Effect of plasma exchange with albumin replacement on albumin functionality and organ dysfunction in acute-on-chronic liver failure

Abstract

Background & Aims

Effective treatments for acute-on-chronic liver failure (ACLF) are a major unmet need. This proof-of-concept pilot study was aimed at evaluating the effects of plasma exchange (PE) with albumin 5% (PE-A5%) on albumin functional capacity and organ dysfunction in patients with ACLF.

Methods

Ten adult patients were enrolled in a single-center phase II, prospective, open-label, non-controlled study. Six PE-A5% sessions were performed in 10 days followed by a 1-month follow-up visit. Albumin functional capacity and circulatory function were assessed, as were renal, cerebral, and liver function, and systemic inflammation. The main safety variable was the percentage of PE sessions associated with at least one procedure-related adverse event (AE).

Results

Patients with ACLF showed lower albumin binding capacity, lower antioxidant capacity, and lower levels of albumin with preserved structure compared to healthy donors (n = 19). From baseline to day 11, PE-A5% treatment increased albumin levels and improved albumin binding capacity to Sudlow site II (15.3±1.6 mg/ml to 18.9±1.7 mg/ml; p = 0.003), fatty acid-binding capacity (8.2±1.4 μM to 3.1±1.5 μM; p = 0.013) and antioxidant capacity (human mercaptalbumin 9.5±1.5 mg/ml to 14.6±1.6 mg/ml; p = 0.001). Native albumin levels were increased throughout day 1-11 PE-A5% sessions (6.5±1.0 mg/ml to 10.2±1.4 mg/ml; p = 0.035). PE-A5% improved systemic hemodynamics (mean arterial pressure, heart rate, cardiac index), renal function (creatinine level, blood urea nitrogen), cerebral function (hepatic encephalopathy grade), liver parameters (transaminases, bilirubin) and inflammatory parameters (C-reactive protein, leukocyte count). All patients had at least one of the 78 AEs reported, mostly mild (product/procedure-related: 36%). Sixteen serious AEs were reported in eight patients (procedure/product-related: none).

Conclusions

PE-A5% was a safe procedure associated with positive effects on albumin functionality, and circulatory, renal, cerebral, and liver function in patients with ACLF.

Impact and implications

Acute-on-chronic liver failure (ACLF) is a clinical condition characterized by severe systemic inflammation, organ failure, and high mortality. Plasma exchange removes patient’s plasma containing pathogenic substances, replacing it with 5% albumin and fresh frozen plasma (PE-A5%). In this study, cirrhotic patients with ACLF were treated with PE-A5%, which was a safe procedure that increased binding and antioxidant capacity of patients’ albumin, while improving circulatory, kidney, brain, and liver functions. These beneficial effects could impact survival in ACLF.

ClinicalTrials.gov Identifier

NCT01201720

EudraCT number

2010-021360-15

Graphical abstract

Highlights

• •This pilot study investigated the effects of plasma exchange with albumin 5% (PE-A5%) in patients with ACLF.
• •PE-A5% increased albumin levels and improved albumin function (binding and antioxidant capacity) in this setting.
• •PE-A5% improved systemic hemodynamics, renal, cerebral, and liver functions, and some inflammatory parameters.
• •Safety, feasibility, and tolerability of PE-A5% in patients with ACLF was confirmed.

Introduction

Acute-on-chronic liver failure (ACLF) is a recently described syndrome characterized by severe acute organ failure(s) in patients with cirrhosis, resulting in high short-term mortality (50-90%). Jaundice, severe hepatic encephalopathy and/or hepatorenal syndrome are manifestations of ACLF,¹ caused predominantly by intense and sustained stimulation of the innate immune system that triggers systemic inflammation.^2,3 Despite its poor short-term prognosis, ACLF is potentially reversible, and approximately 50% of patients with ACLF are admitted to intensive care units for organ support.³ Current medical management of ACLF consists of early recognition, treatment of the precipitating event, and supportive care.³ Although liver transplantation is the only effective treatment for ACLF, the lack of organ donations demands alternative treatment strategies such as albumin dialysis, and plasma exchange (PE).^4,5

PE is an established and safe apheresis procedure that combines two treatment approaches in a single intervention: removal of deleterious toxins and proinflammatory mediators of plasma (e.g., cytokines and reactive oxygen species) and replacement with another solution (e.g., fresh frozen plasma [FFP], colloids, crystalloids).⁶ A multicenter randomized-controlled trial showed that high-volume PE with FFP replacement significantly improved biochemical, hemodynamic parameters and overall and transplant-free hospital survival, without increasing the incidence of adverse events (AEs) in patients with acute liver failure.⁷

The current EASL guidelines recommended the use of PE for the treatment of acute liver failure.⁸ In patients with decompensated cirrhosis, PE offers a liver support system to remove toxins that cause hepatic encephalopathy and hemodynamic instability.[9], [10], [11]

Albumin is a protein with pleiotropic properties that determines plasma oncotic pressure and thereby maintains circulating blood volume.¹² Besides that, it has other non-oncotic functions including antioxidant, anti-inflammatory, and immunomodulatory activity, as well as roles in the transport and detoxification of endogenous and exogenous molecules, hemostasis regulation, and endothelial stabilization.[12], [13], [14] In patients with cirrhosis, the quantity of toxic substances that bind to albumin is increased whereas albumin levels, transport, binding and antioxidant capacities are decreased.¹⁵ Albumin dysfunction (oxidation and reduced fatty acid-binding capacity) is related to the degree of liver failure and is associated with decreased survival.^2,16,17 Intravenous human albumin prevents circulatory dysfunction after large volume paracentesis and spontaneous bacterial peritonitis, and reverses hepatorenal syndrome in association with vasoconstrictors.[18], [19], [20] Long-term albumin treatment also reduces systemic inflammation and improves circulatory dysfunction in patients with decompensated cirrhosis and ascites, indicating that beneficial effects of IV albumin in patients with cirrhosis may be related to its non-oncotic functions.²¹

PE using albumin 5% (PE-A5%) as a replacement fluid is an effective and safe treatment in patients with acute liver failure,²² improving hepatic encephalopathy and survival in this setting.[22], [23], [24] However, there are no studies evaluating the effects of PE-A5% in patients with ACLF. Therefore, this pilot study was aimed at investigating the effects of PE-A5% on the functional capacity of albumin, as well as circulatory, renal, cerebral, and liver function in patients with ACLF.

Patients and methods

Design and objectives

Patients with cirrhosis and ACLF were recruited in a single-center phase II, prospective, open-label, non-controlled, pilot clinical trial (ClinicalTrials.gov Identifier: NCT01201720; EudraCT number: 2010-021360-15).

The primary objective was to assess the effects of PE-A5% in patients with ACLF, using therapeutic albumin (Human Albumin Grifols® 5%/Albutein® 5%; Grifols, Barcelona, Spain) as a replacement solution to restore albumin functional capacity and reverse circulatory dysfunction. Secondary objectives were to assess the effects of PE-A5% on renal, cerebral, and liver function, as well as on the systemic inflammatory response and oxidative stress.

The trial was performed according to the principles of the Declaration of Helsinki, the standards of Good Clinical Practice of the International Conference of Harmonization, and current legal regulations. The protocol was approved by the Ethical Committee of the Hospital Clínic and by the Spanish Agency for Medicines and Health Products (AEMPS). Patients freely gave written informed consent before study enrolment. In case of hepatic encephalopathy, a legal representative signed the informed consent form.

Patients

Both males and females, aged ≥18 years and ≤80 years, with cirrhosis (established by biopsy or clinical, laboratory and ultrasound data) and ACLF were included in the study between March 2011 and July 2013. ACLF was defined as an acute deterioration in hepatic function occurring in 2-4 weeks, with jaundice (total serum bilirubin ≥5 mg/dl) and hepatic encephalopathy (grade ≥2) and/or renal failure (serum creatinine ≥2 mg/dl).³ EF-CLIF Consortium criteria were not used since they were published after this study was designed.³

Main exclusion criteria were: neoplastic disease; active bacterial or fungal infection with hemodynamic instability requiring the use of vasoactive drugs at high dose (>0.5 μg/kg/min of noradrenaline); moderate or severe structural heart disease; chronic renal failure with hemodialysis; chronic lung disease; previous transplant; HIV infection; adult respiratory distress syndrome or acute pulmonary injury; gastrointestinal bleeding in the 72 h before starting treatment; severe coagulopathy (international normalized ratio ≥3) and/or platelet count <30,000/μl; extrahepatic cholestasis; recent hepatobiliary surgery and bilirubin ≥5 mg/dl for more than 4 weeks.

All patients who met the inclusion criteria and performed at least one PE-A5% comprised the enrolled population and were used for safety analysis. Those who completed the six PE-A5% treatments comprised the per protocol population.

Intervention: PE-A5%

All patients were admitted to the intensive care unit during the intervention period. PE-A5% was performed using the Spectra Optia cell separator (Terumo BCT, Zaventem, Belgium), using a citrate solution (ACD-A Grifols®, Grifols) as anticoagulant. Routinely, to prevent citrate toxicity, we infused intravenously calcium chloride (Ca) plus magnesium sulfate (Mg) solution throughout the PE-A5% at a rate of 1 mol of Ca and Mg per 10 mol of citrate.²⁵

Treatment consisted of two initial PE-A5% sessions on consecutive days followed by every 2 days (a total of 6 PE-A5% sessions in 10 days). One follow-up visit was done 1 month after the last PE-A5%. A 1.1 plasma volume was exchanged in each session, where approximately 30% of the replacement fluid was FFP, which was administered at the end of PE-A5%. The volume of therapeutic albumin 5% administered varied according to patient’s weight and hematocrit and was estimated through the Nadler formula.²⁶

To prevent coagulopathy, PE-A5% sessions were spaced every 2 days after the second PE session. To prevent hypogammaglobulinemia and potential infections, intravenous polyclonal immunoglobulin (200 mg/kg) was administered every two PE-A5% sessions at an infusion rate of 0.5 ml/kg/h. Blood samples were obtained for laboratory determinations. For further details in relation to products used in the study, please refer to the supplementary CTAT table.

Assessment of albumin functional capacity

Albumin function was assessed in terms of albumin binding capacity to both Sudlow site II and fatty acids, and albumin antioxidant capacity. The albumin binding capacity to Sudlow site II, which is one of the main drug albumin binding sites,²⁷ was determined by the albumin binding capacity assay, using the specific fluorescent marker dansylsarcosine.²⁸ Results were expressed as albumin concentration, referring to the content of albumin with high binding capacity.

Albumin’s fatty acid-binding capacity was determined by electronic paramagnetic resonance. Samples were incubated with increasing concentrations of the spin probe 16-doxylstearic acid, and results were expressed by the dissociation constant (Kd).

The albumin Cys34 thiol antioxidant capacity was analyzed by anionic-exchange chromatography as detailed elsewhere.²⁹ Results were expressed as the total amount of each albumin form: reduced albumin (human mercaptalbumin; HMA), reversibly oxidized albumin (human non-mercaptoalbumin form 1, HNA1) and irreversibly oxidized albumin (human non-mercaptoalbumin form 2, HNA2).

The total amount of native albumin (nAlb), the primary structure-preserved form of albumin without post-translational modifications, was determined by mass spectrometry. Briefly, albumin-enriched samples were obtained by Cibacron Blue dye and analyzed on the intact mass protein (top-down) by using ultra high-performance liquid chromatography coupled to electrospray ionization mass spectrometer. The identification of albumin post-translational modifications was performed as previously described and expressed as a percentage of relative intensity.²⁹

Plasma samples from 19 age-matched healthy donors (aged 48-73, kindly provided by Ace Alzheimer Center, Barcelona, Spain) were used as a control for albumin functional capacity at baseline.

Assessment of circulatory function

Circulatory function was assessed invasively at baseline and after completing the six PE sessions. Briefly, an introducer was placed in the right jugular vein using the Seldinger’s technique. A Swan-Ganz catheter was advanced into the pulmonary artery to measure cardiopulmonary pressures and cardiac output was determined by thermal dilution. A 7 F balloon catheter was introduced into the right hepatic vein to determine free and wedged hepatic pressure. All hemodynamic measurements were performed in triplicate. Systolic and diastolic blood pressures, mean arterial pressure (MAP), right arterial pressure, pulmonary artery and capillary pressures, free and wedge hepatic pressures and hepatic venous pressure gradient (mmHg), heart rate (bpm), cardiac index (L/min/m²), systolic volume (ml), systemic vascular resistance (dyn.s.cm⁵) calculated as ([MAP-right atrial pressure]/cardiac output x 80), and systemic vascular resistance index (dyn.s.cm⁵/m²) were performed. All of them were measured before the first PE and repeated 24 h after the last PE session.

Circulatory function was also assessed by measuring plasma renin activity,^19,30 aldosterone, noradrenaline plasma concentration, vasopressin plasma levels, atrial natriuretic factor and brain natriuretic peptide at the beginning of the baseline hemodynamic study, and at the start of the second hemodynamic study.

Assessment of renal, cerebral, and liver dysfunction

To assess renal function, plasma levels of blood urea nitrogen, serum creatinine, sodium, potassium, and phosphates were determined daily, before and after each PE-A5% session by standard clinical analysis.

Cerebral function was measured daily by the degree of hepatic encephalopathy, according to the West Haven criteria.³¹

Conventional parameters of liver function were determined daily, before and after each PE-A5%: alanine aminotransferase (ALT), gamma-glutamyltransferase (GGT), alkaline phosphatase (ALP), aspartate aminotransferase (AST), serum albumin, prothrombin index, and serum bilirubin (total, and conjugated). The Child-Pugh score,³² MELD score,³³ APACHE II,³⁴ SAPS II, and the SOFA score³⁴ were also calculated daily. Liver toxins (bile acids, ammonium, and lactate) were determined before and after each PE-A5% session by standard methods.

Assessment of systemic inflammation, endothelial activation, and oxidative stress

Plasma levels of inflammatory markers IL-6, IL-8, TNF-α and CRP were quantified by immunoassay, and leukocyte count by flow cytometry. As endothelial activation parameters, serum metabolites of nitric oxide (nitrates and nitrites), von Willebrand factor antigen (vWF:Ag) and von Willebrand factor ristocetin cofactor (vWF:RCo) were measured by enzyme immunoassay. Plasma levels of oxidative stress parameters alpha-tocopherol, malondialdehyde and ascorbic acid were determined by high-performance liquid chromatography. All of them were determined before and after each PE-A5% session.

Safety

Safety variables were assessments of AEs, vital signs, physical assessments, clinical chemistry, hematology, and coagulation tests. The main safety criterion was the percentage of PE-A5% associated with at least one AE that may be related to the study procedure. AEs were classified by severity (severe, moderate, mild), seriousness (serious adverse events [SAEs]), and causality to the procedure or study treatment (adverse drug reaction).

Statistical analysis

As this was a pilot study, a sample size of 10 patients was deemed acceptable to allow for the evaluation of the study objectives. Continuous variables are presented as mean (SEM). Categorical variables are presented as n (%).

Albumin function comparisons between healthy donors and patients with ACLF were analyzed using unpaired t test with or the Mann-Whitney test comparison. Continuous efficacy variables were analyzed through mixed models for repeated measurements, including time in the model as a categorical factor. The dependent variable was the value at each time-point and differences between all time-points were assessed in an exploratory manner. The main contrasts were the difference between the first observation (either the baseline value or the first scheduled time-point measurement) and the last assessment. The treatment effect was estimated through adjusted least square means (±SEM). A p value <0.05 indicated a statistically significant difference among variables.

Results

Patient characteristics

Ten out of 29 screened patients were enrolled (Fig. 1). Two patients did not undergo the expected six PE-A5% sessions as per protocol. One of them completed five PE-A5% sessions and then stopped due to a nosocomial pneumonia caused by cytomegalovirus and died, while the other patient completed three sessions and then stopped due to hepatic transplant. Four patients did not complete the 1-month follow-up visit. Baseline patient characteristics are shown in Table 1.

Fig. 1.

Flowchart of enrolment and follow-up of patients through the study period.

PE-A5%, plasma exchange with albumin 5%.

Table 1.

Baseline demographic and clinical characteristics of the enrolled population of patients with acute-on-chronic liver failure (n = 10).

Variable	n (%)
Demographic
Sex, male	6 (60)
Age, years, mean (SD)	55 (9.3)
BMI, kg/m2, mean (SD)	29.6 (7.1)
Cirrhosis etiology
Alcohol	6 (60)
HCV	2 (20)
Primary biliary cirrhosis	1 (10)
Alpha-1 antitrypsin deficiency	1 (10)
Previous complications
Spontaneous bacterial peritonitisa	3 (33.3)
Hepatic encephalopathya	6 (66.7)
Gastrointestinal bleedinga	5 (55.6)
Ascites	8 (80)
Varices sizea ≥5 mm	6 (66.7)
Concomitant pathologies with frequency ≥20%
Type 2 diabetes mellitus	3 (30)
Depression	2 (20)
Hypertension	2 (20)
Concomitant medicationb
Furosemide	5 (50)
Spironolactone	3 (30)
Norfloxacin	2 (20)

^aAssessed in nine patients.

^bConcomitant medication related to study pathology.

At inclusion, three patients were receiving norepinephrine (septic shock), two were mechanically ventilated, one received terlipressin (hepatorenal syndrome) and renal replacement therapy. During the study period, nine (90%) patients were treated with systemic antibiotics, including quinolones (n = 6), beta-lactam (n = 7), aminoglycoside (n = 3), and other antibiotics (n = 6). Regarding vasopressors, two (20%) patients received norepinephrine and terlipressin due to septic shock and hepatorenal syndrome.

Baseline albumin functional capacity of patients with ACLF vs. healthy donors

At baseline, albumin binding capacity of patients with ACLF was significantly reduced in comparison to the external cohort of healthy donors, both for Sudlow site II (15.3 ± 1.6 mg/ml vs. 35.0 ± 0.8 mg/ml, p <0.0001, Fig. 2A) and for fatty acids Kd (8.3 ± 1.6 μM vs. 1.4 ± 0.3 μM, p = 0.0008, Fig. 2B). In patients with ACLF, albumin oxidation was markedly increased, showing a decrease of the reduced albumin form compared with healthy donors (HMA 9.5 ± 1.5 mg/ml vs. 27.0 ± 0.8 mg/ml, p <0.0001, Fig. 2C) while both reversible and irreversible oxidation forms were increased (HNA1: 18.8 ± 1.9 mg/ml vs. 12.7 ± 0.5 mg/ml, p = 0.002, Fig. 2D; HNA2: 4.5 ± 0.6 mg/ml vs. 0.7 ± 0.1 mg/ml, p <0.0001, Fig. 2E). Similarly, nAlb was also decreased in patients with ACLF compared to healthy donors (6.5 ± 1.0 mg/ml vs. 19.7 ± 0.4 mg/ml, p <0.0001, Fig. 2F).

Fig. 2.

Albumin functional capacity in patients with ACLF (n = 10) compared to HD (n = 13) at baseline.

(A) ABiC to Sudlow Binding Site II (p <0.0001); (B) fatty acid-binding capacity (p = 0.0008); and (C–F) albumin antioxidant capacity: (C) albumin reduced form (HMA) (p <0.0001); (D) reversibly oxidized form (HNA1) (p = 0.002); (E) irreversibly oxidized form (HNA2) (p <0.0001); (F) nAlb (p <0.0001). Unpaired t test (A-D, F) and Mann-Whitney test (E). Data are expressed as mean ± SEM. ABiC, albumin binding capacity; ACLF, acute-on-chronic liver failure; HD, healthy donor; HMA, human mercaptalbumin; HNA1/2, human non-mercaptoalbumin form 1/2; nAlb, native albumin.

Effects of PE-A5% treatment on albumin functional capacity

In the enrolled population of patients with ACLF (n = 10), albumin binding capacity to Sudlow site II significantly increased over time, from 15.3 ± 1.6 mg/ml (LS means ± SEM) at day 1 to 18.9 ± 1.7 mg/ml after the six PE-A5% sessions (day 11) (p = 0.003) (Fig. 3A). Albumin’s fatty acid-binding capacity significantly decreased (in terms of Kd) from 8.2 ± 1.4 μM (day 1) to 3.1 ± 1.5 μM (day 11) (p = 0.013, Fig. 3B). Albumin antioxidant capacity, in terms of HMA amount, was significantly increased throughout the study period, from 9.5 ± 1.5 mg/ml (day 1) to 14.6 ± 1.6 mg/ml (day 11) (p <0.001) (Fig. 3C). The nAlb levels were also significantly increased throughout PE-A5% sessions from 6.5 ± 1.0 mg/ml (day 1) to 10.2 ± 1.4 mg/ml (day 11) (p = 0.035) (Fig. 3D). Albumin binding and antioxidant capacities, as well as the amount of nAlb, increased after most of the PE-A5% sessions, compared with pre-PE-A5%. Similar results were obtained in the per protocol population (n = 8) (data not shown).

Fig. 3.

Effects of PE-A5% on albumin functional capacity in cirrhotic patients with acute-on-chronic liver failure.

Grey areas denote the start and the end of each of the 6 PE-A5% sessions. (A) Albumin binding capacity (ABiC) to Sudlow Binding Site II (p = 0.003); (B) fatty acid-binding capacity (p = 0.013); (C-D) albumin antioxidant capacity: (C) albumin reduced form (HMA) (p <0.001); (D) nAlb (p = 0.035). Mixed models for repeated measurements analysis. Data are expressed as least squares mean ± SEM (day 1: n = 9; day 11: n = 8). Significant p values (p <0.05) are shown. ABiC, albumin binding capacity; ACLF, acute-on-chronic liver failure; HD, healthy donor; HMA, human mercaptalbumin; PE-A5%, plasma exchange with albumin 5%; nAlb, native albumin.

Effects of PE-A5% treatment on circulatory function

PE-A5% treatment was associated with a significant increase in MAP (baseline, 73.6 ± 3.0 mmHg vs. day 11, 82.0 ± 3.3 mmHg; p = 0.026), diastolic blood pressure (baseline, 56.1 ± 2.2 mmHg vs. day 11, 62.8 ± 2.4 mmHg; p = 0.009), heart rate (baseline, 78.5 ± 5.6 bpm vs. day 11, 95.9 ± 6.1 bpm; p = 0.013) and cardiac index (baseline, 4.8 ± 0.6 L/min/m² vs. day 11, 6.1 ± 0.7 L/min/m²; p = 0.049), from the baseline visit up to day 11, after six PE-A5% sessions (Fig. 4). There were no statistically significant changes in pulmonary artery and capillary pressure, systolic volume, nor systemic vascular resistance. Regarding splanchnic hemodynamics, PE-A5% treatment was not associated with significant changes in hepatic venous pressure gradient. Plasma renin activity, noradrenaline, vasopressin, atrial natriuretic factor and brain natriuretic peptide did not show significant changes during the study. When analyzing hormonal determinations, only aldosterone showed statistically significant differences over treatment, with a decrease from baseline (47.0 ± 10.0 mg/dl) vs. day 11 (30.7 ± 10.6 mg/dl) (p = 0.047) (Table S1).

Fig. 4.

Effects of PE-A5% on circulatory function in patients with acute-on-chronic liver failure at baseline (n = 10) and after the six PE-A5% sessions (day 11) (n = 8).

(A) MAP (p = 0.026); (B) DBP (p = 0.009); (C) heart rate (p = 0.013); (D) cardiac index (p = 0.049). Mixed models for repeated measurements analysis. Data are expressed as least squares mean ± SEM. Significant p values (p <0.05) are shown. DBP, diastolic blood pressure; MAP, mean arterial pressure; PE-A5%, plasma exchange with albumin 5%.

Effects of PE-A5% treatment on renal, cerebral, and liver function

PE-A5% treatment improved renal function, as demonstrated by a significant reduction in creatinine levels over time from 1.8 ± 0.2 mg/dl at baseline to 0.8 ± 0.2 mg/dl at day 10, and 1.3 ± 0.2 mg/dl at month 1 (p <0.001). Similarly, blood urea nitrogen values decreased over time, from 58.5 ± 13.9 at baseline to 33.2 ± 13.9 mg/dl at day 10 and 30.8 ± 15.0 mg/dl at month 1 (p = 0.036) (Fig. 5A, Table S2).

Fig. 5.

Effects of PE-A5% on renal, cerebral, and liver function in patients with acute-on-chronic liver failure.

Grey areas denote the start and the end of each of the six PE-A5% sessions. (A) Creatinine (p <0.001) and BUN (p = 0.036); (B) Hepatic encephalopathy grade (percentage of patients) following West Haven criteria (p = 0.006); (C) Total bilirubin (p <0.001); conjugated bilirubin (p = 0.001); (D) MELD score (p <0.001). Mixed models for repeated measurements analysis. Data are expressed as least squares mean ± SEM (A, C, D) or percentage of patients (B) (baseline: n = 10; day 10: n = 7; month 1: n = 4). Significant p values (p <0.05) are shown. BUN, blood urea nitrogen; PE-A5%, plasma exchange with albumin 5%.

Hepatic encephalopathy grade, measured by West Haven criteria, was significantly reduced over the time (baseline vs. month 1, p = 0.006). At baseline, most patients (n = 5, 55.6%) showed hepatic encephalopathy grade 2, whereas at day 10, four out of seven (57.1%) patients were classified at grade 0 (Fig. 5B). Similar results were observed through time in the post-PE-A5% periods (post-PE-A5%, day 1 to day 10: p = 0.015).

Conventional hepatic function parameters – ALT, GGT, ALP, AST (Table S2), and bilirubin (Fig. 5C) – were significantly diminished from baseline to day 10 and month 1. There were also statistically significant differences across visits between pre-PE-A5% and post-PE-A5% values of ALT, GGT, ALP, AST, total and conjugated bilirubin. Prothrombin index and hepatic toxin levels did not show statistically significant changes during the study whereas serum albumin levels were restored at the end of PE-A5% treatment (Table S2).

MELD score significantly decreased from baseline (33.1 ± 1.5) to day 10 (27.2 ± 1.6) (Fig. 5D). Post-PE-A5% values of SOFA score decreased from day 1 to day 10 (p = 0.05). There were no significant differences in the rest of prognostic scores, APACHE, Child-Pugh and SAPSII, through the study period (Table S2).

Effects of PE-A5% on systemic inflammation, endothelial activation and oxidative stress

There was a significant reduction in CRP (Fig. 6A, post-PE-A5%, p = 0.033) and leukocyte count (Fig. 6B, pre-PE-A5%, p = 0.003) from day 1 to day 10. However, no significant changes in cytokines IL-6, IL-8, and TNF-α were observed (Table S3). Regarding the endothelial function parameters, vWF:Ag levels significantly dropped off after PE-A5% treatment over time (Fig. 6C, pre-PE-A5%, p = 0.005), whereas vWF:RCo levels remained unchanged. PE-A5% had no statistically significant effect on nitric oxide levels (measured by nitrites and nitrates) nor oxidative stress parameters (Table S3).

Fig. 6.

Effects of PE-A5% on systemic inflammation and endothelial activation in patients with acute-on-chronic liver failure.

Grey areas denote the start and the end of each of the 6 PE-A5% sessions. (A) CRP (p = 0.033); (B) leukocyte count (p = 0.003) and (C) vWF:Ag (p = 0.005). Mixed models for repeated measurements analysis. Data are expressed as least squares mean ± SEM (day 1: n = 9; day 10: n = 8). Significant p values (p <0.05) are shown. PE-A5%, plasma exchange with albumin 5%; vWF:Ag, von Willebrand factor antigen.

Safety evaluation

There were a total of 78 AEs and all patients had at least one AE. The most frequent AEs were anemia (9 patients, n = 15 events), hypokalemia (5 patients, n = 6), thrombocytopenia (3 patients, n = 5) and hypofibrinogenemia (3 patients, n = 5) from baseline up to the 1-month follow-up. Regarding severity, AEs were classified as severe (n = 13, 17%), moderate (n = 12, 15%) and mild (n = 53, 68%). Severe AEs were atrial fibrillation, multi-organ failure, acute hepatic failure (2), hepatorenal syndrome, bacterial peritonitis, pneumonia (2), hyponatremia, acute kidney injury, and bronchial hemorrhage.

Thirty-eight AEs (49%) were related to the study procedure and 23 (40%) of PE-A5% procedures were associated with at least one procedure-related AE. Twenty-eight AEs (36%) were related to the study product and study procedure (Table 2). Adverse drug reactions were not considered serious: anemia (8 patients, n = 13 events), hypofibrinogenemia (2 patients, n = 3), thrombocytopenia (2 patients, n = 3), urinary tract infection (2 patients, n = 3), injection site hemorrhage (1 patient, n = 2), bacteremia (1 patient, n = 1), hypokalemia (1 patient, n = 1) and hypomagnesaemia (1 patient, n = 2).

Table 2.

Summary of AEs in the safety population (n = 10).

Variable	Value
Total AEs, n	78
Patients with at least one AE, n (%)	10 (100)
Total PE sessions performed, n	57
Patients with product-related AEs, n (%)	9 (90)
Product-related AEs by PE, n (%)	28 (49)
% of total AEs	36
PE associated with AE, n	38
Patients with procedure-related AEs, n (%)	10 (100)
% of total PE sessions	66.7
% of total AEs	48.7
PE affected, n (% of procedures)	25 (44)
PE associated with AE – 72 h, n	36
Patients with procedure-related AE 72 h, n (%)	10 (100)
% of total PE sessions	63.2
% of total AEs	46.2
PE sessions affected, n (% of procedures)	23 (40)
Total SAEs, n	16
Patients with SAE, n (%)	8 (80)
% of total AEs	20.5
Patients with product/procedure-related SAEs, n (%)	0 (0)
Patients with AE with outcome of deatha	4
% of total AEs	5.1

AE, adverse event; PE, plasma exchange; SAE, serious adverse event.

^aOne patient died before study treatment administration.

There were 16 SAEs reported in eight patients, none of them considered related to the study procedure or study product. Four deaths occurred following commencement of PE-A5% due to progression of ACLF (n = 4).

Regarding vital signs and physical assessments, a post-PE-A5% increase in heart rate was the only statistically significant parameter detected. Hematocrit recovered from 28.1 ± 1.0 % at baseline to 34.0 ± 1.5 % at month 1 (p = 0.015). There was a statistically significant decrease for platelet count (x10⁹/L) from baseline to month 1 (67.2 ± 10.4 vs. 60.8 ± 14.4, respectively: p = 0.037).

Discussion

Albumin levels and its functionality are severely compromised in patients with ACLF.¹⁶ In that sense, albumin dialysis has been used in patients with ACLF to remove albumin-bound components and dysfunctional albumin. MARS (molecular adsorbent recirculating system) improves hepatic encephalopathy but has no impact on survival.⁵ New extracorporeal devices are therefore needed. It is reasonable to think that PE-A5% could also restore the functional capacity of albumin. This therapeutic approach could result in clinical benefits in these sick patients but has never been tested outside cohort studies.^35,36

The current pilot study demonstrates that albumin binding and antioxidant capacities are markedly decreased in patients with ACLF compared to healthy donors. Importantly, PE-A5% almost normalized serum albumin levels, being close to 35-45 g/L at the end of the study period,^15,21 and improved albumin binding and antioxidant capacities. Therefore, the use of albumin in patients with ACLF could go beyond being just a replacement fluid but could be indicated to restore non-oncotic albumin functions such as binding, detoxification and antioxidant activity.

In patients with ACLF, impairment in systemic circulatory function is characterized by reduced MAP and decreased cardiac output, which lead to organ hypoperfusion, multi-organ failure and increased risk of death.^15,37 Our results showed that PE-A5% improves cardiocirculatory function by increasing MAP and cardiac index, thus mitigating organ hypoperfusion. The hemodynamic findings of this study support previous research in which PE increased hepatic blood flow in patients with liver failure,.³⁸ This remarkable beneficial effect of functionally active albumin surprisingly was not attributed to the expansion of central blood volume induced by the protein.³⁹ Markers of effective blood volume remained unchanged after PE-A5%. The decrease in vWF:Ag, a surrogate marker of endothelial dysfunction, was a noteworthy finding that supports the role of PE-A5% on endothelial stabilization, as demonstrated in other studies.^21,40,41

PE-A5% significantly improved liver, renal and cerebral function across the study period. MELD score also improved after PE. Importantly, PE-A5% treatment also modulated the intensity of systemic inflammation, by decreasing serum CRP levels and leukocyte count. Intense systemic inflammatory response is a major contributor to organ failure and a major pathogenic characteristic of ACLF.⁴² PE is an effective detoxification system capable of removing key activators of the inflammation cascade such as damage-associated molecular patterns and pathogen-associated molecular patterns.⁴³ These results suggest that PE-A5% could exert its positive effects in patients with ACLF through two main mechanisms: the attenuation of systemic inflammation and the improvement of cardiocirculatory function.

Regarding safety, most AEs were considered mild or moderate in severity and, importantly, SAEs and deaths occurred during the study period were considered unrelated to the study procedure. The number of PE sessions depends on the patient’s condition and disease but a range between three and six is commonly accepted.⁴⁴ In the current study, six PE-A5% sessions were performed in eight patients (4 and 3 were performed in the others), and less than 50% of them were associated with one procedure-related AE. These results suggest that PE-A5% is a safe procedure, in accordance with previous apheresis data collected from a multinational registry.⁴⁵ The overall safety results of this pilot study indicate that there would be no concern regarding the feasibility and tolerability of PE-A5% in patients with ACLF.

The pilot nature of this study entails a series of limitations, such as being open-label and uncontrolled, and not focused on patient survival. In addition, of the 10 enrolled patients, eight patients completed the intended six PE-A5% sessions, and only four completed the 1-month follow-up visit, which potentially restricts the strength of the study findings. Another limitation of our study is the definition used for the diagnosis of ACLF. A post hoc evaluation of our patients classified all of them as having ACLF according to the EF-CLIF criteria (4 ACLF-1, 3 ACLF-2 and 3 ACLF-3). To confirm the potential clinical relevance of the current results, the pivotal phase III, multicenter, randomized-controlled trial (APACHE trial, NCT03702920; EudraCT 2016-001787-10) is being conducted. This European and North American randomized-controlled trial will include 380 patients with ACLF and excludes patients with severe forms (4 to 6 organ failures) or with ACLF lasting for more than 10 days. It is important to remark that PE was performed using not only 5% albumin, but also FFP (30%) and intravenous polyclonal immunoglobulin (every 2 sessions). The positive effects reported in this study are probably related to the entire treatment, which was evaluated in patients who were also receiving standard medical treatment.

Overall, PE-A5% treatment was a safe procedure associated with a significant improvement in circulatory, renal, cerebral, and liver function, and with an attenuation of systemic inflammatory response. These effects could translate into a potential clinical benefit in terms of survival for patients with ACLF.

Financial support

This study was supported by Grifols (Barcelona, Spain), the manufacturer of Human Albumin Grifols® 5%/Albutein® 5%.

Authors’ contributions

MT, RH, NA, LN, AM, and APe: conceptualization, project administration, formal analysis, supervision, validation, visualization, writing – review & editing; MC and APa: conceptualization, methodology, supervision, writing – review & editing; JF and VA: conceptualization, data curation, resources, investigation, writing – review & editing; ML and JC: data curation, resources, investigation, writing – review & editing. All authors critically revised, edited and approved the final manuscript.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflict of interest

JF has received speaker honorarium from Grifols (not in the last 2 years). MT, NA, LN, RH, AM, APe, MC, and APa are full-time employees of Grifols. ML, JC, and VA declares no conflict of interest related to this manuscript.

Please refer to the accompanying ICMJE disclosure forms for further details.

Supplementary data

The following are the supplementary data to this article: