Use of intraoperative haemoadsorption in patients undergoing heart transplantation: a proof‐of‐concept randomized trial

Abstract

Aims

The aim of this trial was to compare the clinical effects of intraoperative haemoadsorption versus standard care in patients undergoing orthotopic heart transplantation (OHT).

Methods and results

In a randomized, controlled trial, OHT recipients were randomized to receive intraoperative haemoadsorption or standard care. Outcomes were vasoactive‐inotropic score (VIS), frequency of vasoplegic syndrome (VS) in the first 24 h; post‐operative change in procalcitonin (PCT) and C‐reactive protein (CRP) levels; intraoperative change in mycophenolic acid (MPA) concentration; frequency of post‐operative organ dysfunction, major complications, adverse immunological events and length of in‐hospital stay and 1‐year survival. Sixty patients were randomized (haemoadsorption group N = 30, control group N = 25 plus 5 exclusions). Patients in the haemoadsorption group had a lower median VIS and rate of VS (VIS: 27.2 [14.6–47.7] vs. 41.9 [22.4–63.2], P = 0.046, and VS: 20.0% vs. 48.0%, P = 0.028, respectively), a 6.4‐fold decrease in the odds of early VS (OR: 0.156, CI: 0.029–0.830, P = 0.029), lower PCT levels, shorter median mechanical ventilation (MV: 25 [19–68.8] hours vs. 65 [23–287] hours, P = 0.025, respectively) and intensive care unit stay (ICU stay: 8.5 [8.0–10.3] days vs. 12 [8.5–18.0] days, P = 0.022, respectively) than patients in the control group. Patients in the haemoadsorption versus control group experienced lower rates of acute kidney injury (AKI: 36.7% vs. 76.0%, P = 0.004, respectively), renal replacement therapy (RRT: 0% vs. 16.0%, P = 0.037, respectively) and lower median per cent change in bilirubin level (PCB: 2.5 [−24.6 to 71.1] % vs. 72.1 [11.2–191.4] %, P = 0.009, respectively) during the post‐operative period. MPA concentrations measured at pre‐defined time points were comparable in the haemoadsorption compared to control groups (MPA pre‐cardiopulmonary bypass: 2.4 [1.15–3.60] μg/mL vs. 1.6 [1.20–3.20] μg/mL, P = 0.780, and MPA 120 min after cardiopulmonary bypass start: 1.1 [0.58–2.32] μg/mL vs. 0.9 [0.45–2.10] μg/mL, P = 0.786). The rates of cardiac allograft rejection, 30‐day mortality and 1‐year survival were similar between the groups.

Conclusions

Intraoperative haemoadsorption was associated with better haemodynamic stability, mitigated PCT response, lower rates of post‐operative AKI and RRT, more stable hepatic bilirubin excretion, and shorter durations of MV and ICU stay. Intraoperative haemoadsorption did not show any relevant adsorption effect on MPA. There was no increase in the frequency of early cardiac allograft rejection related to intraoperative haemoadsorption use.

Introduction

Orthotopic heart transplantation (OHT) has remained the gold standard therapy for end‐stage heart failure for decades. ¹ However, there has been an expansion in candidate acceptance criteria for OHT over recent years, positioning recipients at higher risk in terms of early post‐transplant complications. ¹ , ² Indeed, the second most frequent cause of death in the first 30 days after OHT is multiple organ failure. ³ The dominant component of post‐transplant multiorgan dysfunction is severe vasoplegia and consecutive haemodynamic instability, which substantially increases the risk of developing more organ dysfunctions. ⁴ , ⁵ The complex pathophysiology of vasoplegic syndrome (VS) involves coexisting pathways of endogenous vasopressin depletion, dysregulated inflammatory response, and endothelial dysfunction resulting in excessive nitric oxide production and loss of vascular tone. ⁴ , ⁶ To date, there are no specific pharmacological treatments which have been shown to be effective for the control or prevention of cardiopulmonary bypass (CPB) associated VS. ⁶

Extracorporeal haemoadsorption is a blood purification technology with confirmed adsorption capacity for cytokines, chemokines, bilirubin, myoglobin, plasma free haemoglobin, and various pharmacological agents. ⁷ , ⁸ , ⁹ Recent clinical investigations have reported reduced sepsis related mortality, less bleeding complications related to adsorption of direct acting oral anticoagulants or P2Y₁₂ inhibitors, and faster recovery of haemodynamics and organ function in patients undergoing complex cardiac surgeries when intraoperative haemoadsorption has been applied. ⁸ , ¹⁰ , ¹¹ , ¹² , ¹³ Similarly, our previous observational study showed that OHT recipients who were treated with haemoadsorption intraoperatively experienced significantly reduced post‐operative vasopressor requirements and favourable trends in clinical outcome. ¹⁴ While the number of randomized and observational studies for the evaluation of haemoadsorption in cardiac surgery has increased in the last 10 years, published results remain controversial regarding clarifying the clinical utility of this intervention in terms of post‐operative morbidity and mortality. ¹⁵ Considering the unique pathophysiological environment of OHT, the presumed benefit of intraoperative haemoadsorption during CPB could be based on control of the immune system dysregulation along with endogenous vasoactive substance overproduction. To date, there have been no published RCTs in field of OHT, which have analysed the relationship between intraoperative haemoadsorption and clinical outcome.

The aim of this randomized controlled trial (RCT) was to compare the effects of intraoperative haemoadsorption versus standard medical care on the severity of early post‐operative haemodynamic instability, frequency of post‐operative organ dysfunctions, early graft rejection, and length of hospital stay in patients undergoing OHT.

Methods

Study design and patients

This prospective, single‐centred, open‐label RCT was approved by the Semmelweis University Regional and Institutional Committee of Science and Research Ethics (approval number: 246/2016) and was conducted in accordance with the Declaration of Helsinki. The study was registered at ClinicalTrials.gov (identifier: NCT03145441) on 9 May 2017, before initiation of patient enrolment. The trial was conducted in accordance with the original protocol. Adult OHT candidates registered on the waiting list (age ≥18 years) with United Network for Organ Sharing (UNOS) Status 6 were eligible for inclusion during the study period between April 2018 and December 2021. Exclusion criteria were ‘high urgency’ status, re‐transplantation, long‐standing hospitalization, inotrope dependence, mechanical circulatory support, and progressive end‐organ failure prior to OHT. Written informed consent was obtained from all participants. Randomization was performed prior to first patient recruitment using a computerized random number generator and the randomization scheme of 60 subjects was concealed from the research group. Eligible patients were allocated into one of two study groups (i.e. control group and haemoadsorption group) in real time before their consent. However, they were blinded for the intraoperative treatment selection at the time of information and consent.

Perioperative patient management

All patients involved in this trial received standardized anaesthetic, surgical and post‐operative intensive care in accordance with the institutional protocol. Induction of anaesthesia consisted of sufentanil, propofol and atracurium, and a combination of sevoflurane and propofol continuous intravenous infusion administered for maintenance. Intraoperative monitoring of patients was based on the standards of cardiac anaesthesia extended with pulmonary artery (PA) pressure monitoring via a PA catheter and continuous transoesophageal echocardiography. Non‐pulsatile, mild hypothermic CPB was applied for all participants using a roller‐pump (SORIN C5 Perfusion System, Sorin Group Deutschland GmbH, Munich, Germany) and a membrane oxygenator (SORIN Inspire P8, Sorin Group Italia Srl, Mirandola, Italy) primed with 1250 mL of Ringer's lactate, 100 mL of mannitol, and 60 mL of sodium bicarbonate 8.4%. Perfusion was maintained with flow rate of 2.4 L/min/m² and a mean arterial pressure (MAP) of 60–80 mmHg throughout the CPB period. The clinical management of unfractionated heparin anticoagulation, haemodynamic, temperature, and metabolic targets during CPB was based on institutional standards.

The basic pharmacological components of haemodynamic management were noradrenaline as first‐line and argipressin as second‐line vasopressors, and dobutamine and milrinone as inotropic agents. Argipressin was indicated in cases where noradrenaline requirements were ≥0.3 μg/kg/min. Inhalational nitric oxide was given routinely from the beginning of CPB weaning and extended for the subsequent post‐CPB/post‐operative period depending on actual pulmonary vascular resistance and right ventricular function. Invasive PA pressure monitoring was regularly continued over the first post‐operative 48 h. Cardiac allograft function follow‐up was performed with echocardiography (transthoracic or transoesophageal) 24 hourly during the first 5 post‐transplant days, and on a weekly basis thereafter. Perioperative coagulopathy was monitored by conventional static and dynamic haemostasis tests (rotational thrombelastometry, ROTEM™, Tem International GmbH, Munich, Germany; multiple electrode aggregometry, Multiplate™, Roche Diagnostics International Ltd, Rotkreuz, Switzerland) and was treated by differential haemostasis management using factor concentrates and blood products. Packed red cell transfusion trigger criteria were defined as a haemoglobin <7.0 g/dL during CPB and <8.5 g/dL for the post‐CPB and post‐operative periods.

Immunosuppression therapy consisted of mycophenolate mofetil (MMF), methylprednisolone, anti‐thymocyte globulin and tacrolimus. The institutional protocol for perioperative immunosuppression of OHT used in this trial is summarized in Table S1 . Cardiac allograft rejection was followed up with endomyocardial biopsy (EMB) weekly during the first month after the OHT. ¹⁶

Intraoperative haemoadsorption treatment

Eligible patients were randomly selected to receive either standard OHT care plus intraoperative haemoadsorption or standard OHT care alone. Haemoadsorption was managed using a CytoSorb™ 300 mL cartridge (CytoSorbents™, Monmouth Junction, NJ, USA) for a one‐cycle treatment during the entire period of CPB. The haemoadsorption cartridge was integrated into the CPB circuit (see Figure S1). Considering the therapeutic drug adsorption potential of CytoSorb™ ¹⁷ , an additional prophylactic dose of antibiotic was administered after the start of CPB. Perioperative dosing of immunosuppressants remained unchanged and were identical in both groups.

Mycophenolic acid measurements

Venous blood was collected from OHT recipients into EDTA tubes for mycophenolic acid (MPA; active metabolite of MMF) assessment at two pre‐defined time points: 5 min prior to CPB initiation and 120 min after CPB start. MPA was quantified by particle‐enhanced turbidimetric inhibition immunoassay (PETINA, Siemens Dimension® System MPAT, Siemens Healthcare GmbH, Erlangen, Germany; detection limit <0.1 μg/mL).

Outcome parameters

The primary outcome of this trial was early post‐operative haemodynamic instability quantified by the vasoactive‐inotropic score (VIS), frequency of VS and length of vasopressor need. VIS was calculated according to the formula: VIS = dopamine dose (μg/kg/min) + dobutamine dose (μg/kg/min) + 100× adrenaline dose (μg/kg/min) + 10× phosphodiesterase inhibitor dose (μg/kg/min) + 100× noradrenaline dose (μg/kg/min) + 10 000× vasopressin dose (U/kg/min) ¹⁸ based on the mean doses in the post‐operative first 24 h for each agent. VIS was considered as ‘high’ if values ≥30 points, representing a higher risk for unfavourable outcomes. ¹⁹ Quantitative criteria of VS were mean noradrenaline requirements ≥0.3 μg/kg/min and need for argipressin supplementation at any dose to achieve a MAP >60 mmHg assessed over the first 24 h.

Secondary outcome parameters were defined as the inflammatory response characterized by a 72‐h change in procalcitonin (PCT) and C‐reactive protein (CRP) levels; duration of MV; surgery associated bleeding and reoperation for bleeding; frequency and severity of acute kidney injury (AKI) classified by applying the KDIGO creatinine‐based definition criteria for the first 5 post‐operative days ²⁰ ; 24‐h per cent change in bilirubin level using the equation: PCB = ([post‐CPB 24‐h bilirubin level (mg/dL)] − [pre‐operative bilirubin level (mg/dL)]/[pre‐operative bilirubin level (mg/dL)]) × 100, frequency of early sepsis screened for in the first 5 post‐operative days; length of ICU and hospital stay; intraoperative change in MPA plasma concentration; early allograft rejection; 30‐day mortality rate and 1‐year survival. Biomarkers of inflammatory response and creatinine clearance as well as the total bilirubin serum concentration were quantified from venous blood samples collected at designated time points using standard validated laboratory measurements.

Statistical analysis

Because of the lack of published RCTs performed in OHT patients with a similar primary outcome, no formal sample size calculation was performed. Based on our regular OHT activity we assumed that including 60 patients (30 per group) in a study over 3 years would be feasible.

All statistical tests were performed using IBM SPSS Statistics for Windows, version 28.0.1.0 (IBM Corp., Armonk, NY, USA). Continuous variables were tested with the Shapiro–Wilk test for normality. Descriptive statistics of data were displayed as median [interquartile range], mean ± standard deviation, and number of patients and frequency where appropriate. Mann–Whitney U test, two‐sample t‐test, χ² test or Fisher's exact test were performed for the univariate analysis of group comparisons. The comparative analyses of within‐subjects changes in the cohort were accomplished with the Wilcoxon signed‐rank test. To evaluate the impact of intraoperative haemoadsorption on the early post‐operative VS a multivariate, logistic regression, backward elimination, likelihood‐ratio method was performed. One year follow‐up was completed for all participants and included an estimated one‐year survival using the Kaplan–Meier method. Equality testing of survival curves was accomplished with a log‐rank test applying the Mantel–Cox method. Statistical significance was defined as a P value of 0.05 in all tests.

Results

During the study period, 165 patients were assessed for eligibility. Sixty patients were randomized to the control (N = 30) and haemoadsorption (N = 30) groups, but five patients from the control group had to be excluded. The reasons for exclusion and details of the study flowchart are summarized in Figure 1 . Baseline clinical characteristics and intraoperative factors were similar in both groups (Table  1 ); however, the pre‐transplant use of amiodarone was less frequent in the control group than in the haemoadsorption group. The demographic and baseline characteristics are depicted in Table 1 .

Figure 1.

Patient selection flowchart. OHT, orthotopic heart transplantation.

Table 1.

Demographic and baseline characteristics of the study population

	Control group (N = 25)	Haemoadsorption group (N = 30)	P‐value
Pre‐operative variables
Recipient age, year	56 [48–60]	56 [47–61]	0.839
Donor age, year	46 ± 9	41 ± 11	0.355
Body mass index, kg/m2	26.9 ± 4.8	25.4 ± 3.3	0.084
Female sex, n	10 (40.0%)	15 (50.0%)	0.458
Diabetes mellitus, n	6 (24.0%)	5 (16.7%)	0.521
Chronic kidney disease, n a	10 (40.0%)	13 (43.3%)	0.803
Chronic anaemia, n b	10 (40.0%)	9 (30.0%)	0.437
ACEI/ARB, n	10 (40.0%)	18 (60.0%)	0.140
ARNI, n	9 (36.0%)	12 (40.0%)	0.761
Beta receptor blocker, n	21 (84.0%)	28 (93.3%)	0.394
Amiodarone, n c	3 (12.0%)	11 (36.7%)	0.061
PVR, Wood unit	2.4 [1.2–3.5]	2.7 [1.9–4.4]	0.257
IMPACT score, point	4 [2.5–5.0]	4 [2.0–7.0]	0.892
Creatinine, μmol/L	104.0 [82.5–149.5]	105.5 [80.3–132.8]	0.742
eGFR, mL/min/1.73 m2	64.2 [42.4–73.6]	61.5 [46.9–76.5]	0.813
Haemoglobin, g/dL	13.4 ± 1.9	13.0 ± 1.3	0.068
Bilirubin, mg/dL	0.56 [0.34–0.98]	0.69 [0.37–0.83]	0.919
C‐reactive protein, mg/L	3.3 [1.8–7.3]	2.3 [0.9–4.8]	0.151
Procalcitonin, μg/L	0.04 [0.03–0.09]	0.04 [0.02–0.07]	0.463
White cell count, G/L	8.2 [6.2–9.7]	8.0 [7.0–9.2]	0.980
Aetiology of end–stage heart failure
Ischaemic cardiomyopathy, n	8 (32.0%)	8 (26.7%)	0.665
Hypertrophic cardiomyopathy, n	1 (4.0%)	3 (10.0%)	0.617
Idiopathic cardiomyopathy, n	12 (48.0%)	15 (50.0%)	0.883
Other, n	4 (16.0%)	4 (13.3%)	1.00
Intraoperative factors
Aorta cross‐clamp time, min	50 [41–79]	72 [43–86]	0.375
CPB time, min	129 [104–169]	133 [116–154]	0.819
Total ischaemic time, min d	173 ± 41	152 ± 45	0.484

Data are presented as median [interquartile range], mean ± standard deviation and number of patients (frequency). N = 55.

ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin‐receptor blocker; ARNI, angiotensin receptor‐neprilysin inhibitor; CPB, cardiopulmonary bypass; eGFR, estimated glomerular filtration rate; IMPACT, Index for Mortality Prediction After Cardiac Transplantation; PVR, pulmonary vascular resistance.

^a Chronic kidney disease was defined as estimated glomerular filtration rate <60 mL/min/1.73 m².

^b Pre‐operative anaemia was defined according to sex–based criteria of World Health Organization: women haemoglobin <12.0 g/dL and men haemoglobin <13.0 g/dL.

^c Frequency of the applied pre‐operative daily amiodarone therapy: 100 mg – control group: 4.0% (1 patient); haemoadsorption group: 3.3% (1 patient). 200 mg – control group: 8.0% (2 patients); haemoadsorption group: 30.0% (9 patient). 400 mg – control group: 0%; haemoadsorption group: 3.3% (1 patient).

^d Total ischaemic time corresponds the ischaemic time of the donor heart.

Patients in the haemoadsorption group had significantly lower VIS than patients in the control group during the first post‐operative 24 h (median VIS: 27.2 [14.6–47.7] vs. 41.9 [22.4–63.2], P = 0.046, respectively). Among the dominant components of VIS, there was a tendency of lower dose of vasopressors in the haemoadsorption group compared to controls, which reached a statistically significant difference in the case of argipressin (Figure  2 ). However, the median dose of inotropes did not differ between the groups (Figure  2 ). According to the a priori definition, the observed rate of VS was 48.0% (12 patients) in the control group versus 20.0% (6 patients) in the haemoadsorption group, P = 0.028. Additionally, the frequency of extreme noradrenaline demand (i.e. ≥0.5 μg/kg/min) during the first post‐transplant 24 h was significantly lower in patients from the haemoadsorption rather than the control group: 3.3% (1 patient) versus 24.0% (6 patients), P = 0.039, respectively. Similarly, patients in the control group experienced a longer median length of vasopressor support compared to subjects in the haemoadsorption group: 3.0 [1.5–5.0] days versus 2.0 [1.0–4.0] days, P = 0.046, respectively. In a multivariate logistic regression model, patients who received intraoperative haemoadsorption had a 6.4‐fold lower odds ratio for developing early post‐operative VS (P = 0.029) than those who received standard intraoperative care. The independent predictors of the early post‐operative VS are presented in Table 2 .

Figure 2.

Major components of vasoactive inotropic score during the first 24 h after orthotopic heart transplantation. Noradrenaline (A); Argipressin (B); Dobutamine (C); Milrinone (D). N = 55. Data are presented as medians and 95% confidence intervals.

Table 2.

Independent predictors of early post‐operative vasoplegic syndrome

Variable	OR	95% CI	P‐value
Intraoperative haemoadsorption	0.156	0.029–0.830	0.029
Pre‐operative amiodarone therapy	6.315	1.032–38.630	0.046
CPB ≥ 180 min	25.776	2.089–318.016	0.011

Multivariable logistic regression, backward elimination likelihood‐ratio, N = 55. Adjusted covariates in the regression model: intraoperative haemoadsorption treatment; female sex; chronic kidney disease; angiotensin‐converting enzyme inhibitor/angiotensin II receptor blocker treatment pre‐transplant; amiodarone treatment pre‐transplant; pre‐operative pulmonary vascular resistance >3.0 Wood units; CPB ≥ 180 min.

CI, confidence interval; CPB, cardiopulmonary bypass; OR, odds ratio.

PCT and CRP levels showed a marked increase post‐operatively with their peaks at 24 h and 48 h, respectively (Figure  3 ). Interestingly, PCT concentrations were significantly lower at each time point of the 72‐h observation period in the haemoadsorption group compared to controls (Figure  3 ). However, CRP concentrations did not differ between the groups (Figure  3 ).

Figure 3.

Post‐transplant change in procalcitonin (A) and C–reactive protein (B). N = 55. Data are presented as medians and 95% confidence interval. *P < 0.05; **P < 0.01.

MPA plasma concentrations decreased considerably after 2 h of CPB compared to pre‐CPB levels in both groups, but its median level was comparable to controls in the haemoadsorption group at each measurement point (Figure  4 ). The time interval between MMF pre‐operative administration and CPB start was 123 ± 48 min in the control group versus 226 ± 44 min in the haemoadsorption group, P = 0.302.

Figure 4.

Intraoperative change in mycophenolic acid. N = 55. Filled circle indicates outlier, while asterisk represents extreme value.

As shown in Table 3 , shorter durations of MV and ICU stay were registered in the haemoadsorption than in the control group. Similarly, patients who had intraoperative haemoadsorption experienced significantly lower rates of post‐operative AKI and renal replacement therapy (RRT) versus subjects in the control group (Table  3 ). In addition, the PCB was significant in the controls, while it was found to be <3.0% in the haemoadsorption group over a 24‐h time frame (Table  3 ). Nevertheless, only one patient from the control group developed post‐operative hyperbilirubinaemia (bilirubin ≥3.0 mg/dL). There was a low rate of 30‐day mortality for the total study cohort (3.6%) which did not show difference between the groups (Table  3 ). Importantly, the follow‐up EMB examinations did not confirm any grade of cardiac allograft rejection on post‐operative day 7 and the frequency of low‐grade allograft rejections were similar in the groups over the subsequent weeks (Table  3 ). The secondary outcome parameters are described in Table 3 . The analysis of cumulative post‐transplant 1‐year survival did not reveal any statistically significant difference between the groups (control group: 88.0% vs. haemoadsorption group: 96.7%, P = 0.210, Figure 5 ). There were no reported device‐related adverse events over the study period.

Table 3.

Comparative analysis of secondary outcome parameters

Parameters	Control group (N = 25)	Haemoadsorption group (N = 30)	P‐value
Post‐cardiotomy ECMO, n	3 (12.0%)	0 (0%)	0.088
Post‐operative bleeding, mL	570 [385–1305]	565 [350–1130]	0.543
Reoperation for bleeding, n	2 (8.0%)	0 (0%)	0.202
PRC/post‐CPB 24 h, unit	4.0 [0–5.5]	2.0 [0–4.0]	0.243
FFP/post‐CPB 24 h, unit	2.0 [0–3.0]	2.0 [0–3.0]	0.571
PLT/post‐CPB 24 h, unit	12.0 [0–16.0]	12.0 [8.0–16.0]	0.597
Post‐operative MV, h	65 [23–287]	25 [19–68.8]	0.025
Acute kidney injury stage 1, n a	15 (60.0%)	9 (30.0%)	0.025
Acute kidney injury stage 2, n a	0 (0%)	1 (3.3%)	1.00
Acute kidney injury stage 3, n a	4 (16.0%)	1 (3.3%)	0.104
Acute kidney injurytotal, n	19 (76.0%)	11 (36.7%)	0.004
Post‐operative RRT, n	4 (16.0%)	0 (0%)	0.037
Per cent change in bilirubin, %	72.1 [11.2–191.4]	2.5 [−24.6–71.1]	0.009
Early sepsis, n b	1 (4.0%)	0 (0%)	0.455
Length of ICU stay, day	12 [8.5–18.0]	8.5 [8.0–10.3]	0.022
Length of hospital stay, day	28 [24–38.5]	25 [22–34.3]	0.232
30‐day mortality, n	2 (8.0%)	0 (0%)	0.202
EMB cellular rejection
Post‐transplant day 7, n	0 (0%)	0 (0%)
Post‐transplant day 14, n	5 (20.0%)	5 (16.7%)	1.00
Post‐transplant day 21, n	5 (20.0%)	5 (16.7%)	1.00
Post‐transplant day 28, n	6 (24.0%)	10 (33.3%)	0.448
EMB antibody‐mediated rejection
Post‐transplant day 7, n	1 (4.0%)	0 (0%)	0.455
Post‐transplant day 14, n	1 (4.0%)	2 (6.7%)	1.00
Post‐transplant day 21, n	1 (4.0%)	3 (10.0%)	0.617
Post‐transplant day 28, n	2 (8.0%)	1 (3.3%)	0.585

Data are presented as number of patients (frequency) and median [interquartile range]. N = 55. Of 36 registered cellular rejections, 35 (97.2%) were confirmed as grade 1R and 1 (2.8%) was confirmed as grade 2R (ISHLT 2005). The registered antibody mediated rejections were confirmed as pAMR 1I (ISHLT 2013).

ECMO, extracorporeal membrane oxygenation; EMB, endomyocardial biopsy; FFP, fresh frozen plasma; ICU, intensive care unit; MV, mechanical ventilation; pAMR, pathologic antibody‐mediated rejection; PLT, platelet transfusion; PRC, packed red cell; RRT, renal replacement therapy.

^a Acute kidney injury was classified according to Kidney Disease Improving Global Outcomes creatinine‐based definition criteria over the first 5 post‐operative days.

^b Early sepsis was screened over the first 5 post‐operative days.

Figure 5.

Kaplan–Meier estimates of cumulative 1‐year survival, according to the intraoperative treatment. Red line represents the haemoadsorption group, while blue line illustrates the control group. P value (log–rank test) shows the difference in survival.

Discussion

This RCT including OHT patients investigated the clinical effects of the use of pre‐emptive, intraoperative haemoadsorption. Our study found that OHT patients who received intraoperative haemoadsorption experienced reduced VIS and less severe post‐operative vasoplegia compared to standard care alone. The use of intraoperative haemoadsorption was associated with a 6.4‐fold decrease in the odds of developing early post‐operative VS, mitigated PCT kinetics, lower rates of post‐operative AKI and RRT, preserved post‐CPB hepatic bilirubin excretion and shorter durations of MV and ICU stay. This trial did not confirm any evidence of a relevant adsorption effect on MPA. Our RCT did not reveal differences in the frequency and severity of early cardiac allograft rejection as well as in mortality between the groups.

Patients undergoing OHT are reported to be at a remarkably higher risk for developing severe vasoplegia with an incidence ranging from 11% to 66% based on previous analyses. ²¹ , ²² , ²³ , ²⁴ Playing a dominant role in post‐transplant haemodynamic instability, VS can substantially contribute to the development of post‐operative multiple organ dysfunction, resulting in prolonged duration of MV and increased ICU and hospital stays. ²⁴ Considering the most relevant pre‐disposing factors for post‐transplant VS such as advancing age, elevated body mass index, chronic kidney disease, and expanded CPB time, the two groups were found to be homogenous. ²² , ²⁴ In this RCT, intraoperative haemoadsorption showed significant associations with reduced post‐operative VIS. The VIS composite score represents the extent of pharmacological support of the cardiovascular system, and can predict post‐operative morbidity and mortality in adult cardiac surgical patients. ¹⁸ , ¹⁹ The median VIS was significantly higher in the control than in the haemoadsorption group, where it was in the range of ≥30 indicating a higher risk for unfavourable outcomes. Among the four major VIS components, decreased vasopressor requirements were the main determinant of the reduced VIS in the haemoadsorption group, however the doses of inotropes did not differ between the groups (Figure  2 ). These results are indicative of the less severe vasoplegia that developed in the haemoadsorption group, and they are also consistent with the less frequent VS and extreme noradrenaline demand, shortened vasopressor need and decrease in the odds of VS found in the same group. To date, only one observational study has investigated the effect of intraoperative haemoadsorption on post‐operative vasopressor need and outcome among OHT patients. ¹⁴ Interestingly, they observed significantly reduced vasopressor requirements linked to haemoadsorption use. ¹⁴ Similarly, in a propensity score matched analysis of high‐risk infective endocarditis patients, the median vasopressor dose on post‐operative day 1 was found to be significantly lower in the haemoadsorption group than in controls. ¹¹ On the other hand, several recent RCTs including intraoperative haemoadsorption in medium‐ to high‐risk cardiac surgical patients reported controversial data on the post‐operative need for vasoactive support. ¹⁰ , ²⁵ , ²⁶ , ²⁷ , ²⁸ The results of our RCT are in line with earlier observational studies confirming a clear relationship between intraoperative haemoadsorption and the moderate manifestation of post‐operative vasoplegia. Most likely, the discrepancies among these results can be explained by the inhomogeneity of the examined patient populations in terms of perioperative risk for severe vasoplegia.

One of the theoretical aims for introducing intraoperative haemoadsorption in OHT recipients is to modulate the dysregulated inflammatory response related to OHT surgery. This trial demonstrated a mitigated post‐operative PCT response at all pre‐defined time points in the haemoadsorption group compared to controls (Figure  3 ). This result is consistent with the findings of less frequent post‐operative organ dysfunction such as severe vasoplegia, respiratory failure and AKI in patients receiving intraoperative haemoadsorption. Additionally, our data indicate a well‐preserved hepatic bilirubin excretion in the interventional (PCB < 3.0%) versus control group, in which a significant post‐operative decline of this hepatic function was shown (PCB > 70.0%). The previous observational study in OHT patients showed similar kinetics in post‐operative PCT in both the haemoadsorption and control groups and established only favourable trends in the length of MV, ICU stay, and rate of AKI. ¹⁴ However, an arbitrary criterion was used to indicate intraoperative haemoadsorption in their investigation, definitely influencing patient selection bias in terms of pre‐operative immune priming level or risk and reversibility of post‐operative organ dysfunction. ¹⁴ It has recently been shown that bilirubin can be removed directly by haemoadsorption treatment integrated into extracorporeal devices. ²⁹ , ³⁰ , ³¹ , ³² In line with these results, a degree of direct bilirubin removal by intraoperative haemoadsorption can be supposed. The preserved hepatic bilirubin excretion in the interventional group correlated with less manifested post‐operative organ dysfunction, associated with reduced VIS and mitigated PCT response as represented in our study group versus controls.

Cardiac allograft rejection early after OHT is among the most severe complications which can negatively affect recipients' long‐term outcomes. ³³ High variability in the immunosuppressive drug concentrations is confirmed to be linked to increased risk for acute allograft rejection. ³⁴ To date, no data exist on interactions between intraoperative haemoadsorption and immunosuppressive drug concentrations in terms of OHT. Interestingly, Lindstedt et al. did not find histopathological signs of acute rejection at 1‐ and 3‐month post‐transplant in patients who received cytokine adsorption during lung transplantation, compared to patients managed without the adsorber. ³⁵ Also, a very recently published large animal study reported on an adsorption rate of less than 5% for immunosuppressive agents such as tacrolimus, cyclosporin A, mycophenolate mofetil, everolimus, and methylprednisolone during 6 h of in vivo extracorporeal haemoadsorption treatment. ³⁶ Data presented in our RCT strongly substantiate these previous investigations. Similar MPA concentrations were measured pre‐CPB and at 2 h of CPB run in the study groups (Figure  4 ), and there were no differences in the frequencies of cardiac allograft rejection over the 1‐month follow‐up period between the groups (Table  3 ). These results demonstrate significant safety information regarding the interaction between intraoperative haemoadsorption treatment and perioperative immunosuppressive therapy of OHT.

In our RCT we involved low risk OHT recipients (median IMPACT score was 4 in both groups, see Table 1 ) with identical pre‐operative inflammatory activity and risk profile for post‐operative organ dysfunctions (Table  1 ). Accordingly, the registered 30‐day mortality rate was 8.0% and 0%, and 1‐year survival was 88.0% and 96.7% in the control versus haemoadsorption groups, respectively. In the light of these favourable survival numbers in both groups, a positive impact of haemoadsorption on mortality was not to be expected. However, our results in terms of proximal endpoints suggest the effectiveness of intraoperative haemoadsorption in controlling the dysregulated inflammatory processes and reducing post‐operative organ dysfunctions. In addition, this method of intraoperative immune modulation of OHT surgery did not show a relationship with an increased rate of adverse immunological events, and the use of intraoperative haemoadsorption was not linked to any complications in our study.

This trial has strengths and limitations. To the best of our knowledge, our investigation is the first RCT to assess the clinical effects of intraoperative haemoadsorption among OHT patients focusing on proximal primary endpoints. Despite the small sample size, a homogeneous cohort of patients was randomized into two similar arms in terms of clinical characteristics and risk profile. However, due to a lack of any previous RCT in this field based on similar primary outcomes, we did not perform a formal sample size calculation. It is a single‐centre study; therefore, the presented results are subject to selection bias requiring external validation by other centres. These limitations in part restrict the interpretation of our results.

In conclusion, the results of this RCT suggest that intraoperative haemoadsorption during OHT is associated with better haemodynamic stability, as indicated by a 6.4‐fold decrease in the odds of developing VS and less frequent VS in the early post‐operative period compared to standard care. Our study also found that patients in the haemoadsorption versus the control group experienced a mitigated PCT response, lower rates of post‐operative AKI and RRT, more stable hepatic bilirubin excretion, and shorter durations of MV and ICU stay. This trial does not confirm any relevant adsorption effect on MPA and more frequent adverse immunological events such as early cardiac allograft rejection and sepsis related to intraoperative haemoadsorption treatment. This study did not discover evidence of any device‐related complications. The promising results of our proof‐of‐concept trial support the need for adequately powered RCTs to clarify the potential benefits of intraoperative haemoadsorption in OHT patients.

Conflict of interest

E.N. reports travel funding and honoraria for presentations from CytoSorbents Europe GmbH, Berlin, Germany, in the past 24 months. E.K. reports honoraria for presentation from CytoSorbents Europe GmbH, Berlin, Germany, in the past 24 months. All other authors declare that they have no competing financial or other interests in relation to their work.

Funding

No funding to declare.

Supporting information

Table S1. Applied immunosuppression protocol of orthotopic heart transplantation during the perioperative period and the first month postoperatively.

Figure S1. Integration method of the haemoadsorption cartridge (CytoSorb™) into the cardiopulmonary bypass.

Nemeth, E. , Soltesz, A. , Kovacs, E. , Szakal‐Toth, Z. , Tamaska, E. , Katona, H. , Racz, K. , Csikos, G. , Berzsenyi, V. , Fabry, S. , Ulakcsai, Z. , Tamas, C. , Nagy, B. , Varga, M. , and Merkely, B. (2024) Use of intraoperative haemoadsorption in patients undergoing heart transplantation: a proof‐of‐concept randomized trial. ESC Heart Failure, 11: 772–782. 10.1002/ehf2.14632.