The Journal of Nutrition Vol. 128 No. 2 February 1998, pp. 198-203

Moderate and Large Doses of Ethanol Differentially Affect Hepatic Protein Metabolism in Humans^1,2

Elena Volpi, Paola Lucidi, Guido Cruciani, Francesca Monacchia, Stefania Santoni, Gianpaolo Reboldi, Paolo Brunetti, Geremia B. Bolli, and Pierpaolo De Feo³

Department of Internal Medicine, Endocrine and Metabolic Sciences, University of Perugia, 06126 Perugia, Italy

	ABSTRACT

Abstract Introduction Methods Results Discussion References

The intake of ~70 g of alcohol impairs liver protein metabolism in healthy humans. To establish the threshold at which alcohol impairs hepatic protein metabolism in humans we compared the effects of 500 mL of water (control study), 300 (28.4 g ethanol) or 750 mL (71 g ethanol) of table wine on hepatic protein metabolism in three groups of healthy nonalcoholic volunteers. Hepatic protein metabolism was estimated (L-[1-¹⁴C]leucine infusion) by measuring the fractional secretory rates of albumin and fibrinogen during the overnight postabsorptive state (basal) and the subsequent administration of water or two different amounts of wine (300 or 750 mL) given with a liquid glucose-lipid-amino acid meal. During the meal, water did not affect fibrinogen fractional secretory rate and increased albumin fractional secretory rate by ~50% (P < 0.01). The 300 mL of wine increased albumin secretory rate by only ~20% (P < 0.01 vs. basal, P < 0.04 vs. water) and did not affect fibrinogen secretory rate. The 750 mL of wine profoundly impaired hepatic protein metabolism, decreasing the fractional secretory rates of albumin (P < 0.01 vs. water and 300 mL wine) and fibrinogen (P < 0.04 vs. water and 300 mL of wine) below the postabsorptive values. These results demonstrate that a moderate dose of alcohol (28 g, ~2 drinks) slightly affects postprandial hepatic protein metabolism by blunting the meal-induced increase in albumin synthesis, whereas it does not interfere with fibrinogen synthesis as do higher doses.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

There is growing evidence that drinking moderate amounts of alcohol may benefit the cardiovascular system by improving the lipid profile (Gaziano et al. 1993), insulin sensitivity (Facchini et al. 1994) and reducing the risk of myocardial infarction (Gaziano et al. 1993). However, before the intake of moderate amounts of alcohol can be recommended to healthy subjects, a "moderate" dose must be established. Although many physicians usually recommend as "moderate" a dose of alcohol <30 g/day, no verification of this assumption exists.

Using labeled leucine to measure the fractional secretory rates (FSR)⁴ of hepatic and extrahepatic plasma proteins in healthy humans, we have recently shown that the acute intake of a moderately large amount of ethanol, ~70 g (750 mL wine), severely impairs the synthesis and/or the secretion of liver proteins in spite of the absorption of an amino acid-enriched meal (De Feo et al. 1995).

The toxic effect of acute alcohol intake on liver protein metabolism is a direct consequence of ethanol oxidation, which takes place mainly in the hepatocytes (Lieber 1980), and results in the production of reducing equivalents and metabolites (French 1989, Lieber 1980, Volpi et al. 1997). This effect alters not only protein synthesis and/or secretion (De Feo et al. 1995, Volpi et al. 1997) but also glucose (Madison et al. 1967) and lipid metabolism (French 1989, Lieber 1980). Although the mechanisms of ethanol toxicity have been extensively studied in animal models and cell cultures (French 1989, Lieber 1980), the overall capacity of the normal human liver to safely metabolize ethanol remains unknown, i.e., the maximum amount of alcohol that can be oxidized without interfering with the metabolism of other substrates. We hypothesize that the daily consumption of an amount of ethanol that does not acutely affect liver metabolism will not lead to the development of alcohol-induced liver injury.

Because liver protein metabolism is very sensitive to acute alcohol intake (De Feo et al. 1995, Volpi et al. 1997), the measurement of the synthesis and/or secretion of secretory liver proteins can be a useful tool to establish the threshold dose for the adverse effects of alcohol on hepatic metabolism. Thus we compared the effects of the intake of water (control) with those of moderate (300 mL) or large (750 mL) amounts of wine (28.4 and 71 g ethanol, respectively) on hepatic protein metabolism in three groups of normal volunteers. To better simulate social drinking, water or wine as given during the intragastric (ig) infusion of a glucose-lipid-amino acid meal that was similar in energy content and composition to a healthy lunch.

	SUBJECTS AND METHODS

Abstract Introduction Methods Results Discussion References

Protocol.  After receiving Institutional Review Board approval, written informed consent was obtained from 15 healthy volunteers (4 women, 11 men), with normal physical examinations, routine blood analyses and without any serological evidence of viral hepatitis. All of them reported to be occasional consumers of moderate amounts of alcoholic beverages (<110 g ethanol/wk). The subjects were divided into water (Control, n = 5), low ethanol (LE, n = 5) and high ethanol (HE, n = 5) groups, matched for age (Control 24 ± 1, LE 25 ± 2, HE 24 ± 1 y) and body mass index (Control 21 ± 1, LE 23 ± 2, HE 23 ± 1 kg/m²). All subjects were studied 3 d after consuming a weight maintenance diet of 146 kJ/(kg·d) containing 55, 30 and 15% carbohydrate, fat and protein, respectively, and without drinking alcoholic beverages. After overnight fasting, at ~0730 h, the volunteers were admitted to the Clinical Research Unit of the Dipartimento di Medicina Interna e Scienze Endocrine e Metaboliche of the University of Perugia. At ~0800 h, a 20-gauge plastic catheter needle was placed in an antecubital vein for the infusion (Harvard syringe pump, Harvard Apparatus, Ealing, South Natick, MA) of [1-¹⁴C]leucine (specific activity 2 MBq/mmol, Amersham International, Buckinghamshire, England) and saline (0.5 mL/min, Vial Médical pump, Grenoble, France). A contralateral hand vein was cannulated in a retrograde fashion with a 19-gauge butterfly needle and the hand was maintained at 65°C in a thermoregulated Plexiglas box to permit intermittent sampling of arterialized venous blood. A feeding nasogastric tube was inserted for ig meal infusion. At ~0900 h (0 min), a primed-constant intravenous infusion of L-[1-¹⁴C]leucine [prime: 333 kBq (9 µCi), infusion rate: 11.1 kBq/min (0.3 µCi/min)], was started and continued for 8 h. At 240 min, after blood and breath sampling, the ig infusion of a liquid mixed meal was started at a rate of 1.75 mL/min and continued throughout the study (Vial Médical). The meal provided 17% of total energy from amino acids, 33% from lipids and 50% from carbohydrates, for a total of 2642 kJ. It was prepared by mixing a complete formula (Isopuramin Plus 10%, Bieffe Medical, Modena, Italy) of non-essential and essential amino acids (unlabeled leucine infusion rate is reported in Table 1) with 84 g glucose and a mixed oil solution (Lipofundin S, B. Braun, Melsungen, Germany). When the ig infusion was started, 140 mL of mineral water (Santa Chiara, Motette srl, Scheggia, Italy) (Control group), 60 (LE group) or 150 (HE group) mL of table white wine (12% v/v ethanol, Pergoleto Lungarotti, kindly offered by Cantine Lungarotti srl, Torgiano, Italy) as given to the volunteers. Then, 40 mL of water or 20 (LE group) or 50 (HE group) mL of wine was given every 15 min from 255 to 420 min, for a total amount of 500 mL of water to the Control group, 300 mL of wine to the LE group and 750 mL of wine to the HE group. Because the meal composition of the water and ethanol studies was the same, the three studies differed in energy content (852 and 2130 kJ, LE and HE groups, respectively) as a result of ethanol (LE 28.4, HE 71 g of ethanol, 30 kJ/g). This variable was preferred to the isocaloric substitution of other fuels with ethanol, because a reduced supply of glucose and/or lipids could directly affect leucine metabolism (De Feo and Haymond 1992), thereby complicating the interpretation of the results.

View this table: [in this window] [in a new window]   Table 1. Tracer infusion rates, plasma -ketoisocaproate (KIC) specific activities, intragastric (ig) unlabeled leucine infusion rates and whole-body leucine kinetics in three groups of normal subjects during the postabsorptive state (Basal) and during the intragastric infusion of a mixed meal (Meal) given with either water (Control) or 300 mL (LE) or 750 mL (HE) wine1,2

Blood (16 mL) and breath samples were collected at 15, 0, 60, 120, 180, 200, 220, 240, 300, 360, 420, 440, 460 and 480 min to measure the plasma concentrations of glucose, lactate, pyruvate, insulin, isoleucine, leucine, -ketoisocaproic acid (KIC), albumin, fibrinogen, the plasma specific activity (SA) of leucine and KIC, the rates of expired total CO₂, total ¹⁴CO₂ and CO₂ specific activity and the SA of leucine deriving from albumin and fibrinogen hydrolysis. Plasma ethanol concentration (1 mL of blood) was measured every 30 min from 240 to 480 min.

Analytical methods.  The plasma concentrations of glucose, insulin, albumin, fibrinogen, ethanol, lactate and pyruvate were determined as previously described (De Feo et al. 1995, Holsen 1971). The plasma concentrations of isoleucine, leucine and KIC were determined by HPLC; the SA of plasma leucine and KIC, and the SA of leucine derived from hydrolyzed albumin and fibrinogen were determined by HPLC separation and liquid scintillation; the excretion and SA of expired CO₂ were determined by means of liquid scintillation (De Feo et al. 1995, Horber et al. 1989).

Calculations.  The estimates of whole-body leucine kinetics were determined from the data obtained during the last hour of each study period (basal postabsorptive state: 180-240 min; absorptive state: 420-480 min) at the isotopic and metabolic steady state, by using the four-compartment model (De Feo and Haymond 1994). The rate of total leucine appearance [Total Ra, µmol/(kg·min)] was calculated by using the following formula:

(1)

where i is the labeled leucine infusion rate [dpm/(kg·min)] and SA_KIC is the plasma KIC specific activity (dpm/µmol). During meal administration, the rate of endogenous leucine appearance [Endogenous Ra, µmol/(kg·min)], an index of whole-body proteolysis, was calculated as follows:

(2)

where D_Leu is the ig leucine infusion rate. The rate of leucine oxidation [Ox, µmol/(kg·min)] was calculated according to the precursor-product model:

(3)

where CO₂ is the ¹⁴CO₂ excretion rate [dpm/(kg·min)] and  is the correction factor for the CO₂ recovery, assuming values of 0.70 and 0.82 in the basal postabsorptive and absorptive states (Volpi et al. 1996), respectively. The rate of nonoxidative leucine disposal [NOLD, µmol/(kg·min)], index of whole-body protein synthesis and endogenous balance [Balance, µmol/(kg·min)] were estimated as follows:

(4)

The fractional secretory rates (FSR, %/h) of plasma proteins were calculated according to the precursor-product model (Schwenk et al. 1985):

(6)

where SA_Leu protein/t is the incorporation rate of labeled leucine into proteins from 180 to 240 min (basal postabsorptive state) and from 420 to 480 min (absorptive state), and SA_KIC plasma is the mean plasma KIC SA during the same time periods. The assumptions for the measurement of plasma protein FSR have been recently discussed in detail (De Feo and Haymond 1994, De Feo et al. 1995, Volpi et al. 1996). Experimental evidence suggests that, under the experimental conditions of this study, plasma KIC SA is a reliable index of the precursor pool SA for hepatic protein synthesis (De Feo et al. 1995).

Statistics.  Statistical analysis was performed with SAS/STAT software release 6.11 (SAS Institute, Cary, NC). The effect of treatment on the response variables in the postabsorptive and absorptive states was analyzed using the General Linear Models procedure with or without correction for repeated measures (Winer 1972) as appropriate. Specific contrast matrices (planned comparisons method) were constructed to evaluate differences among group means. Data are expressed as means ± SEM and percentage of change for albumin and fibrinogen FSR. Linearity of the model describing label incorporation into plasma proteins over time was tested according to the method suggested by Snedecor and Cochran (1980).

	RESULTS

Abstract Introduction Methods Results Discussion References

Plasma concentrations of ethanol, substrates and insulin.  The plasma ethanol concentrations were undetectable at 240 min in the three groups (Fig. 1). With subjects in the absorptive state, the concentrations remained undetectable in the Control group, whereas they increased significantly in the LE and HE groups (P < 0.001 time × group interaction, P < 0.001 vs. Control) with a significant difference between these two groups (P < 0.01 HE vs. LE).

View larger version (24K): [in this window] [in a new window]   Fig 1. Plasma ethanol, glucose and insulin concentrations and lactate/pyruvate ratio in three groups of healthy volunteers in the basal postabsorptive state (0-240 min) and during the intragastric infusion of a glucose-lipid-amino acid meal combined with the intake of either 500 mL of water or 300 or 750 mL of wine (12% v/v ethanol). Values are means ± SEM, n = 5. The plasma glucose and insulin concentrations did not differ among the three groups, whereas the plasma concentrations of ethanol and the lactate/pyruvate ratios were increased by the intake of 300 mL of wine (P < 0.01 vs. Control group) and additionally increased by intake of 750 mL of wine (P < 0.01 HE vs. LE).

With subjects in the basal postabsorptive state, the plasma lactate/pyruvate ratio did not differ in the three groups (Control 11.1 ± 0.3, LE 9.7 ± 0.2 and HE 11.0 ± 0.1 mol/mol). During meal absorption, the plasma lactate/pyruvate ratio increased by 100% during the ingestion of 300 mL of wine and twofold during the intake of 750 mL of wine with a significant difference between these two groups (Control 9.4 ± 0.2, LE 18.9 ± 0.1 and HE 29.6 ± 1.0 mol/mol, P < 0.001 time × group interaction, P < 0.001 Control vs. LE and HE, P < 0.01 HE vs. Control and LE).

The plasma concentrations of glucose and insulin were increased by meal absorption (P < 0.001 vs. basal), without differences among the three groups.

In each group, the plasma concentrations of leucine and KIC were at steady state over the last hour of the basal postabsorptive (180-240 min) and absorptive (420-480 min) periods. During the postabsorptive period, the plasma concentrations of leucine (180-240 min: Control 106 ± 7, LE 122 ± 11 and HE 113 ± 8 mmol/L), isoleucine (Control 50 ± 6, LE 49 ± 7 and HE 52 ± 3 mmol/L) and KIC (Control 41 ± 2, LE 44 ± 2 and HE 43 ± 5 mmol/L) were not different among the three groups. Meal administration increased the plasma concentrations of leucine (420-480 min: Control 160 ± 3, LE 223 ± 5 and HE 211 ± 15 mmol/L, P < 0.001 vs. basal) and isoleucine (Control 95 ± 13, LE 136 ± 21 and HE 144 ± 11 mmol/L, P < 0.001 vs. basal) in the three groups, with a greater increment in the HE and LE groups (P < 0.01 vs. Control). During meal intake, the plasma KIC concentration increased in the LE (60 ± 4 mmol/L) and HE (58 ± 8 mmol/L) groups (P < 0.05 vs. Control) and remained unchanged in the Control group (37 ± 2 mmol/L).

Leucine kinetics.  Expired CO₂, plasma KIC and leucine SA were at steady state over the last hour of the basal postabsorptive (180-240 min) and absorptive (420-480 min) periods (Table 1). In the postabsorptive state, leucine kinetics were not different among the three groups. Meal administration decreased the rate of endogenous leucine appearance (P < 0.001 vs. basal) and increased the rate of nonoxidative leucine disappearance and the endogenous balance (P < 0.02 vs. basal), without significant differences among the three groups. The rate of leucine oxidation was increased by meal infusion in the three groups (P < 0.001 vs. basal); however, wine intake (LE and HE groups) significantly blunted this increment (P < 0.03, LE and HE vs. Control).

Concentrations and FSR of plasma proteins.  The plasma concentrations of albumin and fibrinogen were not different in the three groups and were unaffected by meal administration (data not shown). The SA of leucine derived from hydrolyzed proteins increased linearly (P < 0.001) from 180 to 240 min and from 420 to 480 min. A straight-line fit was adequate to describe the time course of label incorporation for the data from each subject during the three studies. In the postabsorptive state, the FSR of albumin and fibrinogen were not different in the three groups (Table 2). Meal administration increased albumin FSR by 46.5% in the Control group (P < 0.01) and by 19.5% in the LE group (P < 0.01 vs. Control), whereas it decreased albumin FSR by 20.2% below the basal values in the HE group (P < 0.01 vs. Control and LE groups) (Fig. 2). During meal administration, fibrinogen FSR decreased below the basal values by 11.8% in the HE group, whereas it increased by 14.5 and by 13.7% in the Control and LE groups, respectively (P < 0.04 HE vs. Control and LE).

View larger version (25K): [in this window] [in a new window]   Fig 2. Percentage change from basal (overnight postabsorptive) state of the fractional secretory rates (FSR) of albumin and fibrinogen in three groups of healthy volunteers after the intragastric administration of a glucose-lipid-amino acid meal combined with either 500 mL of water or 300 or 750 mL of wine (12% v/v ethanol). Data are means ± SEM, n = 5. *P < 0.01 vs. others, P < 0.01 vs. Control, P < 0.04 vs. Control and LE, P < 0.01 vs. Control and LE.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

These results demonstrate that postprandial hepatic protein metabolism of healthy subjects is slightly influenced by the intake of 300 mL of wine (28.4 g ethanol), but profoundly impaired by that of 750 mL of the same wine (71 g ethanol). In contrast to the high dose of wine, which depressed the FSR of albumin and fibrinogen, the low dose only blunted the physiologic postprandial increase of albumin FSR (Fig. 2). Both low and high wine intakes had similar effects on leucine estimates of whole-body protein metabolism. The oxidation of ethanol and its substrates, probably through a mechanism of substrate competition, reduced leucine oxidation and consequently increased plasma leucine concentration. However, this effect was not sufficient to further improve postprandial protein anabolism because leucine estimates of whole-body protein synthesis and net balance did not differ significantly among the three groups.

Based on in vitro data, the reduction of postprandial albumin and fibrinogen FSR observed after wine ingestion in this and in our previous studies (De Feo et al. 1995, Volpi et al. 1997) is due to a reduction in either the synthesis or secretion of liver proteins (Lieber 1980). In normal nonhabitual alcohol consumers, such as our volunteers, the intra-hepatic oxidation of ethanol is carried out by alcohol dehydrogenase, and the resulting acetaldehyde is subsequently oxidized to acetate by aldehyde dehydrogenase (Lieber 1980). These oxidative processes lead to an excessive production of reducing equivalents in the form of NADH (French 1989, Lieber 1980), thus increasing the NADH/NAD⁺ ratio. The alteration in the hepatocyte redox state decreases the activity of the glycolytic pathway (Berry et al. 1994) and citric acid cycle (Baraona et al. 1980), limiting ATP availability (Akinshola et al. 1991, Masson et al. 1993) for protein synthesis. The concept that the shift in the redox state mediates the effects of ethanol on protein synthesis is supported by the fact that, in isolated hepatocytes, ethanol inhibition of protein synthesis can be prevented by the addition of an acceptor of reducing equivalents (methylene blue), which restores the NADH/NAD⁺ ratio to normal values (Baraona et al. 1980). Furthermore, we have recently confirmed this hypothesis in a recent study performed in vivo in humans (Volpi et al. 1997), in which the administration of nicotinamide during high dose ethanol intake (~70 g ethanol) prevented the reduction of albumin and fibrinogen FSR by ameliorating the hepatic redox state. However, from our data, it is not possible to rule out an effect of ethanol on protein secretion through its oxidative product acetaldehyde, a highly reactive molecule that can form adducts with the sulfhydryl groups of proteins (Tuma et al. 1987). When tubulin, the contractile protein necessary for cell secretion, is involved in adduct formation, the result is an impairment in its polymerization that reduces the secretive process (Tuma et al. 1987). In any case, this possibility is unlikely, because the liver aldehyde dehydrogenase enzyme system is characterized by a high capacity and low K_m, so that hepatic and blood concentrations of acetaldehyde remain very low (Newsholme and Leech 1983).

The effects of ethanol on the NADH/NAD⁺ ratio can be indirectly estimated by measuring the plasma lactate/pyruvate ratio (French 1989). In our control study, the lactate/pyruvate ratio was ~10 during both the postabsorptive and absorptive periods; the low dose of ethanol increased it to ~20, and the high dose of ethanol increased it to ~30. Because the low ethanol group showed just a slight blunting in the prandial increase of albumin FSR, i.e., albumin FSR increased but to a lesser extent than the water group, whereas fibrinogen FSR was unaffected, a lactate/pyruvate ratio of 20 as a result of alcohol consumption might be considered the threshold for ethanol interference with liver protein metabolism.

The differential effects of the low dose of wine on postprandial albumin and fibrinogen FSR might be due to the fact that during meal absorption, in contrast to fibrinogen secretion that proceeds at the same rate, albumin secretion increases by ~50% (Volpi et al. 1996) and consequently should be more sensitive to a decrease in ATP production and/or in the efficiency of the protein export system.

The results of this study suggest that prandial liver protein metabolism is differentially affected by the intake of the two doses of ethanol. The dose of 71 g decreased albumin and fibrinogen FSR, whereas one of 28 g blunted albumin but not fibrinogen FSR. This suggests that, during meal absorption, 28 g is the highest dose of ethanol that the liver of normal subjects can metabolize without substantial interference with local protein metabolism. The fact that all of the numerous factors involved in the development of alcoholic liver injury (e.g., accumulation of triglycerides into the hepatocytes or perivenular hypoxia, reviewed by French 1989, Lieber 1980, Tuma et al. 1987) are related to the interference caused by intrahepatic alcohol metabolism might suggest that a daily intake of ethanol < 28 g (which corresponds to ~2 drinks of alcoholic beverages) should not lead to liver injury because this amount induces only minimal alterations of liver metabolism. However, caution must be used before drawing the conclusion that the intake of this moderate amount of alcohol is safe. In fact, the nature of the long-term effect of blunting (28 g ethanol) and/or suppressing (71 g ethanol) postprandial albumin FSR on liver function remains unknown.

	ACKNOWLEDGMENTS

The authors are indebted to Vania Cesarini and Giampiero Cipiciani for skillful technical assistance.

	FOOTNOTES

Manuscript received 18 July 1997. Initial reviews completed 20 August 1997. Revision accepted 20 October 1997.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Akinshola B. E., Potter J. J., Mezey E. Ethanol increases the formation of NADP⁺ in rat hepatocytes. Hepatology 1991; 13:509-514 [Medline]
• Baraona E., Pikkarainen P., Salaspuro M., Finkelman F., Lieber C. S. Acute effects of ethanol on hepatic protein synthesis and secretion in the rat. Gastroenterology 1980; 79:104-111 ^[Medline]
• Berry M. N., Gregory R. B., Grivell A. R., Phillips J. W., Schon A. The capacity of reducing-equivalent shuttles limits glycolysis during ethanol oxidation. Eur. J. Biochem. 1994; 225:557-564 [Medline]
• De Feo, P. & Haymond, M. W. (1992) Effects of substrates on protein metabolism in humans. In: Protein Metabolism in Diabetes Mellitus (Nair, K. S., ed.), pp. 199-206. Eldred Smith-Gordon, London, UK.
• De Feo P., Haymond M. W. Principles and calculations of the labelled leucine methodology to estimate protein kinetics in humans. Diabetes Nutr. Metab. 1994; 7:165-184
• De Feo P., Volpi E., Lucidi P., Cruciani G., Monacchia F., Reboldi G., Santeusanio F., Bolli G. B., Brunetti P. Ethanol impairs post-prandial hepatic protein metabolism. J. Clin. Invest. 1995; 95:1472-1479
• Facchini F., Chen I.Y.-D., Reaven G. M. Light-to-moderate alcohol intake is associated with enhanced insulin sensitivity. Diabetes Care 1994; 17:115-119 [Abstract]
• French S. W. Biochemical basis for alcohol-induced liver injury. Clin. Biochem. 1989; 22:41-49 [Medline]
• Gaziano M. J., Buring J. E., Breslow J. L., Goldhaber S. Z., Rosner B., VanDenburgh M., Willett W., Hennekens C. H. Moderate alcohol intake, increased levels of high-density lipoprotein and its subfractions, and decreased risk of myocardial infarction. N. Engl. J. Med. 1993; 329:1829-1834 ^{[Abstract/Free Full Text]}
• Holsen C. An enzymatic fluorimetric micromethod for the determination of acetoacetate, -hydroxybutirate, pyruvate and lactate. Clin. Chim. Acta 1971; 33:293-300 ^[Medline]
• Horber F. F., Kahl J., Lecavalier L., Krom B., Haymond M. W. Determination of leucine and -ketoisocaproic acid concentrations and specific activity in plasma and leucine specific activities in proteins using high-performance liquid chromatography. J. Chromatogr. 1989; 495:81-94 [Medline]
• Lieber C. S. Alcohol, protein metabolism, and liver injury. Gastroenterology 1980; 79:373-390 ^[Medline]
• Madison L. L., Lochner A., Wulff J. Ethanol induced hypoglycemia. Mechanism of suppression of hepatic gluconeogenesis. Diabetes 1967; 16:252-258
• Masson S., Desmoulin F., Sciaky M., Cozzone P. J. Catabolism of adenine nucleotides and its relation with intracellular phosphorylated metabolite concentration during ethanol oxidation in perfused rat liver. Biochemistry 1993; 32:1025-1031 [Medline]
• Newsholme, E. A. & Leech, A. R. (1983) Biochemistry for the Medical Sciences, pp. 472-478. John Wiley & Sons, Chichester, UK.
• Schwenk W. F., Beaufrere B., Haymond M. W. Use of reciprocal pool specific activities to model leucine metabolism in humans. Am. J. Physiol. 1985; 249:E646-E650 [Abstract/Free Full Text]
• Snedecor, G. W. & Cochran, W. G. (1980) Statistical Methods, 7th ed. Iowa State University Press, Ames, IA.
• Tuma D. J., Jennett R. B., Sorrell M. F. The interaction of acetaldehyde with tubulin. Ann. N.Y. Acad. Sci. 1987; 492:277-286 [Medline]
• Volpi E., Lucidi P., Cruciani C., Monacchia F., Reboldi G., Brunetti P., Bolli G. B., De Feo P. Contribution of amino acids and insulin to protein anabolism during meal absorption. Diabetes 1996; 45:1245-1252 [Abstract]
• Volpi E., Lucidi P., Cruciani C., Monacchia F., Reboldi G., Brunetti P., Bolli G. B., De Feo P. Nicotinamide counteracts alcohol-induced impairment of hepatic protein metabolism in humans. J. Nutr. 1997; 127:2199-2204 [Abstract/Free Full Text]
• Winer, B. J. (1972) Statistical Principles in Experimental Design, 2nd ed., pp. 261-308. McGraw-Hill, New York, NY.

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences