A Study of Oral Nutritional Support with Oxandrolone in

Malnourished Patients with Alcoholic Hepatitis: Results of a

Department of Veterans Affairs Cooperative Study
CHARLESL. MENDENHALL,'THOMASE. MORITZ,~ GARY A. ROSELLE,'TIMOTHY R. MORGAN,3
BERNARD A. NEMCHAUSKY,~ CARL0 H. TAMBURR0,4 EUGENE R. SCHIFF,5 CRAIG J. MCCLAIN,'
LUIS s. h " 0 , ' JOHNI. &LEN,7 h U N SAhlANTA,' ROBERTE. WEESNER,' WILLIAM HENDERSON,~
PETER GARTSIDE,~ THOMASs. CHEN,' SAMUEL w. FRENCH,' ANTONIO CHEDID,2 AND VETERANS AFFAIRS
COOPERATIVE STUDY GROUP275*
Department of Veterans Affairs Medical Centers: 'Cincinnati, Ohio 45220; 2Hines, Illinois 60141; 3Long Beach, California
90822; 4Louisuille, Kentucky 40292; 5Miami, Florida 33125; 'Lexington, Kentucky 50511; 7Minneapolis,Minnesota 551 47;
'East Orange, New Jersey 07019; and 'Harbor-UCLA Medical Center, Torrance, California 90509

A Veterans Affairs cooperative study involving 273 had no effect on survival. However, adequate caloric
male patients was performed to evaluate efficacy of intake was associated with 19% mortality, whereas
oxandrolone in combination with an enteral food patients with inadequate intake exhibited 51% mor-
supplement in severe alcoholic hepatitis. All patients tality (p = 0.0001). These results indicate that nutri-
had some degree of protein calorie malnutrition. On an tional status should be evaluated in patients with
intention-to-treat basis, only minimal changes in mor- alcoholic hepatitis. When malnutrition is present,
tality were observed. However, in patients with mod- vigorous nutrition therapy should be provided, and in
erate malnutrition mortality on active treatment at 1 patients with moderate malnutrition oxandrolone
mo was 9.4% compared with 20.9% in patients receiving should be added to the regimen. (HEPATOLOGY 1993;17:
placebo. This beneficial effect was maintained so that 564-576.)
after 6 mo on active treatment 79.7% of patients were
still alive, compared with 62.7% of placebo-treated A universally accepted and effective therapy for acute
patients (p = 0.037). Improvements in both the se- severe alcoholic hepatitis (AH) does not exist. Nutri-
verity of the liver injury (p = 0.03) and malnutrition
(p = 0.05) also occurred. No significant improvement tional therapy has been used to treat severe alcoholic
was observed with severe malnutrition. To better liver disease. One of the earliest and largest studies was
determine the effect on therapeutic efficacy, we com- that by Patek et al., first reported in 1941 (1). The study
pared results with those from a nearly identical popu- was expanded in 1948 (2) and ultimately contained
lation (cooperative study 119) treated with oxan- survival data on 345 patients (115 actively treated and
drolone but not given the food supplement. Patients 230 controls). The study reported a significant im-
were stratified according to their caloric intake provement in survival when a diet high in protein and
(greater than 2,500 kcal/day was considered adequate supplemented with B vitamins was provided (controls
to supply energy needs and promote anabolism). For had 61% mortality, 140 of 230, at 1yr vs. 35%mortality,
patients with moderate malnutrition and adequate
caloric intake, oxandrolone treatment reduced 6-mo 40/115, in the nutritionally supplemented patients).
mortality (4% active treatment vs. 28% placebo Unfortunately, this study was flawed in that the control
[p = 0.0021). For patients with moderate malnutrition group was not randomized and its members were treated
and inadequate calorie intake, oxandrolone had no at different hospitals around the city (Boston, MA).
effect on mortality (30% active treatment vs. 33% Since then, several smaller reports have appeared (3-151,
placebo). In cases of severe malnutrition, oxandrolone ranging in number of patients studied from 4 to 52 and
in duration of treatment from 3 to 60 days. Although
individually these reports are of limited value, collec-
tively they strongly indicate that these patients do
*Authorship listing is based on manuscript preparation, data analyses,
patient enrollment and study participation.
tolerate high protein intake and that their nutritional
This work was funded by the Cooperative Studies Program of the Department status may be improved. When viewed collectively, these
of Veterans Affairs Medical Research Service. reports suggest improved survival.
Data were presented in part a t the 42nd Annual Meeting of the American More recently, anabolic steroids have been combined
Association for the Study of Liver Disease, Chicago, IL, Nov 2-5, 1991. with nutritional therapy. In patients with alcoholic liver
Address reprint requests to: Charles L. Mendenhall, M.D., Ph.D., Depart-
ment of Veterans Affairs, Medical Center (151F1, 3200 Vine Street, Cincinnati,
disease, Bonkovsky et al. (16, 17) observed the most
OH 45220 rapid improvement in patients receiving combined ox-
31/1/44739 androlone therapy and parenteral nutrition. The
564
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
HEPATOLOGY
Vol. 17, No. 4, 1993 MENDENHALL ET AL. 565

problem with all of these reports is that their study TABLE

1. Parameters used to evaluate degree of PCM
populations were small and the possibility of missing a Variable Lower limit of normal
small but significant effect was great.
In 1984, we reported our initial observations on the 1. Ideal body weight (%I 100
2. Skin-fold thickness (mm)
short-term (30-day) use of oxandrolone in patients with Triceps 10
AH (18).At that time, the results were encouraging in Biceps 5
that once therapy was completed, the 6-mo conditional Subscapular 17
death rate was significantly diminished in patients with Iliac 18
moderately severe liver disease. Very severe liver 3. Midarm muscle area (em2) 54
disease, as defined by coagulopathy and jaundice, was 4. Creatinine height index (%) 100
not improved. All of these patients had some degree of 5. Visceral proteins
protein calorie malnutrition (PCM), and many ate very Albumin (gmidl) 4.2
poorly during therapy. However, we have not previously Transferrin (mg/dl) 235
analyzed treatment efficacy in these patients with Prealbumin (mg/dl) 29
Retinol binding protein (mg/dl) 4.4
respect to their level of malnutrition and food intake 6. Total lymphocyte count 1460
during therapy. (cells/mm3)
Our objective in this second study was to design a 7. Skin-test responsiveness to 3 mm in duration to 3
therapeutic regmen that would significantly improve a battery of antigens" or more antigens
survival by extending the duration of oxandrolone 8. Muscle strength (hand dy- 91.5 pounds of pressure
administration and by increasing caloric and protein namometry)'
intake to meet estimated requirements. In an attempt to
A PCM score on each patient was derived from variables 2 through
achieve this objective, we provided a high-calorie, high- 7. PCM score was calculated as the sum of the individual parameter
protein food supplement (Hepatic Aid). In this manner, divided by the number of parameters on which data were available. No
we hoped to better provide an adequate calorie and patient was evaluated unless data on a t least five of the seven variables
protein source for anabolism. The impact of these were available. To combine diagnostic parameters, each parameter has
interventions on disease status and survival is the basis been expressed as a percentage of the lower limit of normal. The values
of this report. used as lower limits of normal were determined from a group of 22
nonalcoholic, non-liver-diseased, age-matched comparison subjects.
PATIENTS A N D METHODS "Seven skin-test antigens were used. They were tetanus toxoid
Patient Population and Evaluation. The study population antigen, diphtheria toxoid antigen, Streptococcus antigen, old tuber-
consisted of male patients from eight Department of Veterans culin, candida antigen, Trichophyton antigen and Proteus antigen plus
Affairs Medical Centers (DVAMCs) in whom a diagnosis of AH a glycerin solvent control. Skin-test antigens were available through
was based on history of heavy alcohol intake and the presence the Multitest CMI (Merieux Institute, Inc., Miami, FL).
of laboratory changes characteristic of alcoholic liver injury 'Hand dynamometry was measured with a Jamar dynamometer
(19). To avoid excluding more severely ill patients, histological (Asimow Engineering Co., Santa Monica, CA). Pounds pressure from
confirmation was not an absolute requirement except in those the right and left hands were averaged and expressed as the percentage
instances when findings were atypical of alcohol-related injury of the normal value for the 25th percentile.
Le., very high AST [>400 U/L] or ALT level [>200 UIL],
ALT > AST or unusually high alkaline phosphatase level [ > 5
times normal]). Atypical biochemical liver test results without
histological proof of diagnosis represented an exclusion cri- or greater but less than 90 were classified as representing
terion. Other exclusion criteria included (a) condition that moderately severe liver disease, whereas DF values of 90 or
might alter liver function and affect outcome variables (e.g., greater represented very severe liver disease. Patients were
additional forms of nonalcoholic liver disease, severe con- excluded if other medical conditions were present that might
gestive heart failure, insulin-dependent diabetes mellitus, alter the therapeutic outcome (i.e., recent intravenous drug
prior bypass surgery, malignancy, recent p a r e n t e d drug abuse, positivity for HBsAg or human immunodeficiency virus
abuse [in the preceding 6 mo], HBsAg or human immunode- antibodies, intractable congestive heart failure, neoplasms
ficiency virus positivity, (b) inability or unwillingness to take that commonly metastasize to liver or nonalcoholic liver
oral medications or the food supplement and attend follow-up diseases).
clinic, (c) late identification beyond the fifteenth day of Nutritional examination included assessment of food intake
hospitalization, (d) participation in other therapeutic studies for the month preceding hospitalization and during the
and (e) female gender. Women were not included because treatment period. Calorie intake was assessed at the time of
insufficient numbers were anticipated in the DVAMC system enrollment by the study dietitian, daily during the initial
to yield meaningful statistics. month of hospitalization and weekly during outpatient
The severity of AH was assessed on the basis of the level of treatment with the help of a nutritional diary and recall
coagulopathy and jaundice with the discriminant function history.
(OF) described by Maddrey (20)where DF = 4.6 (prothrombin Nutritional status was quantified with the Blackburn
time [sec]) + bilirubin (mg/dl). Patients with DF values less criteria (21) as outlined in Table 1. Observed values were
than 60 or with bilirubin levels of less than 5 mg/dl were normalized to percentage of lower limits of normal (22) and
considered to have mild AH and were not enrolled for averaged to produce a total PCM score. Using the ranges
treatment. To study severe AH, we enrolled only patients with recommended by Blackburn (21), we considered scores of 80%
clinical jaundice (bilirubin 2 5 mg/dl). These patients were to 99% of normal mild PCM, those 60% to 79% of normal
further stratified on the basis of DF level such that levels of 60 moderate PCM and those less than 60% of normal severe PCM.
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
566 MENDENHALL ET AL. HEPATOLOGY
April 1993

TABLE
2. Comparison of patients in CS 119 and CS 275 with an identical manner. This treatment provided 6.8 gmlday
moderate and severe malnutrition protein and 264 kcallday for inpatients and 5.1 gm protein and
cs 119 C S 275 198 kcallday for outpatients. Standard therapy for any
Parameters (n = 143) (n = 261) complicating conditions (e.g., ascites, encephalopathy, in-
fection) was provided to each patient, as was a standard
Ascites (% of patients) 86 89 hospital diet and a multivitamin and mineral mixture
Portal-systemic encephalopathy 68 61 (Theragran-M; Squibb, Princeton, N J ) to correct any unrec-
(% of patients) ognized micronutrient deficiencies. Additional enteral or par-
Nutrition (PCM score) 67 59 enteral alimentation was used at the discretion of the
Bilirubin (mg/dl) 16 15 primary-care physician for patients with inadequate oral food
AST (UL) 109 136 intake. Our aim was to compare the efficacy of standard
ALT (U/L) 47 51 treatment with that achieved when an anabolic steroid and
Prothrombin time (sec) 4.3 3.6 food supplement are added to standard therapy.
Endpoints. The main endpoint for response to therapy was
survival at 6 mo. Secondary endpoints were based on improve-
ments in results of clinical tests for liver injury, as well as
We recognize that all of the parameters used to assess PCM nutritional status and histological appearance when available.
are not specific for PCM but can be influenced by other disease Each patient was categorized at baseline with respect to his
processes (e.g., urinary creatinine is altered by muscle and liver disease (as measured by DF, where < 60 was mild, 60 to
kidney disease, visceral proteins are affected by liver disease 89 was moderately severe and 2 90 was very severe) and with
and changes in the extravascular space and immune status is respect to his level of malnutrition (as measured by the PCM
affected by hormonal changes and concomitant infections). For score, where 280 was mild, 60 to 79 was moderate and <60
that reason, the complex of nutritional parameters was was severe). Improvement or worsening of a patient’s con-
combined into a total score that gives a more global picture of dition was deemed to have occurred when the changes in status
the nutritional status while diminishing the nonnutritional were of magnitude sufficient to move him to a different
influence. Furthermore, changes in this PCM score have been category from baseline. Histological appearance was assessed
shown to accurately reflect survival in this patient population as described previously (23) by three pathologists who were
(221, which is the primary endpoint of treatment efficacy. blinded to the treatment randomization and who indepen-
However, in view of the nonspecificity of these parameters, we dently evaluated each specimen identifying the category of
also performed a functional assessment of nutritional status alcoholic liver injury (e.g., fatty liver, alcoholic hepatitis,
using hand dynamometry, which was not included in the PCM cirrhosis). They also assigned a pathological severity index
score. As with the other nutritional parameters, hand dyna- score, which quantified the various pathological features
metric readings were expressed as percentages of normal present in each category (e.g., inflammation, necrosis, fibrosis).
values. Changes from baseline in the pathologic category and the
Patients were enrolled in the study between 5 and 15 days severity index constituted the basis for histological im-
after hospitalization to permit patient identification; stabili- provement or worsening.
zation of acute liver injury; and time to obtain the required Comparison with Prior Observations. In our earlier study
clinical, nutritional, laboratory and histological evaluations. (Cooperative Study [CS] 119), we enrolled a nearly identical
Treatment Regimen. The physicians, patients and study patient population into a similar treatment regimen in which
nurses were blind with respect to study treatments. Although the anabolic steroid oxandrolone was administered and calorie
matched for volume, color and flavoring, the high-calorie and intake and nutritional status were assessed (18,22,241. Table
low-calorie supplements were different in weight. Hence the 2 shows the close similarity between the two study popula-
study dietitian a t each hospital who prepared and administered tions. A statistical comparison (CS 119 vs. CS 275) of the
the food supplement was often able to identify the food severity of the liver injury (DF) and degree of malnutrition
supplement. Dietitians were instructed not to reveal this (PCM score), as well as 6-mo mortality confirmed, that the two
information to the nurse or physician collecting study data. patient populations were quite comparable. This permitted us
Treatments (active vs. placebo) were randomly assigned by the to compare patients from both studies on the basis of
coordinating center (Hines DVAMC) so as to be balanced for treatment with oxandrolone and calorie intake. A detailed
the severity of the liver disease (DF) in each hospital. Because description of the inclusion and exclusion criteria, clinical and
the severity of liver disease correlated with the degree of PCM nutritional evaluations and study design and treatment
(221, this also produced balanced distribution of malnutrition. regimen for CS 119 have been reported (18, 22, 241, but
Because of the severity of patients’ liver disease at enrollment, survival analyses based on nutritional status, dietary intake
the first 30 days of treatment (active and placebo) were and oxandrolone treatment have not been previously analyzed.
administered while patients were in the hospital and followed All patients from CS 119 with comparable levels of liver
by 60 days of outpatient care. Active treatment consisted of disease (e.g., jaundice with bilirubin 2 5 mddl and a DF 2 60)
oxandrolone (80 mdday) for 30 days accompanied by a were stratified into subpopulations on the basis of the degree
high-calorie, high-protein food supplement (Hepatic Aid) of PCM present in a manner identical to that employed in CS
containing a preponderance of branched-chain amino acids. 275. Each subpopulation was further subdivided into active-
This supplement provided 60 gm protein and 1,600 kcal/day in treatment and placebo groups (i.e., active = oxandrolone [80
divided feedings. Outpatient therapy consisted of oxandrolone rndday for 30 days]; placebo = oxandrolone placebo for 30
(40 mdday) for 60 days accompanied by Hepatic Aid in days). Caloric intake recorded in hospital during treatment
amounts estimated to provide an additional 1,200kcdday and was determined; survival analyses and statistical comparisons
45 gm proteinlday in divided feedings. were then made between the resulting groups in CSs 119 and
Placebo treatments consisting of placebo tablets (identical in 275.
appearance to the oxandrolone tablets) accompanied by the Data Analyses. Analyses were performed on an intention-
low-calorie, low-protein food supplement were administered in to-treat basis. Comparability of treatment groups was tested
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
HEPATOLOGY
Vol. 17, No. 4, 1993 MENDENHALL ET AL. 567

TABLE
3. Characteristics of the total population
Characteristics Active treatment (n = 137) Placebo (n = 136)

Age O..)" 50 t 10 51 k 9
Daily caloric intake (kcal/dayY 2,830 t 1,573 2,637 t 1,417
Alcohol (96) 50 46
Nonalcohol (%) 50 54
Alcohol intake Cgndday)" 221 t 170 188 t 180
Duration of alcohol intake (yr)" 25 t 12 2 6 t 12
Severity of liver disease (DF)" 86.6 t 12.8 87.0 t 14.3
Moderate (% distribution) 62 58
Severe (% distribution) 38 42
Malnutrition (PCM score: % of normal)" 59.8 t 11.1 60.0 t 10.8
Hand dynamometry (% of normal)" 65.5 t 22.1 66.4 t 22.8
Ascites (% with moderate or severe) 64 68
Portal-systemic encephalopathy (% present) 57 63
Hepatomegdy (% present) 84 83
Splenomegaly (% present) 37 26
Infection (% present) 29 27

"Data expressed as mean t S.D.

with the x2 statistic on qualitative variables. Quantitative the malnutrition was 3.7% mild (10 of 2711, 48.3%
variables were evaluated with Student's t test for normally moderate (131 of 271) and 48.0% severe (130 of 271).
distributed data and the Wilcoxon rank-sum test for abnor- Mean initial PCM score for all patients was 60% 2 11%
mally distributed data. All tests were two-sided. When of normal.
treatment groups were further stratified by level of PCM, the These severely depressed values were supported with
resulting four-group comparisons were made with one-way a functional assessment of nutritional status (hand
ANOVA using Tukey's least-significant-differencetest to
compare groups in pairs. Survival curves were computed by the dynamometry). Adjusting for age and sex using standard
Kaplan-Meier method. The comparison of survival distribu- tables for normal strength (25), we observed that the
tions was made with the log-rank test. Multivariable logistic- median strength for our patients was at the tenth
regression analysis was performed to determine which risk percentile, representing a decrease to 66% of normal
factors and their interactions affected survival. strength. These values showed a significant correlation
(p I 0.0006) with those parameters related to calorie
RESULTS deprivation (e.g., ideal body weight, skin-fold thickness,
Population.Between 1987 and 1991, 273 male pa- midarm muscle area or creatinine height index) and with
tients met the inclusion criteria, signed the informed some of those related to protein deprivation (e.g.,
consent and enrolled for treatment; 136 received placebo albumin [p = 0.041 and transferrin [p = 0.00011) but
and 137 underwent active treatment. not the visceral proteins with short half-lives (e.g.,
Comparability of Treatment Groups a t Baseline. All retinol binding protein or prealbumin) or immune status
patients had severe alcoholic liver injury with some (e.g., total lymphocytes and skin-test responses).
evidence of clinical liver decompensation (as defined by Liver histological samples were available for evalu-
jaundice, coagulopathy, ascites and hepatic encephalop- ation in 121 patients. Of these, 96 were obtained in the
athy). As an inclusion criterion, all had jaundice (bili- first month of treatment; an additional 25 specimens
rubin 2 5 mg/dl; mean = 14.7 2 9.2 mg/dl), 89% had were obtained sequentially after treatment, either on
ascites, 60%had some evidence of portosystemic enceph- autopsy or biopsy. The histological distribution between
alopathy, and 56% had coagulopathy (prothrombin time treatment groups was not significantly different, except
prolonged >3.0 sec after vitamin K). If one further that the more serious combined lesion of AH with
quantifies the severity of the injury with Maddrey's DF cirrhosis was slightly more prevalent in the active-
(20), 164 patients (60%)had moderately severe disease treatment group (i.e., 37.3% of the active treatment
and 109 (40%) had very severe disease. Patients with group and 26.5% of the placebo group). A detailed
mild disease (bilirubin level < 5 mgjdl and DF < 60) were description of the nutritional abnormalities and histo-
excluded from treatment. Overall mean DF was logical changes will appear in other manuscripts.
86.8 ? 13.6. The baseline characteristics at enrollment Correlates with Survival. Survival appeared to cor-
are shown in Table 3. The two treatment groups (active relate with several interrelated parameters. These in-
treatment vs. placebo) were comparable for all baseline clude the status of the liver disease as quantified by
characteristics (p > 0.05). bilirubin level and prothrombin time (DF), nutritional
Nutritional status was abnormal in every patient. The status as quantified by the parameters utilized in the
severity of nutritional abnormalities (PCM score) PCM score and caloric intake during the first month of
ranged from 34% to 92% of normal. The distribution of hospitalization. Figure 1 shows the relationship of
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
568 MENDENHALL ET AL HEPATOLOGY
April 1993

40 I‘
0 1
I I

2 3
, + 4 6
Time (months)

FIG.1. Six-month survival analyses of 273 patients with AH. Severity of the AH was effectively stratified on the basis of the level ofjaundice
(bilirubin [mgidl]) and coagulopathy (prothrombin time [sec]) using a DF (20) where DF = 4.6 (prothrombin time) + bilirubin. o = moderately
severe AH, DF = 60 to 89; + = very severe AH, DF 2 90. Patients with moderately severe AH had significantly better 6-mo survival than did
those with very severe AH (p = 0.0001; log-rank test).

survival to DF. Indeed, patients with moderately severe (Maddrey’s DF [201) is designed to predict outcome.
liver disease (DF = 60 to 89) had a 6-mo cumulative However, a subpopulation based on liver disease severity
mortality of 26% compared with those with very severe (DF) failed to demonstrate a select group that benefitted
liver disease (DF 2 go), who had a mortality of 53% significantly from therapy. Nutritional status also has
(p = 0.0001, log-rank test). Similarly, the level of mal- been shown to be a significant risk factor for mortality
nutrition could be used to predict outcome (Fig. 2). (221, and, because our therapy was directed at improving
When baseline PCM score was moderate (PCM nutritional status, it seemed reasonable to stratify the
score = 60% to 79% of normal), 6-mo cumulative mor- population on the basis of the pretreatment level of
tality was 29%, compared with 45% in patients with malnutrition (i.e., PCM score).
severe PCM (<60%); p = 0.0012, log-rank test. Stratification of the population on the basis of the level
Mild-PCM patients were not included because of the of malnutrition did not result in a biased distribution of
small sample size (n = 10). the treatment groups. As shown in Table 4, separating
Caloric intake during the first month of therapy also the moderately and severely malnourished patients for
was related to outcome. This is shown graphically in active and placebo treatments resulted in four popula-
Figure 3. As caloric intake decreased to less than 1,000 tions whose baseline levels of liver disease were indis-
kcallday, mortality approached 90%, whereas intake tinguishable. Only the levels of malnutrition distin-
greater than 3,000 kcal/day was associated with 0% guished the populations. In the moderate group, the
mortality (p = 0.0001, x2 test). mean PCM score was 67 5 5; in the severe group it was
Further correlation analyses of these three variables 51 -+ 6. Although PCM scores identified the level of
revealed that the caloric intake in the first month of malnutrition, the distribution of patients based on DF
treatment tended to be lower as severity of liver disease was heterogenous, with 54% to 64% of each group
increased (DF vs. kcal/day; p = 0.0001, r = - 0.32) and comprising patients with moderately severe DF levels
as PCM worsened (PCM vs. kcal/day; p = 0.0007, (60 to 89) and 36% to 46% comprising those with very
r = 0.21). severe levels ( 2 90). This distribution closely paralleled
Treatment Eficacy. On a n intention-to-treat basis, that of the entire population.
we saw no significant differences in response to therapy. Table 5 shows dietary caloric and protein con-
The analysis for survival, which was our main endpoint sumption during treatment, as well as the patients’
for response to treatment, revealed cumulative 6-mo renewed alcohol abuse, stratified according to their
mortality rates of 35% for active treatment and 39% for baseline level of nutrition and the treatment provided.
placebo treatment (a 10% difference in mortality that Overall, 40% of the active treatment group and 39% of
was not statistically significant; p = 0.455). This does the placebo group admitted to drinking after discharge
not preclude the possibility that a well-defined subgroup from the hospital. The addition of the food supplement
had benefitted significantly. Various a priori features of Hepatic Aid provided a significant increase in protein
the disease have been established as risk factors for and energy consumption compared with placebo. This
mortality. These were then used to characterize sub- was especially true during the first month of hospital-
groups within the population. Severity of liver disease ization and was seen in all groups, regardless of level of
quantified by the degree of coagulopathy and jaundice baseline malnutrition. The use of the food supplement
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
HEPATOLOGY
Vol. 17, No. 4, 1993 MENDENHALL ET AL. 569

so I
0 20 40 60 SO 100 120 140 160 180

TIME ( Days )
FIG.2. Six-month survival analyses on 259 patients with AH and moderate or severe PCM. Data were insufficient in four patients to determine
the PCM score. o = moderate malnutrition, PCM score = 60 to 79; * = severe malnutrition, PCM score <60. Patients with moderate
malnutrition had significantly better 6-mo survival than did those with severe malnutrition (p = 0.0012; log-rank test).

80

60

20

Caloric Intake (Kcal/d)

FIG.3. Six-month mortality associated with 1-mo daily caloric intake in 245 patients. Insufficient data on caloric intake precluded the inclusion
of 28 patients. Numbers in parentheses denote number of patients. A highly significant relationship was observed (p = 0.0001; x2 test).

resulted in 53%of the active treatment group consuming bloating, diarrhea) were somewhat more frequent in
greater than 2,500 kcdday, compared with only 15%of patients given Hepatic Aid, these complications were
the placebo group. not significantly more prevalent than those occurring
Complications during therapy were frequent in both in the placebo patients. Protein intake exceeded 100
active treatment and placebo treatment patients (Table gm/day for many of the patients receiving active treat-
6). Although gastrointestinal complaints (e.g., nausea, ment; encephalopathy was slightly more prevalent in
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
570 MENDENHALL ET AL. HEPATOLOGY April 1993

TABLE
4. Clinical and laboratory parameters at baseline related to level of malnutrition
Mild PCM Moderate PCM Severe PCM
Active Active Active
treatment Placebo treatment Placebo treatment Placebo
Parameters (n = 5) (n = 5) (n = 64) (n = 67) (n = 67) (n = 63)
_ _ ~

Ascites (%) 60 80 88 87 91 92
Portal-systemic encephalopathy (%) 40 40 53 59 63 68
Nutrition score (%) 87 85 67 67 51 51
Bilirubin (mglday) 9.0 11.5 15.0 15.8 14.8 13.7
ALT (UL) 151 175 144 138 133 129
Alkaline phosphatase (UL) 248 206 232 249 260 232
Prothrombin time (sec) 3.0 2.6 3.5 3.3 3.6 3.9
DF 78.0 79.6 86.6 87.2 86.9 87.5
% Moderate 80 80 59 60 64 54
% Severe 20 20 41 40 36 46

On comparison of clinical and laboratory parameters a t baseline in active and placebo treatment groups, none were significantly different
except for the predefined level of nutrition (PCM score); on comparison of these parameters in the mild, moderate and severe PCM populations,
patients with mild PCM tended to exhibit less severe disease, but the number of patients (five) was small and none of the comparisons was
statistically significant.

TABLE
5. Dietary differences during 3 mo of treatment
Mild PCM Moderate PCM Severe PCM
Active Active Active
Dietary parameters treatment Placebo treatment Placebo treatment Placebo

Alcohol intake
% drinking 67 25 38 44 38 34
gm/day 103 130 121 50 110 102
Daily dietary intake
In hospital
EER 2,735 2,521 2,696 2,709 2,368 2,386
Diet (kcdday) 1,183 1,722 1,368 1,238 1,014 1,357"
Supplement (kcaVday) 1,114 190" 1,207 187" 1,063 178"
Protein (gm/day) 91 73 100 54" 80 59"
Outpatient
Diet (kcdday) 1,826 2,622 1,912 1,627" 1,732 1,804
Supplement (kcallday) 748 159" 739 139" 706 113"
Protein (gm/day) 97 108 101 71" 92 74"
"p < 0.05.
EER = estimated energy requirements as calculated from the Harris-Benedict equation plus an injury factor for the increased metabolic
needs in AH. Indirect calorimetry in patients with AH indicates a 144% increase in resting energv expenditures with moderate AH(29) and a
174%increase for severe AH. The lesser EER seen in the severe-PCM patients resulted from greater wasting, which produced lower body weights.

the placebo group but did not reach statistical signif- enced an allergic skin rash attributed to oxandrolone
icance. (but not confirmed by rechallenge) and one had severe
Although high-dose oxandrolone therapy was em- diabetes mellitus that complicated management.
ployed in a patient population with preexisting severe Therapy was stopped in one placebo patient, also
liver injury, no complications attributable to its use were because of management difficulties posed by severe
observed. Increased cholestasis, which has been asso- diabetes mellitus.
ciated with the use of most 17-alkylated steroids (261, The therapeutic efficacy of active treatment at 6 mo as
was observed with nearly identical frequency (16.1% of estimated on the basis of DF and level of malnutrition
active-treatment patients and 16.2%of placebo patients) (PCM score) was most successful in those patients with
and in each instance was attributed to the natural moderate PCMs at baseline. In this group, active
progression of the disease. treatment was associated with a stabilized level of liver
Side effects of therapy were severe enough to neces- disease (i.e., the DF stayed within the same range of liver
sitate cessation of treatment in eight patients under- injury or improved in 72% of the patients, compared to
going active treatment. Five of these patients had severe 52% in the placebo group [Fig. 41). A n identical per-
nausea, vomiting and bloating: one was allergic to the centage, 72% (not necessarily the same patients) also
chocolate flavoring in the food supplement, one experi- had stabilized or improved PCM scores with active
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
HEPATOLOGY
Vol. 17, NO. 4, 1993 MENDENHALL ET AL. 571

PLACEBO

IMPROVED IMPROVED

-
DF 40 UNCHANGED I..

- 28
W 0R S EN E D
WORSENED
4a

IMPROVED IMP ROVED

A1

-
PCM

UNCHANGED
31
-28
WORSENED
w ~~

WORSENED
47
p=o.os

FIG.4. Six-month changes in severity of the liver disease and nutritional status in patients with moderate PCM a t baseline. Patients were
categorized a t baseline as to levels of liver injury, DF and PCM score (i.e., mild, moderate and severe categories). Improvement or worsening
represented a change in the DF or PCM category. Comparisons between unchanged and improved vs. worsened condition in the active and
placebo treatments were made with x2 test where active treatment was associated with significantly better results (p = 0.03 for DF, 0.05 for
PCM score).

TABLE
6. Complications during therapy
Treatment group
Complications Active" Placebo" p Valueb

Gastrointestinal
Bloating 62.8 55.2 NS
Nausea 50.4 43.4 NS
Anorexia 43.8 41.9 NS
Diarrhea 52.6 41.1 NS
Gastrointestinal bleeding 29.9 24.3 NS
Hepatic
Ascites 29.2 30.2 NS
Encephalopathy 19.0 21.3 NS
Hepatorenal syndrome 21.9 22.8 NS
Cholestasis 16.1 16.2 NS
Miscellaneous
Infection 48.9 44.1 NS
Pancreatitis 8.0 5.2 NS
Azotemia 38.7 36.8 NS
Hyperglycemia 42.3 34.6 NS
Other 23.4 24.3 NS
"Data expressed as percentage occurrence.
bComparisons (active vs. placebo) were not significantly different (p > 0.05).

treatment, compared with 53% receiving placebo score) in patients with severe malnutrition. In this
treatment. This represented a significant difference in group, 38%of the DF values and 39%of the PCM scores
response (p = 0.03 and 0.05, respectively).No difference remained the same or improved on active treatment (vs.
between treatments (active vs. placebo) was observed in 49% and 48%, respectively, with placebo [p = 0.25 for
the liver status (DF) or level of malnutrition (PCM DF and 0.35 for PCM score]).
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
572 MENDENHALL ET AL. HEPATOLOGY
April 1993

S
'O\ \' 1

90

80

70

60

50
0 1 2 3 4 5 6

7 - r e (rrontns)
Active *
- - Placebo Active -*- Placebo
Moderate PCM Severe PCM

FIG.5. Survival distribution for 131 patients with moderate and 130 patients with severe malnutrition (PCM). Because only five active and
five placebo patients had mild PCM, accurate survival analyses for this level of PCM were not possible. Active treatment of patients with
moderate PCM resulted in significantly improved survival ( p = 0.037; log-rank test). Severe-PCM patients with active treatment were not
significantly different from placebo patients (p = 0.39). Changes in the number of patients in each group (n) were 0 : n = 64 at baseline, 51 at
6 mo; 0:n = 67 at baseline, 42 at 6 mo; A: n = 67 at baseline, 34 at 6 mo; A : n = 6 3 at baseline, 37 at 6 mo.

Serial specimens of liver tissue were available early in and CS 275) significantly improved survival at 6 mo
treatment and after treatment in 25 patients. This compared with placebo treatment. For CS 119 (active
number was insufficient to permit separation into both treatment vs. placebo treatment) in patients with
moderate-PCM vs. severe-PCM and active- vs. placebo- moderate PCM, the 6-mo cumulative mortality was 22%
tre9tment groups. Viewing the study population collec- vs. 40% (p = 0.021, with a median survival time of 32 mo
tively, we noted that histological appearance improved vs. 12 mo. In CS 275 (active treatment vs. placebo
without a significant difference between treatment treatment), the 6-mo cumulative mortality was 20% vs.
groups (76.5% improved on active treatment, compared 37%, with a median survival of 49 mo vs. 26 mo
to 75.0% receiving placebo). (p = 0.037). For patients with severe PCM, active
The most striking effect of therapy was its rela- treatment was not associated with improved 6-mo
tionship to survival. This is shown graphically in Figure cumulative mortality in CS 119 or CS 275 (6-mo
5. After 30 days we observed an 11.5% difference in mortality 44% with active treatment vs. 38% with
mortality such that 9.4% had died while receiving active placebo).
treatment, compared with 20.9% receiving placebo Our study design did not differentiate dietary effect
(confidence interval = 0.6 and 23.6, respectively). At 6 from oxandrolone effect. All patients ate ad libitum.
mo this difference was 17%, with a mortality of 20.3% on Some undergoing active treatment with oxandrolone
active treatment compared with 37.3% with placebo consumed inadequate calories despite the use of Hepatic
treatment (confidence interval = 1.8 and 32.2, respec- Aid. Similarly, many patients receiving placebo con-
tively); p = 0.037; log-rank test. In the patients with sumed adequate calories, even without a high-calorie
severe PCM, no difference in survival was noted between food supplement. Therefore, to determine the rela-
therapy groups (p = 0.39). tionship of dietary intake during treatment to survival,
Comparison of Results between CS 119 and CS we further substratified our patients into four groups on
275. When patients were stratified on the basis of the basis of calories consumed (Fig. 6). Using the
baseline malnutrition (PCM score), the results of CS 119 Harris-Benedict equation, we defined adequate intake as
and CS 275 were quite similar. Active therapy in CS 119 greater than 2,500 kcallday. In patients with moderate
differed from that in CS 275 in that patients were given PCM, oxandrolone in the absence of adequate food
no food supplement, only the standard hospital diet, and intake produced results similar to those observed with
were treated for only 1 mo with oxandrolone. For placebo, irrespective of caloric intake (e.g., 28% mor-
patients with moderate PCM (60% to 79% of normal), tality with placebo plus high calorie intake [group 21 vs.
the study population consisted of 115 patients in CS 119 30% mortality with oxandrolone but inadequate calorie
(59 active treatment and 56 placebo treatment) and 131 intake [group 31 vs. 33% placebo and inadequate calorie
patients in CS 275 (64 active treatment and 67 placebo intake [group 41). Only the combination of oxandrolone
treatment). Both active-treatment regimens (CS 119 and adequate calorie intake (group 1)resulted in a highly
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
HEPATOLOGY
Vol. 17, NO. 4, 1993 MENDENHALL ET AL. 573

1 2 3 4 1 2 3 4
Treatment Groups

FIG. 6. Six-month mortality in patients with moderate and severe PCM. Patients with moderate and severe PCM from two DVA CSs (CS 119
and CS 275) are combined and stratified into four treatment groups. Group 1 = oxandrolone with adequate energy intake (n = 75 [source: CS
119 = 13, CS 275 = 621);group 2 = placebo with adequate energy intake (n = 33 [source: CS 119 = 17; CS 275 = 161);group 3 = oxandrolone
with deficient energy intake (n = 98 [source: CS 119 = 42; CS 275 = 561); and group 4 = placebo with deficient energy intake (n = 138 [source:
CS 119 = 35, CS 275 = 1031).Adequate intake was estimated using the Harris Benedict equation corrected for the increased resting energy
expenditure for patients with AH (29). On the basis of these estimates, > 2,500 kcal/day was deemed necessary to provide an adequate energy
supply for recovery and to correct the existing energy deficit (groups 1and 2); 5 2,500 kcaVday was classified as deficient (groups 3 and 4). The
bars indicate the combined percentage mortality from both studies; numbers in parentheses indicate the number of deaths per total group
(fractional mortality [FM]). Each group exhibited the following characteristics. Moderate PCM: group 1: FM for CS 119 = 0 of 11, CS 275 = 2
of 36; DF = 83 2 13; PCM = 68 L 5. Group 2: FM for CS 119 = 4 of 12, CS 275 = 1 of 6; DF = 84 t 15; PCM = 69 t 7. Group 3: FM for
CS 119 = 6 of 32, CS 275 = 10 of 22; D F = 89 t 15; PCM = 69 t 6. Group 4: FM for CS 119 = 8 of 26, CS 275 = 18 of 54; DF = 87 t 15;
PCM = 68 t 5. Severe PCM: Group 1:FM for CS 119 = 0 of 2, CS 275 = 6 of 26; D F = 79 t 15; PCM = 54 2 5. Group 2: FM for CS 119 = 2
*
of 5, CS 275 = 0 of 10; D F = 78 8; PCM = 54 t 4. Group 3: FM for CS 119 = 3 of 10, CS 275 = 23 of 34; D F = 89 t 12; PCM = 51 t 7.
Group 4: FM for CS 119 = 2 of 9; CS 275 = 24 of 49; D F = 90 -c 16; PCM = 51 t 7.
With regard to patients with moderate PCM, group 1 had significantly less mortality than did group 2, 3 or 4 (p = 0.003; X2 test). We saw
no significant difference between groups with respect to PCM or DF.
For patients with severe PCM, oxandrolone did not significantly improve survival (group 1 vs. group 2). However, adequate calorie intake
(groups 1 and 2) was associated with significantly lower mortality (p = 0.002; x2 test). We saw no significant difference between groups with
respect to PCM, but DF was significantly higher (p = 0.001) in patients with deficient energy intake.

significant improvement in mortality: 4% (p = 0.001). in those with more severe liver disease (mean DF for
These differences could not be accounted for on the basis group with adequate caloric intake, 78 5 5 [n = 351 vs.
of distributional bias because, in patients with moderate 90 & 14 [n = 1101 for patients with inadequate intake;
PCM, the pretreatment differences in the liver disease p = 0.001). Hence the difference in mortality may not
(DF) and level of PCM were not significantly different have resulted from poor nutritional intake during active
(active treatment vs. placebo; p > 0.05). treatment but rather may have occurred because these
In patients with severe disease, on comparison of patients who ate inadequate amounts were sicker at the
oxandrolone and placebo groups (group 1vs. group 2 and time of hospitalization. To gain more information on the
group 3 vs. group 4), no improvement was observed with relationship of survival with these four risk factors
oxandrolone. It appeared to be ineffective in patients (baseline liver injury [DF], level of malnutrition [PCM
with severe protein and calorie depletion. However, on score], treatment effect and dietary intake during the
comparison of patients with severe PCM who consumed first month of therapy), a multivariable logistic-
adequate calories ( > 2,500 kcal/day) with those who regression model was developed with backward-
consumed inadequate calories ( I 2,500 kcal/day), a elimination logistic-regression analysis. Candidate vari-
highly significant difference was observed (6-mo mor- ables included these four main variable and all two-way
tality 19%with severe PCM and adequate intake vs. 51% interactions between the four (Table 7). These analyses
with inadequate intake [groups 1 and 2 vs. groups 3 and indicated that, even after adjustment for three of the
4; p = 0.0004). This comparison strongly supports four variables, highly significant relationships still ex-
vigorous therapeutic nutritional intervention in the isted between survival and PCM (p = O.OOOl), food
group with severe PCM. However, in patients with intake (p = 0.0003) and DF (p = 0.0009).
severe PCM, a confounding bias could have been present Alone, treatment showed only a marginal relationship
because dietary intake was more frequently inadequate with survival (p = 0.116). However, in combination
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
574 MENDENHALL ET AL. HEPATOLOGY
April 1993

TABLE
7. Relationship between survival, baseline malnutrition (PCM score), liver dysfunction (DF), treatment and dietary
intake during the initial month of treatment
Predictors of survival p Value Risk ratio Confidence limits

1. PCM score 0.0001 4.22 2.02-8.84

2. Dietary intake (kcaUday) 0.0003 3.38 1.76-6.48
3. DF 0.0009 2.35 1.42-3.89
4. Treatment 0.116 1.76 0.87-3.56
5. Treatment - PCM 0.024 3.14 1.16-8.47
6. PCM - DF 0.069 2.57" 0.93-7.11"
7. Treatment dietary intake
~
0.066 3.30" 0.92-1I .75"
8. Treatment - DF > 0.20 - -
9. Dietary intake - DF > 0.20 - -
10. Dietary intake - PCM > 0.20 - -

Relationship between survival and the four independent variables (predictors 1 through 4) was determined with backward-elimination
logistic-regression analysis on data from patients in CS 119 and CS 275. Variables were evaluated alone as primary risk factors and in
combination to determine independent relationships and interactions in survival. The risk of dying was significantly increased by the level of
malnutrition (PCM), dietary intake and severity of the liver disease (DF). Treatment had marginal effect on survival but in combination with
PCM it had a significant interaction. Marginally significant interactions were present in PCM - DF and treatment - dietary intake.
"Risk ratios and confidence limits represent those observed just before their elimination from the model.

with PCM an interaction occurred that produced a eat the most poorly during recovery. It has been
significant effect on survival such that the risk of death suggested that this is related to increased production of
increased 3.14-fold in patients treated with placebo. cachexin (tumor necrosis factor-a) (27), the level of
Marginally significant interactions were also observed which is increased in this population and is also related
between DF and PCM (p = 0.069) and treatment and to outcome (28).
calorie intake (p = 0.066). Thus, although three of four To assess the relative importance of food intake on
primary factors are independent and significant in the survival independent of the baseline level of PCM or
model, the presence of significant interactions leads to liver injury (DF), which may influence both intake and
the conclusion that these factors are not independent of survival, we performed a backward-elimination logistic-
each other with respect to outcome. regression analysis. This analysis (Table 7) suggests that
the level of PCM, caloric dietary intake and severity of
DISCUSSION the liver injury (DF) are all associated with significant
Our data indicate that when patients with moderate effects on survival. Analysis for interactions showed that
or very severe AH are carefully screened for the presence treatment, combined with the level of malnutrition, also
of PCM using a global approach to diagnosis, some had a significant interrelationship on survival (p =
evidence of nutritional abnormality is present in every 0.024).
patient. These findings closely compare with those on The interrelationships among these three variables
functional assessment for PCM (e.g., muscle grip on survival can best be seen after stratification of
strength). The mechanism for nutritional change before patients on the basis of PCM and food intake (Fig. 6).
hospitalization is more complex than poor food intake In patients with moderate PCM, the action of oxan-
because total calorie intake appears to be sufficient to drolone was seen to be most efficacious when it was used
maintain a normal state of nutrition (Table 2). However, with adequate food intake (group 1). In this population,
alcohol typically contributed approximately 50% of mortality was 4% (group 1>,vs. 28%in the corresponding
caloric intake, resulting in inadequate dietary protein placebo group (group 2). However, when inadequate
and micronutrient intake. In some instances it may calorie intake was present, oxandrolone (group3) did not
produce pathological conditions in other digestive significantly influence outcome compared with placebo
organs (e.g., pancreas). Hence many consider such (group 4). Thus the oxandrolone effect was closely linked
malnutrition a result of the initial alcohol injury rather to calorie intake. This seems reasonable because both
than a cause of the injury. This is difficult to prove. Once energy and protein building blocks (amino acids) are
present, however, the degree of malnutrition appears to essential factors if the increased anabolism associated
parallel outcome such that the more severe the PCM, the with oxandrolone is to be achieved.
poorer the prognosis. This does not prove a cause-and- In patients with severe malnutrition, the critical
effect relationship but establishes PCM as an important factor in survival was calorie intake. Inadequate calorie
factor in outcome. Furthermore, the level of baseline intake (groups 3 and 4) in the presence of severe
malnutrition (PCM) and the severity of liver disease malnutrition was associated with a high mortality
(DF) both appear to be related to ad libitum food intake (51%), whereas adequate intake (groups 1 and 2) was
during therapy. Thus those patients with the most associated with significantly lower mortality (19%)
severe liver disease and greatest degree of malnutrition (p = 0.002).
(the greatest need) have the most severe anorexia and Thus, as shown in Figure 6 and Table 7, achieving
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Vol. 17, No. 4,1993
HEPATOLOGY MENDENHALL ET AL. 575

adequate caloric intake is a significant factor associated William G. Henderson, Ph.D. (chief); Thomas E. Moritz, M.S.
with improved survival in patients with moderate PCM (study biostatistician); Jean Rowe (administrative officer);
(and oxandrolone administration) and severe PCM. Raslan 0.Othman, M.S. (programmer); and Helma R. Weyrich
On comparison of CS 119 and CS 275, the source of (statistical assistant).
Cooperative Studies Program Clinical Research Pharmacy
the calories (hospital diet vs. a high-branched-chain- Coordinating Center (Albuquerque, N2cI). Mike Sather, R.Ph.,
amino-acid supplement such as Hepatic Aid) did not M.S. (chief); Philip L. Day, R.Ph., M.S. (clinical research
appear to be critical. However, if one wishes to avoid pharmacist); Carol L. Fye, R.Ph., M.S. (clinical research
nasogastric and intravenous feeding, then better ad pharmacist); Louis Montano and Francis R. Chacon (produc-
libitum oral intake can be achieved with the addition of tion controllers); John D. Recio, Jr. (research technologist);
a liquid food supplement (e.g., Hepatic Aid). In CS 275, and Sandra Buchanan (administrative assistant).
in which the supplement was employed, adequate intake Psychosocial Testing (Cincinnati, OH). Juris Mezinskis,
( 2 2,500 kcal/day) was achieved in 62% (36 of 58) of the Ph.D.
cases with moderate PCM and 43% (26 of 60) of those Pathology Laboratory. Antonio Chedid, M.D.; Thomas
with severe PCM. Success in achieving adequate intake Chen, M.D.; and Samuel French, M.D.
Central Chemistry Laboratory (Cincinnati, OH). Gary A.
without a high-calorie food supplement (CS 119) was Roselle, M.D.; Susan Rouster, B.S.; and Robin Scott.
significantly less satisfactory, with adequate intake in Special Membrane Laboratory (Minneapolis, MN). Gibson
only 10% (6 of 60) of those with moderate PCM and 17% Wood, Ph.D.
(10 of 59) of those with severe PCM (p = < 0.0001). Plasma Amino Acid Assays Laboratory (Lexington,
Furthermore, when the high-calorie food supplement in KY). Craig J. McClain, M.D.; Jack Schmidt, Ph.D.; and Nancy
CS 275 was discontinued at 3 mo, a decline was seen in Meyer.
many of the nutritional parameters. Whether survival Data Monitoring Board. Carroll Leevy, M.D. (chairman);
and PCM status could have been significantly improved Willis Crocker Maddrey, M.D.; Esteban Mezey, M.D.; and
in those patients consuming inadequate calories by Richard G . Cornell, Ph.D.
forced hyperalimentation and extended duration of Human Rights Committee. Eileen Collins, R.N. (chair-
treatment remains t o be determined. woman); Eileen Hagarty, R.N., M.S.; Lauren Lawson, Ph.D.
(past chairwoman); Rev. Thomas S. Burris, Sr.; Elizabeth M.
We conclude that PCM is a critical complication in this Butler; Walter Dorus, M.D.; Mary Ann Emanuele, M.D.; Rev.
patient population. The initial diagnostic evaluation Martin W. Feldbush; Donna Franklin, Ph.D.; Barbara Harvey;
should include a multifaceted approach to nutritional Robert E. Lee; Thomas M. Schmid, Ph.D.; Rev. Jeffrey
assessment, as outlined by Blackburn (21). In all Stinehelfer; William Boyce Upholt, Ph.D.; and James Wil-
malnourished patients, vigorous nutritional therapy liams, Ph.D.
should be promptly initiated. When an appropriate Cooperative Studies Central Administration (Department
nutritional status is present, oxandrolone should be of Veterans Af3cairs Central Ofice, Washington, D.C.). Daniel
added to the regimen to obtain optimal results. Deykin, M.D. (chief); Ping C. Huang, Ph.D. (staff assistant);
and Janet Gold (administrative officer).
Acknowledgments We thank the following companies for their gracious
Chairman’s Ofice (Cincinnati, OH). Charles L. Men- contributions to the successful completion of this study:
denhall, M.D., Ph.D. (chairman); Deborah Day (technical McGaw, Inc. (Irvine, CA) for Hepatic Aid 11; G.D. Searle & Co.
assistant); Novie E. Fricke (past administrative assistant); (Chicago, IL) for oxandrolone and matching placebo; Merck
Lela Rittmeier and Marie Lucente (secretary). Sharpe & Dohme (West Point, PA) for the recombinant
Executive Committee. Charles L. Mendenhall, M.D., Ph.D. hepatitis B vaccine; and Grand Forks Human Nutrition Center
(chairman); Samuel French, M.D.; Craig J. McClain, M.D.; (Grand Forks, ND) for the GRAND Food Description Master
Thomas E. Moritz, M.S.; Gary A. Roselle, M.D.; Eugene Schiff, Coding Manual and nutrient conversion table dataset.
M.D.; Philip L. Day, R.Ph., M.S. (adhoc member); and William
G. Henderson, Ph.D. (ex officzo).
REFERENCES
Participants. (Including DVAMCs, participating investi-
gators, nurses and dietitians). Cincinnati, OH: R.E. Weesner, 1. Patek A J Jr, Post J. Treatment of cirrhosis of the liver by a
M.D.; D.V. Trulzsch, M.D. (assisting from Dayton, OH); Ellen nutritious diet and supplements rich in vitamin B complex. J Clin
Pinnow, R.N.; Judy Nelson, R.N.; Tanya Jolliffe, R.D., L.D.; Invest 1941;20:481-505.
and Jean Streicher, R.D. East Orange, NJ: Arun Samanta, 2. Patek AJ Jr, Post J , Ralnoff OD, Mankin H, Hillman RW. Dietary
treatment of cirrhosis of the liver. JAMA 1948;139:543-549.
M.D.; Denise Small, R.N.; and Lynn Schellhase, R.D. Hines, 3. Nasrallah SM, Galambros JT. Amino acid therapy of alcoholic
IL: Bernard Nemchausky, M.D.; Aida Ranchero, R.N.; Linda hepatitis. Lancet 1980;2:1276-1277.
Kroker, R.N., M.S.; and Rosemary Robinson, M.S., R.D. 4. Rudman D, Kutner M, Ansley J , Jansen R, Chipponi J , Bain RP.
Lexington, KY: Luis Marsano, M.D.; Linas J. Adams, M.D.; Hypotyrosinemia, hypocystinemia, and failure to retain nitrogen
Deborah Dyer, R.N.; Michele Parker, R.N.; and Renee Phillips, during total parenteral nutrition of cirrhotic patients. Gastroen-
R.D. Long Beach, CA: Tim Morgan, M.D.; Shannon Jackson, terology 1981;81:1025-1035.
R.N.; Monica Cordry, R.N.; and Rene Haas, R.D. Louisville, 5. Smith J , Horowitz J , Henderson M, Heymsfield S. Enteral
KY: Carlo Tamburro, M.D.; Barbara Miller, R.N.; and Diane hyperalimentation in undernourished patients with cirrhosis and
Kinny, R.D. Miami, FL: Eugene Schiff, M.D.; Karen Nordt, ascites. Am J Clin Nutr 1982;35:56-72.
6. Keohane PP, Attrill H, Grimble G , Spiller R, Frost P, Silk DBA.
R.N.; Dorothy Hancock, R.D.; and Gail A. Obenauff, R.D. Enteral nutrition in malnourished patients with hepatic cirrhosis
Minneapolis, MN: John Allen, M.D.; Judi Causier, R.N.; and acute encephalopathy. JPEN 1983;7:346-350.
Patricia Welo, R.N.; and Robin Bougie, R.D. 7. McGhee A, Henderson M, Millikin WJ Jr., Bleier JC, Vogel R,
Biostatistics and Research Data Processing. (Hines Coop- Kassouny M, Rudman D. Comparison of the effects of hepatic-aid
erative Studies Program Coordinating Center; Hines, IL). and a casein modular diet on encephalopathy, plasma amino acids,
 15273350, 1993, 4, Downloaded from https://aasldpubs.onlinelibrary.wiley.com/doi/10.1002/hep.1840170407 by Cochrane Philippines, Wiley Online Library on [16/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
576 MENDENHALL ET AL. HEPATOLOGY
April 1993

and nitrogen balance in cirrhotic patients. Ann Surg 1983;197: T, Seeff L, Sorrel1 M, et al. Short-term and long-term survival in
288-293. patients treated with oxandrolone and prednisolone. N Engl J Med
8. Horst D, Grace ND, Conn HO, Schiff E, Schenker S, Viteri A, Law 1984;311:1464-1470.
D, et al. Comparison of dietary protein with an oral, branched 19. Mendenhall CL. Alcoholic hepatitis. In: Schiff L, Schiff ER, eds.
chain-enriched amino acid supplement in chronic portal-systemic Diseases of the liver. 6th ed. Philadelphia: J.B. Lippincott Co.,
encephalopathy: a randomized controlled trial. HEPATOLOCY 1984; 1987:669-685.
4279-287. 20. Maddrey WC, Boitnott JK, Bedine MS, Weber FL Jr, Mezey E,
9. Mendenhall C, Bongiovanni G, Goldberg S, Miller B, Moore J , White RI J r . Corticosteroid therapy of alcoholic hepatitis. Gastro-
Rouster S, Schneider D, et al. VA cooperative study on alcoholic enterology 1978;75:193-199.
hepatitis 111: changes in protein-calorie malnutrition associated 21. Blackburn GL, Bistrain BR, Maini BS, Schlamm HT, Smith MF.
with 30 days of hospitalization with and without enteral nutri- Nutritional and metabolic assessment of the hospitalized patient.
tional therapy. JPEN 1985;9:590-596. JPEN 19773: 11-22.
10. Christie ML, Sack DM, Pomposelli J , Horst D. Enriched branched- 22. Mendenhall CL, Tosch T, Weesner RE, Garcia-Pont P, Goldberg
chain amino acid formulaversus a casein-based supplement in the SJ, Kiernan T, Seeff LB et al. VA cooperative study on alcoholic
treatment of cirrhosis. JPEN 1985;9:671-678. hepatitis 11: prognostic significance of protein-calorie malnu-
11. Diehl AM, Boitnott JK, Herlong F, Potter JJ, Van Duyn MA, trition. Am J Clin Nutr 1986;43:213-218.
Chandler E, Mezey E. Effects of parenteral amino acid supple- 23. Chedid A, Mendenhall CL, Gartside P, French SW, Chen T, Rabin
mentation in alcoholic hepatitis. HEPATOLOGY 1985;5:57-63. L, and the VA Cooperative Study Group. Prognostic factors in
12. O h t a M, Watanabe A, Nagashima H. Nutritional treatment of alcoholic liver disease. Am J Gastorenterol 1991;86:210-216.
liver cirrhosis by branched-chain amino acid-enriched nutrient 24. Mendenhall CL, Anderson S,Weesner RE, Goldberg SJ, Crolic KA.
mixture. J Nutr Sci Vitamin01 1985;31:291-303. Protein-calorie malnutrition associated with alcoholic hepatitis.
13. Cabre E, Gonzalez-Huix F, Abad-Lacruz A, Esteve M, Acero D, Am J Med 76;211-222, 1984.
Fernandez-Bariares, Xi01 X, e t al. Effect of enteral nutrition on the 25. Kellor M, Kondrasuk R, Iversen I, Frost J , Silberberg N, Hoglund
short-term outcome of severely malnourished cirrhotics: a ran- M. Sister Kenny publication number 721: technical manual: hand
domized control trial. Gastroenterology 1990;98:715-720. strength and dexterity tests. Minneapolis, MN: Sister Kenny
14. Naveau S, Pelletier G, Poynard T, Attali P, Poitrine A, Buffet C, Institute, 1977:14-19.
Etienne JP, et al. A randomized clinical trial of supplementary 26. Zimmerman HJ. Hormonal derivatives and other drugs used to
parenteral nutrition in jaundiced alcoholic cirrhotic patients. treat endocrine disease. In: Zimmerman HJ, ed. Hepatoxicity: the
HEPATOLOGY 1986;6:270-274. adverse effects of drugs and other chemicals on the liver. New
15. Achord JL. A prospective randomized clinical trial of peripheral York: Appleton-Century-Crofts, 1978:436-444.
amino acid-glucose supplementation in acute alcoholic hepatitis. 27. McClain CJ, Cohen DA. Increased tumor necrosis factor pro-
Am J Gastroenterol 1987;82:871-875. duction by monocytes in alcoholic hepatitis. HEPATOLOGY 1989;9:
16. Bonkovsky H, Jafri I, Singh R, Cotsonis G, Slaker D. Treatment 349-351.
of alcoholic hepatitis with parenteral nutrition and oxandrolone: 28. Felver FE, Mezey E, McGuire M, Mitchell MC, Herlong HF, Veech
a randomized controlled trial. 11. Effects on nitrogen metabolism GA, Veech RL. Plasma tumor necrosis factor alpha predicts
[Abstract]. HEPATOLOCY 1990;12(suppl):978. decreased long-term survival in severe alcoholic hepatitis. Alcohol
17. Bonkovsky H, Fiellin D, Smith G, Slaker D, Simon D, Galambos Clin Exp Res 1990;14:225-259.
J. Treatment of alcoholic hepatitis with parenteral nutrition and 29. John WJ, Phillips R, Ott I, Adams W, McClain CJ. Resting energy
oxandrolone: a randomized controlled trial. I. Short-term effects expenditure in patients with alcoholic hepatitis. JPEN 1989;13:
on liver function [Abstract]. HEPATOLOGY 1990;12(supp1):870. 124-127.
18. Mendenhall CL, Anderson S, Garcia-Pont P, Goldberg S, Kiernan