Therapeutic Ef ficacy of L-Ornithine-L-Aspartate Infusions in

Patients With Cirrhosis and Hepatic Encephalopathy:

Results of a Placebo-Controlled, Double-Blind Study
GERALD KIRCHEIS,
1,2
RU

DIGER NILIUS,
1,3
CONSTANTIN HELD,
4
HANS BERNDT,
4
MICHAEL BUCHNER,
5
ROMAN GO

RTELMEYER,
2
RU

DIGER HENDRICKS,
6
BERTRAM KRU

GER,
7
BODO KUKLINSKI,
5
HELMUT MEISTER,
8
HANS-JU

RGEN OTTO,
8
CHRISTOPH RINK,
1,9
WOLFGANG RO

SCH,
10
AND SABINE STAUCH
10
prandial venous ammonia concentrations in the OA- One hundred twenty-six patients with cirrhosis, hy-
treated group showed improvements in comparison perammonemia (50 mmol/L), and chronic (persistent)
withplacebo. Inaddition, venous fasting bloodammonia hepatic encephalopathy (HE), which developed sponta-
concentration (P .01), mental state gradation (P neously without the existence of known precipitating
.001), andPSEI (P.01), whichincludes the mental state factors, were enrolled in a randomized, double-blind,
gradation, NCT-A time, and postprandial venous ammo- placebo-controlled clinical trial of intravenously admin-
nia in this trial, improved to a much higher degree in istered L-ornithine-L-aspartate (OA). Patients with sub-
the OA group than in the placebo group. In subgroups clinical (grade 0, West-Haven criteria) hepatic encepha-
retrospectively classified according to their initial men- lopathy (SHE), characterized by a prolonged number
tal state gradation, OA showed differential but uni- connectiontest A(NCT-A) time, andmanifest HE(grades
formly significant efficacies in patients with manifest I andII, West-Havencriteria) were includedinthe inves-
HE with respect to ammonia-lowering, improvement in tigation. The trial was plannedas aconfirmatory clinical
NCT times, and mental state gradation. In patients with trial. OA administered in a dose of 20 g/d, as well as
initial SHE, OA revealed differences between the medi- placebo, were dissolved in 250 mL of 5% fructose and
cations in the psychometric test used. Adverse events infused intravenously for a period of 4 hours during 7
consisting of mild gastrointestinal disturbances were consecutive days with a superimposed protein load at
observed in 3 of the OA-treated patients (5%). OA infu- the end of the daily treatment period. Primary variables
sion appears to be a safe, effective treatment of chronic were postprandial venous ammonia and NCT-A perfor-
(persistent) manifest HE in cirrhotic patients. Addi- mance time measured following OA or placebo infusions
tional investigations are required to assess the efficacy to evaluate the net effect of the treatment on the preven-
of OA in patients with SHE, as well as in patients with tion of the protein-induced hyperammonemia, and on
more severe grades of HE. (HEPATOLOGY 1997; 25:1351- parameters such as NCT-A influenced by hyperammo-
1360.) nemia. Mental state gradation, portal systemic encepha-
lopathy index (PSEI), and fasting ammonia levels were
estimated as additional efficacy parameters. The data
Hepatic encephalopathy (HE) is one of the major complica-
presented are based on the total study sample (intent-to-
tions of cirrhosis. Five years after the diagnosis of cirrhosis,
treat analysis), whichincluded63 patients inthe placebo
the probability of developing at least one episode of this spe-
group and 63 patients in the OA group. Of the 126 pa-
cific form of decompensated cirrhosis is in the range of 26%.
1
tients, 114 met all the criteria for inclusion and com-
Once clinical decompensation has occurred, however, the
pletedthe trial andtreatment as outlinedin the protocol
prognosis (16%
1
to 22%
2
probability of survival at 5 years)
(treated-per-protocol analysis). During baseline, the pla-
compared with a survival probability of 55%
2
to 70%
1
in cir-
cebo and treatment groups were homogeneous with re-
rhotic patients without HE is very poor.
1,2
Therefore, preven-
gard to mental states, NCT-A performance time, fasting
tion and effective treatment of HE may have important prog-
venous blood ammonia levels, and Child-Pugh criteria.
nostic implications in cirrhotic patients.
Although a slight improvement occurred in the placebo
Hyperammonemia
3,4
by a variety of postulated mecha-
group, NCT-A performance times (P .001) and post-
nisms
3-8
is felt to be one of the primary pathogenetic factors
in the development of HE.
8
The majority of therapeutic mea-
sures currently in use are therefore directed at reducing blood
Abbreviations: HE, hepatic encephalopathy; SHE, subclinical hepatic encephalopathy;
ammonia levels,
9-13
mainly by diminishing enteric ammonia
OA, L-ornithine-L-aspartate; GS, glutamine synthetase; NCT-A, number connection test A;
PSEI, portal systemic encephalopathy index. production. Otherwise, it is known that liver and muscle play
From the
1
Martin-Luther-University Halle-Wittenberg, Department of Internal Medi-
an important part within the ammonia detoxification system,
cine, Halle;
2
Merz Clinical Research Department and Merz Biostatistics Department,
especially in cirrhosis. Irrespective thereof, increased extra-
Frankfurt/Main;
3
Hospital Max Uibeleisen, Department of Internal Medicine, Bad Kis-
intestinal ammonia production and reduced ammonia detoxi-
singen;
4
Humboldt University, Department of Internal Medicine Theodor Brugsch, Char-
ite, Berlin;
5
Hospital, Department of Internal Medicine, Rostock;
6
Hospital, Department of fication capacity seen in patients with cirrhosis
14-16
and
Internal Medicine, Heide;
7
Hospital Fra nkische Schweiz, Department of Internal Medi-
chronic (persistent) manifest HE, as well as subclinical he-
cine, Ebermannstadt;
8
Hospital St. Vincenz, Department of Internal Medicine, Heiligens-
patic encephalopathy (SHE), have, to date, not been targets
tadt;
9
Hospital, Department of Internal Medicine, Aue; and
10
Hospital Nordwest, Depart-
of therapeutic agents within randomized, placebo-controlled
ment of Internal Medicine, Frankfurt/Main, Germany.
Received January 4, 1995; accepted February 25, 1997. trials.
Address reprint requests to: Gerald Kircheis M.D., Merz Clinical Research Department,
Results of animal experiments
13-23
suggest that L-orni-
Eckenheimer Landstrasse 100-104, 60318 Frankfurt/Main, Germany. Fax: /49 691503-
thine-L-aspartate (OA) may have a favorable influence on HE
409.
by virtue of its blood ammonialowering effects. A series of
Copyright 1997 by the American Association for the Study of Liver Diseases.
0270-9139/97/2506-0010$3.00/0 investigations in human subjects that started more than 25
1351
AID Hepa 0001 / 5p21$$$$$1 05-08-97 16:59:50 hpta WBS: Hepatology
1352 KIRCHEIS ET AL. HEPATOLOGY June 1997
(4 ampules of 10 mL each in 250 mL 5% fructose) for 7 consecutive years ago has confirmed these suggestions and has shown
days. The placebo solution contained riboflavin and polyethylene gly-
that OA can reduce blood ammonia and improve the symp-
col to make the control solution identical to the OA infusate in ap-
toms and laboratory features of HE.
24-27
pearance, behavior, and viscosity. All patients were given a nutri-
Experimental and clinical findings in liver, muscle, and
tious diet containing 1 g meat-vegetable dietary protein per kg body
brain have provided new insights into the mechanisms of
weight per day. The quantity of protein was divided at the test days
action of these ammonia-lowering amino acids, especially
among three main meals of the day in such a way that, as in a
in the therapy of HE. These tissues are important target
previous study,
27
the quantity of protein consumed with breakfast
organs in the removal of ammonia in hyperammonemia.
14-
was 0.25 g/kg body weight and that consumed with lunch was 0.5 g/
kg body weight. In this way, ingestion of an identical amount of
16,20-22,28-32
It has been proposed that, whereas aspartate and
protein before determination of the postprandial ammonia, the NCT- citric dicarboxylates under pathological condition in cirrhotic
A times, and the mental state at 1:00 PM was guaranteed.
patients may serve as a carbon source for the impaired gluta-
Concomitant medications that the patients were receiving and
mine synthetase (GS) flux in the perivenous scavenger hepa-
that were continued throughout the study included diuretic drugs
tocytes,
15,31-36
ornithine improves the flux through the im-
(OA: 24 patients; placebo: 18 patients) and/or aldosterone antago-
paired urea cycle enzyme system, especially through
nists (OA: 18 patients; placebo: 16 patients), b-adrenergic blocking
carbamylphosphate synthetase, localized in periportal hepa-
agents (OA: 6 patients; placebo: 4 patients), and digitalis-containing
tocytes.
20-23
In contrast to hepatic ammonia removal by urea
drugs (OA: 4 patients; placebo: 5 patients). Long-termadministration
and glutamine synthesis,
14-16,20-23,31,32
recent findings suggest of either clomethiazole (4 patients), a drug used for treatment of
alcohol-withdrawal syndrome, or a tranquilizer (1 patient) was also fundamentally different regulatory mechanisms for GS
27,37-39
continued. and removal of excess ammonia in muscle
27,37
and brain
38,39

tissues that are devoid of an effective urea cycle.

30
Therapeutic Endpoints
These concepts are the basis for the use of amino acids
such as ornithine to stimulate the impaired flux of the urea
The size of comprehensive test batteries renders precise, defined
cycle
20-23
and to interact with brain glutamine,
38,39
as well as
therapeutic endpoints impractical for use in the clinical situation
dicarboxylates and amino acids such as a-oxoglutarate or and for follow-up observation in large therapeutic studies. Conse-
quently, the participants of the study group decided to use a limited aspartate, benzoate, and malate to stimulate the impaired
number of simple, validated, sensitive test procedures. flux of glutamine synthesis in liver
34-36
and muscle
27,37
of cir-
Primary variables were postprandial venous ammonia and NCT-
rhotic patients.
A performance time representing the net effect of the previous 4-
To assess the clinical efficacy of OA in the treatment of HE,
hour infusion in combination with the effects of the previous 2 to 7
a randomized, double-blind, placebo-controlled, multicenter
days of infusion on the prevention of the protein-induced (12:00 PM
trial with intravenously administered OA in cirrhotic pa-
meal [0.5 g protein per kilogram of body weight]) hyperammonemia
tients with hyperammonemia and chronic (persistent), mani-
and on parameters such as NCT-A time, influenced by hyperammo-
fest HE or SHE was performed. The aim was to evaluate in
nemia. Mental state gradation and portal systemic encephalopathy
a practice-adapted design the influence of OA on the improve- index (PSEI), assessed at the same time as the primary variables,
as well as fasting ammonia levels, representing the impact of the ment of number connection test A (NCT-A) times and of men-
previous 2 to 7 days of infusion, were estimated as additional efficacy tal states as a consequence of the ammonia-lowering effect.
parameters. Any difference between the OA-and the placebo-treated
group with respect to the above-indicated parameters (NCT-A time, PATIENTS AND METHODS
NH
4
/
, mental state gradation, PSEI) could be interpreted in terms
Inclusion criteria were assessed 16 to 24 hours before the start of
of a prevention of protein-induced hyperammonemia.
the first infusion. Patients who met the following inclusion criteria
Inclusion criteria were checked at a nonfixed time point at the
assessed at the randomization (day 01) were eligible for the study:
time of randomization (day 01) before any treatment or superim-
1) chronic (persistent), manifest HE (mental state grade I or II, ac-
posed protein load. Therefore, these parameters could not be used
cording to the West-Haven criteria
9,11,40
developed spontaneously
as baseline parameters for the analysis of primary and secondary
without existence of specified precipitating factors or SHE (mental
response variables. The baselines of postprandial NH
4
/
, NCT-A, and
state grade 0 and an NCT-A performance time 30 seconds);
9,11,40
HE-grading used in efficacy analyses were obtained after the first
2) hyperammonemia (venous ammonia concentration 50 mmol/
OA- or placebo-infusion (day 0; 1:00 PM). From this point of view,
L);
9,11
and 3) cooperative, hospitalized, adult patients with histologi-
the therapeutic effects after 2, 4, and 7 days of treatment will be
cally proven cirrhosis or overt, clinically proven cirrhosis confirmed
analyzed for these parameters at postprandial time points (1:00 PM).
ultrasonographically who volunteered to remain in the hospital for
Because the fasting ammonia level (day 0; 8:00 AM) is the only unin-
the full 7 days of the investigation.
fluenced baseline level, the therapeutic effects of OA- or placebo-
Patients were included if they remained stable with respect to HE
infusion on this parameter will be evaluated after 2, 4, and 7 days
grade, existence of hyperammonemia, and increased NCT-A times
in the fasting state (8:00 AM).
between randomization (day 01) and first treatment day (day 0).
NCT-A Performance Time in Seconds. We used the NCT-A, a stan-
Exclusion criteria were the following: 1) active, major complica-
dardized psychometric test, for the diagnosis of SHE.
9,11,40
The avail-
tions of portal hypertension, such as gastrointestinal hemorrhage,
ability of four validated parallel forms of NCT-A and the inclusion
hepatorenal syndrome, or spontaneous bacterial peritonitis; 2) acute
of a trial run of a simpler form of NCT-A before the first use of the
superimposed liver injury; 3) serious nonhepatic diseases such as
test at the randomization point was the most reliable way of ensuring
decompensated heart failure and/or respiratory failure, decompen-
that learning effects were minimized.
sated diabetes mellitus, renal insufficiency, or electrolyte imbal-
Venous Ammonia Concentration. The ammonia determination was
ances; 4) Acute (recurrent), severe HE (mental state grade 3 or
performed according to the enzymatic determination of ammonia
4);
9,11,40
5) pregnancy or lactation; 6) age 18 years; 7) contraindica-
with glutamate dehydrogenase in a rapid and interference-free pho-
tions to intravenous infusion of fructose such as fructose intolerance,
tometric determination (340 [334] nm) of NH
4
/
in native blood
hyperhydration, hypotonic dehydration, hypokalemic alkalosis, or
plasma according to Da Fonseca-Wollheim.
41
For safety reasons,
acidosis; 8) administration of anti-HE medications such as neomycin,
blood after withdrawal was immediately taken by refrigerated trans-
lactulose, lactitol, branched-chain amino acids, etc.; and 9) any addi-
port to the laboratory for immediate (within 15 minutes of with-
tional precipitating factors such as high protein intake (additional
drawal of the blood) determination of NH
4
/
. Because the same stan-
high-protein meals), constipation or intake of psychostimulants, sed-
dardized technique was used at all the centers, and because all the
atives, antidepressants, benzodiazepines, or benzodiazepine-antago-
laboratories were subject to continual quality control with method
nists.
control and checks in accordance with Good Laboratory Practice,
NH
4
/
was determined under qualitatively consistent conditions at
Study Medications
all centers. Because of ethical considerations, we were not able to
use arterial blood in this investigation. Study medications consisted of intravenous infusions of 20 g OA
(4 ampules of 10 mL each) in 5% fructose (250 mL) administered Mental State Grade. The mental state was graded on a 0-to-4 scale
of severity.
9,11,40
The interindividual variation of its assessment is daily over 4 hours (8:00 AM until 12:00 PM) or the placebo solution
AID Hepa 0001 / 5p21$$$$$1 05-08-97 16:59:50 hpta WBS: Hepatology
HEPATOLOGY Vol. 25, No. 6, 1997 KIRCHEIS ET AL. 1353
TABLE 1. Baseline Data of 126 Cirrhotic Patients with consistently of the order of 5%.
9
At all centers, two physicians partici-
pating in this trial were trained in the correct use of mental state Hyperammonemia and Chronic (Persistent) Manifest Hepatic
Encephalopathy or Subclinical Hepatic Encephalopathy (SHE), grading according to West-Haven criteria introduced by Conn and
Lieberthal
9
; the mental state of an individual patient was assessed Included in the Trial (Intent-to-Treat Sample)
by the same observers and the grade defined by a consensus decision.
OA Group Placebo Group
PSEI. The PSEI according to the West-Haven criteria
9,11,40
re-
quires the systematic semiquantitative estimation of five compo-
No. (%) Mean SD No. (%) Mean SD
nents of the HE syndrome, each graded on a 0-to-4 scale of sever-
ity.
9,11,40
Each of the parameters is arbitrarily weighted separately Patients 63 63
and in proportion to its importance. Because calculations permit the Male 45 (71) 46 (73)
presentation of the percentage improvement noted in an individual Female 18 (29) 17 (27)
patient or in groups of patients, the use of the index is acceptable Age (yr) 53.9 12.4 52.3 13.3
even when measurements of specific symptoms are missing. In the Body weight (kg) 75.1 15.5 72.0 15.1
presented clinical trial, the PSEI contains venous postprandial am- Body height (cm) 170.6 9.0 173.2 7.4
monia, mental state gradation, and NCT-A time, whereas electroen- Body mass index 25.7 3.8 24.0 4.4
cephalogram monitoring and asterixis are missing. All analyzed pa- Duration of cirrhosis
rameters were estimated at the same time (1:00 PM). (mo) 58.9 49.7 60.2 56.0
Safety Parameters. Safety parameters included blood tests (hemo- Etiology
globin, hematocrit, white blood cell count, and thrombocytes) and alcohol 49 (78) 51 (81)
liver function tests (alanine aminotransferase, aspartate amino- posthepatitic 12 (19) 8 (13)
transferase, g-glutamyltransferase activity, serum bilirubin, serum others 2 (3) 4 (6)
albumin concentrations, cholinesterase activity, prothrombin time, Child-Pugh grade
and partial thromboplastin time) on days 0 and 7 (8:00 AM). The A 30 (48) 34 (54)
same standard methods were used at all centers. All laboratories B 28 (44) 22 (35)
were subject to constant quality control by means of suitable method C 5 (8) 7 (11)
checks. HE grade
Compliance, Tolerance, and Adverse Events. The consumption of 0 0 0
28 ampules of OA- or placebo-infusion corresponds to 100% compli- SHE 26 (41) 27 (43)
ance, a minimum consumption of 26 ampules of infusion concentrate I 26 (41) 27 (43)
being regarded as acceptable (95% compliance). All adverse events II 11 (17) 9 (14)
were assessed by the investigator according to traditional clinical- Ammonia (inclusion) 63 (100) 78.6 23.5 63 (100) 90.0 45.4
therapeutic methods. Any pathological clinical or laboratory findings NCT-A (inclusion) 62 (98) 61.8 22.8 63 (100) 60.2 19.8
observed during the trial were monitored and documented until their
normalization was observed or their cause and their correlation with NOTE. The medication groups in the intent-to-treat sample were homoge-
neous with regard to anamnestic and diagnostic criteria. the investigational medication could be explained. In the event of a
serious adverse event, the investigator was entitled to open the de-
coding envelope prepared for the individual patient. None of the
prepared decoding envelopes were opened.
beyond the nonspecific effects of nonpharmacotherapeutic origin
Statistical Methods
within the framework of a placebo-controlled trial that is universally
Clinical data management was performed using a data bank estab- accepted.
43,44
As such, the new drug must be tested against a pla-
lished in INGRES version 6 (Relational Technology Inc., Alameda, cebo.
43,44
Consequently, the ethical commission agreed with a pla-
CA). The statistical evaluation was performed using SAS procedures cebo-controlled trial in the treatment of chronic (persistent) HE or
of version 6.06 (SAS Institute Inc., Cary, NC). The medication differ- SHE.
ences at the end of treatment of the target variables NCT-A and The study was performed in accordance with the revised Helsinki
postprandial ammonia concentration were tested by the Mann-Whit- Declaration, the stipulations (sections 40-42) of the German Drug
ney Utest (Wilcoxon two-sample test). The global level of significance Law, and the requirements of the EEC guidelines on Good clinical
was a 0.05 (nominal value). The m-correction that became neces- practice for trials on medicinal products in the European Commu-
sary for the assessment of the medication differences at any point nity. With respect to these standards, all patients had to declare
of measurement (except for day 0) was performed according to Hoch- voluntary participation in this trial by signing an informed-consent
berg and Benjamini.
42
Further inferential analyses of additional vari- form. Further documentation included: the vote of the ethics commis-
ables (fasting ammonia concentration, mental state grade, PSEI, sion of the Hesse State Medical Association of 15.01.1990, and the
etc.), as well as subgroup analyses of the indicated parameters, were registration of the clinical test according to section 67, paragraph 1,
made by the Wilcoxon two-sample test, Wilcoxon signed rank test, of the German Drug Law.
x
2
test, Kruskal-Wallis test, Fishers two-tailed exact test, and a test
for homogeneity of odds ratios. The error probabilities of these tests
Treatment Groups
are descriptive significances.
A total of 126 cirrhotic patients fulfilled the criteria for inclusion
Sample size calculation was based on the results of the ammonia-
in this investigation performed between April 1990 and May 1991.
lowering effect of OA in a previous randomized, placebo-controlled,
They were randomly assigned to receive placebo (n 63) or OA (n
double-blind trial.
27
The calculation was based on the clinical as-
63). The two groups were similar in demographic characteristics,
sumption that 20 g OA/d during a 7-day infusion period was at least
etiology, duration, and severity of cirrhosis as determined by the
equally effective in reducing blood ammonia concentration as a single
Child-Pugh criteria. (Table 1). In 12 patients, violations with respect
infusion of 20 or 40 g OA. The required number of patients per group
to the inclusion and exclusion criteria, to adverse events, adverse
was estimated as n 18 (a* of 0.05/2 0.025; b 0.10). The a was
drug reactions, or shortened duration of treatment occurred (for de-
corrected with regard to the repeated tests for two primary variables.
tails, see Results). Four of these patients who had been taking prohib-
Randomization was performed with the randomnumber generator
ited medications and had been entered into the study erroneously
of G. Marsaglia and T. A. Bray in random permuted size blocks of
were retrospectively excluded for the treated-per-protocol analysis.
four to avoid a center- and treatment-related bias. The treatment
Afifth patient whose vision prevented himfromperforming the NCT-
forms (ampules of placebo and OA) were appropriately numbered.
A (inclusion criteria) was also withdrawn from the intent-to-treat
In accordance with the randomization list, 63 patients each were
analysis even though he had received the full course of therapy.
allocated to the groups treated either with OA infusion concentrate
The medication groups in the intent-to-treat sample as well as
or a placebo. The code of the randomization list was not broken until
those in the treated-per-protocol sample were homogeneous with re-
the last Case Report forms had reached the Biostatistical Depart-
gard to the anamnestic and diagnostic criteria. Discrepancies in the
ment and the database was frozen.
composition of the two groups with respect to precipitant factors
Ethical Considerations
could be minimized, because the patient population was well defined
by inclusion and exclusion criteria. Testing a newsubstance against a reference substance is permissi-
ble only if the latter has been shown to induce effects above and The primary analysis of the data on efficacy is based on the intent-
AID Hepa 0001 / 5p21$$$$$1 05-08-97 16:59:50 hpta WBS: Hepatology
1354 KIRCHEIS ET AL. HEPATOLOGY June 1997
FIG. 1. Effects of treatment with OA or placebo on performance time in NCT-A. Results (median, mean { SD) based on the evaluation of 126 cirrhotic
patients. Fifty percent of the observed NCT-A values are located within the 25th and 75th percentile (P 25 to P 75). The group differences between the
treatments were evaluated by means of the Wilcoxon two-sample test.
to-treat sample (total sample of 126 patients). In the intent-to-treat of patients treated with OA or placebo (Table 2). After 7 days,
analysis, changing samples at any time point were recorded due to
there were no longer any differences in final values attained
dropouts, missing data, and retrospectively excluded or withdrawn
among the different subgroups treated with OA therapy. The
patients. Treatment-per-protocol analysis was performed in 114 pa-
NCT-A times attained were in the vicinity of the normal for
tients who showed no violation of inclusion or exclusion criteria and
noncirrhotic patients. In all of the three subgroups with an
no deviations of protocol. This analysis showed no changing samples.
initial HE gradation of SHE, HE I, HE II, the change of the
Both statistical analyses showed virtually identical results. The data
pre-/post-difference in the NCT-A showed descriptive statisti-
presented here are based on the total study sample of 126 patients.
cal significances in the OA-treated patients (.001 P .01) Subgroup Analysis. All 126 patients in the study were also retro-
spectively classified in subgroups (SHE, HE I, HE II) according to (Table 2) with the aid of the Wilcoxon signed rank test. Medi-
the initial HE gradation (West-Haven criteria). Fifty-three of the
cation differences appear from comparison of the levels at-
patients were assigned to the group of patients with initial SHE (OA
tained by the subgroups with OA therapy with those for the
therapy: 26; placebo therapy: 27); 53 to the group with initial HE
corresponding placebo subgroups (Table 2). With respect to
classification I (OA therapy: 26; placebo therapy: 27); and 20 to the
the treatment-related pre-/post-differences in NCT-A, de-
group with initial HE classification II (OA therapy: 11; placebo ther-
scriptive significances between OA and placebo could be
apy: 9). The need for and effectiveness of therapy was compared
found in the subgroup of patients with initial SHE(P .0499)
and evaluated separately for this set of patients in the different
as well as in HE I (P .0034). The most important pre-/post- subgroups, according to the medication administered. Because the
original sample size calculation was directed to the whole study pop- differences in NCT-A were found in the subgroups of patients
ulation, the sample size in each of the subgroups is relatively small
with initial mental state grade II (Table 2).
and might have introduced a type II error within the subgroup analy-
Venous Blood Ammonia Concentration. The pretreatment
sis.
fasting venous ammonia concentrations (mean { SD) were
similar in both treatment groups (OA: 81 { 38 mmol/L; pla-
RESULTS
cebo: 83 { 43 mmol/L) (Fig. 2). On average, the fasting venous
ammonia concentration decreased after 7 days by 17 { 37
Efficacy and Response to Therapy
mmol/L in the OA group and by 6 { 32 mmol/L in the placebo
group (pre-/post-differences). Significant differences between
NCT-A. NCT-A performance times on the first day of treat-
the treatments in favor of OA (P .05, descriptive signifi-
ment (after 4 hours infusion and a superimposed protein load)
cance) were observed after 4 (P .0115) and 7 (P .0183)
were similar in OA- and placebo-treated groups (58 { 23
days (Fig. 2).
seconds vs. 55 { 19 seconds) (Fig. 1). The mean decrease in
The initial postprandial venous ammonia concentration on NCT-A score after 7 days was 19 { 19 seconds in the OA
the first day of therapy (i.e., after 4 hours of infusion and group and 7 { 12 seconds in the placebo group. The compari-
12:00 PM protein load) was 82 { 37 mmol/L in the group sons of treatments showed significant group differences on
treated by OA versus 91 { 48 mmol/L in the group treated days 4 (P .0078) and 7 (P .0006) (Fig. 1) in favor of OA.
by placebo. After 7 days of treatment, a mean reduction of The NCT-A performance times increased as the HE changes
that existed initially increased in the individual subgroups 16 { 40 mmol/L in the OA group versus 10 { 36 mmol/L in
AID Hepa 0001 / 5p21$$$$$1 05-08-97 16:59:50 hpta WBS: Hepatology
HEPATOLOGY Vol. 25, No. 6, 1997 KIRCHEIS ET AL. 1355
TABLE 2. Comparison of the Effectiveness of OA with Placebo in the Treatment of Cirrhotic Patients with Respect to NCT-A,
NH
/
4 (Fasting/Postprandial) and PSEI in Consideration of the Initial HE Gradation (SHE, HE I, HE II)
OA Placebo
Pre-/Post Pre-/Post
Initial 7 Days Difference Initial 7 Days Difference
NCT-A SHE N { SD 25 22 22 27 27 27
51.16 { 19.44 36.32 { 15.67 12.27 { 13.74 46.59 { 16.17 41.15 { 19.94 5.44 { 11.72*
HE I N { SD 25 25 24 27 27 27
60.84 { 27.31 37.48 { 19.63 20.38 { 20.90 56.00 { 15.02 51.48 { 21.29 4.52 { 12.62
HE II N { SD 11 9 9 9 8 8
65.27 { 17.41 32.44 { 6.80 31.33 { 18.84 74.78 { 21.83 51.38 { 17.32 17.63 { 8.25*
NH
/
4 (fasting) SHE N { SD 26 23 23 26 27 26
76.73 { 47.77 69.87 { 50.88 5.22 { 32.94 73.00 { 27.94 67.07 { 29.25 5.12 { 36.57
HE I N { SD 26 25 25 27 26 26
82.12 { 27.12 59.20 { 26.35 22.76 { 40.91 93.81 { 57.61 87.62 { 55.14 7.50 { 31.05
HE II N { SD 11 9 9 9 8 8
90.73 { 34.35 48.33 { 22.90 32.78 { 30.23* 81.56 { 19.58 75.25 { 31.29 7.13 { 25.85
NH
/
4
(postprandial) SHE N { SD 26 23 23 27 27 27
69.65 { 38.73 72.04 { 52.96 0.174 { 37.00 78.78 { 34.28 68.30 { 37.71 10.48 { 34.91
HE I N { SD 25 25 24 27 26 26
88.68 { 31.81 64.92 { 29.86 24.5 { 39.44 102.56 { 60.10# 93.27 { 60.72# 9.35 { 41.98
HE II N { SD 11 9 9 9 8 8
97.00 { 35.51 54.44 { 24.19 34.11 { 39.44* 92.78 { 34.20 85.88 { 28.63 6.75 { 22.40
PSEI SHE N { SD 25 22 22 27 27 27
0.185 { 0.065 0.110 { 0.107 0.076 { 0.091 0.184 { 0.059 0.144 { 0.095 0.040 { 0.092*
HE I N { SD 24 25 23 27 26 26
0.296 { 0.062 0.146 { 0.101 0.140 { 0.119 0.302 { 0.073 0.247 { 0.127 0.055 { 0.097
HE II N { SD 11 9 9 9 8 8
0.468 { 0.064 0.164 { 0.055 0.286 { 0.088 0.472 { 0.067 0.363 { 0.169 0.100 { 0.136
The table includes the respective number of investigated patients (N) as well as the mean level ({SD) of each subgroup during the course of this 7-day
treatment. The high SD in the NH
/
4 data in the subgroup of patients with initial HE I treated with placebo results from one patient (patient 19) with an
extremely high fasting ( day 0: 317 mmol/L; day 7: 254 mmol/L) as well as postprandial (# day 0: 274 mmol/L; day 7: 147 mmol/L) NH
/
4 level. The Wilcoxon
signed rank test was used for testing descriptive statistical significances (*P .05; P .01; P .001) for the pre-/post-differences between the initial level
and the level after 7 days of treatment within each medication group. To find a difference between the medication groups (OA vs. placebo) with respect to the
pre-/post-differences, the Wilcoxon-Mann-Whitney U test (Wilcoxon two-sample test) was applied.
the placebo group could be seen. In a comparison of the final sults: in the placebo group, the mean mental state grade was
0.91 { 0.48 versus 0.97 { 0.53 in the OA group. During values after 2 (P .012) and 4 (P .013) days of therapy,
significant differences between both groups in favor of OA the 7 days of placebo administration, the mean mental state
improved, but the mean scores after 2, 4, and 7 days of pla- were observed after 2 and 4 days of therapy (P .013). On
day 7, a tendency toward statistical significance (P .078) cebo treatment showed only minor differences from basal
(day 0) levels: 0.86 { 0.45, 0.82 { 0.47, and 0.72 { 0.52. was shown (Fig. 2). Because of differences in the initial post-
prandial ammonia levels, the comparison of the pre-/post- Mean scores in the OA-treated group fell progressively and
to a higher degree than in the placebo group. The mean val- differences between the treatments by means of the Wilcoxon
two-sample test did not achieve significant effects (day 2: P ues on days 2, 4, and 7 were 0.81 { 0.46, 0.64 { 0.42, and
0.42 { 0.33, respectively. .053; day 4: P .065).
The pretreatment fasting and postprandial NH
4
/
levels in On the basis of frequencies of change in the mental state
grade, which may be more appropriate indicators for a rank the individual subgroups increased with the severity of the
initially existing HE in both medication groups (Table 2). The scale such as HE grading, a clear difference between OA and
placebo was observed (P .001, descriptive significance). initial data for the NH
4
/
values for the subgroups with HE
I and HE II, respectively, for the placebo group, which appar- Improvement in the mental state grade was seen more fre-
quently in the OA group than in the placebo group (59% vs. ently do not agree with this postulate, are caused by the
extreme value of one patient with an initially determined HE 32%). The most important effects can be seen in the subgroup
of patients with initially overt HE (grade I and II) (Table I (Table 2). The most effective ammonia reduction occurred
in the OA-treated subgroup of patients with HE grade II, 3). Additionally, the ratio of improved versus not improved
mental state grade as well as the odds ratios for the sub- which initially showed the highest ammonia changes (HE II/
NH
4
/
/[fasting]/OA/initial: 90.7 { 34.4 mmol/L; day 7: 48.3 { groups with initial manifest HE (mental state grade I or II)
(Table 3) were calculated. As can be seen in Table 3, there 22.9 mmol/L; pre-/post-difference: 32.8 { 30.2 mmol/L; HE II/
NH
4
/
/[postprandial]/OA/initial: 97.00 { 35.31 mmol/L; day 7: are differences in both odds ratios between the subgroups.
The test for homogeneity of the odds ratios was not significant 54.44 { 24.19 mmol/L; pre-/post-difference: 34.1 { 39.4 mmol/
L). There are also distinct differences between the final val- (P .2193), which indicates that both subgroups share an
odds ratio of 4.78 (95% confidence interval: 1.76; 36.08). In ues attained when placebo therapy is compared with 7 days
of treatment with OA for subgroups HE I and HE II, respec- the two-sided test, the common odds ratio is significantly
different by 1 (P .0026 [RGB variance] or P .0016 [M-H tively (Table 2). There are no detectable changes when com-
paring the two medication groups in the subgroups with ini- variance]).
Thirty-seven OA-treated patients manifested a reduction tially existing SHE (Table 2).
Mental State. For better comparison to other published by at least one grade in mental state during the treatment
period; 26 patients showed no improvement. Seventeen of the data, the mean values of the treatment groups are provided
first. On day 0, both treatment groups gave comparative re- latter (65%) had SHE during the whole treatment period.
AID Hepa 0001 / 5p21$$$$$1 05-08-97 16:59:50 hpta WBS: Hepatology
1356 KIRCHEIS ET AL. HEPATOLOGY June 1997
FIG. 2. Effects of treatment with
OA or placebo on fasting, as well as
postprandial venous blood ammonia
concentrations. Results based on the
evaluation of 126 cirrhotic patients
(median, mean { SD) participating
in this 7-day, double-blind, placebo-
controlled clinical trial. P 25 and P 75
represent the 25th and 75th percen-
tile of an empiric distribution. Fifty
percent of the observed values are
within this range. The group differ-
ences between the treatments were
calculated on the basis of the pre-/
post-differences (Wilcoxon two-sam-
ple test). The pre-/post-differences
within each group were compared by
means of the Wilcoxon signed rank
test.
Only 9 of these 26 patients with initial SHE (35%) showed or only SHE (mental state grade 0; NCT-A time 30 seconds)
at the end of the OA treatment period. In contrast, 20 of improvement, which is defined as normalization of the ini-
tially prolonged NCT-A test time below the 30-second border- the placebo-treated patients showed improvement in mental
state grade compared with 43 patients who showed no re- line in the case of mental state grade 0 after 7 days of OA
treatment. After 7 days of treatment, more than 76% of the sponse. Of these 43 patients, 21 had originally exhibited
SHE. Six of the 27 placebo-treated patients who had SHE OA-treated patients with initial grade I and II exhibited im-
provement. Seventy-five percent of these patients had no HE initially showed improvement (22%).
PSEI. On day 1, both groups (OA- and placebo-treated)
showed the same PSEI score (mean { SD: 0.28 { 0.12; N
60 [OA], N 63 [placebo]), which, in this trial, includes the
TABLE 3. Results in the Patient Subgroups
three estimated, above-mentioned parameters, namely men-
Not tal state, NCT-A performance time, and postprandial blood
Improved Improved Total Ratio
ammonia concentration based on the West-Haven crite-
Improved/ Odds
ria.
9,11,40
At the end of therapy, a mean reduction of 0.14 {
Medication n (%) n n Not Improved Ratio
0.13 (day 7: 0.135 { 0.10, N 56) was observed in the OA
Initial HE grade II group versus 0.05 { 0.1 (day 7: 0.22 { 0.14; N 61) in the
OA 9 (82) 2 11 4.50 15.75
placebo group. The comparison of both treatments on day 7
Placebo 2 (22) 7 9 0.286
(P .0011) as well as the comparison of the pre-/post-differ-
Total 11 9 20
ences in OA and placebo-groups (P .0003) indicated a con-
Initial HE grade I
siderable difference in favor of OA (Wilcoxon two-sample
OA 19 (73) 7 26 2.71 3.39
test). The PSEI reached after OA treatment was closer to
Placebo 12 (44) 15 27 0.80
normal than after placebo (P .01 after 4 and 7 days).
Total 31 22 53
The initial scores in the PSEI also increased in the individ-
AID Hepa 0001 / 5p21$$$$$1 05-08-97 16:59:50 hpta WBS: Hepatology
HEPATOLOGY Vol. 25, No. 6, 1997 KIRCHEIS ET AL. 1357
FIG. 3. Comparison of the effectiveness of OA with placebo treatment in consideration of the initial HE gradation (SHE, HE I, HE II). The changes of the
mean levels of the PSEI of the respective subgroups are presented during the course of this 7-day treatment. The presentation of the upper and lower 95%
confidence limit of the mean (CLM) for each time point allows the conclusion of descriptive, statistically significant differences in or between the subgroups.
The subgroup differences between the treatments were calculated on the basis of the pre-/post-differences (Wilcoxon two-sample test). Medication differences
appear in the subgroups with initially manifest HE grade I (P .0153) and HE grade II (P .02) in favor of OA. No statistical group differences could be
proven between the subgroups of patients with initial SHE.
ual subgroups as the initially existing HE changes increased in the study and was withdrawn before therapy started. One
of the seven patients in the placebo group developed progres- in both medication groups (Table 2, Fig. 3). There are no
significant differences between the initial values for the par- sive heart failure and was withdrawn from the study after
the study medication had been started. ticular medication groups (OA and placebo) when the match-
ing subgroups (SHE, HE I, HE II) are compared. No differ-
DISCUSSION
ences can be detected between the final values (mean { SD)
of the scores (0.11 { 0.11 [SHE] to 0.16 { 0.06 [HE grade The treatment with OA, a stable salt of the natural L-
II]) for the different subgroups after 7 days of OA therapy amino acids ornithine and aspartic acid, results in significant
(Table 2, Fig. 3). This tendency cannot be detected in the reductions in blood ammonia levels. This ammonia-lowering
subgroups treated with placebo (Fig. 3). Medication differ- effect of OA and other ornithine or aspartate conjugates has
ences with respect to the pre-/post-differences after 7 days been known for many years from both animal experiments
of treatment appear in a comparison of the subgroups with and clinical observations. Clinical findings were mainly
initially manifest HE grade I and II treated with OA(P .05) based on the results obtained in uncontrolled studies.
24-26
and those treated with placebo (Table 2, Fig. 3). No statistical Therefore, the present study was aimed at quantifying the
group differences (Wilcoxon two-sample test) could be proven therapeutic efficacy of intravenously administered OA in a
between the subgroup of patients with initial SHE. randomized, placebo-controlled, double-blind trial in a de-
fined patient population of cirrhotic patients with hyperam-
Adverse Events and Adverse Drug Reactions
monemia and chronic (persistent) manifest HE and SHE.
Although we refrained from including patients with severe OA was well tolerated in 86% of the treated patients. Pla-
cebo was tolerated well in 100% of the patients. Seven pa- HE, and despite the relatively short 1-week treatment, a
clear clinical benefit of OA was observed. This statement is tients showed adverse events or adverse drug reactions and
a shortened treatment period. Two of these patients were based on results proven within the whole study population
included as required by the inclusion and exclusion criteria, treated for 4 days. None of the other five had received treat-
ment for longer than 2 days. Six of these seven patients had as well as within the subgroup analyses. Improvements oc-
curred using OA in both postprandial and fasting venous received OA. Three had exhibited upper gastrointestinal and
central nervous symptoms such as nausea and/or vomiting ammonia levels compared with placebo. The postprandial
ammonia level estimated at 1:00 PM represents the reduction and were withdrawn from the study (5%). In none of the 63
patients in the placebo group were nausea and/or vomiting of the postprandial hyperammonemia, induced by the protein
intake of 0.5 g protein/kg body weight of the midday meal, observed. One patient treated with OA developed an acute
abdomen, secondary to a penetrating ulcer, and was with- after a 4-hour OA or placebo infusion in the morning (8:00
AM to 12:00 PM). It therefore represents the net effects of 4 drawn. One cirrhotic patient developed hepatorenal syn-
drome and was withdrawn. One patient refused to participate hours of infusion on hyperammonemia in combination with
AID Hepa 0001 / 5p21$$$$$1 05-08-97 16:59:50 hpta WBS: Hepatology
1358 KIRCHEIS ET AL. HEPATOLOGY June 1997
the previous 2 to 7 days of OA infusion. The fasting ammonia and HE II, it was found that the deviation from the normal
clearly increased as the initially existing mental state became level estimated at 8:00 AM represents the reduction of the
fasting ammonia level after 2, 4, and 7 days after daily 4- more pronounced, with respect to the changes detected in the
NCT-A, the changes in the NH
4
/
balance, and the changes hour OA or placebo infusion, in contrast to the uninfluenced
baseline level, as a result of the OA- or placebo-induced ef- in the PSEI. Thus, the worst initial conditions occurred in
the groups with the greatest limitations of mental function. fects. It therefore represents the impact of the previous 2 to
7 days of OA infusion. No differences among the different subgroups could be de-
tected in the final values attained after 7 days of therapy It should be borne in mind that the relevant initial post-
prandial ammonia levels used to calculate the difference be- with OA, whether for NCT-A, the NH
4
/
values (postprandial,
fasting), or the PSEI. The conformity of the improvements tween OA and placebo was determined in patients undergo-
ing OA or placebo treatment. In a previous study,
27
it was with respect to the final values attained in the OA group
during the 7-day therapy indicates the efficacy of the OA established that, at an identical protein intake of 0.50 g/kg
protein with the midday meal, a significant ammonia-reduc- therapy in all the subgroups. Regardless of the severity of
the initial condition in these mild forms of HE, it is possible ing effect occurs during a 4-hour infusion of 20 g as well
as 40 g OA. Therefore, it can be assumed that the initial with the OA therapy to reach a statewith respect to the
mental condition (HE gradation), the results in the psycho- postprandial ammonia levels on day 0 in the OA-treated pa-
tients had already been therapeutically influenced by the metric tests (NCT-A), and the ammonium (NH
4
/
) that is
near or in the vicinity of the normal for all the subgroups. intravenously administered OA. This is reflected by the lower
postprandial ammonia levels in the OA group (mean: 82 { Distinct differences from the placebo therapy can be shown
here. The improvements in the placebo groups are indeed 37 mmol/L) as compared with the placebo group (mean: 91 {
48 mmol/L). Because the baseline data in the fasting ammonia small, on the order of 10% to 15% spontaneous improvement
rates, but they are clearly different from the improvements levels confirm the comparability of both groups, ammonia
reduction in the OA group may be considerably greater than attained under OA. The more pronounced the initial changes
in HE gradation, NCT-A, or ammonium, the more pro- that reflected by the present data. The significant effect of
OA on fasting ammonia levels confirms this conclusion. nounced are the reactions to effective therapy in this sub-
group. The smaller the initial changes in HE gradation, NCT- A significant therapeutic benefit was also observed in favor
of OAin the NCT-A, which is an accepted and reliable psycho- A, or ammonium, the more difficult it is to show an effective
therapeutic treatment in this subgroup. Because of this, no metric test for the assessment of mental function in cirrhotic
patients with HE.
9,11,40
Psychometric tests were performed at significant improvements in comparison with placebo could
be shown regarding mental state gradation and ammonia- the same time as the postprandial ammonia measurement
and the evaluation of the patients mental state. The aim was lowering in patients with initially existing SHE.
Another reason is due to a possible type II error that arises to evaluate if the impact of the previous 4-hour infusion on
the protein-induced (12:00 PM meal [0.5 protein/kg body from the sample size calculation, which, as planned, was
based on the whole study population. That also applies to weight]) hyperammonemia in combination with the previous
2 to 7 days of infusion resulted in improvement of NCT-A specific analyses in the subgroup of patients with initial men-
tal state grade II. Because our definition of SHE was based and mental state grading. With regard to the whole study
population, a clear treatment effect in favor of OA could be on the 30-second borderline in NCT-A, we cannot exclude an
erroneous diagnosis of SHE in a number of patients. At the shown regarding improvement in NCT-A test time. Accord-
ingly, a superiority of OA in comparison with placebo was same time, the evidence of treatment differences because of
erroneously included patients could have been more difficult. shown in the subgroups with manifest HE (HE I, HE II).
Interestingly, the evaluation of the patients with initial SHE It is also clear from the results that SHE and overt HE have
different significance with respect to the necessity of the characterized by a prolonged NCT-A test time (30 seconds)
in the absence of manifest HE (grade 0) shows the only signif- treatment, but not with respect to its efficacy.
Adverse effects were not observed in patients in the placebo icant treatment-related effect in favor of OA in this subgroup
population with regard to improvement in NCT-A test time. group and in only 5% (3 patients) in the OA group. None
of the OA-induced side-effects were severe. These findings Otherwise, it could be shown that the response to OA in
patients with initial SHE is smaller than in those patients confirm the results of uncontrolled open trials with OA in
over 4,500 patients that show that it is a safe, well-tolerated with initially manifest HE (mental state grade I or II).
Decreases in mean mental state grading occurred twice as therapeutic agent.
This investigation confirms the beneficial results of intra- frequently in the OAgroup as in the placebo-treated patients.
These findings show that OA administration approximates venous OA in several previous clinical trials. Henglein-Ot-
termann showed improvement in ammonia levels in a dou- the therapeutic results obtained with lactulose/lactitol and
neomycin.
10-13
On the other hand, no improvement in mental bleblind crossover comparison of placebo and OA in normal
subjects and cirrhotic patients with hyperammonemia in- state grade was observed in 26 OA-treated patients, 17 of
whom (65%) had SHE at the beginning of treatment. The duced by infusions of ammonium chloride.
26
They found that
5 g OA given over 1 hour caused a rapid decrease in blood separate evaluation of these patients with SHE also shows,
with regard to mental state gradation, that the response to ammonia. Holm et al. studied various doses of OA infused
over 8 hours in cirrhotic patients with hyperammonemia in- OA is smaller than in those patients with initial manifest
HE (mental state grade I or II). Based on the odds ratios, it duced by ingestion of a liquid protein diet.
27
A single 8-hour
infusion of OA given in doses of 5 and 20 g diminished the was observed that the greater the initial mental state grada-
tion, the greater the effect of OA. These results indicate at hyperammonemia in a dose-related manner. Forty grams of
OA completely abolished the increase in blood ammonia con- least for the subgroups of patients with initial manifest HE
a significant improvement with regard to their mental state centration. In addition, this large dose of OA increased the
branched chain amino acidaromatic amino acid ratio.
27
gradation. We would anticipate on the basis of the relation-
ship shown by the odds ratio that patients with mental state The mechanism by which ornithine-aspartate works is of
much theoretical and practical interest. With respect to the grades 3 and 4 would respond even more dramatically. It is
probable that, in the case of initial SHE, the responder rate proven functional heterogeneity of the hepatocytes,
15,16,31-33
evidence was presented showing that glutamate, aspartate, with regard to HEgradation will improve with further contin-
uation or higher dose of OA treatment. a-oxoglutarate, and citric dicarboxylates such as malate and
benzoate were taken up almost exclusively by perivenous he- In the individual subgroups of patients with SHE, HE I,
AID Hepa 0001 / 5p21$$$$$1 05-08-97 16:59:50 hpta WBS: Hepatology
HEPATOLOGY Vol. 25, No. 6, 1997 KIRCHEIS ET AL. 1359
4. Hawkins RA, Jessy J. Hyperammonemia does not impair brain function
patocytes.
34-36
It was postulated that, under a pathological
in the absence of net glutamine synthesis. Biochem J 1991; 277:697-703.
condition such as cirrhosis, aspartate and also dicarboxylates
5. Norenberg MD, Baker L, Norenberg LOB, Blicharska J, Bruce-Gregorius
may serve as a carbon source and are consistent as a determi-
JH, Neary JT. Ammonia-induced astrocyte swelling in primary culture.
Neurochem Res 1991; 16:833-836. nant for the impaired GS flux in the perivenous scavenger
6. Dombro RS, Hutson DG, Norenberg MD. The action of ammonia on
cells.
34-36
As has been shown, the flux through the impaired
astrocyte glycogen and glycogenolysis. Mol Chem Neuropathol 1993; 19:
urea cycle enzyme system, especially through carbamylphos-
259-268.
phate synthetase, is increased by the administration of orni-
7. Takahashi H, Koehler RC, Brusilow SW, Traystman RJ. Inhibition of
brain glutamine accumulation prevents cerebral edema in hyperammo- thine, taken up by the periportal hepatocytes.
20-23
nemic rats. Am J Physiol 1991; 261:H825-H829.
In contrast to hepatic GS, recent findings suggest funda-
8. Ha ussinger D, Laubenberger J, Vom Dahl S, Ernst T, Bayer S, Langer M,
mentally different regulatory mechanisms for GS and the
Gerok W, et al. Proton magnetic resonance spectroscopy studies on human
removal of excess ammonia in muscle and brain. Being devoid
brain myo-inositol in hypo-osmolarity and hepatic encephalopathy. Gas-
troenterology 1994; 107:1475-1480. of an effective urea cycle,
30
these tissues rely on glutamine
9. Conn HO, Lieberthal MM. The hepatic coma syndromes and lactulose.
synthesis via GS for the removal of excess ammonia. Whereas
Baltimore: Williams and Wilkins, 1979:1-121.
GS activities in brain are decreased by portocaval shunting,
45
10. Bircher J. Treatment of chronic portal-systemic encephalopathy with lac-
enzyme activities in homogenates of skeletal muscle were
tulose. Lancet 1966; 1:890-893.
11. Atterbury CE, Maddrey WC, Conn HO. Neomycin-sorbitol and lactulose found to be significantly increased after shunt surgery.
30
It
in the treatment of acute portal-systemic encephalopathy. Dig Dis 1978;
is conceivable, as reported by a previous study, that the addi-
23:398-408.
tion of branched-chain amino acids to skeletal muscle homo-
12. Riggio O, Balducci G, Ariosto F, Merli M, Tremiterra S, Ziparo V, Capoc-
genates resulted in further significantly increased GS activi-
cacia L. Lactitol in the treatment of chronic hepatic encephalopathya
randomized cross-over comparison with lactulose. Hepatogastroenterology ties in muscle tissue.
37
1990; 37:524-527.
Recent evidence suggests that ornithine may act in the
13. Uribe M, Conn HO. Dietary management of portal-systemic encephalopa-
central nervous system to ameliorate HE in animal models.
38
thy. In: Conn HO, Bircher J, eds. Hepatic Encephalopathy: Syndromes
In addition to the stimulation of the periportal urea synthesis
and Therapies. Bloomington, IL: Medi-Ed Press, 1994:331-350.
14. Kaiser S, Gerok W, Ha ussinger D. Ammonia and glutamine metabolism in the liver, the proven increased transport of ornithine
in human liver slices: newaspects of the pathogenesis of hyperammonemia
across the blood-brain barrier indicates the central nervous
in chronic liver disease. Eur J Clin Invest 1988; 18:535-542.
system as a second target for ornithine. Both mechanisms
15. Ha ussinger D, Gerok W. Hepatocyte heterogeneity in ammonia metabo-
and targets afford reasonable explanations for the reduction
lism: Impairment of glutamine synthetase in CCl4-induced liver cell necro-
sis with no effect on urea synthesis. Chem Biol Interact 1984; 48:191-193. of increased blood ammonia concentration by ornithine asso-
16. Gebhardt R, Reichen J. Changes in distribution and activity of glutamine
ciated with lower concentrations of glutamine and lactate in
synthetase in CCl4-induced liver cirrhosis in the rat: potential role in hy-
brain.
39
These changes may be involved in the prevention of
perammonemia. HEPATOLOGY 1994; 20:684-691.
brain edema and may explain the simultaneous improvement
17. Salvatore F, Scoppa P, Cozzolina D. Protective effect of ornithine and
aspartic acid in chronic carbon tetrachloride intoxication. Clin Chim Acta of HE gradation shown by ornithine administration in these
1959; 4:728-732.
models.
39
Further clinical and experimental investigations on
18. Salvatore F, Cimino F, dAyello-Caracciolo M, Cittadini D. Mechanism of
the role of skeletal muscle and brain in the possible mecha-
the protection by L-ornithine-L-aspartate mixture and by L-arginine in
nisms of beneficial action of OA in HE are in progress.
ammonia intoxication. Arch Biochem Biophys 1964; 107:499-508.
19. Sioya A, Kuraishi K, Kakimoto M, Tamama Y. Pharmacological study on In summary, OA infusion resulted in significant, improved
L-ornithine-L-aspartate. Jap J Pharmacol 1964; 14:201-214.
NCT-A times, mental state grading, and PSEI scores, as well
20. Cohen NS, Chia-Wei Chung, Rijman L. The effects of ornithine on mito-
as in decreased fasting and postprandial venous ammonia
chondrial carbamyl phosphate synthesis. J Biol Chem 1980; 255:10248-
concentrations compared with placebo. In subgroups retro-
10255.
spectively classified according to their initial HE gradation 21. Goodman MW, Zieve L, Konstantinides, Cerra FB. Mechanism of arginine
protection against ammonia intoxication in the rat. Am J Physiol 1984;
(SHE, mental state gradation I and II), OA showed descrip-
247:G290-G295.
tive significances in the subgroups with manifest HE with
22. Gebhardt R, Beckers G, Gaunitz F, Haupt W, Jonitza D, Klein S, Scheja
respect to ammonia-lowering, improvement in NCT times,
L. Treatment of cirrhotic rats with L-ornithine-L-aspartate enhances urea
and mental state gradation. Subgroup analyses in patients synthesis and lowers serum ammonia levels. J Pharmacol Exp Ther 1997
(submitted).
with initial SHE revealed medication differences in the psy-
23. Zieve L, Lyftogt C, Raphael D. Ammonia toxicity: comparative protective
chometric test used in favor of OA. Consistently, OA appears
effect of various arginine and ornithine derivatives, aspartate, benzoate,
to be a safe and effective treatment of chronic (persistent)
and carbamyl glutamate. Metab Brain Dis 1986; 1:25-35.
manifest HE. Based on the results seen in the subgroup of
24. Leonhardt H, Bungert HJ. Therapie der schweren Hyperammonia mie.
Med Klin 1972; 67:1052-1056. patients with SHE, it is fair to assume that an effect of OA
25. Mu ting D, Reikowski J. Kontrollierte Studie hber die Wirkung einer oral
will be proven if the type II error could be excluded and more
verabreichten ammoniaksenkenden Aminosa ure Ornithin-Aspartat auf
sensitive diagnostic criteria could be used.
Leber- und Pankreasfunktion bei Leberzirrhosekranken. Therapiewoche
All studies of OA to date were performed in patients with
1980; 30:5990-5996.
26. Henglein-Ottermann D. Der Einflub von Ornithin-Aspartat auf die exper- SHE or mild, manifest HE (mental state grade I or II). Addi-
imentell erzeugte Hyperammonia mie. Klinisch-experimentelle Studie.
tional investigations in patients with SHE as well as with
Ther Gegenw 1976; 115:1504-1518.
more severe HE (grades III and IV) are in progress.
27. Staedt U, Leweling H, Gladisch R, Kortsik C, Hagmu ller E, Holm E. Ef-
fects of ornithine aspartate on plasma ammonia and plasma amino acids
Acknowledgment: The authors thank the physicians and
in patients with liver cirrhosis. A double-blind, randomized study using a
staff of all the author-affiliated hospitals and laboratories for
four-fold crossover design. J Hepatol 1993; 19:424-430.
their cooperation in the performance of the trial and for mak- 28. Ganda OP, Ruderman NB. Muscle nitrogen metabolism in chronic hepatic
insufficiency. Metabolism 1976; 25:427-435.
ing their clinical and laboratory data available to us. The
29. Butterworth RF, Girard G, Gigue`re JF. Regional differences in the capac-
authors gratefully acknowledge the secretarial assistance of
ity for ammonia removal by brain following portocaval anastomosis. J
Ms. Annette Dreher.
Neurochem 1988; 51:486-490.
30. Girard G, Butterworth RF. Effect of portocaval anastomosis on glutamine
REFERENCES
synthetase activities in liver, brain, and skeletal muscle. Dig Dis Sci 1992;
37:1121-1126. 1. Gines P, Quintero E, Arroyo V, Teres J, Bruguera M, Rimola A, Caballerria
31. Ha ussinger D. Hepatocyte heterogeneity in glutamine and ammonia me- J, et al. Compensated cirrhosis: natural history and prognostic factors.
tabolism and the role of an intercellular glutamine cycle during ureogen- HEPATOLOGY 1987; 7:122-128.
esis in perfused rat liver. Eur J Biochem 1983; 133:269-274. 2. Rink CH, Haeting J, Nilius R. Prognosis assessment in patients with liver
32. Gebhardt R, Mecke D. Heterogeneous distribution of glutamine synthetase cirrhosis [Abstract]. Hepatogastroenterology 1990; 37(suppl II):A86, 514.
among rat liver parenchymal cells in situ and in primary culture. Euro- 3. Jessy J, Mans AM, DeJoseph MR, Hawkins RA. Hyperammonemia causes
pean Molecular Biological Organisation Journal 1983; 2:557-560. many of the changes found after portocaval shunting. Biochem J 1990;
272:311-317. 33. Gaasbeek-Janzen JW, Lamers WH, Moorman AFM, DeGraaf A., Los AJ,
AID Hepa 0001 / 5p21$$$$$1 05-08-97 16:59:50 hpta WBS: Hepatology
1360 KIRCHEIS ET AL. HEPATOLOGY June 1997
Charles R. Immunohistochemical localization of carbamoylphosphate syn- cacy and mechanism studied by in vivo brain
1
H-MR spectroscopy [Ab-
stract]. Proceedings of the Society of Magnetic Resonance and the thetase (ammonia) in adult rat liver: evidence for a heterogeneous distribu-
tion. J Histochem Cytochem 1984; 32:557-564. European Society for Magnetic Resonance in Medicine and Biology, Nice
1995; 1759. 34. Stoll B, McNelly S, Buscher HP, Ha ussinger D. Functional hepatocyte
heterogeneity in glutamate, aspartate and a-ketoglutarate uptake: a his- 40. Conn HO. Trailmaking and number connection tests in the assessment of
mental state in portal systemic encephalopathy. Am J Dig Dis 1977; 22: toautoradiographical study. HEPATOLOGY 1991; 13:247-253.
35. Stoll B, Haussinger D. Functional hepatocyte heterogeneity: vascular oxo- 541-550.
41. Da Fonseca-Wollheim F. Direkte Plasmaammoniakbestimmung ohne En- glutarate is almost exclusively taken up by perivenous glutamine-synthe-
tasecontaining hepatocytes. Eur J Biochem 1989; 181:709-716. teiweissung. J Clin Chem Clin Biochem 1973; 11:421-431.
42. Hochberg Y, Benjamini Y. More powerful procedures for multiple signifi- 36. Stoll B, Ha ussinger D. Hepatocyte heterogeneity in uptake and metabo-
lism of malate and related dicarboxylates in perfused rat liver. Eur J cance testing. Stat Med 1990; 9:811-818.
43. Kleijnen J, De Craen A, Van Everdingen J, Krol L. Placebo effect in double- Biochem 1991; 195:121-129.
37. King PA, Goldstein L, Newsholme EA. Glutamine synthetase activity of blind clinical trials: a reviewof interactions with medications. Lancet 1994;
344:1347-1349. muscle in acidosis. Biochem J 1983; 216:523-525.
38. Albrecht J, Hilgier W, Januszewski S, Kapuscinski A, Quack G. Increase of 44. Rothman KJ, Michels KB. Sounding board. The continuing unethical use
of placebo controls. N Engl J Med 1994; 331:394-397. the brain uptake index for L-ornithine in rats with hepatic encephalopathy.
Neuroreport 1994; 5:671-673. 45. Girard G, Gigue`re JF, Butterworth RF. Region-selective reductions in
activities of glutamine synthetase in rat brain following portocaval anasto- 39. Vogels BAPM, Karlsen OT, Bove e WMMJ, Chamuleau RAFM. Orni-
thine treatment of hyperammonemia induced encephalopathy: the effi- mosis. Metab Brain Dis 1993; 8:207-215.
AID Hepa 0001 / 5p21$$$$$1 05-08-97 16:59:50 hpta WBS: Hepatology