First Clinical Use of a Novel Bioartificial Liver Support System (BLSS)


  • Partial support for this study was provided by Excorp Medical, Inc., Oakdale, MN, USA.


The first clinical use of the Excorp Medical Bioartificial Liver Support System (BLSS) in support of a 41-year-old African-American female with fulminant hepatic failure is described. The BLSS is currently in a Phase I/II safety evaluation at the University of Pittsburgh/UPMC System. Inclusion criteria for the study are patients with acute liver failure, any etiology, presenting with encephalopathy deteriorating beyond Parson's Grade 2. The BLSS consists of a blood pump; a heat exchanger to control blood temperature; an oxygenator to control oxygenation and pH; a bioreactor; and associated pressure and flow alarm systems. Patient liver support is provided by 70–100 g of porcine liver cells housed in the hollow fiber bioreactor. The patient exhibited transient hypotension and thrombocytopenia at initiation of perfusion. The only unanticipated safety event was a lowering of patient glucose level at the onset of perfusion with the BLSS that was treatable with intravenous glucose administration. Moderate changes in blood biochemistries pre- and post perfusion are indicative of liver support being provided by the BLSS. While the initial experience with the BLSS is encouraging, completion of the Phase I/II study is required in order to more fully understand the safety aspects of the BLSS.


Fulminant hepatic failure (FHF) is an acute, catastrophic, usually rapidly fatal illness that results from severe hepatocyte damage or massive liver necrosis (1). The swiftness of acute liver damage is sufficient to cause rapid onset of encephalopathy and coagulopathy followed by multiorgan system failure. Unlike progressive liver failure from underlying cirrhosis, FHF occurs in previously healthy individuals. While some agents that cause FHF are known, no identifiable cause is found for up to 50% of FHF cases. Approximately 2000 FHF cases are reported annually in the United States (2). FHF has an 80% mortality rate in the absence of liver transplantation (3).

The clinical manifestations of FHF are varied, but almost always include advancing encephalopathy, blood coagulopathy, elevated ammonia, lactate, and bilirubin levels, and frequently include hemodynamic instability, renal failure, and respiratory distress (1). Care of FHF patients presents a tremendous challenge in the intensive care unit (ICU). Standard-of-care treatment is primarily supportive while awaiting the availability of a donor liver for transplantation.

Only four bioartificial liver devices for support for FHF patients have entered sustained clinical trials. The devices all rely on hollow fiber membranes to isolate hepatocytes from direct, shear contact with patient fluids. They differ in the source and treatment of hepatocytes prior to use with a patient and in the choice of perfusate: plasma or whole blood. The Circe Biomedical HepatAssist® System (4, 5), currently in Phase II/III efficacy trials for fulminant hepatic failure or primary nonfunction of a transplanted liver graft, perfuses patient plasma through a bioreactor that contains freshly thawed, cryopreserved, primary porcine hepatocytes. The VitaGen ELAD™ bioartificial liver device (6, 7), currently in a Phase II safety/efficacy trial for fulminant hepatic failure or primary nonfunction of a transplanted liver graft, perfuses patient plasma through a bioreactor that uses the C3A derivative of the HepG2 human hepatoblastoma cell line. The Gerlach BELS (8, 9), which recently finished a Phase I/II safety evaluation in Europe, perfuses patient plasma through a unique bioreactor configuration that uses cultured, primary porcine (more recently, human) hepatocytes.

The Excorp Medical Bioartificial Liver Support System (BLSS), which is the subject of this report, is the most recent system to enter a Phase I/II clinical safety evaluation (10, 11). Like the Gerlach BELS, the BLSS uses cultured, primary porcine hepatocytes. In contrast to the other three systems which perfuse patient plasma through the bioreactor, the BLSS perfuses whole blood through the bioreactor. A recent analysis suggests that, given bioreactors of equal potency, whole-blood perfusion has a significant advantage over plasma perfusion in affecting the patient outcome (12). BLSS safety considerations gained from experience with the first four patients are reported elsewhere (11).

BLSS description

A schematic diagram of the Excorp Medical Bioartificial Liver Support System (BLSS) is provided in Figure 1. The BLSS consists of a blood pump, heat exchanger to control blood temperature, oxygenator to control blood oxygen tension and pH, and bioreactor, with associated pressure and flow alarm systems. The bioreactor is a hollow fiber cartridge that has cellulose acetate membranes with a nominal 100-kDa diffusive size cut-off. Approximately 70–100 g of primary porcine hepatocytes, harvested by a modified two-step collagenase digestion technique, are infused into the extralumenal space of the bioreactor cartridge. The hepatocytes are harvested under an approved Institutional Animal Care and Use Committee protocol.

Figure 1.

Schematic flow diagram of the BLSS. Dotted lines are the blood-flow path; solid lines are gas-flow paths; dashed lines are nutrient media-flow path.

After loading with hepatocytes, the bioreactor is maintained under physiological tissue culture conditions in an incubator prior to use with a patient. A chemically defined medium (Williams Medium E enhanced with: 100 U/L insulin; 0.4 mg/L glucagon; 0.33 mg/L dexamethasone; 5 μg/L epidermal growth factor; 6.25 mg/L transferrin; 2.9 μg/L selenium) is perfused through the blood path. The composition of the nutrient medium that is percolated through the hepatocyte cell mass is similar in that it is further augmented with 500 mg/L albumin but has no insulin. Oxygenation and pH control is maintained with the use of gas mass flow controllers that control the blend of O2/CO2/N2 through the oxygenator. Prior to use with a patient, the lumen flow path is purged with PlasmaLyte, a physiologic buffer.

Phase I/II clinical safety evaluation protocol

The Phase I/II clinical safety evaluation protocol has University of Pittsburgh Institutional Review Board and Food and Drug Administration approval. Inclusion criteria for the study are age between 18 and 65 years with acute liver failure (FHF or acute decompensation on chronic liver disease) accompanied by encephalopathy deteriorating beyond Parson's Grade 2. Primary exclusion criteria include clinical evidence of brainstem herniation or intracranial hemorrhage; inoperable liver cancer or cancer metastatic to the liver; respiratory distress defined as PaO2/FiO2 < 200; pressor support requirement > 0.15 μg/kg/min norepinephrine equivalent; uncontrolled active hemorrhage from any site; uncontrolled septic shock; and platelet count < 50 000/μL that is uncorrectable. Fifteen supported patients will be enrolled in this study.

As encephalopathy is a required inclusion criterion for enrollment into this study, patients are incapable of providing informed consent on their own. Informed surrogate consent is thus sought for patients admitted to the Liver Transplant Intensive Care Unit (ICU) at UPMC Presbyterian University Hospital who meet the safety study inclusion criteria and who are not excluded according to the exclusion criteria. A full explanation and discussion of the study risks, including those risks associated with xenotransplantation, is provided to surrogate consent providers prior to obtaining consent.

A single study period under the protocol consists of 12 h of preperfusion, baseline monitoring, 12 h of extracorporeal blood perfusion through the BLSS, and 12 h post perfusion baseline monitoring. Physiological parameters such as blood pressures, heart rate, temperature, and respiratory function are monitored continuously during the study period. Blood chemistries and hematologies are obtained every 6 h during pre- and post perfusion monitoring and every 4 h during perfusion. If the patient tolerates the first perfusion well and remains within the inclusion/exclusion criteria, he/she may receive a second perfusion at the discretion of the attending physician.

Case Report

A 41-year-old African-American female presented with left upper quadrant abdominal pain, headache, lower chest pain, fever, moist cough, pancytopenia, fatigue, and elevated liver enzymes. The patient had a 6-year history of systemic lupus erythematosis and during the month prior to developing FHF had been treated with pulse steroids and high-dose cytoxan for Stage IV diffuse proliferative glomerular nephritis. She was being treated with daily prednisone at the time of admission. She was also thrombocytopenic and bacteremic with Enterococcus faecalis.

The patient had no prior diagnosed history of liver failure at the time of the present admission. She acknowledged use of extra-strength acetaminophen on a regular basis, three or four times per day, but indicated no alcohol abuse or intravenous drug abuse. Blood screens for acetaminophen, alcohol, and drugs were negative. Likewise, screens for hepatitis virus, cytomegalovirus, Epstein-Barr virus, and herpes simplex virus were also negative. The patient was deeply icteric, had asterixis, and had mild tenderness on the left side of the abdomen and left lower ribs. The skin did not show signs of chronic liver disease. An abdominal ultrasound scan showed normal echogenicity of the liver with no intrahepatic biliary duct dilatation. Computer tomography of the abdomen revealed a liver volume of 1522 mL and a normal pancreas. Liver enzymes, as noted in Figure 6, were extremely elevated.

Figure 6.

Total bilirubin, aspartate aminotransferase (AST), and alanine aminotransferase (ALT) levels as a function of time post start of the first perfusion. Vertical bars enclose perfusion periods (solid: first perfusion; dotted: second perfusion).

The patient was admitted to the ICU with the diagnosis of FHF of unclear etiology. Her level of consciousness deteriorated shortly after arrival. Standard-of-care treatment included acyclovir, timentin, fluoconazole, and lactulose administration. Fresh frozen plasma was used to correct bleeding from invasive procedures and coagulopathy. Following a transjugular liver biopsy that showed acute hepatitis with centrilobular-to-panlobular hemorrhagic necrosis affecting 50–60% of the sample parenchyma (suggesting a toxic insult to the liver), the patient was listed for orthotopic liver transplantation (OLTx) as UNOS Status I. The liver biopsy showed no evidence of chronic injury – either fibrosis or cirrhosis.

Surrogate consent was sought and obtained and the patient was enrolled in the study. Baseline preperfusion monitoring began approximately 36 h after admission. Perfusion was initiated approximately 12 h later and was relatively uneventful. The patient was intubated part way through the first perfusion to protect her airway. As the patient remained within inclusion/exclusion criteria and tolerated the first perfusion well, a second study period was initiated using the post perfusion monitoring period of the first perfusion as the preperfusion monitoring period for the second perfusion. The second perfusion was also well tolerated. Although the BLSS has provision for heparin administration to maintain coagulation parameters in a specified range, no heparin administration was used or was necessary with this patient.

Although standard of care for FHF patients, an intracranial pressure probe was not performed because the patient was bacteremic with Enterococcus. Rather, we assessed blood-flow velocity in the middle cerebral artery using transcranial Doppler (TCD). After the first perfusion TCD demonstrated normal systolic and diastolic flow velocity. TCD demonstrated a marginal increase in systolic acceleration after the second perfusion. Diastolic flow velocity was preserved, suggesting that intracranial pressure was not significantly elevated. A computer tomography scan of the head demonstrated minimal dedifferentiation of gray and white matter. This was probably age appropriate, but may also be observed in acute ischemic injury.

The patient remained stable post second perfusion and was extubated 6 days later. The patient died 10 days post perfusion from septic shock due to resistant Pseudomonas aeruginosa infection unrelated to the BLSS perfusion. The organism was also isolated from an implanted catheter tip in addition to coagulase-negative Staphylococcus and Enterococcus faecium. Blood cultures drawn 6 h post second perfusion and again 7 d post second perfusion were negative for growth. Although the patient was listed as UNOS Status I for her entire hospitalization, no suitable liver was available for OLTx during that time.


The primary emphasis in the current study is a safety evaluation of the Excorp Medical BLSS in support of patients with acute liver failure. Primary measures of the safety evaluation are hemodynamic parameters, which are measured continuously, and vital blood chemistries and hematologies, which are measured periodically during perfusion. Continuous systolic, diastolic, and mean arterial pressure measurements during the entire study period are shown in Figure 2. Other than a transient period of hypotension noted at the start of each perfusion, patient hemodynamic parameters remained stable. We anticipated transient hypotension after starting perfusion from our preclinical experience (10), and treated it with volume expansion with crystalloid and colloid. The decrease in arterial pressure was attenuated with the second perfusion, as the patient volume was expanded prior to starting.

Figure 2.

Patient systolic, diastolic, and mean arterial pressures as a function of time post start of first perfusion. Vertical bars enclose perfusion periods (solid: first perfusion; dotted: second perfusion).

Platelet and white blood-cell depletion is a concern with any extracorporeal perfusion and an especial concern with FHF patients. Figure 3 displays the platelet and white blood-cell counts throughout the study period and for 4 d post the end of the second perfusion. The patient platelet count was corrected to > 50 000/μL at the start of each perfusion. During perfusion, platelet count dropped, but was correctable post perfusion. Our preclinical experience (10) found approximately 10% reduction in platelets with initiation of perfusion with return to preperfusion levels after an hour. Thus, the loss in platelets and constant need for replacement is believed to be due to the natural progression of the disease (13). The white blood-cell count was relatively unaffected by perfusion. As noted in Table 1, patient red blood-cell count remained fairly constant at 3·106/μL throughout the perfusion periods. Also, serum hemoglobin, a measure of hemolysis from the perfusion, actually decreased through the perfusion periods.

Figure 3.

Platelet and white blood cell (WBC) counts as a function of time post start of the first perfusion. Vertical bars enclose perfusion periods (solid: first perfusion; dotted: second persusion).

Table 1. : Selected parameters at admission, 6 h pre first perfusion, 6 h post first perfusion/pre second perfusion, and 6 h post second perfusion
ParameterAdmission6 h pre6 h post 1st6 h post
  first6 h pre 2ndsecond
  1. PT: prothrombin time; INR: International Normalized Ratio; RBC: red blood cells.

Albumin, mg/dL 2.0 2.3 2.8 2.2
PT, s 19.617.015.8
INR  1.7 1.5 1.4
RBC, 106/μL 3.63 3.19 3.17 3.08
Serum hemoglobin, mg/dL 23.516.012.5

An unexpected finding was the severity of hypoglycemia which attended each perfusion (Figure 4). We treated this with boluses of dextrose and a continuous infusion of dextrose to prevent further episodes of hypoglycemia. The mechanism for this hypoglycemia is unclear. The nutrient medium that is percolated through the bioreactor during patient perfusion has no insulin. The maximum amount of insulin that might be resident in the hepatocyte cell space at the initiation of perfusion is calculated at 2 U. Thus the likelihood of insulin from the bioreactor transiting the membrane to the patient's bloodstream affecting glucose sequestration in the patient is low. Blood urea and creatinine levels (Figure 4) are unchanged during the perfusion as no ultrafiltration is performed.

Figure 4.

Glucose, blood urea nitrogen (BUN), and creatinine levels as a function of time post start of first perfusion. Vertical bars enclose perfusion periods (solid: first perfusion; dotted: second perfusion).

Although the primary focus of the study is safety, measures of efficacy are also of interest. Figure 5 shows the changes in ammonia and lactate through the perfusion periods and 4 d post the second perfusion. Both ammonia and lactate were rapidly cleared for this patient. Figure 6 shows changes in total bilirubin, aspartate aminotransferase (AST), and alanine aminotransferase (ALT) with time. While the BLSS can clear bilirubin through hepatocyte activity, changes in the AST and ALT levels more likely reflect the passage of time from the liver injury than any effect from the BLSS. Finally, gas exchange was not compromised by the perfusion. In fact, the PaO2/FIO2 ratio actually improved during and post perfusion (Figure 7).

Figure 5.

Ammonia and lactate levels as a function of time post start of first perfusion. Vertical bars enclose perfusion periods (solid: first perfusion; dotted: second perfusion).

Figure 7.

PaO2/FiO2 ratio as a function of time post start of first perfusion. Vertical bars enclose perfusion periods (solid: first perfusion; dotted: second perfusion).

Neurological status was monitored with a modified Glasgow Coma Score (GCS) (appropriate verbal response 1–5 points; best motor response 1–6 points; eye opening 1–4 points; total 3–15 points). The composite and part scores are displayed in Figure 8. The patient was sedated during BLSS perfusion. leading to lowered scores during perfusion. Also, the patient was intubated partway through the first BLSS perfusion, resulting in a verbal response score of 1 while intubated. As indicated in Figure 8, the patient remained neurologically stable post perfusion with maximum motor response and eye-opening scores. The patient was extubated 6 d post perfusion.

Figure 8.

Glasgow Coma Scores (GCS) as a function of time post start of first perfusion. Vertical bars enclose perfusion periods (solid: first perfusion; dotted: second perfusion).

Bioreactor activity evaluation post perfusion

A new bioreactor was prepared for each patient perfusion. Each bioreactor was prepared according to protocol and met clinical acceptance standards. Evaluation of the bioreactor conjugation activity, as measured by 4-methyl-umbelliferone (4-MU) conversion, and P450 enzyme activity, as measured by lidocaine conversion, post perfusion are shown in Figures 9 and 10, respectively, for the bioreactors from both perfusions. The activity evaluation follows the clearance of 4-MU and lidocaine from a 1-L reservoir over a 2-h period while recirculating the reservoir contents through the bioreactor. Fractional conversion or clearance is defined as the fractional amount (relative to the initial amount) of the species removed from the reservoir. Also provided is the activity seen with a typical fresh bioreactor prior to being used for patient perfusion. As seen in Figure 9, 4-MU conjugation clearance remains unchanged after 12 h of perfusion with FHF patient blood. The 4-MU was converted almost quantitatively (data not shown) to 4-methyl-umbelliferone-glucuronide with only trace amounts of 4-methyl-umbelliferone-sulfate. The bioreactors retain approximately 75% of lidocaine clearance activity after 12 h of perfusion with FHF patient blood, as seen in Figure 10. The major lidocaine metabolite was monoethylglycinexylidide with lesser amounts of 3-hydroxy-lidocaine being formed (data not shown). No statistical significance can be attached to the single experimental evaluations depicted in Figures 9 and 10.

Figure 9.

Fractional 4-methyl-umbelliferone (4-MU) conversion as a function of time post injection of a bolus after 12 h of perfusion with FHF patient blood.

Figure 10.

Fractional lidocaine conversion as a function of time post injection of a bolus after 12 h of perfusion with FHF patient blood.


Fulminant hepatic failure is seen as a formidable challenge in the ICU environment, given the conventional supportive treatments presently available (1). The high rate of mortality in the absence of liver transplantation has encouraged the search for and development of novel techniques for improved support of this population. The scarcity of liver donors makes management of FHF patients even more challenging because the indications for and timing of the decision for OLTx change frequently during the evolution and progression of FHF. The presence of encephalopathy and usual concurrent appearance of brain edema associated with FHF further complicates patient management.

The liver can be regarded as a complex biochemical reactor that performs several functions: oxidative detoxification; intermediate metabolism and supply of nutrients; toxin and waste excretion through the bile; protein and macromolecule synthesis; and modulation of immune and hormonal systems. The Excorp Medical BLSS is one of the new generation of liver assist devices that are designed to augment these functions in patients with acute liver failure through the use of hepatocytes housed in a bioreactor.

The initial clinical experience with the BLSS was relatively uneventful from a safety evaluation perspective. Although provision for heparin administration to help prevent clotting in the bioreactor is available with the BLSS, no heparin was used with the first patient. The patient's activated clotting time (ACT) remained above the desired minimum of 200 s without anticoagulation administration. An expected transient hypotension at the initiation of perfusion was noted and treated satisfactorily with fluid bolus. An unanticipated depletion of glucose was noted and treated satisfactorily with intravenous glucose administration. Platelet reduction, whether from BLSS perfusion or through the normal course of FHF failure, was correctable with platelet administration.

Associated with BLSS perfusion were improvements in blood chemistries associated with intermediate metabolism (ammonia and lactate reduction, Figure 5), improvements in blood chemistries associated with toxin and waste excretion (total bilirubin reduction, Figure 6), and indications of protein and macromolecule synthesis as evidenced by reduction in coagulopathy as measured by prothrombin time (Table 1). An improvement in respiratory function as evidenced by improving PaO2/FiO2 ratio (Figure 7) was also noted. The chemical and physiological improvements correlated with clinical improvement, allowing the patient to be successfully extubated.

Our assessment of the safety profile of the BLSS in this first treated patient is positive. We encountered problems which were easily resolved, as noted. We volume expanded the patient to correct rapidly the hypotension which developed within minutes of starting perfusion. Thrombocytopenia, which likely reflected marrow depression from cytoxan, was easily corrected and possibly not related to the BLSS. Hypoglycemia, once recognized, was easily treated and prevented. The patient remained coagulopathic throughout both perfusions with ACT > 200 s and we did not use heparin anticoagulation.


The authors gratefully acknowledge the efforts of the Liver Transplant Intensive Care Unit nursing staff at UPMC Presbyterian Hospital, who contributed substantially to the performance of this study, and the efforts of the staff at the University of Pittsburgh Central Animal Facility, who have also contributed substantially.