Liver Support by Extracorporeal Blood Purification: A Clinical Observation Jan Stange,* Steffen R. Mitzner,* Sebastian Klammt,† Jens Freytag,* Piotr Peszynski,* Jan Loock,* Heiko Hickstein,* Gero Korten,* Reinhardt Schmidt,* Jörg Hentschel,‡ Martin Schulz,‡ Matthias Löhr,† Stefan Liebe,† Wolfgang Schareck,§ and Ullrich T. Hopt§ Liver failure associated with excretory insufficiency and jaundice results in an endogenous accumulation of toxins involved in the impairment of cardiovascular, kidney, and cerebral function. Moreover, these toxins have been shown to damage the liver itself by inducing hepatocellular apoptosis and necrosis, thus creating a vicious cycle of the disease. We report a retrospective cohort study of 26 patients with acute or chronic liver failure with intrahepatic cholestasis (bilirubin level > 20 mg/dL) who underwent a new extracorporeal blood purification treatment. A synthetic hydrophilic/hydrophobic domain-presenting semipermeable membrane (pore size < albumin size, 100-nm thick) was used for extracorporeal blood detoxification using dialysis equipment. The opposite side was rinsed with ligandin-like proteins as molecular adsorbents that were regenerated online using a chromatography-like recycling system (molecular adsorbent recirculating system [MARS]). Bile acid and bilirubin levels, representing the previously described toxins, were reduced by 16% to 53% and 10% to 90% of the initial concentration by a single treatment of 6 to 8 hours, respectively. Toxicity testing of patient plasma onto primary rat hepatocytes by live/dead fluorescence microscopy showed cell-damaging effects of jaundiced plasma that were not observed after treatment. Patients with a worsening of Child-TurcottePugh (CTP) index before the treatments showed a significant improvement of this index during a period of 2 to 14 single treatments with an average of 14 days. After withdrawal of MARS treatment, this improvement was sustained in all long-term survivors. Ten patients represented a clinical status equivalent to the United Network for Organ Sharing (UNOS) status 2b (group A1), and all survived. Sixteen patients represented a clinical status equivalent to UNOS status 2a, and 7 of these patients survived (group A2), whereas 9 patients (group B) died. We conclude that in acute excretory failure caused by a chronic liver disease, this treatment provides a therapy option to remove toxins involved in multiorgan dysfunction secondary to liver failure. (Liver Transpl 2000;6: 603-613.)
I
n liver failure associated with excretory insufficiency, endogenous toxins known to participate in the progression of disease accumulate and cannot be removed sufficiently by established techniques for blood detoxification, such as dialysis, adsorption, and conventional plasma exchange.1 Dialysis as a biocompatible method failed to remove the majority of toxins accumulating in liver failure be-
cause of their high protein-binding rate that prevented their passage through protein-impermeable dialysis membranes. More invasive detoxification methods, such as hemo- or plasmasorption, enabled greater clearance of protein-bound toxins because of direct contact between the sorbent and albumin-toxin complex and the large surface of these sorbents.2 However, negative side effects arising from plasma-sorbent contact representing an unsatisfying biocompatibility outweighed the basically positive effects of this treatment.1,3 Another method is the removal of the patient’s plasma by centrifugation or filter devices and replacement with fresh frozen plasma.4 This method limits the amount of the removed toxins to that contained in the convection-filtered plasma in contrast to adsorption methods based on diffusive removal of toxins. Therefore, significant toxin removal requires a high volume of exchanged plasma of approximately 8 L. Ongoing clinical trials are evaluating whether such side effects as thrombocyte loss, citrate load, coagulation, and high costs are justified by the therapeutic effect in patients with acute liver failure who may be eligible for highurgency (HU) transplantation. Bioartificial liver assist devices based on the use of bioreactors containing living hepatocytes in contact with the patient’s plasma have been under development
From the *Department of Internal Medicine, Division of Nephrology; †Department of Internal Medicine, Division of Gastroenterology; ‡Department of Internal Medicine, Division of Intensive Care Medicine; and the §Department of Abdominal and Liver Surgery, University of Rostock, Rostock, Germany. Supported in part by the German Ministry for Research and Technology, Gambro Dialysatoren GmbH Hechingen, and Teraklin AG, Rostock, Germany. Address reprint requests to Jan Stange, Department of Internal Medicine, E-Heydemann Str 06, Rostock D 18055, Germany. Telephone: 49 171 787 3769; FAX: 49 381 494 7354; E-mail:
[email protected] Copyright © 2000 by the American Association for the Study of Liver Diseases 1527-6465/00/0605-0108$3.00/0 doi:10.1053/jlts.2000.7576
Liver Transplantation, Vol 6, No 5 (September), 2000: pp 603-613
603
604
Stange et al
Table 1. Patient Characteristics Infection
Patient Age Group/No. Sex (yr) Primary Liver Disease A1/1 A1/2 A1/3 A1/4 A1/5
M M F M M
34 45 56 46 33
A1/6
F
40
A1/7 A1/8 A1/9
F F F
30 52 36
A1/10 A2/1 A2/2 A2/3 A2/4
F F M F M
42 50 25 36 46
A2/5
F
27
A2/6 A2/7 B/1
M F M
37 36 33
B/2 B/3
F F
60 46
B/4
M
44
B/5 B/6 B/7 B/8 B/9
F F M F M
29 38 33 30 39
WBC (GPT/L)
Cirrhosis†/WoA Cirrhosis†/WoA Cirrhosis?/WoA Cirrhosis†/WoA Based on sonography, hepatic steatosis with signs of cirrhosis, clinical grade IV esopharyngeal varices Based on sonography, hepatic steatosis with signs of cirrhosis Cirrhosis?/unknown Cirrhosis†/WoA Based on sonography, hepatic steatosis with signs of cirrhosis, clinical grade II esopharyngeal varices Cirrhosis†/WoA Cirrhosis†/drug induced Cirrhosis†/WoA Cirrhosis†/WoA Cirrhosis†/WoA
Fever (°C)
Bacterial pneumonia, 31 No, Norm No, 18.9 No, 14.7 Fungal urinary tract infection, Norm
38.5 Norm Norm Norm Norm
Fungal pneumonia, 15
Norm
Staphylococcal sepsis, 18 Bacterial urinary tract infection, 23 Bacterial urinary tract infection, 35
39 Norm 38.5
Esophageal candidosis, 35 Bacterial urinary tract infection, 30 Bacterial urinary tract infection, 26.4 Staphylococcal sepsis, 46.7 Bacterial pneumonia, esophageal candidosis, 30 Based on sonography, hepatic steatosis with signs Fungal pneumonia, staphylococcal of cirrhosis (WoA), clinical grade II esopharyngeal sepsis, 15 varices Cirrhosis†/WoA Clinical signs of catheter sepsis, 32.9 Cirrhosis (hist)/Hep C Bacterial urinary tract infection, 30 Cirrhosis (hist)/PSC Bacterial urinary tract infection, candida sepsis, 34 Cirrhosis†/WoA Bacterial pneumonia, 20 Cirrhosis (hist)/WoA Bacterial and fungal urinary tract infection, 18 Cirrhosis†/WoA Esophageal candidosis, bacterial pneumonia, 6.4 Cirrhosis (hist)/WoA Bacterial pneumonia, 32 Cirrhosis†/WoA Fungal urinary tract infection, 37 Cirrhosis (hist)/WoA Fungal pneumonia, 22 Alcoholic hepatitis (hist)/WoA Bacterial pneumonia, 14 Cirrhosis†/WoA Fungal pneumonia, bacterial urinary tract infection, 22.7
Norm No 38.5 38.5 38.5 38.5
Norm 38.5 39 38.5 39.5 Norm 38.5 Norm 38.5 38.5 38.5
Abbreviations: WoA, without other obvious reason than alcohol; N/A, not applicable; WBC, white blood cells; Norm, normal; GI, gastrointestinal; HC, hepatic coma; HRS, hepatorenal syndrome; SIRS, systemic immune response syndrome; Hep C, hepatitis C; PSC, primary sclerosing cholangitis; GPT, gigaparticles; progredient ascites, increasing ascites despite standard treatment; hist, histological; Reanimation, resuscitation. *Group A1, survivors of UNOS 2B patients; group A2, survivors of UNOS 2A patients; and group B, nonsurvivors, all fulfilling UNOS 2A criteria. †Diagnosis based on numerous sonographic signs and clinical signs.
for nearly 30 years to support the complex liver functions in a biological way. The first clinical applications in patients with acute hepatic failure who were eligible for HU transplantation have shown encouraging results.5 Patients could be
successfully bridged until a transplant was available for HU transplantation. However, mortality in acute or chronic hepatic failure remains high because of less availability of urgent transplants for these patients and thus, the needed bridging time is much longer.6
605
Liver Support—Clinical Observation
Table 1. (Continued) Patient Characteristics Patient Group/No.
Other Precipitating Event
Other Complications
UNOS Status
Reanimation, mechanical ventilation, renal failure, HC 3-4 HRS, therapy-resistant ascites, HC 2
2B
A1/3 A1/4 A1/5 A1/6
Trauma, liver ischemia during cardiac arrest Profound epipharyngeal bleeding N/A N/A N/A N/A
2B 2B 2B 2B
A1/7 A1/8 A1/9 A1/10 A2/1 A2/2 A2/3
N/A N/A N/A N/A Except GI bleeding N/A N/A N/A
A2/4
N/A
A2/5
N/A
A2/6 A2/7
N/A Except epipharyngeal bleeding N/A
B/1
N/A
B/2
N/A
B/3
N/A
B/4 B/5
GI bleeding, successfully stopped N/A
HC 2-3 HC 2, therapy-resistant progredient ascites HC 2, therapy-resistant progredient ascites HC 2, one-sided total atelectasis, intermittent cardiac arrhythmia HC HC 1 HC 3, SIRS HC 3, SIRS HC 2, HRS, therapy-resistant ascites HC 3-4, SIRS, diffuse bleeding HRS, GI fungal infection, antibioticinduced bone marrow suppression, pancytopenia, diffuse bleeding, HC 3-4 HRS, therapy-resistant progredient ascites, pneumonia, severe epipharyngeal bleeding HC 3-4, mechanical ventilation, HRS, diffuse bleeding HC 2, therapy-resistant ascites, HRS, SIRS Drug-induced allergic reaction, HRS, HC 2-3 HC 3, HRS, SIRS, pericarditis with pericardial fluid HC 2, HRS, SIRS, mechanical ventilation, therapy-resistant ascites HC 1, therapy-resistant ascites, renal insufficiency, SIRS HC 2, HRS, therapy-resistant ascites
B/6
N/A
B/7
N/A
B/8 B/9
N/A N/A
A1/1 A1/2
We report the clinical use of a liver support system developed to support excretory liver function in patients with liver insufficiency caused by acute or chronic liver insufficiency to remove the toxic load from the liver and enable it to compensate with remaining liver function.
Pulmonary infarction, pneumonia, toxic digoxin assay days after MARS, brain bleeding HRS, pulmonary edema, 2 reanimation, mechanical ventilation, SIRS, severe lower GI bleeding HRS, therapy-resistant ascites, fungal pneumonia, HC 2 HRS, SIRS, ARDS, HC 3-4 HC 1, therapy-resistant ascites, HRS
2B
2B 2B 2B 2B 2A 2A 2A
2A
2A 2A 2A 2A 2A 2A 2A 2A
2A
2A 2A 2A
In the healthy liver, hepatocytes overtake plasma protein-bound toxins from the space of Disse by binding them to soluble intracellular transport proteins (e.g., ligandin), which are then regenerated by the hepatocellular detoxification apparatus, which consists of phase I and/or phase II biotransformation and excretion.
606
Stange et al
To mimic this detoxification in liver failure, the patient’s blood is roller pump– driven along semipermeable dialysis membranes that are impermeable to albumin because of a pore size smaller than its molecular size, but that are able to loosen the albumin-toxin complex by physicochemical interaction. Thus, the toxins are adsorbed by the membrane surface, whereas the proteins remain in the bloodstream. Through diffusion on the surface of the polymer pores, which have hydrophilic and hydrophobic domains, even lipophilic toxins are able to pass through the membrane and come into contact with the dialysis side. In contrast to conventional dialysis, the dialysate solution contains albumin, as well as electrolytes, glucose, and bicarbonate buffer, that is regenerated online by sequential dialysis, anion exchange chromatography, and charcoal perfusion, thus representing free binding sites and a greater affinity for the toxic ligands than the patient’s plasma albumin, which is overloaded with protein-bound toxins. This creates a continuous concentration gradient for protein-bound toxins toward the dialysate side. To enable this process, a membrane has to fulfill specific physicochemical and structural requirements that have been previously published.7 All water-soluble substances with a molecular size smaller than 30,000 d, such as lactate, creatinine, and ammonia, are able to pass through the membrane because it enables normal highflux hemofiltration in addition to the removal of protein-bound toxins. In parallel, ultrafiltration, pH, electrolyte, and glucose balance are possible. The clinical use of this process for 13 patients with acute or chronic hepatic failure was reported by Stange et al.8 These preliminary clinical results suggested a recovery of hepatic function after detoxification of jaundiced patients. We report the extracorporeal removal of such albumin-bound toxins as bile acids and bilirubin and the effect of this detoxification procedure on hepatotoxicity of the patient’s plasma. In addition, clinical parameters for the documentation of hepatic recovery were evaluated in 26 patients with excretory liver failure caused by a chronic liver disease.
Patients, Materials, and Methods Patients and Design Twenty-six patients (11 men, 15 women; age, 25 to 60 years) with a chronic liver disease that represented an acute, cholestatic decompensation associated with coagulopathy and enhanced aminotransferase levels were included after showing a continuous clinical worsening despite standard care. All patients with a history of alcohol abuse represented a high risk
for mortality because of an enhanced discriminant function (⬎93), introduced by Maddrey et al9 based on enhanced prothrombin time and severe jaundice. Hepatic encephalopathy and ascites were also present as further mortality risk factors. During follow-up, the patients who survived (group A) were divided into those fulfilling clinical criteria according to the United Network for Organ Sharing (UNOS) status 2b as subgroup A1 and those who fulfilled criteria according to UNOS status 2a as subgroup A2. All nonsurvivors fulfilled the clinical criteria according to UNOS status 2a and were defined as group B. Patient characteristics are listed in Table 1, and the hospitalization periods before, during, and after the molecular adsorbents recirculating system (MARS) treatments are shown in Figure 1. Posthepatic cholestasis was excluded by sonography. In patients with acute or chronic hepatic failure caused by alcoholic cirrhosis, the chance of successful abstinence therapy was evaluated in discussions with the patient and, if not possible because of hepatic encephalopathy, with the family. Consent of the patient and/or the family was obtained. The patients usually were placed on wards in departments of surgery and medicine and in the respective intensive care unit if required, e.g., because of mechanical ventilation. All patients received standard treatment, including selective digestive tract decontamination. In addition, antibiotic treatment was administered if infection was present, as listed in Table 1. Coagulation factor substitution was performed with fresh frozen plasma to achieve a prothrombin time greater than 40% and with antithrombin III to achieve a value greater than 50%. The substitution policy was not different before, during, or after the MARS treatment period.
Extracorporeal Treatment Albumin dialysis to remove protein-bound toxins was performed by a technical procedure called MARS.8 The extracorporeal blood circuit was driven using dialysis machine equipment (AK10 IC; Gambro AB, Lund, Sweden, or Fresenius 2008; Fresenius Medical Care AG, Bad Homburg, Germany). An albumin-impregnated highly permeable dialyzer (test device P3S; Gambro Dialysatoren Hechingen, Germany, for patients 1 through 18 and MARS-Flux; Teraklin AG, Rostock, Germany, for patients 19 through 26) was used as the membrane enabling the removal of protein-bound toxins for blood treatment. A closed-loop 10% ⫾ 5% commercial human serum albumin (as used for intravenous substitution) containing dialysate solution was used to guarantee the removal of the toxins on the dialysate side. The albumin dialysate was recirculated at a mean flow rate of 150 mL/min through the dialysate compartment of the dialyzer and regenerated online by perfusion over an anion exchanger column, an uncoated charcoal column, and through a low-flux dialyzer for dialysis.8 A mean of 6- to 8-hour treatments per patient depended on the clinical situation and are shown in Figure 1. Treatments were performed every day or every other day, depending on the response to the detoxification therapy, monitored by bilirubin measuring.
Liver Support—Clinical Observation
Figure 1. Hospitalization for groups A1 and A2 and before, during, and after (until discharge from the hospital in groups A1 and A2 or death in group B) MARS treatment. Because the nonsurvivor group (B) showed a more dramatic worsening, the decision for treatment was made earlier.
Clinical Scoring The hepatic encephalopathy score was evaluated according to the procedure of Schafer and Jones10 before and after the series of MARS treatments. Based on this information, the presence of ascites, and the determination of coagulation, bilirubin, and albumin parameters, the Child-Turcotte-Pugh (CTP) index of the patients was determined before and after MARS treatment according to the guidelines provided by UNOS. Finally, the outcome was documented.
Statistical Methods Statistical analysis of data before and after the MARS treatment was performed using Student’s t-test because the data were normally distributed. Sigma Plot 5.0 (SPSS Inc, Richmond, CA) was used as the software for the t-test.
Decrease of Cholestatic Markers and Investigation of Plasma Hepatotoxicity In 20 single treatments, 1 mL of heparinized plasma of patients before and after treatment were incubated (37°C, 5% carbon dioxide) with 106 primary isolated rat hepatocytes attached to 10-cm2 polystyrene wells in 6 separate experiments. Total bile acid and bilirubin blood levels were measured before and after these single treatments. The procedure to prepare and use primary isolated hepatocytes for a plasma toxicity assay has been described before by Hughes et al.11 The hepatocytes were monitored for 24 hours by videomicroscopy. After a 24-hour incubation, a fluorescence viability test was performed by incubating the primary isolated hepatocytes after contact with jaundiced plasma and
607
MARS-treated plasma for another 20 minutes with 4 mol/L of ethidium bromide homodimer staining dead cells and 2 mol/L of Calcein AM staining living cells (Life Dead Test; Molecular Probes, Eugene, OR). Ethidium bromide is only able to penetrate damaged hepatocellular membrane and stain damaged DNA by 40-fold enhanced red fluorescence after incorporporation into the helix. Calcein AM can only be taken up by viable cells and requires metabolization to present green fluorescence. In addition to the microscopic viability test, cell viability was tested using the XTT test (Boehringer Mannheim GmbH, Mannheim, Germany). This test is based on the cleavage of the yellow tetrazolium salt, XTT (3'-[1-phenylaminocarbonyl]-3,4tetrazolium]-bis[4-methoxy-6-nitro]benzene sulfonic acid hydrate) to form an orange formazan dye by metabolic active cells. This conversion occurs only in viable cells. Therefore, the XTT test is widely used for the evaluation of cytotoxicity. After a 24-hour incubation with plasma from patients before and after MARS treatment, the cells were incubated for another 24 hours with 0.5 mg of XTT in 1.5 mL of RPMI 60 Cell Culture Medium (Sigma-Aldrich Chemie GMBH, Deisenhofen, Germany). The formation of the orange metabolite was detected by spectrophotometric extinction measurement at 450-nm wavelength (reference wavelength, 690 nm). This cell viability test represents the viability of all cells from a culture well and avoids bias that may arise from microscopic cell viability assays that focus on selected areas.
Results Effect of Single Treatments on Plasma Cytotoxicity Because bile acids and bilirubin represent albumin-bound endogenous toxins in liver failure, the kinetics of the single treatments (6 hours) are shown in Figures 2 and 3. The reduction rates by a single 6-hour treatment ranged from 16% to 53% of the initial bilirubin concentration or from 10% to 90% of the initial bile acid concentration in vivo, depending on body weight and initial plasma concentrations. Hepatocytes cultivated in patient plasma before and after detoxification are shown in Figure 4. Hepatocytes incubated with plasma, representing greater concentrations, represent greater rates of vacuolization (Fig. 4A). This corresponded to a visible staining of hepatocytes with ethidium bromide homodimer representing membrane damage (Fig. 4C). Vacuolization appeared to be less in cells incubated in plasma after treatment (Fig. 4B). These cells represented bright green fluorescence, proving active metabolization of Calcein AM (Fig. 4D). In parallel, the formation of the orange formazane metabolite in the cultures incubated with plasma from patients after MARS treatment was significantly (P ⬍ .05, paired t-test) greater (extinction, 1.76 ⫾
608
Stange et al
Figure 2. Decrease in serum bile acid concentrations during a 6-hour treatment in single treatment sessions.
0.105; n ⫽ 20) than in cultures incubated with plasma from the same patients before MARS treatment (absorbance, 1.427 ⫾ 0.22; n ⫽ 20). Clinical Course All patients representing a clinical status in accordance with UNOS status 2b (group A1; n ⫽ 10) survived after a mean number of 4 treatments despite not showing a response to conventional treatment before the start of MARS treatments. Seven of 16 patients representing additional complications (e.g., hepatorenal syndrome and/or ascites without response to therapy), thus fulfilling clinical criteria in accordance with UNOS status 2a, underwent an average of 7 treatments and were longterm survivors (group A2; n ⫽ 7). The remaining 9 patients in this group underwent an average of 8 treatments and died after a mean survival time of 15.66 ⫾ 10.30 days (group B; n ⫽ 9). Only 1 patient from group B died within 4 days after fulfilling the UNOS status 2a standards and matched the expected survival time, which was estimated to be less than 7 days in groups A2 and B, who all fulfilled UNOS status 2a criteria. The long-term survival rate (⬎6 months of observation) in UNOS status 2a patients was 43%, and the overall survival rate for all patients was 62%. Only 1 patient underwent transplantation 1 year later; another patient was placed on the elective waiting list but died a year later of massive epipharyngeal bleeding. The mean treatment period (from the first treatment until the day after the last MARS treatment) was 14.5 ⫾ 5.6 days for group A1, nearly identical with 14.5 ⫾ 5.57 days for group A2, and 12.7 ⫾ 9.7 days for group B. There was no significant difference among these treatment periods. Within this period, a significant reduction in biliru-
bin level as a marker of albumin-bound toxins could be achieved in the total survivor group A (Table 2), as well as in the nonsurvivor group B (Table 3). In parallel, there was a reduction in creatinine level and ascites in both groups that was highly significant in the survivor group (Table 2), but not in the nonsurvivors (Table 3). Although substitution with fresh frozen plasma, coagulation factors, and antithrombin III was already started before the first MARS treatments without significant dosage difference between the groups, a significant improvement in albumin level, antithrombin III level, and thromboplastine time value (Quick, Thromboril S-Kit; Behring AG, Marburg, Germany) was only observed in the survivor group A (Table 2); the improvement was not significant in group B (Table 3). This significant improvement in synthetic function was maintained after withdrawal of MARS therapy and any substitution until discharge from the hospital in the survivor group (Table 2). This was accompanied by a further slight improvement in kidney function, ascites, and bilirubin elimination. In the nonsurvivor group after withdrawal of extracorporeal treatment, bilirubin and creatinine levels increased again and ascites remained. Values for albumin, prothrombin time, and antithrombin III required further substitution until the patients died. Retrospectively, the nonsurvivor group represented a more aggressive increase in bilirubin and creatinine levels before the first MARS treatment (Table 3), whereas these parameters did not show significant changes in the survivor group before the start of therapy (Table 2). A summary of the CTP index and hepatic encephalopathy score as one of the CTP parameters of the 26 patients is shown in Figures 5 and 6. During conventional treatment in the hospital, there was a significant increase in CTP and hepatic encephalopathy scores in
Figure 3. Decrease in serum bilirubin concentrations during a 6-hour treatment in single treatment sessions.
Liver Support—Clinical Observation
609
Figure 4. Photographs of hepatocytes from patients’ plasma before and after MARS treatment. (A) Videomicroscopic picture of primary rat hepatocytes after 24-hour incubation with jaundiced plasma before MARS treatment. The cells show vacuolization. (B) Videomicroscopic picture of primary rat hepatocytes after 24-hour incubation with plasma after MARS treatment. Vacuolization appears to be less. (C) Fluorescence microscopic picture of primary rat hepatocytes after 24-hour incubation with jaundiced plasma before MARS treatment. Cells are stained with the Life Dead Test that comprises 20-minute incubation with 4 mol/L of ethidium bromide homodimer and 2 mol/L of Calcein AM. Ethidium bromide is only able to penetrate damaged hepatocellular membrane and stain damaged DNA by 40-fold enhanced red fluorescence (excitation with 492-nm blue light) after incorporation into the helix. Calcein AM can only be taken up by viable cells and requires metabolization to present green fluorescence (excitation with 492-nm blue light). A significant number of cells is shown to represent dead stain. (D) Fluorescence microscopic picture of primary rat hepatocytes after 24-hour incubation with plasma after MARS treatments. Cells are stained with the Life Dead Test, described in (C). All cells are shown to be viable, representing a bright green fluorescence.
the total population. After the start of MARS therapy, CTP and hepatic encephalopathy scores could be reduced significantly. In the total analysis of the patients, this improvement continued after withdrawal of therapy until the end of hospitalization.
Discussion The present data show that in jaundiced patients with acute or chronic liver failure, bile acids and other protein-bound toxins can be removed by MARS treatment. The clearance mechanisms for water-soluble (e.g., creatinine, ammonia) and protein-bound, usually non-
dialyzable toxins (e.g., bile acids, benzodiazepines), have been recently published.8,12 All survivors (Table 2) did not undergo substitution of plasma proteins at the time of hospital discharge. Therefore, data from column 8 in Table 2 should represent the function of the liver itself. If increased bile acids in cholestasis have hepatotoxic effects, as reported in the literature,13,14 and reduction of these substances by MARS treatment, which has been described previously8 and in this report (Figs. 2 and 3), reduces this effect, it could be hypothesized that an improvement in liver function could be a consequence of reduced plasma hepatotoxicity caused by ex-
610
Stange et al
Table 2. Biochemistry Data and Presence of Ascites for All Survivors Column No. 2
3
4
5
Admission to Hospital (n ⫽ 17)
P (2 v 4)
Start of MARS Treatment
P (4 v 6)
6
7
End of MARS P Treatment (6 v 8)
Antithrombin III (80%-120%) 36 ⫾ 12.7 ⬍.001 52.9 ⴞ 17.9 ⬍.01 Prothrombin time value (Quick) (70%-120%) 47.3 ⫾ 14.6 NS 45.3 ⴞ 15.3 ⬍.05 Albumin (39-50 g/L) 30.1 ⫾ 5.4 NS 29.6 ⴞ 4.2 ⬍.005 Bilirubin (3-22 mol/L) 431.7 ⫾ 115.8 NS 429 ⴞ 100 ⬍.0001 Creatinine (62-133 mol/L) 224.6 ⫾ 291.1 NS 195 ⴞ 191 ⬍.05 Presence of ascites (CTP score, 1) 2.2 ⫾ 0.60 ⬍.05 2.7 ⴞ 0.46 ⬍.0001
64.5 ⫾ 17.3
NS
56.5 ⫾ 15.3 ⬍.05 37.1 ⫾ 4.8 NS 288 ⫾ 94 140 ⫾ 142
8 Hospital Discharge (n ⫽ 17)
9
10
11
P P P (4 v 8) (2 v 6) (2 v 8)
73.8 ⴞ 21.7
