Nutritional Therapy in Cirrhosis or Alcoholic Hepatitis
We identified 1593 potentially eligible references in the electronic and manual searches (Fig. 1). After excluding duplicates and references that clearly did not fulfil our inclusion criteria, we retrieved 24 potentially eligible references. Eight of these references were excluded because they referred to trials on patients with viral hepatitis (none had cirrhosis) or assessed nutritional interventions that fulfilled <75% of the daily requirements. One references referred to a trial that assessed nutrition vs. steroids for alcoholic hepatitis. The patient inclusion criteria and the criteria for the nutritional regimen were met. We included the trial in a post hoc sensitivity analysis, but excluded the trial from the primary analyses.
(Enlarge Image)
Figure 1.
Trial flow chart.
We included 13 randomized controlled trials published during 1980–2012. The trials included 663 patients. In total, 559 patients had cirrhosis. In three trials, all patients had alcoholic hepatitis. The total number of patient with alcoholic hepatitis was 184 (104 of these patients did not have cirrhosis). The mean age ranged from 43 to 65 years and 60% were men. Included patients were hospitalized in 10 trials and outpatients in three trials. In six trials, more than 80% of the included patients had ascites at baseline (range 36–100%). Overt hepatic encephalopathy was reported in nine trials and ranged from 11 to 40%. Nine trials assessed enteral (oral or via an enteral tube) nutrition and four trials assessed intravenous (parenteral) nutrition. The nutritional interventions contained between 71 and 150 g of protein/day and 2115 to 4260 kcal/day. Ten trials reported the actual intake. The difference in the actual protein intake between the nutritional supplement and no intervention control group ranged from −7 to 75 g protein/day. The daily calorie intake was 185 to 2260 kcal higher in the intervention group than in the control group. The duration of therapy ranged from 3 to 365 days (Table 1).
All trials were classed as having a high or unclear risk of bias based on one or more of the components assessed (Table 2).
In total, 72 of 329 patients in the intervention and 97 of 334 controls died (Fig. 2). Random-effects meta-analysis showed that the nutritional therapy reduced mortality 0.80 (95% CI, 0.64 to 0.99 I = 0%). There was no evidence of small study effects (P = 0.47). In the sequential analysis, the monitoring boundary was not crossed suggesting that the results are not robust to multiple testing. In a post hoc analysis, we included data on long-term follow-up from a trial on nutrition vs. steroids. Random-effects meta-analysis including the trial found that nutrition reduced mortality (0.80 95% CI 0.67 to 0.96, I = 0%). In subgroup analyses, there were no differences between trials on enteral or parenteral administration (test for subgroup differences P = 0.44) or trials with alcoholic hepatitis compared to trials with cirrhosis (test for subgroup differences P = 0.07).
(Enlarge Image)
Figure 2.
Nutritional therapy vs. no intervention. Outcome; Mortality overall (random effect analysis).
In the random-effects metaregression analysis, the intervention effect was not predicted by the mean age (P = 0.34), the proportion of men (P = 0.56), the daily amount of calories (P = 0.08), the daily amount of protein (P = 0.11), the duration of the intervention (P = 0.15) or the duration of follow-up (P = 0.26).
In fixed-effect meta-analysis, nutritional therapy reduced the risk of developing overt hepatic encephalopathy 0.73 (95% CI, 0.55 to 0.96, I = 20%, Fig. 3); and infections 0.66 (95% CI, 0.45 to 0.98, I = 41%, seven trials). The beneficial effects were not confirmed in random-effects meta-analyses (overt hepatic encephalopathy 0.76 (95% CI, 0.54 to 1.07) and infections 0.68 (95% CI, 0.37 to 1.26)). Random-effects meta-analysis showed no beneficial or detrimental effects on resolution of overt hepatic encephalopathy 1.29 (95% CI, 0.74 to 2.22, I = 0%; five trials), resolution of ascites 0.78 (95% CI, 0.54 to 1.13, I = 56%; four trials), incidence of ascites 1.06 (95% CI, 0.54 to 2.08, I = 57%; four trials) or gastrointestinal bleeding 0.96 (95% CI 0.67 to 1.38, I = 51%; six trials).
(Enlarge Image)
Figure 3.
Nutritional therapy vs. no intervention. Outcome; Overt hepatic encephalopathy (fixed-effect analysis).
Nutrition reduced bilirubin levels with 0.51 g/dl (95% CI, −0.54 to −0.48 g/dl, I = 41%, 10 trials) in the fixed-effect analysis, but not in the random-effects meta-analysis (−0.54 95% CI, −1.10 to 0.03). There was no effect on alanine amino transferase; 0.82 U/L (95% CI, −11.08 to 12.72 U/L, I = 30%; five trials), alkaline phosphatase −19.1 U/L (95% CI, −145.5 to 107.3 U/L, I = 67%, three trials), albumin; 0.13 g/dl (95% CI, −0.08 to 0.19 g/dl, I = 75%, 10 trials) or other nutritional parameters as triceps skinfold test (TSF); 1.36 mm (95% CI, −0.87 to 3.58 mm, I = 0%; two trials), mid arm circumference (MAMC); 0.92 cm (95% CI, 0.34 to 1.50 cm, I = 0%; three trials). None of the included trials reported data that allowed analyses of hepatorenal syndrome, body mass index or grip strength.
Results
We identified 1593 potentially eligible references in the electronic and manual searches (Fig. 1). After excluding duplicates and references that clearly did not fulfil our inclusion criteria, we retrieved 24 potentially eligible references. Eight of these references were excluded because they referred to trials on patients with viral hepatitis (none had cirrhosis) or assessed nutritional interventions that fulfilled <75% of the daily requirements. One references referred to a trial that assessed nutrition vs. steroids for alcoholic hepatitis. The patient inclusion criteria and the criteria for the nutritional regimen were met. We included the trial in a post hoc sensitivity analysis, but excluded the trial from the primary analyses.
(Enlarge Image)
Figure 1.
Trial flow chart.
We included 13 randomized controlled trials published during 1980–2012. The trials included 663 patients. In total, 559 patients had cirrhosis. In three trials, all patients had alcoholic hepatitis. The total number of patient with alcoholic hepatitis was 184 (104 of these patients did not have cirrhosis). The mean age ranged from 43 to 65 years and 60% were men. Included patients were hospitalized in 10 trials and outpatients in three trials. In six trials, more than 80% of the included patients had ascites at baseline (range 36–100%). Overt hepatic encephalopathy was reported in nine trials and ranged from 11 to 40%. Nine trials assessed enteral (oral or via an enteral tube) nutrition and four trials assessed intravenous (parenteral) nutrition. The nutritional interventions contained between 71 and 150 g of protein/day and 2115 to 4260 kcal/day. Ten trials reported the actual intake. The difference in the actual protein intake between the nutritional supplement and no intervention control group ranged from −7 to 75 g protein/day. The daily calorie intake was 185 to 2260 kcal higher in the intervention group than in the control group. The duration of therapy ranged from 3 to 365 days (Table 1).
All trials were classed as having a high or unclear risk of bias based on one or more of the components assessed (Table 2).
In total, 72 of 329 patients in the intervention and 97 of 334 controls died (Fig. 2). Random-effects meta-analysis showed that the nutritional therapy reduced mortality 0.80 (95% CI, 0.64 to 0.99 I = 0%). There was no evidence of small study effects (P = 0.47). In the sequential analysis, the monitoring boundary was not crossed suggesting that the results are not robust to multiple testing. In a post hoc analysis, we included data on long-term follow-up from a trial on nutrition vs. steroids. Random-effects meta-analysis including the trial found that nutrition reduced mortality (0.80 95% CI 0.67 to 0.96, I = 0%). In subgroup analyses, there were no differences between trials on enteral or parenteral administration (test for subgroup differences P = 0.44) or trials with alcoholic hepatitis compared to trials with cirrhosis (test for subgroup differences P = 0.07).
(Enlarge Image)
Figure 2.
Nutritional therapy vs. no intervention. Outcome; Mortality overall (random effect analysis).
In the random-effects metaregression analysis, the intervention effect was not predicted by the mean age (P = 0.34), the proportion of men (P = 0.56), the daily amount of calories (P = 0.08), the daily amount of protein (P = 0.11), the duration of the intervention (P = 0.15) or the duration of follow-up (P = 0.26).
In fixed-effect meta-analysis, nutritional therapy reduced the risk of developing overt hepatic encephalopathy 0.73 (95% CI, 0.55 to 0.96, I = 20%, Fig. 3); and infections 0.66 (95% CI, 0.45 to 0.98, I = 41%, seven trials). The beneficial effects were not confirmed in random-effects meta-analyses (overt hepatic encephalopathy 0.76 (95% CI, 0.54 to 1.07) and infections 0.68 (95% CI, 0.37 to 1.26)). Random-effects meta-analysis showed no beneficial or detrimental effects on resolution of overt hepatic encephalopathy 1.29 (95% CI, 0.74 to 2.22, I = 0%; five trials), resolution of ascites 0.78 (95% CI, 0.54 to 1.13, I = 56%; four trials), incidence of ascites 1.06 (95% CI, 0.54 to 2.08, I = 57%; four trials) or gastrointestinal bleeding 0.96 (95% CI 0.67 to 1.38, I = 51%; six trials).
(Enlarge Image)
Figure 3.
Nutritional therapy vs. no intervention. Outcome; Overt hepatic encephalopathy (fixed-effect analysis).
Nutrition reduced bilirubin levels with 0.51 g/dl (95% CI, −0.54 to −0.48 g/dl, I = 41%, 10 trials) in the fixed-effect analysis, but not in the random-effects meta-analysis (−0.54 95% CI, −1.10 to 0.03). There was no effect on alanine amino transferase; 0.82 U/L (95% CI, −11.08 to 12.72 U/L, I = 30%; five trials), alkaline phosphatase −19.1 U/L (95% CI, −145.5 to 107.3 U/L, I = 67%, three trials), albumin; 0.13 g/dl (95% CI, −0.08 to 0.19 g/dl, I = 75%, 10 trials) or other nutritional parameters as triceps skinfold test (TSF); 1.36 mm (95% CI, −0.87 to 3.58 mm, I = 0%; two trials), mid arm circumference (MAMC); 0.92 cm (95% CI, 0.34 to 1.50 cm, I = 0%; three trials). None of the included trials reported data that allowed analyses of hepatorenal syndrome, body mass index or grip strength.
Source...