Conference Coverage

Novel liver dialysis device may safely curb ACLF



An investigational liver dialysis device (DIALIVE) was associated with significantly greater survival of patients with acute-on-chronic liver failure (ACLF), compared with the standard of care in a multicenter randomized study.

Among 30 evaluable patients with ACLF from alcoholic cirrhosis randomized to treatment with the DIALIVE system or standard of care, two-thirds of patients assigned to DIALIVE had both survived and experienced resolution of ACLF by 28 days, compared with one-third of patients assigned to standard of care, reported Banwari Agarwal, MBBS, MD from the Royal Free Hospital in London at the meeting sponsored by the European Association for the Study of the Liver.

Different from MARS

The DIALIVE system differs from the Molecular Adsorbent Recirculating System (MARS) liver dialysis system in that DIALIVE removes and replaces albumin, including proinflammatory albumin, rather than filtering and recirculating it, he explained.

“It addresses systemic inflammation, which wasn’t quite the case with MARS,” he said in the question-and-answer portion of his presentation in a general session.

In patients with ACLF, the risk of 28-day mortality increases substantially as the grade of ACLF increases.

“ACLF, however, is potentially reversible, and the initial grade at presentation undergoes changes over time during the natural course of the illness, with some patients deteriorating, some improving, and some even achieving complete ACLF resolution. The final grade is reached by days 3-7, and it is this final grade which determines their future outcome trajectory. I therefore propose that ACLF resolution in itself is an important therapeutic target,” he said.

Study details

Dr. Agarwal and coinvestigators from eight centers in six European countries enrolled patients with a history indicative of alcohol-related cirrhosis, at least one acute decompensation event, and progression to ACLF grades 1, 2, or 3a.

Patients with an international normalized ratio above 3 were excluded, as were those with more than three organ failures, uncontrolled infections, patients with primary respiratory organ failure, and those with hemodynamic instability refractory to volume resuscitation and low-dose vasopressors.

A total of 32 patients, of whom 30 were evaluable, were randomized to receive liver dialysis in three to five DIALIVE sessions lasting 8-12 hours each (15 evaluable patients) or to standard of care at participating institutions (15 patients).

The investigators looked at safety of the device (the primary endpoint) in all patients who received at least one DIALIVE treatment (safety population), and a modified safety population of patients who received at least three DIALIVE treatments.

The median patient age in each arm was 49 years, and all patients had alcoholic cirrhosis, with alcoholic hepatitis accounting for at least one decompensation event. In addition, about 25% of patients in each arm had decompensation with infections and/or sepsis as precipitating factors.


Serious adverse events on days 1-10 occurred in 11 of 17 patients in the DIALIVE arm, and in 8 in the standard-of-care arm. In the DIALIVE arm, there were seven treatment-related serious device events, three unexpected serious device events (anemia, septic shock, and hypotension), and one patient discontinued dialysis after having unsafe levels of thrombocytopenia.

Four patients in the DIALIVE arm died on study. The first two died on day 1 one from hypotension, coagulopathy, and multiorgan failure, and this prompted a change in the protocol mandating that DIALIVE be conducted only in an ICU setting with more invasive monitoring and more frequent lab analysis of clotting and other biochemical parameters. Of the two other patients in the DIALIVE arm who on died on study, one died from non-MI cardiac arrest on day 8, and one patient with ACLF grade 3 and a European Foundation for the study of chronic liver failure (CLIF)–ACLF score of 68 died from multiorgan failure.

“I must emphasize that even this very sick patient tolerated the device very, very well,” Dr. Agarwal said.

In the standard-of-care arm, two patients died from progressive liver failure on days 17 and 27, respectively, and one died on day 17 from bacterial infections, bleeding, and progressive liver failure.

There were eight instances of filters clotting out of 64 filters used in total, and four episodes of device deficiency, including two instances where tubing could not be disconnected from an Oxiris filter during setup of the DIALIVE circuit, requiring use of new DIALIVE kits; one use of an incorrect dialysis fluid; and one incorrect setup of the DIALIVE circuit.

Significant improvements in many scores

In the DIALIVE group, there were significant improvements over baseline at day 10 in both liver scores (P < .05) and brain scores (P < .001). In contrast, in the standard-of-care group there were no improvements in individual organ scores, and respiration scores were significantly worse (P < .01).

DIALIVE was also associated with significant improvements in CLIF-C organ failure scores, compared with standard of care at day 5 and day 10 (P = .021 and .001, respectively); CLIF-C–ACLF scores at days 5 and 10 (P = .045 and .023); and Model for End-Stage Liver Disease scores at day 5 (P = .028).

In the DIALIVE group, ­40% of patients had ACLF resolution by day 5, and 66.7% had resolution by day 10. In the standard-of-care arm, 15% had resolution on day 5, and 33.3% had resolution on day 10. DIALIVE was also associated with a significantly faster median time to resolution, compared with standard of care (10 days vs. not reached; P = .0307). At 28 days, 10 of 15 evaluable patients were alive and had resolution of ACLF with DIALIVE versus 5 of 15 with standard of care (P = .0281).

Dr. Agarwal said that the data justify the implementation of late-phase clinical trials of the liver dialysis device.

‘Hopeful’ findings

“It’s very early, but we’re really desperate in finding something to bridge to transplantation,” commented Tobias Boettler, MD, from the University of Freiburg (Germany), who was not involved in the study.

“I think this is very hopeful,” said Dr. Boettler, who moderated the briefing where Dr. Agarwal summarized the study findings.

In the question and answer following the talk in a general session, moderator Philip N. Newsome, MD, from University Hospitals Birmingham (England) asked whether patients who were not treated should have been included in the analysis.

Dr. Agarwal replied that “the whole idea behind this study was to understand what this device does to these patients, and how these patients react to this device, so really not looking at the efficacy.”

The study was supported by the European Union’s Horizon 2020 initiative. Dr. Agarwal received a study grant from the initiative, but had no other relevant disclosures. Dr. Boettler and Dr. Newsome had no disclosures relevant to the study.

Next Article: