeartWare™ (Medtronic, MN, USA) LVAD devices were eligible. Intraoperatively, the Cytosorb filter was used at the discretion of the operators depending on the following clinical criteria at the day of surgery: fever, severely elevated inflammatory markers, and/or hemodynamic instability requiring inotropic support. Cytosorb filter was included via side arm directly into the CPB circuit and the filtration was maintained over the whole CPB-time.

Data were obtained from a prospectively collected institutional LVAD database that includes detailed information on patients’ demographics, baseline clinical characteristics, and their laboratory and hemodynamic parameters, as well as intraoperative variables and postoperative outcomes. Patients were followed up based on information available in their electronic medical records, as well as through telephone interviews. Minimal period of follow-up for all patients was until discharge or death.

The study population was divided into two groups: Group 1—patients who received CytoSorb therapy (n = 72) and Group 2—patients who did not receive CytoSorb therapy (n = 135). Baseline patient characteristics and intraoperative data were collected and compared between groups. Laboratory parameters including white blood cell (WBC) count, C-reactive protein (CRP), creatinine, blood urea nitrogen (BUN), total bilirubin, alanine transaminase (ALT), lactate dehydrogenase (LDH), procalcitonin (PCT), and interleukin 6 (IL-6) were also collected, both preoperative and postoperative (within 7 days following surgery).

2.2 Outcome measures

The primary endpoint was overall survival after LVAD implantation. Secondary endpoints were major adverse events in-hospital and during follow-up. Changes in laboratory parameters were also investigated.

2.3 Statistical analysis

Continuous variables were checked for normality using the Shapiro-Wilk test and the difference between groups was tested with the t-test or Mann-Whitney U-test accordingly. Categorical variables are expressed as frequency and proportion, and differences were assessed with the chi-square test. To account for baseline differences between groups, genetic propensity score matching was performed using the R package “MatchIt” (v 3.0.2.). Matching factors included all unbalanced variables (p ≤ .1), that is, coronary artery disease, hyperlipidemia, intra-aortic balloon pump, operative time, CPB time, HeatWare LVAD model, tricuspid valve surgery, ventricular septum defect closure, preoperative creatinine, and preoperative blood urea nitrogen. Two-way ANOVA was used to determine the effect of LVAD implantation and CytoSorb therapy on the levels of laboratory parameters. Kaplan–Meier curves were used to study the effect of CytoSorb therapy on survival during follow-up, with the log-rank test used for statistical significance in this estimate. To study the occurrence of major adverse events during follow-up while taking into account the different follow-up of the two cohorts, incidence rates per person-year were calculated and compared. A p-value of < .05 was considered statistically significant. All analyses were performed using R Statistical Software (version 4.0.2. 2020-06-22, Foundation for Statistical Computing, Vienna, Austria).

3.1 Baseline characteristics and intraoperative data

A total of 207 patient with end-stage heart failure underwent implantation with a LVAD at our institution between August 2010 and January 2020. Baseline characteristics and intraoperative data are given in Tables 1 and 2, respectively. Before propensity score matching, patients in the CytoSorb group had a lower proportion hyperlipidemia (36.1% vs. 55.6%, p = .008) and were less often on intra-aortic balloon pump (5.6% vs. 17.0%, p = .019). They also had shorter operative times [174 min (149–238) vs. 214 min (185–256), p = .001] and CPB time [72 min (62–99) vs. 85 min (73–108), p < .001]. Patients who received CytoSorb also more commonly received Heartmate III (22.2% vs. 3.7%, p < .001) and had concomitant tricuspid valve surgery (6.9% vs. 0.7%, p = .011). Propensity matching resulted in a selection of 72 patients in the CytoSorb group and 40 patients in the non-CytoSorb group; baseline and intraoperative variables were comparable between the two groups.

3.2 Laboratory parameters

Pre- and postoperative laboratory values are presented in Table 3. A graphical presentation with pairwise comparisons is given in Figure 1. All laboratory parameters were similar between both groups (Figure 1). Likewise, two-way ANOVA showed that CytoSorb therapy had no influence of laboratory parameters. LVAD implantation was associated with a significant increase in WBC (p = .033) and CRP (p = .001), and IL6 (p < .001), whereas all other parameters remained stable. No significant interaction terms were found between CytoSorb and timepoint (pre- vs. postoperative), indicating that CytoSorb therapy did not modify the response of the studied laboratory parameters to LVAD implantation.

3.3 Overall survival

In-hospital mortality occurred in 19 (26.4%) patients who received CytoSorb, in contrast to 6 (15.0%) patients in the matched cohort of patients who did not receive CytoSorb (p = .250) (Table 4). Over a median follow-up of 1.18 years (interquartile range (IQR) 0.30–3.04 years) in the CytoSorb group and 2.67 years (IQR 0.37–5.33, p = .021) in the non-CytoSorb group, there was no difference in overall survival between both groups (log rank: p = .580) (Figure 1). Causes of death were similar between both groups (Table 4).

3.4 Major adverse events

In-hospital major adverse events are given in Table 5. In the matched analyses, the incidence of respiratory failure (54.2% vs. 30.0%, p = .024), the need for mechanical ventilation for longer than 6 days post implant (50.0% vs. 27.5%, p = .035), and tracheostomy (31.9% vs. 12.5%, p = .040) were all higher in patients who received CytoSorb therapy (Table 6). No other significant differences were observed with regard to major adverse events (Figure 2).

4.1 Main findings

In this study, we analyzed the outcomes of 112 propensity score-matched patients with advanced heart failure who underwent LVAD implantation between August 2010 and January 2020. We found that WBC, CRP, and IL-6 significantly increased with LVAD implantation and that CytoSorb did not influence this response. In-hospital mortality and overall survival during follow-up were similar with CytoSorb. However, patients treated with CytoSorb were more likely to develop respiratory failure, need mechanical ventilation for longer than 6 days post implant, and require tracheostomy during hospitalization. During follow-up, no significant differences were observed with regard to major adverse events. Overall, our results showed that CytoSorb might not convey a significant morbidity or mortality benefit for patients undergoing LVAD implantation.

4.2 Comparison with the literature

The use of LVADs in patients with end-stage heart failure who are not eligible for heart transplantation has been increasing over the last decades.¹² While this increase has been accompanied by an improvement in LVAD outcomes and a reduction in LVAD-related complications, this technology still carries some potentially major risks. Extensive research in the field has underlined the increased inflammatory profile which patients with heart failure present with.^{13, 14} The release and the levels of pro-inflammatory cytokines in heart failure have been found to be directly proportional to cardiac performance and deterioration, allowing to follow its progression.^{9, 10, 15-17} The pre-existing inflammatory profile of LVAD recipients, coupled with the inflammation derived from LVAD implantation and from CPB, expose the patients to a plethora of inflammatory outcomes, thus offering some theoretical context to the use of CytoSorb in this cohort of patients.^{3, 18-20} Indeed, in line with the published literature, our results illustrate an increase in inflammatory markers, namely CRP (p = .001) and IL-6 (p < .001), coupled with an increase in WBC (p = .033) following LVAD implantation.

The outcomes of CytoSorb application in cardiac surgery have been reported in different trials, leading to inconsistencies in results and failing to demonstrate its efficacy in improving patient outcomes.^{11, 14, 15} Our results have been consistent with the latest published trials in cardiac surgery, illustrating that CytoSorb did not modify the increase in postoperative inflammatory markers following LVAD implantation. In their pilot randomized control trial, Poli et al assessed differences in cytokines levels in 30 elective cardiac surgery patients.¹¹ Levels of several cytokines including IL-1α, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-10, TNF-α, and IFN-γ were measured both intraoperatively and postoperatively, with no significant decrease reported in the CytoSorb group.¹¹ Similarly to our findings, reported levels of in-hospital mortality were also unvaried by CytoSorb use. These results have also been supported by a more recent randomized trial by Taleska Stupica et al comparing the use of CytoSorb versus glucocorticoids during CPB surgery.¹⁴ Indeed, when compared to the control group, the CytoSorb group did not present with any significant changes in 30-day mortality and inflammatory markers including TNF-α, IL-1β, IL-6, and IL-8. Similar findings were also reported by a previous trial published by Bernardi et al who in 32 patients undergoing CPB found no significant differences in perioperative values of IL-6, IL-18, IL-1β, TNF-α, and HMGβ1.¹⁵ The only exception in the study were postoperative levels of IL-10, which were increased in the control group 24- and 48-h post-bypass but not in the CytoSorb group, although no elaboration on its possible implications was offered.¹⁵ It is worth to note that although our results align with the previously mentioned three trials, our study was only able to include WBC, CRP, IL-6, and PCT levels, thus potentially not providing a full inflammatory picture of our patients.

Nevertheless, the outcomes of severely ill patients undergoing cardiac surgery with concomitant CytoSorb seem to suggest the existence of a potential advantage. In patients with infective endocarditis, Träger et al in a case series¹³ and Haidari et al in a retrospective study of 58 patients¹⁷ found levels of IL-6 to be reduced postoperatively with CytoSorb. Sepsis-related death was also found to be significantly lower with the use of CytoSorb, although overall 30-day mortality was unvaried.¹⁷

With regard to overall survival and mortality, data in the literature on the use of CytoSorb in LVAD patients scarce and results inconsistent. Our results show no significant difference in mortality and overall survival with the use of CytoSorb in LVAD implantation. Taleska Stupica et al also reported 1-year mortality of their patients receiving CytoSorb not be significantly different when compared to the control group.¹⁴ Nevertheless, unlike the previously mentioned randomized control trials, we found that patients treated with CytoSorb were more likely to develop respiratory failure, need mechanical ventilation for longer than 6 days post-implant, and require tracheostomy during hospitalization. This contrasts with reports of the use of CytoSorb in non-cardiac surgical setting, where it has been suggested to potentially help improve lung function through pulmonary metabolism modulation.^{18, 19} Furthermore, potential advantages of CytoSorb use with continuous renal replacement therapy in COVID-19 patients has been reported, with improved lung function (paO₂/FiO₂ ratio) 30-day post-treatment.²⁰ However, the validity and accuracy of these small-scale studies results is low and makes it difficult to generalize them.

4.3 Study limitations

This study was a retrospective analysis from a single medical center over a span of 10 years. Results might therefore have been affected by the decision criteria used to select patients for LVAD implantation, which were likely influenced by the growing experience of the center with these procedures and their management. Additionally, the indications for Cytosorb implementation, duration of treatment, and patients’ selection were at the discretion of the operators and were unfortunately not available in the database. Furthermore, given the long timeframe of this study, changes in surgical, anesthesiology, and nursing staff might have had an influence on our results. Our study also did not take into consideration the wide range of ILs, and pro- and anti-inflammatory markers which form part of inflammatory processes.