Antioxidants versus corticosteroids in the treatment of severe

alcoholic hepatitis—A randomised clinical trial*

Martin Phillips, Howard Curtis, Bernard Portmann, Nora Donaldson,

Adrian Bomford, John O’Grady*

Institute of Liver Studies, King’s College Hospital, Denmark Hill, London SE5 9RS, UK

0168-8278/$32.00 q 2005 Published by Elsevier B.V. o

doi:10.1016/j.jhep.2005.11.039

Received 24 August 2005; received in revised form 2

accepted 18 November 2005; available online 20 Decem* The authors who have taken part in this study declar

not a relationship with the manufacturers of the drugs inv

past or present and did not receive funding from the manu

out their research.* Corresponding author. Tel.: C44 207 346 3252.

E-mail address: john.o’grady@kcl.ac.uk (J. O’Grady

See Editorial, pages 633–636

Background/Aims: Severe alcoholic hepatitis is associated with high morbidity and short-term mortality.

Corticosteroids are the only widely used therapy but established contraindications to treatment or the risk of serious

side-effects limit their use. The perceived need for alternative treatments together with the theoretical benefits of anti-

oxidant therapy triggered the design of a randomised clinical trial comparing these treatment modalities.

Methods: One hundred and one patients were randomized into a clinical trial of corticosteroids or a novel antioxidant

cocktail with a primary endpoint of 30-day mortality.

Results: At 30 days there were 16 deaths (30%) in the corticosteroid treated group compared with 22 deaths (46%) in

the antioxidant treated group (PZ0.05). The odds of dying by 30 days were 2.4 greater for patients on antioxidants

(95% confidence interval 1.0–5.6). A diagnosis of sepsis was made more frequently in the AO group (PZ0.05), although

microbiologically proven episodes of infection occurred more often in the CS group (P!0.01). The survival advantage

for corticosteroid treated patients was lost at 1 year of follow-up (PZ0.43).

Conclusions: This study has shown that corticosteroids in the form of prednisolone 30 mg daily are superior to a

broad antioxidant cocktail in the treatment of severe alcoholic hepatitis.

q 2005 Published by Elsevier B.V. on behalf of the European Association for the Study of the Liver.

Keywords: Acute alcoholic hepatitis; Corticosteroids; Antioxidants; Sepsis; Renal failure

1. Introduction

Acute Alcoholic Hepatitis (AAH) is the most aggressive

form of alcoholic liver disease with a short-term mortality as

high as 65%, rising to 95% when complicated by renal

failure [1]. Unlike other causes of liver failure, liver transplan-

tation is not considered an option for these patients [2]. Most

treatment modalities that have been studied in clinical trials in

AAH have disappointed, including colchicine, propylthiouracil,

n behalf of the European

4 October 2005;

ber 2005

ed that they have

olved either in the

facturers to carry

).

insulin and glucagon, anabolic steroids, and penicillamine [3–7].

Pentoxifylline, an inhibitor of TNF-a production, was associated

with a significant short-term survival advantage in one study [8].

The most widely used therapy for AAH is corticosteroids

and 14 randomized trials have been reported [6,9–21]. Five

reported short-term survival benefit from corticosteroids but the

largest trial was not confirmatory [6]. A similar conclusion is

reached by meta-analyses (four published in full [22–25] and a

fifth in abstract form [26]), The majority of studies excluded

patients with sepsis, recent gastrointestinal bleeding, renal

failure and pancreatitis, leading to concern that the results were

not necessarily generalisable to patients with severe AAH.

Concern over the safety of corticosteroids has tempered their

widespread adoption in AAH despite the supportive data.

Enthusiasm for the use of antioxidants in AAH is driven

by the belief that oxidative stress plays a central role in the

pathogenesis of AAH [27]. Oxidative stress is intimately

Journal of Hepatology 44 (2006) 784–790

www.elsevier.com/locate/jhep

Association for the Study of the Liver.

M. Phillips et al. / Journal of Hepatology 44 (2006) 784–790 785

linked to pro-inflammatory cytokines, while reactive

oxygen species activate the nuclear transcription factor

NF-kB which in turn switches on pro-inflammatory

cytokine production including interleukin-1, interleukin-6

and tumour necrosis factor-a (TNF-a) [28]. These cytokines

are overproduced in AAH and plasma and hepatic levels are

proportional to the severity of disease [29–33].

One trial of antioxidants in AAH studied 56 patients with

moderate or severe disease in a placebo controlled trial of

vitamin E 600 mg, selenium 200 mg, and zinc 12 mg [33].

Mortality was 6.5% in the antioxidant group compared to

40% in the placebo group. However, another study of 51

patients with mild to moderate AAH found no clinical

benefit with vitamin E 1000 IUs daily [34]. Some benefit

from N-acetylcysteine was detected with hepatorenal

syndrome including patients with AAH [35].

This study tested the hypothesis that aggressive

antioxidant therapy had a superior risk/benefit profile than

corticosteroids in severe AAH resulting in improved 30-day

survival rates.

Table 1

Antioxidant regimen (total daily doses)

Antioxidant Route Dose

b-Carotene Oral 12 (mg)

Vitamin C (ascorbic acid) Oral 1100 (mg)

Vitamin E (D-a-tocopherol) Oral 388 (mg)

Selenium Oral 300 (mg)

Methionine Oral 1600 (mg)

Allopurinol Oral 300 (mg)

Desferrioxamine Parenteral 30 (mg/kg)

N-Acetylcysteine Intravenous 150 (mg/kg)

2. Methods

2.1. Patients

The study protocol conforms to the ethical guidelines of the 1975Declaration of Helsinki and was approved by the hospital’s ethicscommittee. Recruitment to the study commenced in 1997 and 1-yearfollow-up of all patients was completed in 2002. Written informed assent orconsent was obtained from all patients (or their next of kin inencephalopathic or intubated and sedated patients). All patients had ahistory of heavy alcohol consumption defined as greater than 80 g alcoholper day for men, or greater than 60 g alcohol per day for women, prior to theonset of illness of at least 1-month duration. In such patients featuresconsidered consistent with a diagnosis of severe AAH were:

(a) absence of alternative aetiology of liver disease,(b) serum bilirubin O100 mmol/L,(c) serum AST !300 IU/L,(d) serum IgA O5 g/L,(e) white cell count O20!109/L,(f) ultrasound evidence of hepatic fatty infiltration,(g) hepatomegaly (clinical or radiological).

At the time of assessment, the presence of all of the first three featuresplus one of the remaining features was considered compatible with thediagnosis of severe AAH. Histological confirmation of AAH was not a pre-requisite for entry to the trial as it was considered that this would delayrandomisation and bias the study population to patients with less severedisease. Nevertheless, every effort was made to obtain histologicalconfirmation during or after participation in the study.

Exclusion criteria were as follows:

(1) Active sepsis defined as positive microbiological culture, ascitic whitecell count O500 cells/mL (or ascitic neutrophil count O250 cells/mL),or radiological appearances consistent with pneumonia. Patients wereeligible for trial entry if treated with appropriate antibiotics for at least48 h prior to randomisation.

(2) Active significant gastrointestinal haemorrhage within the previous 48 hdefined a requirement of R4 unit blood transfusion or haemodynamicinstability (pulseO120/min or systolic blood pressure !100 mmHg).

(3) Shock necessitating inotropic support.(4) Evidence of coexisting non-alcoholic liver disease.(5) Pregnant or lactating women.(6) Patients with a history of allergy to any component of the regimen.

(7) Previous randomisation to the study.(8) Previous treatment with one of the trial drugs (within 12 months).(9) Patients clearly improving spontaneously prior to trial entry.

Patients were stratified according to renal function at randomisation.Patients were divided into two groups—serum creatinine !200 mmol/L,and serum creatinine O200 mmol/L. Severity of AAH was assessed usingthe discriminant function (DF) first described by Maddrey DFZ4.6(prologation in prothrombin time in seconds)Cbilirubin [(mmol/L)/17)]with a DF O32 indicative of severe AAH.

All patients were treated with intravenous vitamin K 10 mg daily,intravenous multivitamins (Pabrinexw) once daily for 3 days and oralthiamine 300 mg daily. Nutritional status was assessed and energy andnitrogen requirements estimated using the Schofield equation [35,36].Nutritional supplements were administered as sip feeds or nasogastricenteral feeding.

2.2. Treatment regimens

Eligible patients were randomised to either corticosteroid therapy (CS)or antioxidant therapy (AO) as soon as possible, and preferably within 72 hof admission, using consecutive envelopes produced by computer-generated randomisation. CS therapy was given as oral prednisolone30 mg once daily or methylprednisolone (24 mg) intravenously. Treatmentwas given for 28 days and then tapered over a 2 weeks period. The AOregimen administered is shown in Table 1 and this combination was givenfor 28 days. In addition, this group received 500 ml of 10% intralipid every5 days as a membrane stabiliser.

2.3. Plasma samples

Blood was drawn prior to study drug therapy and at weekly intervals upto 28 days in the fasting state. Blood samples were taken in EDTA tubes andlithium–heparin tubes and cooled immediately on ice, centrifuged at3500 rpm for 10 min. Plasma was separated into three aliquots and frozenimmediately at K80 8C until assay. Vitamins A and E levels were measuredby high performance liquid chromatography [37]. Selenium levels weremeasured by atomic absorption spectrophotometry.

2.4. Endpoints

The primary endpoint was 30 day survival. The secondary endpoint was1-year survival.

2.5. Statistical analysis

The trial set out to detect, with 80% power and 5% significance level, achange in the 30-day mortality from 50% in the CS group to 30% in the AOgroup. A conventional analysis with logistic regression required a samplesize of 93 patients per group. A Cox’s regression model fitted to the times todeath was also used to compare the efficacy of the two treatments.

Sequential monitoring was also planned for this trial [38]. The powerrequirement of the sequential design yielded an expected sample size of 122patients at termination. An interim analysis was triggered by clinicalconcern after a total of 101 patients were recruited. The accumulating

Table 2

Baseline clinical characteristics of patients at randomisation

Antioxidants (48) Corticosteroids

(53)

P

Male 32 (67%) 28 (53%) 0.16

Age (years) 45 (25–71) 43 (26–76) 0.66

Days to

randomisation

3 (1–13) 3 (0–13) 0.26

Hepatic 12 (26%) 17 (32%) 0.43

M. Phillips et al. / Journal of Hepatology 44 (2006) 784–790786

evidence of the advantage of AO was summarized in terms of the oddsratio. The triangular design was used for the sequential monitoring to detectno difference between the treatments. The estimation of the odds ratios andtheir significance is adjusted for the sequential nature of the trial.

The computer package SPSS v.10 is used for the computations of theconventional methods and derivation of the stopping boundaries andadjustments to account for the sequential analyses were obtained using thecomputer package PEST3, developed by the Planning and Evaluation ofSequential Trials project at the University of Reading [39]. Survivalanalyses were performed on an ‘intention to treat’ (ITT) and ‘per protocol’bases.

encephalopathy

Recent GI

haemorrhage

5 (10%) 12 (23%) 0.10

Recent sepsis 2 (4%) 4 (8%) 0.68

3. Results Renal impairment 9 (19%) 13 (25%) 0.48

(CreatinineO200mmol/L)

Values shown as median (range), or value (%).

3.1. Patient characteristics

At the time of analysis 213 patients had been referred for

consideration of trial entry, 112 were excluded and 101

patients were enrolled with 48 patients (32 men and 16

women) randomised to AO and 53 patients (28 men and 25

women) to CS (Fig. 1). Demographic and clinical

characteristics were similar for both treatment groups at

randomisation (Table 2).

Liver histology was available for 29 (60%) of the AO

group and 36 (68%) of the CS group and cirrhosis was

present in 94%. Median time to biopsy was 14 days after

randomisation (7.5 days in the AO group and 17 days in

Assessed for eligibility (213)

Excluded (112)

not AAH (62)

AAH (44)no consent (1)spontaneous improvement (6)pre-treatment with trialdrug (14)other exclusions (23)

Randomised (101)

Allocated AO (48) Allocated CS (53)Received AO (46) Received CS (53)Withdrew consent (2)

Discontinued AO; Discontinued CS;withdrew consent (4) withdrew consent (1)painful phlebitis (2) psychosis (1)protocol violation (2) sepsis (6)

protocol violation (1)lost to follow up (1)

Analysed (n=48) Analysed (n=53)

Fig. 1. Trial profile.

the CS group). 53 patients (84%) were biopsied within

60 days of randomisation AAH was confirmed in 52

patients (98%) and was histologically severe in 35

patients (66%). Many biopsies showed evidence of mild

or moderate iron deposition (grades 1 and 2) on Perls

stain, but two had grade 3 or 4 siderosis.

Median duration of treatment was 14 days for AO

treated patients and 18 days for CS treated patients.

Treatment was completed (either 28 days of therapy or a

clinical response justifying premature termination of

therapy) in 13 patients on AO therapy and 33 patients

on CS therapy. Treatment was prematurely discontinued

in three patients (two patients on AO and one patient on

CS). Three patients were withdrawn from the study

because of protocol violation (two patients in the AO

group had positive hepatitis C serology and one patient in

the CS group had co-existing carcinoma of the head of the

pancreas). In the AO group two patients were intolerant

of treatment due to painful phlebitis. In the CS group one

patient developed severe psychosis on day 9 after

randomisation triggering withdrawal of therapy. In six

other patients on CS treatment was withdrawn due to

clinical sepsis, four within 4 days of randomisation and

Table 3

Laboratory parameters at randomisation

Antioxidants

(48)

Corticosteroids

(53)

P

Haemoglobin (g/dL) 10.5 (7.3–14.3) 10.2 (7.0–13.7) 0.68

White cell count

(!109/mL)

14.9 (4.0–47.0) 14.3 (3.9–46.5) 0.97

Platelets (!109/mL) 150 (36–524) 127 (45–354) 0.45

INR 1.63 (1.00–2.90) 1.51 (1.00–3.64) 0.20

Creatinine (mmol/L) 98 (55–522) 109 (46–1774) 0.36

Albumin (g/dL) 26 (15–35) 27 (15–38) 0.07

Bilirubin (mmol/L) 422 (106–744) 473 (104–786) 0.08

AST (IU/dL) 104 (43–345) 101 (45–346) 0.59

Child-pugh score 12 (7–15) 12 (8–15) 0.97

Discriminant function 61.1 (7.3–163.5) 60.7 (14.7–199.6) 0.13

Values shown as median (range).

Table 4

Stratified analysis and estimation of the 30 days survival (adjusted for sequential monitoring and stratified for renal failure)

Group Patient number Deaths Mortality rate Odds ratio 95% confidence

interval

P

Steroids 53 16 0.27

Antioxidants 48 22 0.47 2.4 (1.0, 5.6) 0.05

Total 101 38

0.60.70.80.91.0

urv

ival

p=0.05

M. Phillips et al. / Journal of Hepatology 44 (2006) 784–790 787

two after 16 and 19 days of treatment. Eight patients

withdrew consent after randomisation.

3.2. Laboratory parameters

Laboratory parameters at randomisation are shown in

Table 3. There were no statistically significant differences

between the two groups. By day 7, the median serum bilirubin

had fallen to 349 mmol/l in the AO group and 321 mmol/l in

the CS group. This early fall in serum bilirubin was seen in

71% of the AO group and 66% of the CS group. Serial

evaluation of the laboratory parameters at 7 day intervals

revealed no differences between the two groups with

the exception of serum albumin at day 7 (AO 25 vs CS

28 g/l, PZ0.01) and white cell count at day 28 (AO 12.2 vs

17.4, PZ0.05).

3.3. Morbidity and survival

At 30 days there were 22 deaths (46%) in the AO group

and 16 deaths (30%) in the CS group (PZ0.05). The odds

of dying by 30 days were 2.4 greater for patients on AO

(95% confidence interval 1.0–5.6) (Table 4). The sample

path of the triangular design crossed the boundary towards

no treatment difference. AO did not prove to be superior to

CS. Cox’s regression analysis adjusted for renal failure

and gender (but unadjusted for the sequential analysis)

showed that the most important determinants of 30 day

mortality were baseline bilirubin, white cell count, and the

presence of encephalopathy (Table 5). After adjusting for

these variables, the treatment difference remains signifi-

cant with the hazard of dying within 30 days whilst on AO

therapy being over three times that for patients treated

with CS (PZ0.003; hazard ratio 3.22, 95% CI, 1.48–7.03).

The Kaplan–Meier survival curves assessing 30 day

Table 5

Cox regression analysis (stratified by renal function) for 30 day

survival

Hazard ratio CI P

Age 1.02 0.99–1.05 0.18

Male sex 1.98 0.93–4.19 0.08

INR 1.77 0.89–3.49 0.10

Bilirubin 1.0 1.0–1.01 0.03

Encephalopathy 2.53 1.13–5.66 0.02

WCC 1.05 1.01–1.10 0.03

Treatment 3.22 1.48–7.03 0.003

mortality and utilising an ITT analysis are shown in

Fig. 2. Similar data based on a ‘per protocol’ analysis

showed greater divergence of the curves with 41.2%

survival in the AO group at 30 days (14 patients ‘at risk’)

as compared with 76.7% in the CS group (33 patients ‘at

risk’).

Causes of death are outlined in Table 6 and there was no

significant difference between the groups (PZ0.96). Eleven

patients had a total of 15 gastrointestinal bleeds during the

study but varices were the site of haemorrhage in only four

patients (two in each group) and all were managed with

band ligation.

Renal stratification divided the group so that nine

patients (19%) in the AO group had renal impairment at

randomisation, as compared to 13 patients (25%) in the CS

group. The 30 day survival in patients who did not exhibit

renal impairment at the time of randomisation was 62% in

the AO group and 80% in the CS group (PZ0.07). Five of

these patients in the AO group and eight patients in the CS

group received renal replacement therapy. Renal failure

developed during the study period in 17 patients, 11

treated with AO and six treated with CS and all but one of

these patients died. In total, four patients received

terlipressin for hepatorenal syndrome (AO 1, CS 3). At

30 days only two patients in the AO group were still alive,

whereas five patients survived in the CS group (PZ0.42).

There were 108 episodes of sepsis within the first

30 days of randomisation in 59 patients (Table 9).

0 5 10 15 20 25 300.00.10.20.30.40.5

Antioxidants

Corticosteroids

Days since randomization

cum

ula

tive

s

No. at risk 53 50 49 46 42 39 37 (Survival 69.8%)48 45 38 35 29 27 26 (Survival 54.2%)

Fig. 2. 30 day survival of patients with severe alcoholic hepatitis treated

with antioxidants or corticosteroids.

Table 6

Causes of death

Cause of death Antioxidants Corticosteroids

Deaths within 30 days

Number of deaths 22 16

Sepsis 11 7

Hepatorenal

syndrome

7 5 PZ0.96

GI haemorrhage 3 3

Progressive liver

failure

1 1

Deaths between 30 days and 1 year

Number of deaths 7 13

Sepsis 2 3

Hepatorenal

syndrome

0 2 PZ0.353

GI haemorrhage 1 2

Progressive liver

failure

4 2

Other 0 4

0 60 120 180 240 300 3600.00.10.2

0.30.4

0.50.60.7

0.80.9

1.0 Antioxidants

Corticosteroids

p=0.43

Days since randomisation

Cu

mu

lati

ve S

urv

ival

No at risk 53 37 32 28 26 26 26 26 (Survival 49%)48 26 24 22 21 20 19 19 (Survival 40%)

Fig. 3. 1-year survival of patients with severe alcoholic hepatitis treated

with antioxidants or corticosteroids.

M. Phillips et al. / Journal of Hepatology 44 (2006) 784–790788

A diagnosis of sepsis was made more frequently in the AO

group but microbiologically proven episodes of infection

occurred more often in the CS group. A diagnosis of sepsis

without microbiological confirmation was mainly due to

cellulites, suspicious chest radiology or spontaneous

bacterial peritonitis (SBP) diagnosed on the ascitic white

cell count.

At 1 year there had been 29 deaths (60%) in the AO

group and 29 deaths (55%) in the CS group (PZ0.43)

(Table 7). The sample path crossed the boundary of no

treatment difference in the triangular design. The odds of

dying before 1 year appear to be, on average, 40% greater

for the AO group. This indicates conclusive evidence of

no difference in the 1-year survival between AO and CS.

Cox’s regression analysis, adjusted for renal failure status

and gender (but unadjusted for the sequential analysis)

indicated the hazard of death by 1 year was three times

Table 7

Stratified analysis and estimation of the 1 year survival, adjusted for

sequential monitoring

Group Patients Deaths Mortality

rate

Odds

ratio

95%

confidence

interval

P

Steroids 53 27 0.51

Anti

oxidants

48 29 0.60 1.4 (0.6, 3.2) 0.43

Total 101 56

Table 8

Cox regression analysis (stratified by renal function) for 1 year survival

Hazard ratio CI P

Renal failure 3.0 1.7–5.3 !0.001

Treatment 1.5 0.9–2.5 0.15

greater for patients with renal impairment at randomisation

(Table 8). The Kaplan–Meier survival curves are presented

in Fig. 3. Six patients in each group resumed alcohol

consumption after discharge from hospital while 10 AO

patients and 12 CS patients remained abstinent. The

drinking pattern is unknown for seven patients (Table 9).

3.4. Antioxidant levels

Vitamins A and E, and selenium levels in plasma were

assayed in paired plasma samples from 38 patients, 19

patients from each treatment group, at the time of

randomisation and 7 days after commencement of

treatment. Baseline antioxidant levels were also compared

to 11 normal volunteer controls. Plasma vitamins A, E and

selenium levels were all significantly depressed in the trial

patients at randomisation (Table 10). At randomisation

there were no differences in antioxidant levels between the

groups. Plasma vitamin E levels, adjusted for lipid levels,

increased significantly after 7 days of AO treatment (p!0.001) to that of the normal controls, but were unchanged

Table 9

Episodes of sepsis during treatment

Antioxidants

(48)

Corticosteroids

(53)

PZ

Patients with sepsis 35 24 0.05

Deaths from sepsis 9 (31.4%) 12 (45.8%) 0.63

Episodes of sepsis 54 54

Microbiologically

proven

21 (38.9%) 42 (77.8%) !0.001

Site of microbiologically proven sepsis

Septicaemia 8 12

Pneumonia 4 5

SBP 3 10

Urine 4 4 0.492

Intravenous line 0 5

Stool 1 2

Other 1 4

Table 10

Plasma antioxidant levels at randomisation (day 0) and after 7 days

treatment (day 7)

Vitamin E/total

lipids

Vitamin A

(mmol/L)

Selenium

(mg/L)

AO group (nZ19)

Day 0 8.41 0.6 34.5

Day 7 16.21 0.55 78*

P!0.001 PZ0.296 P!0.001

CS group (nZ19)

Day 0 7.67 0.56 34

Day 7 9.08 1.15 47*

PZ0.904 P!0.001 PZ0.001

Controls (nZ11) 17.11** 2.78# 84#

Median values shown, *comparison of day 7 selenium levels P!0.001,

**PZ0.014 compared to trial patients, #P!0.001 compared to trial

patients.

M. Phillips et al. / Journal of Hepatology 44 (2006) 784–790 789

in the CS group (PZ0.904). Plasma vitamin A levels were

unchanged after 7 days of AO therapy but did increase

significantly in the CS group (P!0.001). Selenium levels

increased significantly in both groups after 7 days

treatment. However, day 7 levels were significantly higher

in the AO group where levels approached those of the

normal controls.

4. Discussion

The original hypothesis that AO therapy would result in a

survival advantage compared to CS therapy at 30 days was

not proven as the interim analysis showed a statistically

significant survival advantage in favour of CS and the odds

of dying by 30 days were 2.4 greater for patients on AO. The

difference was even more striking in the patients who

completed the intended course of treatment. On this basis

the trial was terminated. At 1 year of follow-up the survival

advantage of CS had been lost and this is in keeping with

previous reports that also found that the benefit of CS was

short-term and as such was useful only as a bridge to further

therapy e.g. liver transplantation.

The design of the current trial was intended to be both

pragmatic and inclusive of cases of severe AAH as

encountered in hospital practice. Although a double blind

study is ideal, it was not considered possible to develop an

intravenous placebo. Patients with features of severe disease

were deliberately not excluded except for very recent sepsis

and gastrointestinal bleeding. The severity of the disease is

also manifest in the Maddrey disciminant function scores

which are higher than in all other published trials in AAH

except one trial [16]. The decision to defer histological

verification of the diagnosis of AAH was intended to reflect

clinical reality and permit early initiation of therapy.

Ultimately, this approach was vindicated by the high

diagnostic accuracy seen in the cohort that was biopsied

and although not always assessed as severe, histological

changes of AAH are known to resolve rapidly in patients

treated with CS [40]. In addition, a subset analysis of this

cohort with histological confirmation of AAH was similar to

the findings in the overall study.

The design of this study reflected the belief that a single

agent, e.g. N-acetylcysteine was unlikely to be effective,

particularly as so many antioxidants have been shown to be

deficient in AAH. The combination selected was derived

from several sources and previous work including a cocktail

used in neonatal hemochromatosis [41]. Desferrioxamine is

very effective in ameliorating free radical-induced hepato-

cyte damage in experimental animal models [42]. Allopur-

inol has been shown to be a potent antioxidant and prevents

alcohol-induced liver injury in animal models [43].

Intralipid may be beneficial to the microcirculation by

inducing post-capillary venodilatation [44]. It is feasible

that some of the components of the AO therapy may have

had deleterious effects. Desferrioxamine may increase

susceptibility to infection [45], while vitamin A has been

linked to hepatotoxicity [46]. However, low doses of

vitamin A were used and plasma levels were not altered

by the AO treatment.

The alternative interpretation is that this study simply

confirms the beneficial effects of CS in severe AAH as

suggested by the earlier data. The reluctance to accept CS as

standard of care because of possibly aggravating co-existing

morbidity may be misplaced, either because the holistic

effect of therapy is superior or the actual risk of perceived

complications is lower than conventionally considered [47].

Acknowledgements

We acknowledge the support of Robin Hughes, Peter

Langley, Paul Cheeseman, Kishor Raja and Roy Sherwood,

amongst others.

References

[1] Hardison W, Lee F. Prognosis in acute liver disease of the alcoholic

patient. N Engl J Med 1966;275:61–66.

[2] McCullough AJ, O’Connor JF. Alcoholic liver disease: proposed

recommendations for the American College of Gastroenterology. Am

J Gastroenterol 1998;93:2022–2036.

[3] Akriviadis EA, Steindel H, Pinto PC, Fong TL, Kanel G,

Reynolds TB, et al. Failure of colchicine to improve short-term

survival in patients with alcoholic hepatitis. Gastroenterology 1990;

99:811–818.

[4] Halle P, Pare P, Kaptein E, Kanel G, Redeker AG, Reynolds TB.

Double-blind, controlled trial of propylthiouracil in patients with

severe acute alcoholic hepatitis. Gastroenterology 1982;82:925–931.

[5] Bird G, Lau JY, Koskinas J, Wicks C, Williams R. Insulin and

glucagon infusion in acute alcoholic hepatitis: a prospective

randomized controlled trial. Hepatology 1991;14:1097–1101.

[6] Mendenhall CL, Anderson S, Garcia-Point P, Goldberg S, Kiernan T,

Seeff LB, et al. Short-term and long-term survival in patients with

alcoholic hepatitis treated with oxandrolone and prednisolone. N Engl

J Med 1984;311:1464–1470.

M. Phillips et al. / Journal of Hepatology 44 (2006) 784–790790

[7] Resnick RH, Boitnott J, Iber FL, Makopour H, Cerda JJ. Preliminary

observations of d-penicillamine therapy in acute alcoholic liver

disease. Digestion 1974;11:257–265.

[8] Akriviadis E, Botla R, Briggs W, Han S, Reynolds T, Shakil O.

Pentoxifylline improves short-term survival in severe acute alcoholic

hepatitis: a double-blind, placebo-controlled trial. Gastroenterology

2000;119:1637–1648.

[9] Helman RA, Temko MH, Nye SW, Fallon HJ. Alcoholic hepatitis.

Natural history and evaluation of prednisolone therapy. Ann Intern

Med 1971;74:311–321.

[10] Porter HP, Simon FR, Pope II CE, Volwiler W, Fenster LF.

Corticosteroid therapy in severe alcoholic hepatitis. A double-blind

drug trial. N Engl J Med 1971;284:1350–1355.

[11] Campra JL, Hamlin Jr EM, Kirshbaum RJ, Olivier M, Redeker AG,

Reynolds TB. Prednisone therapy of acute alcoholic hepatitis. Report

of a controlled trial. Ann Intern Med 1973;79:625–631.

[12] Blitzer BL, Mutchnick MG, Joshi PH, Phillips MM, Fessel JM,

Conn HO. Adrenocorticosteroid therapy in alcoholic hepatitis. A

prospective, double-blind randomized study. Am J Dig Dis 1977;22:

477–484.

[13] Lesesne HR, Bozymski EM, Fallon HJ. Treatment of alcoholic

hepatitis with encephalopathy. Comparison of prednisolone with

caloric supplements. Gastroenterology 1978;74:169–173.

[14] Shumaker JB, Resnick RH, Galambos JT, Makopour H, Iber FL. A

controlled trial of 6-methylprednisolone in acute alcoholic hepatitis.

With a note on published results in encephalopathic patients. Am

J Gastroenterol 1978;69:443–449.

[15] Maddrey WC, Boitnott JK, Bedine MS, Weber Jr FL, Mezey E, White

Jr RI. Corticosteroid therapy of alcoholic hepatitis. Gastroenterology

1978;75:193–199.

[16] Depew W, Boyer T, Omata M, Redeker A, Reynolds T. Double-blind

controlled trial of prednisolone therapy in patients with severe acute

alcoholic hepatitis and spontaneous encephalopathy. Gastroenterol-

ogy 1980;78:524–529.

[17] Theodossi A, Eddleston AL, Williams R. Controlled trial of

methylprednisolone therapy in severe acute alcoholic hepatitis. Gut

1982;23:75–79.

[18] Bories P, Guedj JY, Mirouze D, Yousfi A, Michel H. Treatment of

acute alcoholic hepatitis with prednisolone, 45 patients. Presse Med

1987;16:769–772.

[19] Carithers Jr RL, Herlong HF, Diehl AM, Shaw EW, Combes B,

Fallon HJ. Methylprednisolone therapy in patients with severe

alcoholic hepatitis. A randomized multicenter trial. Ann Intern Med

1989;110:685–690.

[20] Ramond MJ, Poynard T, Rueff B, Mathurin P, Theodore C,

Chaput JC, et al. A randomized trial of prednisolone in patients

with severe alcoholic hepatitis. N Engl J Med 1992;326:507–512.

[21] Mathurin P, Duchatelle V, Ramond MJ, Degott C, Bedossa P,

Erlinger S, et al. Survival and prognostic factors in patients with

severe alcoholic hepatitis treated with prednisolone. Gastroenterology

1996;110:1847–1853.

[22] Imperiale TF, McCullough AJ. Do corticosteroids reduce mortality

from alcoholic hepatitis? A meta-analysis of the randomized trials

Ann Intern Med 1990;113:299–307.

[23] Daures JP, Peray P, Bories P, Blanc P, Yousfi A, Michel H, et al.

Corticoid therapy in the treatment of acute alcoholic hepatitis. Results

of a meta-analysis. Gastroenterol Clin Biol 1991;15:223–228.

[24] Christensen E, Gluud C. Glucocorticoids are ineffective in alcoholic

hepatitis: a meta-analysis adjusting for confounding variables. Gut

1995;37:113–118.

[25] Mathurin P, Mendendall CL, Carithers RL, Raymond MJ,

Maddrey WC, Garside P, et al. Corticosteroids improve short-term

survival in patients with severe alcoholic hepatitis (AH): individual

data analysis of the last three corticosteroids in severe AH. J Hepatol

2002;36:547–548.

[26] Poynard T. Corticosteroids reduce mortality from alcoholic hepatitis

in patients without encephalopathy. A meta-analysis of randomized

controlled trials (RCTs) including French trials. Hepatology 1991;14:

234A.

[27] Sun AY, Ingelman-Sundberg M, Neve E, Matsumoto H, Nishitani Y,

Minowa Y, et al. Ethanol and oxidative stress. Alcohol Clin Exp Res

2001;25:237S–2243.

[28] Hill DB, Brave S, Joshi-Barve S, McClain C. Increased monocyte

nuclear factor-kappaB activation and tumor necrosis factor production

in alcoholic hepatitis. J Lab Clin Med 2000;135:387–395.

[29] McClain CJ, Cohen DA, Dinarello CA, Cannon JG, Shedlofsky SI,

Kaplan AM. Serum interleukin-1 (IL-1) activity in alcoholic hepatitis.

Life Sci 1986;39:1479–1485.

[30] McClain CJ, Cohen DA. Increased tumor necrosis factor production

by monocytes in alcoholic hepatitis. Hepatology 1989;9:349–351.

[31] Sheron N, Bird G, Goka J, Alexander G, Williams R. Elevated plasma

interleukin-6 and increased severity and mortality in alcoholic

hepatitis. Clin Exp Immunol 1991;84:449–453.

[32] Hill DB, Marsano L, Cohen D, Allen J, Shedlofsky S, McClain CJ.

Increased plasma interleukin-6 concentrations in alcoholic hepatitis.

J Lab Clin Med 1992;119:547–552.

[33] Sheron N, Bird G, Koskinas J, Portmann B, Ceska M, Lindley I,

et al. Circulating and tissue levels of the neutrophil chemotaxin

interleukin-8 are elevated in severe acute alcoholic hepatitis, and

tissue levels correlate with neutrophil infiltration. Hepatology

1993;18:41–46.

[34] Wenzel G, Kuklinski B, Ruhlmann C, Ehrhardt D. Alcohol-induced

toxic hepatitis—a ‘free radical’ associated disease. Lowering fatality

by adjuvant antioxidant therapy. Z Gesamte Inn Med 1993;48:

490–496.

[35] Mezey E, Potter JJ, Rennie-Tankersley L, Caballeria J, Pares A. A

randomised placebo controlled trial of vitamin E for alcoholic

hepatitis. J Hepatol 2004;40:40–46.

[36] Holt S, Goodier D, Marley R, Patch D, Burroughs A, Fernando B,

et al. Improvement in renal function in hepatorenal syndrome with

N-acetylcysteine. Lancet 1999;353:294–295.

[37] Schofield WN. Predicting basal metabolic rate, new standards and

review of previous work. Hum Nutr Clin Nutr 1985;39:5–41.

[38] Elia M. Artificial nutritional support. Med Int 1990;82:3392–3396.

[39] Catignani GL, Bieri JG. Simultaneous determination of retinol and

alpha-tocopherol in serum or plasma by liquid chromatography. Clin

Chem 1983;29:708–712.

[40] Whitehead J. Planning and evaluation of sequential trials. 2nd ed.

Chichester: Wiley; 1994.

[41] Brunier H, Whitbread J. The PEST (planning and evaluation of

sequential trials). Reading, MA: Department of Applied Statistics,

University of Reading; 1996.

[42] Spahr L, Rubbia-Brandt L, Pugin J, Giostra E, Frossard JL, Borisch B,

et al. Rapid changes in alcoholic hepatitis histology under steroids:

correlation with soluble intercellular adhesion molecule-1 in hepatic

venous blood. J Hepatol 2001;35:582–589.

[43] Sigurdsson L, Reyes J, Kocoshis SA, Hansen TW, Rosh J,

Knisely AS. Neonatal hemochromatosis: outcomes of pharmacologic

and surgical therapies. J Pediatr Gastroenterol Nutr 1998;26:85–89.

[44] Shaw S, Jayatilleke E. The role of aldehyde oxidase in ethanol-

induced hepatic lipid peroxidation in the rat. Biochem J 1990;268:

579–583.

[45] Kono H, Rusyn I, Bradford BU, Connor HD, Mason RP,

Thurman RG. Allopurinol prevents early alcohol-induced liver injury

in rats. Pharmacol Exp Ther 2000;293:296–303.

[46] Holtzmann A, Schmidt H, Gebhardt MM, Martin E. Propofol-induced

alterations in the microcirculation of hamster striated muscle. Br

J Anaesth 1995;75:452–456.

[47] Conn HO, Poynard T. Corticosteroids and peptis ulcer: meta-anaysis

of adverse events during steroid therapy. J Int Med 1994;236:

619–632.