REVIEW ARTICLE Nutrition Vol. 15, No. 4, 1999

Conditionally Essential Fatty AcidDeficiencies in End-Stage Liver Disease

PETER A. BURKE, MD,* P. R. LING, MD,† R. A. FORSE, MD, PHD,*AND BRUCE R. BISTRIAN, MD, PHD†

From the Departments of *Surgery and †Medicine, Beth Israel Deaconess Medical Center,Harvard Medical School, Boston, Massachusetts, USA

Date accepted: 19 March 1998

ABSTRACT

Patients with end-stage liver disease (ESLD) manifest a wide variety of functional abnormalities that lead eventually to theirdeath. Such patients also often have low levels of long-chain polyunsaturated fatty acids (PUFA) of carbon length 20 or greaterin plasma total lipids, triacylglycerols, cholesterol esters, and phospholipids. We hypothesize that, due to hepatic damage, thereis an impairment in de novo synthesis of very long-chain (20–22) carbon PUFA from their essential fatty acid 18 carbon dietaryprecursors that normally takes place principally in the liver. This results in a “conditional” essential fatty acid deficiency thatmay, in fact, be responsible for some of the pathophysiologic effects in ESLD. We propose that direct supplementation with verylong-chain PUFA will provide a unique advantage in the correction of this “conditional” essential fatty acid deficiency in patientswith ESLD and lead to improvements in their clinical condition.Nutrition 1999;15:302–304. ©Elsevier Science Inc. 1999

Key words: end-stage liver disease, essential fatty acids, polyunsaturated fatty acids, arachidonic acid

Very low levels of arachidonic (C20:4v6, AA) and eicosa-pentaenoic acid (C20:5v3, EPA) are a well-established featureof advanced liver disease. We and others1–3 have demonstratedthat patients with end-stage liver disease (ESLD), particularlythose in Child-Pugh class C, characteristically have very lowlevels of AA, EPA, and docosahexaenoic acid (C22:6v3,DHA), in plasma lipids, and presumably, but not yet proven,lipids in tissue membranes compared with hospitalized controlswithout liver disease.

These very long-chain polyunsaturated fatty acids (PUFA)with 20 or greater carbons and two or more double bonds in theirstructure can be directly obtained from food, usually in limitedamounts, but are largely produced from their essential fatty acidprecursors, linoleic acid (LA) anda-linolenic acid (a-LA),through desaturation and elongation in the liver. These latter fattyacids, LA anda-LA, are considered to be essential, because AA,EPA, and DHA, which serve critical roles in human metabolism,can only be synthesized from these precursors. A poor essentialfatty acid intake in the setting of a generally inadequate foodintake has been considered as one of the major factors responsiblefor the reduced levels of AA and EPA found in patients withESLD. A lack of essential precursors certainly would be expectedto lead to a deficit in very long-chain PUFA analogous to the

production of classic essential fatty acid deficiency. Indeed, theseverity of the deficiency in the essential fatty acids and verylong-chain PUFA is related to the degree of overall malnutrition inthese patients.1 By this reasoning, adequate nutritional supplemen-tation of LA anda-LA should be able to correct both the essentialfatty acid deficiency and their more distal products, the verylong-chain PUFA. In ESLD patients, we recently observed thatintravenous nutritional support administered for 1 wk as totalparenteral nutrition supplemented with substantial amounts ofLA anda-LA was able to significantly increase the concentrationsof LA from 9.9 6 2.8 to 20.76 7.5% anda-LA from 0.0 6 0.0to 1.1 6 1.7% of the total plasma triacylglycerols. However,the levels of very long-chain PUFA, including AA, EPA, andDHA, remained at very low or undetectable levels (D. Duerksenand B. Bistrian, unpublished observations). For these patients,the total energy intake was 22–25 kcalz kg21 z d21 and1.5 gz kg21 z d21 of protein with 30% of total energy intake fromthe fat source. LA anda-LA comprised 54.5 and 8.3% of the totalfat calories respectively, which is substantially more than what isusually required to replete the essential fatty acid-deficient patient.A similar lack of effect on AA level has been reported after anacute intravenous infusion of linoleic acid that significantly in-creased the serum levels of this fatty acid but did not change the

Correspondence to: Peter A. Burke, MD, Beth Israel Deaconess Medical Center, 110 Francis Street, 8C, Boston, MA 02215, USA.

Nutrition 15:302–304, 1999©Elsevier Science Inc. 1999 0899-9007/99/$20.00Printed in the USA. All rights reserved. PII S0899-9007(99)00002-7

level of AA in the serum of patients with cirrhosis and ascites.4

These findings prompted another look at the likely factors oper-ating to produce the very long-chain PUFA deficiency in ESLD aswell as an evaluation of thepossibility that the clinical mani-festation of essential fatty acid deficiency could occur despitean adequate intake of LA anda-LA in ESLD. Given the manyimportant aspects of AA and EPA function as second messen-gers, either directly, or indirectly as eicosanoid precursors, it iscertainly possible that AA, EPA, and DHA could be condition-ally essential fatty acids in this setting.

Several lines of evidence support the proposition that thecritical factor determining the low levels of very long-chain PUFAin ESLD is not the dietary intake of LA anda-LA. First, toprevent essential fatty acid deficiency, the diet need only includeapproximately 0.5% of the total caloric requirement asa-LA and1.0% as LA or approximately 1 ga-LA and 2 g LA perday. Theaverage American diet provides 10–20 times more than theserequired amounts.5 Although the total energy intake in our seriesof ESLD patients was only 70% of control, the essential fatty acidintake was sufficiently high enough to prevent essential fatty aciddeficiency.3 A further piece of evidence pointing to a defect in thetheory that elongation and desaturation as the likely mechanismsfor low levels of AA, EPA, and DHA is the failure to identifyincreased amounts of Mead acid (C20:3,v9), which is produced inclassic essential fatty acid deficiency by elongation and desatura-tion of oleic acid (C18:1,v9) when LA anda-LA are limiting.Finally, the ratio of LA to AA in plasma triacylglycerols andcholesterol ester can be seen as a rough measure of the activity ofthe elongation/desaturation pathway. In the ESLD patients, thisratio was significantly higher than that seen in controls,3 suggest-ing an impairment in this pathway. For these several indirect butcomplementary reasons, we believe as do others that very long-chain PUFA deficiency in ESLD patients may be more closelyrelated to rates of very long-chain PUFA biosynthesis rather thaninadequate intakes of LA anda-LA.

The activity of liverD-6 andD-5 desaturases responsible forthe conversion of LA to AA,a-LA to EPA, and oleic acid to Meadacid is highly regulated and sensitive to both metabolic andendocrine influence. There are limited studies of desaturase activ-ity in man. Liver biopsies of prospective organ donors who hadbeen declared brain dead had low levels ofD-6 desaturase activity.Interestingly, the lowest activity was seen in those patients whoalso carried a diagnosis of alcoholism.6 There is other indirectevidence of impaired desaturase activity. Patients with chronichepatic encephalopathy, reflecting severe liver damage, have thelowest levels of AA and DHA in plasma compared with normalcontrols and ESLD patients without encephalopathy.7 Cabre etal.7 also have found increasedD-6 andD-5 desaturase precur-sor/product molar ratios in plasma lipids in human cirrhosis.8

Thus, the activity of liver desaturases is probably decreased bythe development of ESLD, although this has not been directlymeasured. Other mechanisms could contribute to the lowerlevels of AA and PUFA observed in ESLD. For instance,pathologic shunting bypassing the liver in ESLD could reducethe conversion of dietary LA anda-LA to AA or other long-chain PUFA but the effect should be corrected by intravenouslipid therapies. AA levels could be decreased secondary toincreased utilization through cyclooxygenase pathways due tothe disease itself. However, given the very limited concentra-tions of eicosanoids, this seems an unlikely explanation. In-creases in lipid peroxidation and AA metabolism by the cyto-chrome P450 system secondary to liver disease could also

contribute to changes in AA levels.9,10 However, the lack ofMead acid production and the reduction in AA, EPA, and DHAseen in ESLD patients make a central synthetic defect morelikely. The preceding chain of evidence led us to hypothesizethat very long-chain PUFA become conditionally essential fattyacids in ESLD patients. There is an analogous clinical condi-tion, that of preterm infants and neonates, where very long-chain PUFA are limited in production despite adequate LA anda-LA intake because of an immature enzyme system. There areassociated defects in intellectual and visual function that can beprevented by providing an ample supply of EPA and DHA.11

We hypothesize that a similar supplementation not only of EPAor DHA, or both, but also AA will similarly bypass the desatu-ration impairment in patients with ESLD.

What are the possible consequences of conditionally essentialfatty acid deficiency and its correction in ESLD? Very long-chainPUFA are prominent components of all cell membranes and serveas precursors for the production of various eicosanoids that areubiquitous second messengers as well as for the production of AA,which is itself a second messenger. The clinical importance of AAis illustrated by studies in preterm infants where lower AA levelsinduced by feeding EPA only, which reversibly impairs desatura-tion of LA, slowed growth,11 and in liver disease patients wherethe presence of AA deficiency was significantly correlated withmortality risk.12 DHA is particularly concentrated in the synapticmembranes and the microsomes of gray matter, and in the mem-branes of the rod at the outer segments of the retina where it isimportant to the normal development of vision and brain growth13

but presumably quite resistant to change in adulthood due to slowturnover of brain lipids. However, AA deficiency may contributeto the increased susceptibility to infection and impaired woundhealing characteristic of both classic essential fatty acid deficiencyand ESLD. Other characteristics of ESLD such as a poor proteinanabolic response, renal dysfunction, changes in fluid homeosta-sis, and increased caloric expenditure are also seen in classicessential fatty acid deficiency. These similarities are consistentwith the hypothesis of a conditional essentiality of very long-chainPUFA in ESLD.

Previously, there has been a limited ability to fully test thishypothesis because concentrated natural sources of AA were notavailable. However, the recent production of enriched sources oflong-chain PUFA (EPA and DHA from fish oil or fungal oil andAA from fungal oil) now make this possible. Previously, it hasbeen shown in experimental models of liver disease in the rat fedfat-free ethanol-containing diets that the addition of small amountsof AA can prevent the decrease in AA plasma levels.14 Morerecently, we have shown that the EPA-induced reduction in AA inserum phospholipids can be corrected by AA supplementation asa triacylglycerol with functional consequences as well.15 It hasbeen stated that the enteral provision of long-chain PUFAs tomalnourished cirrhotic patients in amounts as high as 6% of totalcalories for 3 wk failed to change fasting PUFA plasma levels.1

However, the details of these studies are not available and makecomment difficult. We believe that the repair of the very long-chain PUFA deficiency in ESLD patients should lead to significantimprovements in various physiologic functions, which might beexpected to have a favorable impact on clinical status and poten-tially on survival. If proven correct, a more general application ofthese principles might provide a novel nutritional therapy tomodify the clinical management as well as the course of chronicliver disease.

REFERENCES

1. Cabre E, Gassull MA. Polyunsaturated fatty acid deficiency in liverdiseases: pathophysiological and clinical significance. Nutrition 1996;12:542

2. Gonzalez J, Periago JL, Gil A, et al. Malnutrition-related polyunsat-urated fatty acid changes in plasma lipid fraction of cirrhotic patients.Metabolism 1992;41:954

LIPIDS IN END-STAGE LIVER DISEASE 303

3. Palombo JD, Lopes SM, Jenkins RL, et al. Effectiveness of ortho-topic liver transplantation on the restoration of cholesterol metab-olism in patients with end stage liver disease. Gastroenterology1987;93:1170

4. Rimola A, Gines P, Cuso E, et al. Prostaglandin precursor fatty acidsin cirrhosis with ascites: effect of linoleic acid infusion in functionalrenal failure. Clin Sci 1988;74:613

5. James P, Norman K, Rosenberg I. The nutrition of fat. Meetingsummary. Nutr Rev 1992;50:68

6. Spielmann D. Metabolism of unsaturated fatty acids role of n-3 andn-6 fatty acids in clinical nutrition. In: Kinney JM, Borum PR, eds.Perspectives in clinical nutrition.Munich: Urban and Schwarzenberg,1989:351

7. Cabre E, Periago JL, Gonzalez J, et al. Plasma polyunsaturated fattyacids in liver cirrhosis with or without chronic hepatic encephalopa-thy: a preliminary study. JPEN 1992;16:359

8. Cabre E, Periago JL, Gonzalez J, et al. Plasma fatty acid profile inadvanced cirrhosis: unsaturation deficit of lipid fractions. Am J Gas-troenterol 1990;85:1597

9. French SW, Morimoto M, Reitz RC, et al. Lipid peroxidation,

CYP2E1 and arachidonic acid metabolism in alcoholic liver disease inrats. J Nutr 1997;127:907S

10. Daikh BE, Lasker JM, Raucy JL, Koop DR. Regio and stereoselectiveepoxidation of arachidonic acid by human cytochromes P450 2C8 and2C9. J Pharmacol Exp Ther 1994;27:1427

11. Carlson SE, Werkman SH, Tokey EA. Effect of long-chain n-3 fattyacid supplementation on visual acuity and growth of preterm infantswith and without bronchopulmonary dysplasia. Am J Clin Nutr 1996;63:687

12. Cabre E, Abad Lacruz A, Nunez MC, et al. The relationship of plasmapolyunsaturated fatty acid deficiency with survival in advanced livercirrhosis, multivariate analysis. Am J Gastroenterol 1993;88:718

13. Martinez M. Tissue levels of polyunsaturated fatty acids during earlyhuman development. J Pediatr 1992;120:S129

14. Goheen SC, Larkin EC, Rao GA. Severe fatty liver in rats fed afat-free ethanol diet and its prevention by small amounts of dietaryarachidonate. Lipids 1983;18:285

15. Ling PR, Boyce P, Bistrian BR. Role of arachidonic acid in theregulation of the inflammatory response in TNF-treated rats. JPEN1998;22:S1

LIPIDS IN END-STAGE LIVER DISEASE304