Applicability and safety of discontinuous ADVanced Organ Support (ADVOS) in the treatment of patients with acute-on-chronic liver failure (ACLF) outside of intensive care

Background ADVanced Organ Support (ADVOS) is a novel type of extracorporeal albumin dialysis and holds promise to sustain liver function and recovery of patients with acute-on-chronic liver failure (ACLF). Previously, ADVOS was tested as continuous treatment for intensive care patients with liver failure. Data related to the applicability and safety as discontinuous treatment outside of ICU is not available. Aim Evaluation of ADVOS as discontinuous treatment for patients with ACLF outside intensive care unit and comparison with a matched historic cohort. Methods and results In this retrospective study, 26 patients with ACLF and the indication for renal replacement therapy related to HRS-AKI were included. Majority of patients were male (65%) with alcoholic cirrhosis in 88% and infections as a trigger of ACLF in 96%. Liver function was severely compromised reflected by high median MELD and CLIF-C ACLF scores of 37 (IQR 32;40) and 56.5 (IQR 51;60), respectively. Patients were treated discontinuously with ADVOS over a median time of 12 days (IQR 8.25;17) and received 8 (IQR 4.25;9.75) treatment cycles on average. No treatment related adverse events were recorded, and safety laboratory parameters remained constant during the observation time. After 16 h cumulative dialysis therapy, ADVOS significantly reduced protein-bound bilirubin (14%), creatinine (11.8%) and blood urea nitrogen (BUN, 33%). Using a matched cohort with ACLF treated with hemodialysis, ADVOS achieved a stronger decrease in bilirubin (p = 0.01), while detoxification of water-soluble catabolites’ including creatinine and BUN was comparable. The 28-days mortality in the ADVOS group was 56% (14/26) and was not inferior to predicted survival (predicted median 28-days mortality was 44%, IQR 30; 59). Conclusion Discontinuous ADVOS treatment was safe and effective in patients with ACLF outside intensive care and outperformed hemodialysis in reducing protein-bound metabolites.

Introduction Acute-on-chronic liver failure (ACLF) is a life-threatening condition which might occur in patients with liver cirrhosis. According to the European Association for the Study of the Liver-Chronic Liver Failure (EASL-CLIF) Consortium ACLF is characterized by an acute decompensation of the cirrhotic liver and includes organ failure and high short-term mortality. Organ failure manifest as hepatic and extra hepatic abnormalities in ACLF including acute kidney failure (56%), hyperbilirubinaemia (44%) followed by coagulopathy (27%), hepatic encephalopathy (24%), circulatory-(17%) or respiratory constrains (9%) [1]. ACLF arises in up to 40% of the patients with decompensation and is more frequent in young patients [2,3]. In Europe, the most common triggers for ACLF are ongoing alcohol consumption (75%), viral hepatitis related (10%), while in some patients (~15%) the cause remains cryptogenic [1,4]. The European Association for the Study of the Liver Chronic Liver Failure (EASL-CLIF) consortium set up the CANONIC (EASL-CLIF ACLF in Cirrhosis) study and developed the CLIF-C ACLF score to stratify patients with ACLF and predict 28-days mortality [5,6].
The therapeutic options for ACLF are limited as donor organ allocation for liver transplantation is competitive and frequently not an option when e.g. alcohol use is ongoing. Thus, novel therapeutic approaches for ACLF aim to support organ function and recovery to bridge especially acute phases of decompensation are urgently needed. Here, Extracorperal liver support (ELS) is an appealing concept to sustain liver function as bridge-to-transplant or even bridge-to-recovery [7]. Hemodialysis, as a renal replacement therapy, primarily eliminates water-soluble toxins, while protein-bound toxins are retained, increasing the risk of secondary organ failure [8]. In contrast, extracorporeal albumin dialysis (ECAD) closes this gap and removes both water soluble and protein-bound toxins from the blood. Several ECAD concepts were already tested in clinics including single pass albumin dialysis (SPAD), Molecular Adsorbent Recirculating System (MARS) and fractionated plasma separation (Prometheus) [9,10]. ECADs improved secondary endpoints (hemodynamics and hepatic encephalopathy), while primary endpoints (survival probability) and other endpoints were not met [10][11][12]. However, outcomes may be biased by confounders and insufficient stratified patient cohorts. Further data of studies, which applied the latest definition of ACLF (e.g. CLIF-C ACLF score) are yet missing.
Previous ECAD systems consumed significant amounts of human albumin solution (HAS), a scarce resource, making the procedure expensive and unattractive. The ADVanced Organ Support (ADVOS) is a novel type of ECAD, facilitating multi-organ support. The optimized detoxification and continuous regeneration of toxin-binding albumin reduces the needed HAS to a minimum. In a tertiary circuit, the regeneration of HAS is achieved by alteration of the pH and temperature, modulating the conformation of the protein and, thus, it's binding capacities to toxins (Fig 1).
In preliminary studies, ADVOS proved to be safe and effective as continuous treatment for patients with liver and multi-organ failure in intensive care, but applicability of ADVOS as discontinuous treatment on peripheral ward has not yet been evaluated [7,13]. The current study explored the applicability and safety of ADVOS as discontinuous treatment in a regular dialysis unit and compared outcome of ADVOS in a case-control study with intermittent hemodialysis (HD) with matched ACLF patients.

Study design
This is a retrospective study evaluating ADVOS as discontinuous treatment for patients with ACLF outside of an intensive care unit. According to clinical standard, patients were treated with ADVOS from July 2018 to November 2020 on a normal ward at the Cirrhosis Center Mainz, University Medical Center Mainz, Germany. Medical history, dialysis parameter and laboratory data were retrieved by chart review or from the local laboratory system, respectively.

Ethics
The study was conducted according to the ethical guidelines of the 1975 Declaration of Helsinki and its later amendments and in line with the guidelines of the local ethics committee (Ethik-Kommission, Landesärztekammer Rheinland-Pfalz). The study was approved by the local ethics committee. Informed consent was waived due to the retrospective design of the study. Medical data of HD patients was obtained from a local ACLF registry as agreed before by the local ethics committee (reference number: Nr. 837.232. 17 [11066]).

Matching
Detoxification effect of ADVOS vs. regular HD was compared by reduction of the laboratory parameters creatinine, bilirubin and blood urea nitrogen (BUN). Therefore, ADVOS patients were matched 1:1 to ACLF patients who received regular hemodialysis. Matching was done 1:1 based on the following criteria: Sex, age, etiology of cirrhosis, cumulative dialysis time and liver function as reflected by MELD-and CLIF-C ACLF-score. Laboratory data of HD patients was retrieved from a local ACLF registry, containing 136 patients. Baseline characteristics of patients of the registry are shown in S1 Table.

Inclusion and exclusion criteria
Patients were included based on the following criteria, while data was retrieved by chart review Table 1:

ADVOS
The ADVOS system is an ECAD device, facilitating combined liver and kidney support ( Fig  1). ADVOS is approved as treatment for patients with reduced detoxification with impaired liver, kidney and lung organ function in Europe. It consists of three communicating circuits. The first circuit circulates the blood over high flux filters (Elisio-19H, Nipro, Ratingen, Germany) back to the patient through conventional double lumen dialysis catheter. In the second circuit, a dialysate containing 2-4% albumin solution runs in parallel to the first circuit only separated by a semipermeable membrane when running through the dialyzer, loading the protein solution with protein-bound toxins and eliminating water-soluble toxins. The third circus represents metaphorically speaking the heart of the ADVOS system, achieving the detoxification of the dialysate and, thus, recycling of the toxin-loaded albumin solution. In this circuit, the dialysate is separated in two branches where either acids or alkalis are substituted. By alteration of the pH, the albumin changes its confirmation, reducing its affinity to the loaded toxins. The effect is further enhanced by lowering the temperature down to~28˚C. The discharged toxins are removed by a tertiary filtration process. The acidified and alkalized albumin dialysates converge, resulting in a physiological pH after reunion. The temperature of the dialysate is re-warmed up to body temperature. Both parameters, pH and temperature, can be modulated if this is considered useful for therapy. Finally, the regenerated albumin dialysate re-circles to the second circuit for the next cycle. No additional renal replacement therapy, neither serial nor parallel, was performed besides ADVOS. No pre-and post-dilution for ADVOS was conducted.

Anticoagulation
In line with our clinical standards, anticoagulation was performed by regional anticoagulation wit citrate independent of the patient's prothrombin time.

ADVOS treatment
Patients received a median of 8 (IQR 7.25; 9.75) ADVOS cycles as discontinuous treatment over the median period of 12 days (IQR 8.25; 17) on a peripheral ward. Median duration of the first two treatment cycles was 8 h (IQR 7; 8) and was adapted to the specific needs of each patient after the initial treatment cycles. Median blood flow rate was 150 ml/min (IQR 150; 150) and median ultrafiltration rate was 150 ml/h (IQR 110; 250, S2 Table). No adverse events or alarming changes in laboratory parameter were documented (S3 Table).

Treatment efficacy of ADVOS
The detoxification effect of ADVOS was assessed after the first two treatment cycles (16 h (Table 3). In a subgroup analysis, we found that the elimination of bilirubin was concentration dependent. Reduction tended to be more efficient at higher bilirubin levels (�20 mg/dl) as shown in S4 Table.

Comparison of ADVOS vs. hemodialysis in a case-control study
The detoxification effect of ADVOS was compared to regular hemodialysis (HD) based on the reduction of the laboratory parameters creatinine, bilirubin and BUN in ACLF patients. ADVOS treated patients were matched 1:1 to patients of a local ACLF registry. Beside the need for renal replacement therapy because of HRS-AKI, pairs were found based on sex, age, etiology of cirrhosis, liver function as reflected by MELD score and CLIF-C ACLF score and cumulative dialysis time (Table 4) [17]. In total, 25 matches were found based on the defined criteria, while statistical testing between the matched cohorts revealed no significant difference, indicating successful matching. For comparison of ADVOS vs. HD, blood parameters were assessed before and after 16 h cumulative dialysis time from day one of renal replacement therapy. ADVOS patients received two treatment cycles with 8 h per day on two consecutive days, while HD patients were treated for 4 h per day on four consecutive days, summing up to 16 h dialysis time for both modalities. Reduction of bilirubin was higher in the ADVOS group compared to patients treated with regular HD (median reduction -2.48 vs. 1.01 mg/dl, p = 0.01). In contrast, no difference was found for creatinine and BUN. (Fig 2).

28-days mortality of patients treated with ADVOS vs intermittent hemodialysis
28-days mortality of the matched cohort was analyzed between ADVOS vs. HD treated patients. Despite both cohorts had comparable grade of liver failure (MELD-and CLIF-C ACLF score) and baseline characteristics (Table 4), in the ADVOS treated cohort 11 of 25 (56%) patients died, while 15 of 25 (40%) patients in the HD treated cohort did not survived for 28-days mortality (Fig 3). In addition, predicted survival by the CLIF-C ACLF score was 44% (IQR 30; 59, Table 2) in the ADVOS cohort, suggesting that ADVOS does not negatively impact survival. However, there was no substantial difference of long-term survival between ADVOS vs. HD treated patients, while only one patient in the ADVOS cohort received liver transplantation (S1 Fig).

Discussion
ADVOS represents a novel type of extracorporeal nonbiologic liver support system and holds promise to work as a bridging or destination therapy for patients with ACLF either awaiting liver transplantation or recovery of liver function. In contrast to previous albumin dialysis systems, for ADVOS only a minimum amount of human albumin (200 ml) is needed, saving this scarce resource in clinics. Beyond liver support, ADVOS corrects severe metabolic and respiratory acid-base disequilibrium, which is relevant for multi-organ failure therapy [10,13].
In this current study, we were able to demonstrate the applicability and safety of ADVOS in a discontinuous treatment setting for patients with ACLF and the need for renal replacement therapy outside of an intensive care unit.  Included patients had a severe liver dysfunction reflected by high MELD-and CLIF-C ACLF scores and liver transplantation was an option only for one patient, while for the majority allocation of donor organ was not available due to ongoing alcohol abuse. Thus, our cohort represents typical cases of ACLF, where patients cannot qualify for allocation of a donor organ and therapeutic options to support their liver function are scarce.
By now, there were only two studies published, which have tested ADVOS as continuous treatment in intensive care. The first study was published by Huber et. al. and evaluated the applicability of ADVOS for patients with ACLF or secondary liver failure. All 14 included patients were critically ill, diagnosed with HRS-AKI and the need for renal replacement therapy. The studied cohort was comparable to our current cohort with a mean MELD score of 34 (SD±7), age of 54 (SD±13) and 57% of men, being comparable to our cohort ( Table 2). They found that ADVOS reduced significantly mean bilirubin (-8.3 mg/dl SD±6.5), creatinine (-0.6 mg/dl SD±0.6) and BUN (-18 mg/dl SD±16) after a mean time of dialysis of 9.6 h (SD±3.2). While our data confirm these findings that ADVOS decreases the levels of serum bilirubin, creatinine and BUN, there are still some differences between both studies. The most striking one is the reduction of bilirubin, which was more than 2-fold higher in the study by Huber et al. when compared to our data (-3.4 vs. -8.3 mg/dl). In contrast, the reduction of BUN was nearly 2-fold higher in our patients than in the study by Huber et al (-33.7 vs. -18 mg/dl, Table 3). We hypothesize that the lower reduction protein-bound bilirubin in our study could be explained by the absence of vasopressors and, thus, lower mean arterial pressure during dialysis in our cohort (S3 Table). Because dialysis efficacy also relies on sufficient mean arterial pressure, which might become more evident for protein-bound compounds for albumin dialysis [18]. Despite Huber et al. did not declare the exact number of patients who received vasopressors in their study as well as mean blood pressure during dialysis, their exclusion criteria excluded 1.) patients who received an excessive amount of vasopressor support (Dopamine >15 μg/kg/min, epinephrine >0.1 μg/kg/min or norepinephrine >0.1 μg/kg/min); 2.) patients with mean arterial pressure �50 mmHg despite conventional medical treatment, which would include vasopressor therapy. Therefore, it seems obvious that most of the patients received vasopressor treatment during ADVOS, which could have stabilized blood pressure and, thus, promoted a more efficient dialysis. Further, we found that elimination of bilirubin is less effective when anticoagulation is performed with citrate instead of heparin as seen in other ongoing pilot studies (Weinmann-Menke J et al., unpublished data of our group). This might have also contributed to the lower reduction of bilirubin in our cohort as we have applied citrate anticoagulation, while Huber et al. used heparin as anticoagulant during ADVOS [10].
-48%, creatinine of -11.8% vs. -24% (Table 3). In line with our data, they found that the reduction of bilirubin was more efficient at higher starting levels of bilirubin (S4 Table).
Beside detoxification, safety must be ensured to recommend ADVOS for discontinuous use outside of intensive care units. In our current study we did not detect any safety issues. Treatments were well tolerated, and no treatment related adverse events were documented during ADVOS. Strikingly, we observed no bleeding events despite primary and secondary hemostasis were severely compromised (low platelet count and high prothrombin time-INR,  Table 2). In contrast, Fuhrmann et al. reported three major bleeding complications during ADVOS treatment with heparin anticoagulation.
However, it must be acknowledged that our study was retrospective in design. Therefore, we were only able to evaluate documented safety issues and patients were not followed after a predefined protocol. Therefore, these findings have to be interpreted with caution and must be validated in future prospective trials.
In a preliminary survival analysis, ADVOS showed non-inferiority in 28-days morality compared a matched cohort treated with regular hemodialysis (11 vs . 15 patients survived). Additionally, the 28-days mortality of our ADVOS patients was lower as predicted by the CLIF-C ACLF score (44 vs. 56%). However, 28-days and long-term survival analysis between ADVOS and HD did not reach significance (Fig 3, p = 0.225; S1 Fig, p = 0.08) and again, these results have to be interpreted with caution due to the retrospective design of this study. Though, our findings may be a first hint that ADVOS may have a beneficial effect on survival of patients with ACLF and may be helpful for designing future randomized-controlled trials.
Our study has limitations that need to be acknowledged. First, as mentioned before this is a retrospective, case-control study and therefore the results have to be interpreted according to the study's design. Moreover, our study cohort is small and precludes further detailed analyses. Last, this study lacks a randomized control group. Despite a 1:1 matching of ADVOS and HD patients according to defined criteria, selection bias might have occurred.
In conclusion, we report for the first time the feasibility and safety of ADVOS as a discontinuous treatment in patients with ACLF outside of an intensive care unit. Additionally, we found a similar detoxification effect of ADVOS as discontinuous dialysis compared to previous data, where ADVOS was performed in a continuous dialysis setting. We found that ADVOS was not inferior in 28-days morality compared to regular hemodialysis in patients with ACLF, future randomized trials are needed to investigate its prognostic effect compared to regular HD.