Nutritional therapy in alcoholic liver disease


Dr H. K. Seitz, Salem Medical Center, Zeppelinstrasse 11–33, D-69121 Heidelberg, Germany.


Chronic alcohol consumption may lead to primary and secondary malnutrition. In particular, protein energy malnutrition not only aggravates alcoholic liver disease but also correlates with impaired liver function and increased mortality. Therefore, in these patients, adequate nutritional support should be implemented in order to improve their prognosis. Clinical trials addressing this issue have shown that nutritional therapy either enterally or parenterally improves various aspects of malnutrition, and there is increasing evidence that it may also improve survival. Therefore, malnourished alcoholics should be administered a diet rich in carbohydrate- and protein-derived calories preferentially via the oral or enteral route. Micronutrient deficiencies typically encountered in alcoholics, such as for thiamine and folate, require specific supplementation. Patients with hepatic encephalopathy may be treated with branched-chain amino acids in order to achieve a positive nitrogen balance. Fatty liver represents the early stage of alcoholic liver disease, which is usually reversible with abstinence. Metadoxine appears to improve fatty liver but confirmatory studies are necessary. S-adenosyl-l-methionine may be helpful for patients with severe alcoholic liver damage, since various mechanisms of alcohol-related hepatotoxicity are counteracted with this essential methyl group donor, while a recent large trial showed that the use of polyenylphosphatidylcholine is of limited efficacy.


In industrialized countries, between 60 and 70% of the entire population above the age of 18 years consume alcohol. Of these, only a minority reveal patterns of pathologic alcohol intake leading to either acute or chronic physical, mental or social compromise. All over Europe, more than 45 million individuals display signs of alcohol-related organ damage, of which alcoholic liver disease comprises the largest group accounting for approximately 50% of all chronic liver diseases.1 Taken together, in Europe and the United States advanced alcoholic liver damage is responsible for more than 50 000 annual deaths due to cirrhosis and associated complications.2–5

The term alcoholic liver disease includes liver pathologies of various degrees due to direct and indirect effects of continuous alcohol ingestion. Hepatic manifestations of alcoholism comprise, in increasing severity, alcoholic fatty liver, alcoholic steatohepatitis, as well as alcohol-induced hepatic fibrosis and cirrhosis, either with or without inflammation.6, 7 Furthermore, chronic alcohol consumption in patients with liver cirrhosis is a risk factor for the development of hepatocellular carcinoma.8 Ninety to 100% of all heavy drinkers reveal fatty liver, but only a minority of these patients progress to more severe liver damage, such as alcoholic hepatitis and cirrhosis. Between 10 and 35% of alcoholics show more or less pronounced features of hepatic inflammation, and 10–20% eventually develop cirrhosis.9 The overall 5 year survival rate of alcoholic cirrhotics drops to 35% if they continue to drink, a figure which approaches that of many cancers. The high prevalence of alcohol consumption and the rising social, mental and financial challenge emphasizes the urgent need for prevention and therapeutic measures in these patients. While alcoholic fatty liver is usually reversible with abstinence, the more advanced forms of alcoholic liver disease usually require vigorous intervention. Therapeutic options for alcoholics with clinical signs of chronic liver disease include psychotherapy, social support, nutritional supplementation and, increasingly, pharmaceuticals that may counteract some of the pathogenic events in alcoholic liver disease.10 Furthermore, symptomatic treatment of complications is needed as soon as advanced liver disease is established. Finally, liver transplantation may be an option for selected patients who fulfil criteria of abstinence and compliance.11 However, nutritional therapy is by far the most extensively investigated treatment and has proved effective with regard to important clinical endpoints including nutritional status, rate of infections, liver function and survival.

The aim of the present review is to summarize the current knowledge on malnutrition in alcoholic liver disease and to emphasize the need for early nutritional intervention.

Pathogenesis of malnutrition in alcoholic liver disease

A large proportion of individuals with advanced alcoholic liver disease are malnourished and for a long time, it was believed that malnutrition was the cause rather than the result of liver damage in alcoholics. It is now undisputed that events associated with alcohol metabolism and its first metabolite, acetaldehyde, among other causes, play an important role in inducing the hepatic damage found in alcoholics.12 In heavy drinkers, the entire nutritional status is impaired due to primary and secondary malnutrition. For example, anorexia is a frequent cause of primary malnutrition in alcoholic liver disease, resulting from diminished food intake. Alcohol damages the mucosa of the upper gastrointestinal tract, and heavy alcohol drinking may lead to acute oesophagitis and gastritis with abdominal pain, nausea and vomiting.13 Furthermore, there is evidence that chronic alcohol intake influences several regulatory factors of appetite and inflammation, which may reduce the demand for eating. Thus, tumour necrosis factor-alpha and leptin are often up-regulated in alcoholics, both of which are proinflammatory and inhibit appetite and food intake.14–16 Another important factor contributing to malnutrition is maldigestion and malabsorbtion due to morphological and functional alterations of the intestinal mucosa.17, 18

Alcohol is not simply a psychotropic substance but is also a source of calories; 1 g of alcohol provides 7.1 kcal/g of energy, which is more than that of carbohydrates (4.1 kcal/g). While regular alcohol consumers who do not fulfil criteria of alcohol abuse are often overweight due to added calories from alcohol consumption, heavy drinkers often replace a substantial proportion of nutrient-derived calories by alcohol. The former phenomenon is termed ‘alcohol addition’ and usually leads to truncal obesity19, while the so-called ‘alcohol substitution’ results in weight loss and prominent protein energy malnutrition (reviewed in 20). The exocrine pancreatic function is also affected by chronic alcohol exposure, and impaired secretion of pancreatic enzymes with resulting malabsorbtion of fat and proteins commences long before clinical, biochemical or morphological signs of chronic pancreatic damage can be found.21

Chronic alcohol consumption may also lead to profound disturbances affecting the metabolism of numerous macro and micronutrients. With regard to the nutritional status of severe alcoholics, effects of alcohol on protein metabolism are crucial since patients with an inadequate dietary protein intake have a poor survival prognosis.22 Acute and chronic alcohol consumption causes impaired hepatic amino acid uptake and protein synthesis, such as that of lipoproteins, albumin and fibrinogen, reduced protein synthesis and secretion from the liver, and increased catabolism in the gut due to increased cell regeneration.23

Substrate utilization of lipids and carbohydrates in alcoholics is also deeply compromised, due to an excess of reductive equivalents (e.g. NADPH) and impaired oxidation of triglycerides with consecutive ‘trapping’ of fat in hepatocytes and increased peripheral triglyceride levels.24 Another frequent sequel is the early development of insulin resistance in alcoholic cirrhotics, which impairs glucose uptake into muscle cells. This leads to reduced glycogen production with consecutive lack of energy stores, while metabolically inefficient anaerobic lactate formation from glucose is preserved. In 15–37% of all alcoholic cirrhotics, insulin-dependent diabetes develops as an indicator of poor prognosis.25, 26

Uptake and metabolism of a wide array of micronutrients, including water- and fat-soluble vitamins as well as trace elements, are strikingly influenced by concomitant heavy alcohol consumption. Folate deficiency is frequent among alcoholics, and low serum folate and red blood cell folate levels can be found in up to 60% of heavy drinkers.27, 28 This is most probably due to the poor nutritional intake, impaired absorption18, altered storage and metabolic activation29 disturbance of the enterohepatic circulation of folate30, and increased urinary excretion.31 Pyridoxal-5′-phosphate, the biologically active coenzyme of vitamin B6, is frequently deficient in alcoholics28, a result of inadequate dietary intake but also of interactions between alcohol and pyridoxal-5′-phosphate metabolism. Thus, it has been suggested that acetaldehyde displaces protein-bound pyridoxal-5′-phosphate and thereby exposes the coenzyme to inactivating phosphatases which are up-regulated in alcoholics.32 In almost 80% of alcoholics, vitamin B1 (thiamin) levels are decreased, and alcoholism is the most important predictor of thiamine deficiency.33 The alcohol-induced Wernicke–Korsakoff syndrome, as well as the myocardial Beri-Beri disease, is caused by thiamine depletion. Clinical signs of Wernicke–Korsakoff syndrome can be considered a nutritional emergency with an immediate demand for high-dose thiamine supplementation.

A frequently discussed issue refers to the interaction between alcohol consumption and vitamin A (retinol). Chronic alcohol consumption affects several aspects of vitamin A metabolism, including decreased retinol absorbtion, enhanced degradation in the liver, and an increased mobilization of retinol from the liver to other organs.34, 35 Alcoholic liver disease is associated with severely decreased hepatic vitamin A levels, even when liver injury is only moderate and when blood levels of retinol-binding protein and pre-albumin are still unchanged.36 Clinically, this may lead to important consequences, such as night blindness in acute states of deficiency, to alterations in cell regeneration and differentiation, and even tumour development.8, 23 On the other hand, alcohol consumption may enhance vitamin A hepatotoxicity since the alcohol-related induction of the cytochrome P450 2E1 isoenzyme leads to the formation of hepatotoxic polar metabolites from retinoids.37 In addition, it has been shown in alcohol-fed rats that reduced hepatic retinoic acid levels may also contribute to alcohol-associated hepatocarcinogenesis, and that the restoration of hepatic retinoic acid concentrations by dietary retinoic acid supplementation suppresses ethanol-induced hepatocyte proliferation by inhibiting c-Jun overexpression.38

Levels of other micronutrients, such as the B-vitamins riboflavin and cobalamine, and vitamins C, D, E and K, as well as trace elements such as selenium, zinc, copper and magnesium, are also often reduced, albeit to a lesser extent (reviewed in 23).

Assessment and prevalence of malnutrition

In order to obtain valid data on the prevalence and degree of malnutrition among alcoholics with alcoholic liver disease, reliable and sensitive methods to assess nutritional status are required. Easily applicable techniques include anthropometric measurements such as body mass index, triceps skinfold thickness and mid-arm muscle area. Twenty-four hour creatinine excretion related to a reference population has been suggested as an indirect measure of body muscle mass, as 1 g of excreted creatinine was related to 18.5 kg of muscle mass.39 Other approaches are the technique of bioelectric impedance analysis for assessing body composition, and the determination of resting energy expenditure using the Harris and Benedict equation (Table 1) or by indirect calorimetry (reviewed in 22). Each technique reveals considerable limitations and may be faulty in alcoholic cirrhotics. For metabolic studies, direct measurements are needed. In a recent study, different methods for the assessment of body composition in patients with liver cirrhosis were compared.40 All applied techniques, including bioelectric impedance analysis, total body nitrogen content, dual X-ray absorptiometry and four-site skin-fold measurements, could detect advanced protein malnutrition, but differed significantly in patients with high volumes of extracellular fluid such as ascites and peripheral oedema. Although these mistakes may be avoided with more precise approaches, such as in vivo neutron activation analysis and isotope dilution techniques41, application of these methods is time-consuming and costly and restricted for research purposes. Therefore, the indirect assessments of 24 h urinary creatinine excretion and mid-arm muscle area can be used for patients without ascites, while in those with ascites and/or oedema, the creatinine-height-index appears to be reliable.41

Table 1.  Harris and Benedict equation for the noninvasive assessment of resting energy expenditure
GenderResting energy expenditure (REE)
Female66.5 + (9.56 × body weight [kilogram]) + (1.85 × height [centimetres]) – 4.676 × age (years)
Male66.5 + (13.75 × body weight [kilogram]) + (5.0 × height [centimetres]) – 6.75 × age (years)

Numerous cohort and epidemiological studies in alcoholics from North America, Asia and Europe have demonstrated that a poor nutritional condition is a widespread and crucial clinical problem which aggravates the severity of alcoholic liver disease.28, 42–46 Apart from individuals with isolated alcohol-induced steatosis without inflammation, virtually all patients with chronic alcoholic liver disease exhibit one or more signs of malnutrition.47 In cirrhotics, the prevalence of protein energy malnutrition increases from 20% in patients in cirrhosis of Child–Pugh stage A to 60% in stage C.48 In particular, protein energy malnutrition is worrisome, since it correlates strongly with the patient's severity of liver disease and, consequently, survival.

Effects of malnutrition on prognosis and survival

The nutritional status of patients with liver cirrhosis was part of the initial version of the Child–Turcotte index, but because of the lack of commonly applicable diagnostic tools to precisely diagnose malnutrition in clinical practice, the index was later reduced to easily obtainable parameters, which include clinical signs of encephalopathy and ascites, and the laboratory parameters serum bilirubin, serum albumin and prothrombin index.

Protein calorie malnutrition is frequent in alcoholic liver disease and shows a strong association with severe complications of alcoholic cirrhosis such as infections, encephalopathy, development of ascites and variceal bleeding.47, 49, 50 Though not being an independent prognostic factor for patients' survival, malnutrition correlates significantly with liver dysfunction both in alcoholic hepatitis and alcoholic cirrhosis.43, 51

Nutritional therapy of alcoholic liver disease

Different approaches have been undertaken in terms of endpoints, selection of patient groups, applied nutritional formulae, duration of therapy and the route of administration. Furthermore, some studies have combined nutritional therapy with specific drugs, such as oxandrolone, with the aim of enhancing the efficacy of nutritional support.

In addition, the present clinical data on specialized amino acid formulae such as branched-chain amino acids, and defined drugs such as S-adenosyl-l-methionine, will be addressed. Other so-called ‘supernutrients’, drugs that contain polyunsaturated phospholipids such as polyenylphopsphatidylcholine or metadoxin, will be briefly discussed.

The following description of clinical trials on nutritional therapy in alcoholic liver disease is structured according to either alcoholic hepatitis or alcoholic cirrhosis, and to the route of administration.

Alcoholic hepatitis

Oral and/or enteral nutrition. The clinical outcome of alcoholic hepatitis is highly variable and depends on its severity. While patients with mild courses usually recover after they stop drinking, severe cases are associated with a short-term mortality (within the first 4 weeks) higher than 40%.52 As assessed in a large United States trial from Veterans Affairs Hospitals, the prevalence of malnutrition reaches 100% in patients with severe alcoholic hepatitis, and a significant correlation of malnutrition with short- and long-term survival was established.43 So far, five studies have been published that investigated the efficacy of oral or enteral nutritional support in patients with alcoholic hepatitis.53–57 The details of the studies are displayed in Table 2. The design, sample size, duration and nutritional therapies administered differed substantially and hence, it is not surprising that the outcomes were variable. In the first four studies, nitrogen balance and/or serum albumin levels were improved along with nutritional support, while mortality was unaffected. This may have been due to the small sample size in all studies, which could have caused a type II error masking a beneficial effect of nutritional therapy and supporting an ongoing need of trials. In this line, the study of Cabréet al.57 merits attention. In this randomized multicentre trial, 71 patients with severe alcoholic hepatitis were enrolled to receive either 40 mg/day prednisolone and a diet containing 2000 kcal/day (n = 36), or total enteral nutrition via a nasoduodenal tube also providing 2000 kcal/day (n = 35), for 4 weeks. Severity of alcoholic hepatitis was confirmed using the Maddrey's discriminant function (4.6 × [patient-control prothrombin time in seconds] + serum bilirubin [mg/dL]). Although the differences in both the short-term mortality and the 1 year probability of survival between the two groups were not statistically different, several important issues became apparent.

Table 2.  Studies on therapy of alcoholic hepatitis with oral or enteral nutrional supplements
AuthorDesignPatients (n)Duration (days)Experimental treatment (EXP)Control treatment (CTR)MortalitySecondary endpoints
  1. Abbreviations are as follows: AA,, amino acids; BCAA, branched-chain amino acid; FU, follow-up; N.S., not significant; PRED, prednisolone group; TEN, total enteral nutrition group.

Galambos et al. 197954Open label1616–42Oral (standard hospital diet) or intravenous supplement (51.6–77.4 g protein)NoneNot assessedNitrogen balance + albumin improved in EXP, CTR not assessed
Mendenhall et al. 198555Historical controls5730Standard hospital diet (2500 kcal/day) + 2200 kcal/day BCAA,Standard hospital dietN.S.Improvement of albumin, transferrin, RBP
Calvey et al. 198556Randomized, controlled6421Standard diet (∼2000 kcal/day) + 65 g standard AA or BCAA,Standard diet, 80 g protein/dayN.S.Positive nitrogen balance in EXP, delayed hypersensitivity improved
Soberon et al. 198757Crossover14 6Nasoduodenal tube, 35 kcal/kg/day, fat/carbohydr. /protein 45/40/15%3 days standard hospital diet (35kcal/kg/day)0/6 contrs. 3/8 treatm.Nitrogen balance improved fivefold at 2 week.
Cabréet al. 200058Randomized, controlled7128Nasogastric tube, 2000 kcal/day, 72 g protein/day, 31% BCAAStandard diet (1 g protein/kg) + 40 mg/day prednisolone11/35 TEN 9/36 PRED N.S. FU: 2/24 TEN 10/27 (P = 0.04)No dropouts in PRED, 8 dropouts in TEN; equal improvements of albumin, Child score, Maddrey score; equal rate of infections

1 There was a significantly better outcome during the 1 year follow-up after 4 weeks of treatment for the patients treated with total enteral nutrition (P = 0.04, intention-to-treat analysis), with only two patients dying (8%), whereas 10 patients (37%) who were treated with prednisolone alone died during the 1 year follow-up period, mostly due to infections within the first 6 weeks of follow-up.

2 Furthermore, there was a high number of drop-outs in the total enteral nutrition group, resulting from the lack of patient acceptance of the nasogastric tube rather than from serious adverse events. From these data, it may be possible that a combination of total enteral nutrition and steroids may yield even better clinical results than either therapy alone.

Two further studies with a reasonable number of patients showed no treatment benefit for patient survival, but surrogate parameters of nutritional status, such as serum albumin, retinol-binding protein and nitrogen balance, improved.54, 55 The studies by Galambos et al. and Soberon et al. only included 16 and 14 patients, respectively, which appears to be an insufficient sample size to draw firm conclusions.

In two multicentre trials of patients from United States Veterans Administration Hospitals, Mendenhall et al. attempted to enhance the efficacy of enteral nutrition by adding oxandrolone, a synthetic anabolic steroid.58, 59 The study characteristics are listed in Table 3. In the first study, oxandrolone therapy was compared against prednisolone, with nutritional support not being the primary intervention in a cohort of 57 patients with alcoholic hepatitis. In the second study, 273 patients with alcoholic hepatitis on active treatment received a standard hospital diet in addition to a nutritional formula preparation containing 1600 kcal/day (Hepatic Aid), thereby providing a supplement of 60 g/day of proteins during a period of 30 days while patients were in the hospital, together with oxandrolone at 80 mg/day. After discharge, patients received 1200 kcal/day of Hepatic Aid, containing 45 g/day of proteins in addition to 40 mg/day of oxandrolone, for the following 60 days as out-patients. In control patients, oxandrolone was replaced by placebo tablets accompanied by a low calorie, low protein food supplement providing 6.8 g/day of proteins or 264 kcal/day while in the hospital, and 5.1 g/day or 198 kcal/day as out-patients. Because of the large sample size, subgroups were analysed, such as those with either severe or moderate malnutrition at inclusion into the study. Thereby, it was demonstrated that patients with moderate malnutrition experienced a significant improvement through active treatment, as estimated on the basis of Maddrey's discriminant function, and the level of malnutrition defined as the degree of protein energy malnutrition (P = 0.03 and 0.05, respectively). The most important finding, however, was the relationship of therapy to survival, which showed that at 30 days of treatment, 9.4% of patients on active treatment had died while this figure reached 20.9% in the placebo group. After 6 months, this difference was even more pronounced, showing that 20.3% of actively treated patients had died, while mortality in the placebo group was 37.3% (P = 0.037). Another important result was that the treatment benefit could only be observed with successful nutritional therapy and not with oxandrolone medication. Patients from both studies were recently subjected to a meta-analysis, in which the authors concluded that the efficacy of oxandrolone treatment resulted from sufficient nutritional therapy rather than from the administration of oxandrolone.47

Table 3.  Studies on therapy of alcoholic hepatitis with oral nutritional support and the anabolic steroid oxandrolone
AuthorDesignPatients (n)Duration (days)Experimental treatment (EXP)Control treatment (CTR)MortalitySecondary endpoints
  1. Abbreviations are as follows: BCAA, branched-chain amino acid; N.S., not significant; OXA, oxandrolone; CTR, control group; RBP, retinol-binding protein.

Mendenhall et al. 198460Historical controls 5730Standard hospital diet (2500 kcal/day) + 2200 kcal/day BCAA 80 mg OXA/dayStandard hospital diet + 40 mg/day prednisoloneN.S.Improvement of albumin, transferrin, RBP
Mendenhall et al. 199361Randomized, controlled27390OXA 80 mg/day × 30 days; then 40 mg/day × 60 days, standard diet + 1200–1600 kcal/d supplement providing 45–60 g/day protein (46% BCAA)Placebo, standard diet + supplement providing 5.1–6.8 g/day protein and 198–264 kcal/dayEXP 6/64, CTR 14/67 at 30 days; EXP 13/64, CTR 25/67 at 6 months P = 0.037Nitrogen balance and visceral proteins not determined, Improvement only achieved in moderately malnourished patients

Parenteral nutrition — alcoholic hepatitis and cirrhosis. In this section of our review, patients with alcoholic hepatitis and cirrhosis will be analysed together, since in most studies, a clear distinction between the two entities was not made. This reflects the fact that it is notoriously difficult to precisely distinguish between these two liver pathologies without obtaining a liver histology from all patients. In particular, many of these patients are at a high risk of biopsy-induced haemorrhage due to hepatic coagulopathy. A series of studies addressing parenteral nutrition in patients with alcoholic hepatitis and cirrhosis was performed and altogether, 244 patients were treated with different study protocols.60–65 The details of these trials are shown in Table 4. Only randomized controlled trials were published and the treatment duration ranged from 21 to 30 days. Unfortunately, none of the studies could demonstrate an improved survival, although the first study by Nasrallah and Galambos60 provoked this expectation. All trials showed an improvement of visceral protein as assessed by serum albumin levels, and three studies reported a positive nitrogen balance in the treatment groups. Surprisingly, no long-term studies exist which address intermediate or long-term effects on survival in patients with alcoholic hepatitis or cirrhosis, and the most recent trial which investigated the treatment benefit of parenteral nutrition in patients undergoing hepatic resection for hepatocellular carcinoma dates back to 1994.66 However, the proportion of patients in the study population with alcoholic liver disease, i.e. alcoholic hepatitis and cirrhosis, was not reported. To our knowledge, no study on total parenteral nutrition in exclusively alcoholic cirrhotics has been performed.

Table 4.  Randomized, controlled trials on treating alcoholic hepatitis with parenteral nutritional therapy
AuthorDesignPatients (n)Duration (days)Experimental treatment (EXP)Control treatment (CTR)MortalitySecondary endpoints
  1. Abbreviations are as follows: AA, amino acids; CV, central vein; N.S., not significant; PV, peripheral vein; RBP, retinol-binding protein; StD, standard diet.

Nasrallah & Galambos 198061Randomized, controlled342870–85 g standard AA/d via PV, StDStD (actual intake 1400 kcal/day)EXP 0/17 CTR 4/18 (P = 0.06)Serum albumin improved
Diehl et al. 198562Randomized, controlled153052 g standard AA/day, 130 g glucose/day via PV, StDStD (actual intake ∼3000 kcal/day)EXP 0/5 CTR 0/10 (N.S.)Positive nitrogen balance in EXP, prealbumin and RBP equally improved
Naveau et al. 198663Randomized, controlled402888 g standard AA/day ∼2800 kcal/day via CV, StDStD (actual intake ∼2100 kcal/day)EXP 1/20 CTR 1/20 (N.S.)Serum albumin improved only in CTR, transferrin, prealbumin RBP in both
Achord 198764Randomized, controlled282142.5 g standard AA/day, via PV, StDStD (actual intake ∼1200 kcal/day)EXP 1/14 CTR 3/14 (N.S.)Serum albumin imroved only in EXP
Simon & Galambos 198865Randomized, controlled342870 g standard AA/day 100 g glucose/day 50 g lipids/day via PV StD + liquid formulaStD + liquid formula (actual intake not stated)EXP 4/10 CTR 3/12 (N.S.)Serum albumin and transferrin improved only in EXP
Mezey et al. 199166Randomized, controlled543052 g standard AA 130 g glucose/day via PV, StDStD + 130 g i.v. glucose (actual intake 1400 kcal/day)EXP 6/28 CTR 5/26 (N.S.) Serum albumin and prealbumin improved in both; RBP and transferrin improved only in EXP
Bonkovsky et al. 199168, 69Randomized, controlled392170 g standard AA 100 g glucose/day via PV, StDStD (actual intake ∼1800 kcal/day)EXP 0/19 CTR 0/20 (N.S.)Improvement of albumin, transferrin, prealbumin in both groups

In patients with alcoholic hepatitis, only one trial investigated whether beneficial effects on nutritional status may be achieved by the combination of intravenous nutritional therapy and oral oxandrolone.67, 68 Thirty-nine patients with clinical criteria of alcoholic hepatitis, but without confirmatory liver biopsy, were allocated to four groups receiving: (1) nutritionally adequate standard diet; (2) standard diet and oxandrolone 80 mg/day; (3) standard diet plus 2 L of 3.5% branched-chain amino acid enriched solution/day (Aminosyn II, Abbott); or (4) standard diet and intravenous branched-chain amino acid plus 80 mg oxandrolone/day for 21 days. While the first paper of the study described improvements of liver function tests such as galactose elimination capacity, antipyrine clearance, as well as liver and spleen sizes, following parenteral nutrition, regardless of oxandrolone administration, a second analysis of the same study demonstrated a significant improvement of nutritional status, nitrogen balance and visceral proteins in the groups receiving parenteral nutritional therapy with and without oxandrolone. Oxandrolone administration again failed to result in further improvement of measured parameters with the exception of serum pre-albumin and transferrin, which were highest in the patients treated with parenteral nutrition plus oxandrolone. Tolerability of nutritional therapy and oxandrolone was apparently good, since all patients completed the trial. However, in addition to the fact that the study groups were relatively small, the significance of liver function tests to reflect the patients' prognosis is questionable and long-term survival was not reported.

Due to the lack of clear-cut evidence, oxandrolone, either in combination with nutritional therapy or alone, cannot be advised for patients with alcoholic hepatitis.

Alcoholic cirrhosis

Oral and/or enteral nutrition. So far, data are available from 13 studies that tested the hypothesis of whether oral or enteral nutritional supplementation is beneficial for patients with alcoholic cirrhosis.69–81 Details of the study designs and their results are given in Table 5. The vast majority of trials that studied total enteral nutritional therapy in alcoholic cirrhotics are hampered by small sample size and by insufficient duration of therapy, being as short as 3–12 days, which excluded both a relevant treatment effect as well as a valid statistical evaluation. For none of the trials was sample size calculations carried out prior to the start of patient enrolment. Moreover, only seven studies were randomized while the remaining ones were open labelled (n = 6). Considering study design, duration of therapy and sample size, only five studies seem to be adequate and of these, only the study by Cabréet al.76 demonstrated a significant treatment effect on survival. In this study, 35 cirrhotic patients, of which 23 were alcoholics, were either treated with total enteral nutrition through a fine-bore nasogastric feeding tube providing 2115 kcal/day and containing 71 g of protein, or a standard low-sodium hospital diet at 2200 kcal/day with a protein content of 70–80 g. Neither group showed an improvement in nutritional parameters, with the exception of a significant elevation of mean serum albumin concentrations in total enteral nutrition patients. However, the clinical outcome was better in the total enteral nutrition group, showing a significant reduction in Child score and a significantly better survival (2 vs. 9, P = 0.02). Since the incidence of major complications of liver cirrhosis, i.e. gastrointestinal haemorrhage and severe infections, was similar in both groups, it appeared that the total enteral nutrition patients were more likely to survive an incident of cirrhosis-related complication than the control subjects.

Table 5.  Studies on treating alcoholic cirrhosis with oral and enteral nutritional therapy
Author DesignPatients (n)Duration (days)Experimental treatment (EXP)Control treatment (CTR)MortalitySecondary effects
  1. Abbreviations used are as follows: BCAA, branched-chain amino acid; CTR, control group; EXP, experimental group; HRF, hepatorenal failure; N.S., not significant.

Smith et al. 198270Open label1010–60Three different formulae: Oral 76–143 g protein, 2000–3716 kcal/dayNoneNonePositive nitrogen balance, improved albumin, transferrin, creatinine/ height, midarm muscle, fat areas
Keohane et al. 198371Open label103–23 (mean 7.3)Oral BCAA formula, 80 g Protein/day through Nasogastric tubeNone1 death (HRF)Positive nitrogen balance, increased albumin
McGhee et al. 198372Randomized, double-blind, crossover41120 g casein + 30 g BCAA formula50 g casein/dayNoneEXP equal to CTR Positive nitrogen balance
Christie et al. 198573Randomized, double-blind, crossover812BCAA (50%) formulaStandard diet (18% BCAA)1 death (infection)EXP equal to CTR
Okita et al. 198574Open label10440 g protein + 40 g BCAA formula/day2100 kcal/day 80 g protein/dayNoneEXP equal to CTR Positive nitrogen balance
Bunout et al. 198975Randomized, controlled362850 kcal/kg, 1.5 g protein/dayStandard dietEXP 2/17 CTR 5/19 (N.S.)No differences
Cabréet al. 199076Randomized, controlled35 (23 alc.)23–252115 kcal/day incl. 71 g BCAA formulaStandard dietImproved (P = 0.02)Child score improved Albumin improved
Marchesini et al. 199077Randomized6490Standard diet + BCAA supplement (0.24 g/kg)Standard diet + casein supplem. (0.175 g/kg)NoneNitrogen balance improved in both, BCAA better than Standard diet
Kearns et al. 199278Randomized3128Casein supplem. (1.5 g protein/d, 40 kcal/kg/day)Standard dietN.S.Both groups improved nitrogen balance and albumin
Hirsch et al. 199379Randomized controlled5112 (mo)Standard diet + casein supplem. (1000 kcal/day, 34 g protein/day)Standard dietEXP 3/26 CTR 6/25 (N.S.)Fewer hospitalizations, improved albumin and visceral protein
Nielsen et al. 199580Open label1538Increasing amounts of protein via standard diet (1.0–1.8 g/kg/day)NoneNoneIncreased protein retention through gradual elevation of protein intake
Campillo et al. 199581Open label2630Standard dietNoneNoneAnthropometric ratios improved
Hirsch et al. 199982Open label316 (mo)Standard diet + casein supplem. (1000 kcal/day, 34 g protein/day)None6 deaths/31Increased albumin, improved cellular immunity

In the study by Kearns77, 16 alcoholic patients were randomly allocated to receive an enteral nutritional supplement through a plastic tube in addition to a regular diet providing 1.5 g/kg of protein and 40 kcal per day, while 15 patients received a control diet without the tube-fed supplement for 4 weeks. The nutritional intake was significantly different between the two groups, with the tube-fed patients receiving 200% of the calories and protein consumed by the controls. This resulted in a significant improvement in nitrogen balance and a significantly higher serum albumin in treated individuals at 3 weeks of the treatment interval. Clinically, the mean grade of encephalopathy improved in the tube-fed group while a deterioration was noticed in the controls. However, mortality was not affected either during the feeding period or during follow-up. Adverse effects of tube feeding were not reported in detail, but the authors stated that replacement of dislocated tubes was required an average of three times in each patient. This small study demonstrated that tube feeding may significantly improve nutritional status and encephalopathy in alcoholic cirrhotics, mainly because of an increased intake of protein and total calories.

In the study by Hirsch et al.78, out-patients were treated for a period of 1 year. Twenty-six patients received a daily food supplement of 1000 kcal and 34 g of protein given as an oral casein-based nutritional formula, and 25 controls received a placebo capsule. Again, mean calorie and protein intake was significantly higher in case patients than in controls. However, this did not result in an increase of body weight, anthopometric indices or serum albumin levels. Mortality was not significantly higher in control subjects, but case patients showed lower rates of hospital admission due to a significant reduction of complications associated with cirrhosis, such as infections, gastrointestinal haemorrhage, ascites or encephalopathy. This study demonstrates that nutritional support in alcoholic out-patient cirrhotics may decrease the number of hospital referrals, due mainly to a reduction of complications.

In the other two randomized trials74, 76, mortality was not significantly influenced by nutritional therapy, which, in the case of the study by Bunout et al., just failed to reach statistical significance, possibly due to the relatively low overall mortality rate.

So far, there are sufficient data supporting enteral/oral nutritional therapy in alcoholic cirrhotics to improve nutritional status and complications of cirrhosis. However, due to the limitations of most of the trials mentioned above, larger studies are warranted to answer precisely the question of whether vigorous nutritional support improves survival in this sub-group of severely ill alcoholics.

New approaches using specialized formulae

With the increasing understanding of the pathogenesis of alcoholic liver disease and the associated metabolic disturbances, new treatment approaches have emerged that pay tribute to critical pathophysiological events in the development of alcohol-induced liver damage. One such event is that of oxidative stress generated during ethanol metabolism via cytochrome P4502E1. In order to counteract these potentially cytotoxic mechanisms, liver cells require defense mechanisms. Among these, the hepatic antioxidative capacity, represented by compounds that possess reductive capacities such as glutathione, are highly important. The liver's glutathione pool is maintained either through synthesis of glutathione from homocysteine through the transsulphuration pathway catalysed by the enzymes cystathionine-β-synthase and cystathionase, or by restoration of oxidized glutathione via the enzyme glutathione-S-transferase. Either pathway is influenced by alcohol, which interacts at various sites of glutathione synthesis through interaction with precursors or essential coenzymes, and by directly damaging mitochondria which harbour the hepatic glutathione pool. Moreover, malnutrition in general is a major cause of glutathione deficiency. Below, the effects of alcohol on glutathione synthesis are described and strategies to counteract glutathione depletion are delineated.


Increasing insight has been gathered with respect to the role of certain nutritional factors in alcoholic liver disease, and treatment strategies are based on the concept that some of these factors become essential in the diseased liver. One such example is methionine, an essential sulphur-containing amino acid. In order to exert its metabolic actions beyond the incorporation into proteins, methionine requires activation to S-adenosyl-l-methionine, a reaction which is mediated by the activity of the enzyme methionine adenosyltransferase. In chronic liver disease, the capability of activating methionine to S-adenosyl-l-methionine is suppressed along with the activity of methionine adenosyltransferase.82 The inhibition of methionine adenosyltransferase activity is related to oxidative stress which is generated along with the degradation of alcohol. The group of Dr. Mato has convincingly demonstrated that reactive oxygen species and nitric oxide may switch methionine adenosyltransferase to an inactive conformation through S-nitrosylation and oxidation, causing a post-translational inactivation of methionine adenosyltransferase.83 Several experimental studies in various animal species have unequivocally shown that chronic ethanol administration markedly decreases hepatic S-adenosyl-l-methionine (SAM) concentrations while hepatic S-adenosylhomocysteine (SAH) concentrations are increased, leading to an impairment of the hepatic methylation capacity expressed as the S-adenosyl-l-methionine/SAH ratio.84–86 Therefore, supplementation of both methionine and choline is ineffective and may even be toxic.87, 88 Methyl group donors, however, are essential for the functioning of numerous biological processes, including polyamine synthesis, one-carbon group transfer to DNA and methylation of membrane phopsholipids. In this situation, S-adenosyl-l-methionine becomes an essential nutritional substitute, since it acts as the major methyl group donor for virtually all biological transmethylation reactions. Furthermore, S-adenosyl-l-methionine exerts regulatory functions for the synthesis of the most important reductive compound, glutathione, from cysteine, as high hepatic S-adenosyl-l-methionine levels activate the flow of homocysteine through the transsulphuration pathway to glutathione, whereas low S-adenosyl-l-methionine levels lead to the preservation of methionine.89 Perhaps the most important function of S-adenosyl-l-methionine is that of global or gene-specific DNA methylation. The importance of S-adenosyl-l-methionine as a modulator of gene regulation comes from recent studies demonstrating reduced mRNA expression of the methionine adenosyltransferase gene following hypermethylation of its promotor. Interestingly, a decrease in methionine adenosyltransferase mRNA levels could be detected both in alcoholic cirrhosis and in hepatocellular carcinoma.90 In addition, alcohol interferes with various other sites of methyl group transfer. Considering the potential toxicity of high-dose methionine therapy, it seems logical to administer the metabolically active and entirely innocuous methionine adenosyltransferase.

The therapeutic benefit of methionine adenosyltransferase for treatment of certain liver diseases has been demonstrated. For example, S-adenosyl-l-methionine was shown to relieve pruritus and to lower elevated liver enzymes in patients with cholestatic liver disease such as intrahepatic cholestasis, primary biliary cirrhosis and chronic viral hepatitis.91 So far, three clinical trials have suggested a favourable effect of S-adenosyl-l-methionine treatment in alcoholic liver disease. Thus, oral administration of S-adenosyl-l-methionine improved alcohol-induced glutathione deficiency in red blood cells and the liver92, 93, while the most impressive data come from a randomized, placebo-controlled, double-blind multicentre trial in alcoholic cirrhotics.94 In this trial, 123 patients with various stages of liver cirrhosis as assessed by the Child–Pugh index were treated with 1200 mg/day of oral S-adenosyl-l-methionine for 2 years. Death from liver-related complications and liver transplantation were chosen as primary endpoints. The overall mortality at the end of the trial was 30% in the placebo group and 16% in the patients receiving S-adenosyl-l-methionine. This result just failed to reach significance (P = 0.077), but after exclusion of Child C patients, the reduction in the rate of mortality or liver transplantation became significant (29% vs. 12%, P = 0.025). It should be emphasized that patients' compliance was excellent and adverse treatment effects were equal to placebo level. Although these encouraging results require confirmation in other clinical trials, they suggest that S-adenosyl-l-methionine represents an effective, relatively cheap and nontoxic drug for the treatment of chronic alcoholic liver disease.


As mentioned above, deficiency of S-adenosyl-l-methionine and consequently, of hepatic methylation capacity, may lead to hypomethylation of phospholipids, in particular phosphatidylethanolamine, with the result of decreased (methylated) phosphatidylcholine levels in the liver. This situation is further aggravated by the known reduction of the activity of phosphatidylethanolamine methyltransferase by acetaldehyde, thereby inhibiting the methyl transfer to phosphatidylethanolamine.95 Most recently, a disturbed phospholipid metabolism has been confirmed in alcoholic liver disease using in vivo-proton-decoupled 31P-magnetic resonance spectroscopy, showing an increase of phosphoethanolamine compared with phosphocholine and indicating impaired methylation of phosphoethanolamine.96 Impaired methylation of phospholipids is a major factor in the disruption of the integrity of cellular membranes of hepatocytes and consequently, of cell damage. In 1994, Lieber and coworkers observed that lecithin preparations extracted from soybeans and rich in polyunsaturated phospholipids were able to inhibit the development of alcohol-induced fibrosis and cirrhosis in baboons.97 The same group showed that polyenylphosphatidylcholine inhibits the activation and proliferation of hepatic stellate cells, the major source of collagens in fibrogenesis, enhances hepatic collagenase activity and antagonizes at least some of the oxidative stress produced by alcohol.98 Since then, several studies using polyenylphosphatidylcholine as a treatment of alcoholic liver disease have been carried out in experimental animals, mostly rats and baboons, which have shown that the administration of polyenylphosphatidylcholine may counteract most alcohol-related alterations in the distribution of phopsholipids. However, few clinical studies have been performed with polyenylphosphatidylcholine in liver disease. Panos et al. demonstrated a positive trend towards the improvement of laboratory findings, rate of complications and histology in patients with acute alcoholic hepatitis.99 Earlier, largely uncontrolled investigations described a decrease of serum liver enzyme levels and histology, and improvement of patients' well-being.100, 101 These studies were compromised by considerable limitations, such as small sample size, insufficient characterization of patients and lack of control of alcohol consumption during the study and follow-up.

Recently, the results of a large randomized, controlled multicentre trial from the United States were presented, in which the impressive number of 789 patients with biopsy-proven alcoholic liver disease, including fatty liver and fibrosis, were included. Only patients without complete cirrhosis were recruited. Intervention consisted of intensive counseling with respect to medical problems of alcoholism, and either 1500 mg of polyenylphosphatidylcholine daily or placebo for 4–6 years. While the overall comparison of the two groups failed to show a significant treatment effect by polyenylphosphatidylcholine, a subgroup of patients who continued to consume at least six drinks/day (n = 52) showed a cessation of histological progression with polyenylphosphatidylcholine, whereas those on placebo progressed histologically.102 This small benefit, which might well be the result of chance, seems rather disappointing in the light of the high expectations initially linked to polyenylphosphatidylcholine. It can therefore be concluded that there is currently no indication to treat alcoholic liver disease with polyenylphosphatidylcholine until more convincing data support its use.


A single double-blind randomized multicentre trial has been carried out with metadoxine (pyridoxol L, 2 pyrrolidone-5-carboxylate) involving 136 chronic active alcoholic patients diagnosed with fatty liver.103 Patients received 1500 mg/day of metadoxine (n = 69) or placebo (n = 67) for 3 months. Liver enzymes such as aminotransferases and gammaglutamyl transpeptidase improved significantly after only 1 month of treatment. Likewise, the percentage of patients with sonographic signs of steatosis decreased significantly in the metadoxine group after the treatment period (28% vs. 70%, P < 0.01). Among patients who continued to drink, the prevalence (45% vs. 92%, P < 0.05) and the degree of steatosis were also significantly lower in the metadoxine group. However, these encouraging results have to be interpreted with caution, since no histological confirmation of fatty liver was obtained, while the sonographic evaluation of the degree of steatosis remains subjective. Therefore, these study results need confirmation, and the treatment of choice for patients with alcoholic fatty liver should be the reduction or the cessation of alcohol consumption.


Results from clinical trials demonstrate that nutritional therapy is beneficial for patients with alcoholic liver disease. Although treatment benefit seemed to be restricted predominantly to parameters of nutritional status, increasing evidence also indicated improved survival, in particular in patients with alcoholic hepatitis. Whenever possible, nutritional therapy should be administered orally or via enteral tube feeding, since patients appear to benefit from topical nutritional factors in the gut. Another argument favouring oral nutrition is the lack of infections which may occur with central venous catheters. Concerns such as precipitating variceal haemorrhage along with the insertion of the feeding tube have not been confirmed in clinical trials.55

Various international nutritional associations have formulated recommendations for nutritional therapy in alcoholic liver disease which are summarized in Table 6.10, 104 Herein, treatment aims are defined for each stage of alcoholic liver disease, which reflect the sometimes limited prognosis of patients with advanced disease. However, achieving some of these nutritional goals may serve as a so-called ‘bridge-to-transplant’ in severely diseased patients in whom marasmus and nutrient deficiencies could lead to pre-emptive death.

Table 6.  Guidelines for the nutritional therapy of alcoholic liver disease of various stages, and their complications. This table shows a summary of recommendations issued by various international societies10, 104
Disease/ complicationAimsCalories (kcal/KG)Protein (g/KG)Carbohydrates (g/KG)Lipids (g/KG)Fluid/ electrolytesVitamins/ trace elements
  1. Abbreviations used are as follows: BCAA: branched-chain amino acids; MCFA: medium chain fatty acids; PEM, protein energy malnutrition; n.R. no recommendation.

Fatty liverAbstinencen.R.n.R.n.R.n.R.n.R.n.R.
Reduction of alcohol intake      
Alcoholic hepatitisPrevention of PEM401.5–2.04.0–5.01.0–2.0n.R.Supplementation of deficiencies
Prevention of encephalopathy hypoglycaemia, inflammation      
Liver cirrhosis without malnutritionPrevention of PEM351.3–1.54.0–5.01.0–1.5  
Liver cirrhosis with malnutritionPrevention of malnutrition and decompensation35–401.5–2.03.0–4.02.0–2.5Fluid restriction 2.0–2.5 L/dayB-vitamins folate, thiamine, vitamin C and K
Protein intoleranceRecompensation Prevention of PEM≥ 250.3–0.5 BCAA2.5–3.51.0–1.5Fluid restriction 2.0 L/dayB vitamins folate, thiamine, Vitamin B12
Intrahepatic cholestasisImprovement of cholestasis   50% MCFA Fat-soluble vitamins (A, D, E, K)
Prevention of malnutrition      
Ascites and oedemaRecompensation    1.0–1.5 
Prevention of infection    Sodium restriction 


F.S. is recepient of a research fellowship of the Interdisziplinäres Zentrum für Klinische Forschung (IZKF) of the University of Erlangen-Nuremberg. The present work has been in part supported by a research grant by the Fonds für Forschung und Lehre (ELAN), No., University of Erlangen-Nuremberg.