Does a "moderate" alcohol intake damage the brain?

Journal ofNeurology,

Neurosurgery,

and

PsYchiatrY 1988,51:909-913

Does

a

"moderate" alcohol intake

damage the

brain?

CLIVE HARPER,*t JILLIAN KRIL,* JOHN DALY:

FromtheDepartmentof Pathology,* TheUniversityofSydney, Sydney, NSW,DepartmentofAnatomical Pathology, RoyalPrinceAlfredHospital,t NSW, Departmentof Neuropathology, Royal PerthHospital,4 Perth, WA, Australia

SUMMARY

A

range

of volumetric and neurochemical

analyseswerecarriedoutonthe brains of 14

"moderate" drinkers. These data

were compared with previous studies from controls and

alcoholics.

Volumetric

measurements

consistently

suggest a loss of cerebral tissue although the

differences

were not

statistically significant. Changes in

the lipid and water content ofthe white matterwere

noted and the significance of

thesechanges in relationtothe pathogenesisof reversible

brain

shrinkage caused by alcohol abuse is discussed.

There is aconsiderable volume ofclinical,1 2

neuro-radiological,34

neuropsychological'

and

neuropath-ological6- dataconcerning thelongtermeffects of

alcohol abuse on the human central nervous system.

A large proportion ofthese data, particularly with

regard to neuropathological changes, deals with

severechronic alcoholism. Theinterpretation of such

data is difficult becauseof thefrequency ofoccurrence

of attendant disorders such as cirrhosis of the liver

and nutritional deficiency disorders (Wernicke's

encephalopathy)inthe moreadvancedcasesof

alco-holism. Ofparticular interesttothe general publicis

the long term effects ofa moderate alcohol intake

whichmany would equate with "social" drinking.

Materials and methods

Casesfor the studies cited in this paperwereselectedfrom thePerthCity Coroner's Department and the Royal Perth and Royal Prince Alfred Hospitals.,The authors classified thecasesaccording tothefollowing criteria:

1 Controls: teetotalorless than 20 grams of ethanol per day.

2 Moderate: 30-80grams perday.

3 Alcoholic: greaterthan 80 grams perday but without additionalmedicalproblems suchascirrhosisorWernicke's encephalopathy(WE).

4 Alcoholic and cirrhosis: as for (3) pluspathologically confirmed alcoholic cirrhosis(without WE).

5 Alcoholic and WE: as for (3) plus pathologically

confirmedWernicke's encephalopathy (without cirrhosis). (Approximately lOgof ethanol is contained in one public

Addressforreprintrequests:Professor CGHarper,Department of Pathology, TheUniversityofSydney, Sydney, NSW 2006, Australia.

Received 17 November 1987 and in revisedform18February1988.

Accepted7March 1988

house serveof alcohol that is, halfapint (284 ml) of beer, 4floz(110ml)wine, Ifloz(28 ml)spirits.10)

Categorisation of individualcases wasbasedon a

com-pilation of clinical and pathological data. These data included clinicalnotesfrom theteaching hospitalsas aresult of previous admissions, detailed questionnaires on the alcohol intakeprovided by the relatives of the subjects and

acomplete necropsy withmicroscopic examination of tissues including the liver and brain. Ourdefinition of the moderate drinking group (20-80g ofalcohol/day) is not meant to

correlatewith oftquoted clinical definitions of moderateor

social drinking. In the most recent literature figures vary from 30to70grams perday." 12 In the recentreporton

"Is therea safe level ofdaily consumption of alcohol formen

and women?" from the Australian National Health and Medical Research Council, the authorsrecommendthat a

safe daily intake is 40 grams for men and 20 grams for women.'3 From a pathological point of view it is an

extremely difficult groupto categoriseonthe basis of the retrospective collectionof data ondrinkinghabits. Thus it

was necessary to utilise a wider range of alcohol intake (30-80g/day) than is generally quoted in clinical papers. Patients withahistoryorpathological evidence ofa

neuro-logicaldisease, other than those associated withalcoholism,

wereexcludedas were anypatients with evidence of head injury. All the cases in these studies were male. Data on

femalecaseshave been collected but there isaninadequate numberofcases for statisticalanalyses. Separation of data

on a sexbasisisimportantasthereis evidence from human studies andanimal models of alcoholtoxicitythat males and females may beaffecteddifferently.214

Theprocedures andtechniquesutilised indeterminingthe

variousparametershave beendescribedin previous papers relatingtochangesinthe alcoholicgroups.91 -20Samples

from thefrontal,parietalandoccipitalwhite matter of the

righthemispherewere used for the various neurochemical assays. Thewater content wasdeterminedby weighingthe tissue before and afterdryingto a constantweight.2' The

dried residuewasaciddigestedand thesodium and potas-sium concentration determinedby flamephotometry.22The 909

910

lipids were extracted into a chloroform/methanol mixture

andaliquots used foreachanalysis.The totallipidwas

deter-minedbydrying23andthephospholipidand cholesterolby

commercial assaypackages.24Thespecificgravitywas mea-suredon akerosene/bromobenzene gradientcolumn.24 The

water soluble protein was determined by the Biuret

method.26 The methods are all standard laboratory pro-ceduresandeach is accurate andreproducible.2-25

The lefthemispherewasfixed and then sectioned at 3 mm

intervals in the coronal plane and morphometrsc

quan-titationof the volumeof the cerebral cortex, white matter anddiencephalicstructures("basal ganglia") performed.An errorof lessthan2% is involved in thedetermination of these

volumes.26Thepericerebralspace(PICS)wasdeterminedby measuring the intracranial cavity volume with a poly-urethanecast.'6Theerrorin thistechnique is less than 2%.16 Although there is a considerable amount of additional information included in the tabular data theprincipal aim of this paper istoaddress the issue of thepathologicalchanges in the brains ofmenregularlydrinkingbetween 30 and 80

gramsof alcohol perday.The controland alcohol dataare

used forcomparison.

Results

Two hundred and forty cases have been studied to

date but it was necessary to exclude many of these

because the clinicalandpathological informationdid

notenableus to be confidentinclassifying thecases

into oneof thefivegroups as outlined inMaterialsand Methods.

Table 1 summarises the numbers of cases, themean Table 1 Meanage,body weight andheight for eachofthe control, moderate and alcoholic groups

Meanage Bodyheight Body weight

(years) (cm) (kg)

n

mean (SD) mean (SD) mean (SD)

Control 41 53 (20) 172 (6) 70 (13) Moderate 14 53 (16) 173 (11) 73 (17) Alcoholic 28 55 (14) 171 (9) 70 (17) Alcoholic + cirrhosis 14 58 (8) 173 (8) 69 (8) Alcoholic + WE 13 62* (10) 170 (7) 65 (12) *=p<0-05.

Harper, Kril, Dal} Table3 Cerebralgrey matter,whitematterandbasal ganglia volumeas apercentageof the cerebralhemisphere

volume

Cortex Whitematter Basalganglia*

(%) (%) (%)

mean (SD) mean (SD) mean (SD)

Control 54-6 (2-6) 40-4 (2-8) 5-0 (0-4) Moderate 55-3 (2-9) 396 (2-8) 5 1 (04) Alcoholic 54-7 (2-6) 39-7 (2-4) 5-0 (0-4) Alcoholic + cirrhosis 55-4 (3-8) 38-6 (44) 5 0 (04) Alcoholic +WE 575f (30) 37-2t (2-8) 5 0 (0-7)

*=Basalganglia is thediencephalic structures including; basal ganglia,thalamus, hypothalamus and amygdala.

t=p<0-01.

ageand mean bodyheight and weight in each ofthe

groups. The mean body heightand weight for each group are notstatistically different. Undernutrition is therefore not responsiblefor the changes found. Loss of brain tissue or brain shrinkage is reflected in a variety of measurableparameters. These include fresh brain weight, cerebellar weight, ventricular volume and ameasurement of thedifference betweenthe brain volume and the intracranial volume; this latter volume has beencalled thepericerebral space (PICS) and is themostaccurateindicatorof brainshrinkage orswelling for anindividual case.16 The mean values of each parameter for each of the five groups arelisted intable 2.Although there is no statistical difference betweenthe moderate and control groupsthere is a trend suggesting loss of cerebral hemisphere tissue with a reduced brain weight, increased ventricular volume and increased PICS value. The mean cere-bellarweightwasthe same in control and moderate groups. It was only significantly reduced in the WE group.

Inorder to localisethe tissue loss to specific regions of thebrain, morphometric quantitation ofthe vol-umesof cerebral cortex, white matter and basal

gan-gliawerecarried out.These are shown in table 3. As

might be anticipated there is no significant loss of

tissue fromanyofthe regions in the moderate group

but the trend appears to followthe alcoholic groups

Table 2 Brain andcerebellarweight, ventricularvolume andpericerebral space for each oftheGroups Fresh brainweight(g) Cerebellarweight(g) Ventricularvolume(%) PICS* (%)

mean (SD) mean (SD) mean (SD) mean (SD)

Control 1438 (102) 186 (15) 1-29 (0-7) 7-3 (39)

Moderate 1415 (129) 187 (20) 1-65 (0-8) 9-8 (44)

Alcoholic 1352$ (144) 177 (13) 2-40$ (1-3) 11-0$ (5-8)

Alcoholic + cirrhosis 1377 (140) 176 (15) 2.96 $ (08) 13-6§ (29)

Alcoholic +WE 1310§ (147) 165T (12) 2-34$ (0-6) 159§ (5-8)

*PICS=Pericerebralspace aspercentageof intracranialspace. t=p<005.

t=p<0-01.

Table4 Meanpercentage water, total lipid, cholesterol andphospholipid content foreachgroup

Watercontent(%) Total lipid (% D W) Cholesterol(umol/gD W) Phospholipid

(,nol/g

DW)

mean (SD) mean (SD) mean (SD) mean (SD)

Control 70 4 (1-4) 18 3 (1-3) 100-8 (11-6) 25-7 (2 8) Moderate 696 (17) 19.2* (14) 1020 (109) 268 (11) Alcoholic 70 6 (14) 18 3 (1 9) 1011 (14.9) 25 1 (2 5) Alcoholic+ cirrhosis 71 4 (1 9) 18 2 (1 1) 94-8 (8-6) 23 7 (2.1) Alcoholic + WE 72-1 (30) 173 (20) 905 (84) 246 (2.1) DW =dry weight. *=p<0-05.

Table5 Specificgravity,electrolyte concentration and water soluble proteincontent ofthewhitematter foreachgroup

Sodium Potassium

Specific gravity (mEqIlOOgdryweight) Protein(mg/g WW)

mean (SD) mean (SD) mean (SD) mean (SD)

Control 10404 (0-002) 17-6 (2-7) 20-8 (1.5) 22 8 (2 3) Moderate 1-0406 (0-003) 15-7 (2-3) 19.7* (1-5) 22 8 (2 8) Alcoholic 10411 (0002) 17 3 (3-0) 19.7* (1-5) 23 2 (24) Alcoholic +cirrhosis 1.0386* (0002) 18 5 (3-2) 177t (1-7) 198t (24) Alcoholic +WE 1.0384* (0002) 17 7 (0-7) 18.9* (2-7) 22 7 (1.8) WW=wet weight. *=p < 0-05. t=p<0 01. $=p<0-001.

inwhichithasbeenpreviouslyshown that the lossof

tissue islargelyfrom the whitematter.7

Topursuethefindingofareductioninthe volume

ofthewhite matter,blocksweretaken from the fresh

hemisphere for neurochemical evaluation. Water,

lipidandproteincontent, specific gravityand

electro-lyteconcentrationweredetermined ineachcase. The

mean results for each group ofcases are shown in

tables 4 and 5. Changes are relatively small in the

moderategroup but thepoint of interest is thatthe

waterandlipid profilesappear to move in theopposite

direction to those ofthe alcoholic groups(table

6).

Thefigure demonstrates, graphically, thechangesin

water andlipidcontentwithincreasingalcohol intake.

The significance values forthese results arelistedin

table4. Itshouldbenotedthatwehave studied

only

a relatively small number of"moderate" cases and

these

preliminary

results must be followed up by a

larger study in order to confirm or refute these findings.

Table6 Patternsof changein the water and lipid content

ofthecerebral whitematter

Water

content Phospholipid Cholesterol

Normalmaturation * * *

Ageing *

Moderate drinkers * *

Alcoholics *

Alcoholtoxicitymodels NA NA *

Walleriandegeneration *

Multiplesclerosis * NA =data notavailable.

73 72 c Go c 71 8 .5 i 70 69 68 21 - 20-C c 8 19-a a. I--a 18 -17 -16

,/0

J0' 0% N1. '0o <209 30-809 >809 +Cir +WE Control Mod Alcoholic*...

alcoholicintake

Fig Changesinlipidand water contentofthe white matter with increasingintakeofalcohol.

911 Doesa "ttio(lerate"alt-oholintake

(laniage

the braiii?

912

Theelectrolyte concentration (sodium and

potas-sium) of the white matter, the specific gravity and

watersoluble proteincontentwerealsodetermined in

allcases. Themean resultsare shown in table 5. The

potassium value in themoderategroupissignificantly

different from the control group. The reduction in

potassium concentration issimilartothatseeninthe

alcoholic groups. The protein content and specific gravity are unchanged from those of the control group.

Discussion

Brainshrinkage, reduced whitemattervolume, loss of

neurons in the frontal lobes and cerebellar vermis

have been previously documented in the alcoholic groups.7920 Theinterest in this study wasto try to

identify changes in the brains ofagroupofcaseswho

could be defined as moderate or perhaps even

"social" drinkers. A previous CT study of "social" drinkers has shown a high proportion to have significant brain shrinkage.4 The results of these studiescould be saidtobe inconclusive becausemost

of thechanges are not statistically significant.

How-ever,there isaconsistent trend in that all volumes and

weightssuggestalossof cerebral tissue. Moreover in

the context of the brain shrinkage measurements

which are already well documented in alcoholics720

and the fact that there appears to be a pattern of severity of changes within the alcoholicgroups(those withcirrhosisorWE) (seetable 2), it isinterestingto note that the moderate drinkers seem tofit between the controls and the medically uncomplicated alco-holicgroup.Howeverthe neurochemical studies show adifferent trend. Thewater contentrisesand thelipid

contentfalls in the alcoholicgroupswhereasthewater

content appearsto fall and the lipidcontent rises in the moderate group (see figure and table 6). How

couldthis beexplainedonanatomicalorpathological

grounds? Itis obvious that foragiven sample of brain

tissue the lipid and water content should move in

concert. Whitematteriscomposed of approximately 70% water, 20% lipids and 10% protein.27 Asmost

of the lipid and protein is combined to form

mem-branes(myelininparticular), if there isanincrease in

lipidcontentthere will beadecrease inwatercontent

and vice versa. Neurochemical studies of the white

matterfrom demyelinating lesions showanincreasein

watercontentandadecreaseinlipidcontent.28 How-ever, themagnitude of the changes is much greater

than in the moderate drinkers. A similar pattern is

seen in Wallerian degeneration29 although most of these studies have utilised experimental models. During maturation of the brain with active myelination thepatterns are reversed: adecrease in

the water content and an increase in the lipid

con-Harper,

Kril,

Dalb'

tent.27Thelipidandwaterchangesnoted in the

mod-erate and uncomplicated alcoholic groups may, therefore, be an indication of some subtle structural change in the white matter.

Few otherstudies ofthe effects of chronic alcohol abuse on the structurallipids in the human brain have been carried out. Lesch and hiscolleagues30 studied

eight cases with hepatocerebral degeneration; six of

these cases were alcoholics. They showed asignificant loss of phospholipids and cholesterol in the myelin

rich areas of the brain. In relatively acute alcohol toxicity animal models it has been shown that the contentof oneof the membrane lipids, cholesterol, is

increased.3'

Manyclinical'andneuroradiological3 studies have

shown that clinical deficits and brain shrinkage are reversible in a proportion of alcoholics following a prolonged period (months) of abstinence from

alcohol. Theseimprovements are more evident in the younger patients with shorter drinking histories. There is to date, no proven pathological explanation for this reversible brain damage. However, it can be stated categorically that neuronal loss is not revers-ible.Therefore we must look elsewhereforan expla-nation. In viewofourfindings ofchanges in the white matter in the alcoholic groupsandin theknowledge

thatthere is a constant turnover ofmyelin andthat

remyelination can occur to a limited extent in the

CNS, further studies of myelin structure and metabo-lism should becarriedout.There may well prove to be twodifferentpathological changesinthe whitematter as aresult ofalcohol abuse: anirreversiblecomponent due to neuronal death and subsequent Wallerian

degeneration,and a reversiblecomponent which may

relate to subtle structural changes which cannot be

identified by normal subjective histological

exam-ination of human material. The basic cause of the

damagehas also to beelucidatedasthose cases with

the most severe brain shrinkage usually have addi-tional disorders such as WE and/orcirrhosis of the

liver. These may well play an important part in the

degenerative processes although the presence of

abnormalities in the "alcoholic" group who had

neitherWEnorcirrhosis,suggeststhat thealcohol per

seprobably plays a significant role. Animal models of alcohol toxicity and thiamine deficiency in various combinationsmay prove useful inclarifyingsome of theseissues.

Conclusions

Theresults of thevolumetricstudies of themoderate group could be said to be inconclusivebecause most of thechangesare notsignificantly different from con-trols. However, in the context of the brainshrinkage measurementsalready documented in alcoholics and

Does a "moderate" alcohol intakedamage thebrain? the fact that thereappears to be a grading of cerebral shrinkage within the alcoholic group (those with cir-rhosis orWernicke'sencephalopathy), results of mod-erate drinkers seem to fit between the controls and alcoholics. Sincemoderate drinkers are more likely to have reversible changes than severe chronic alco-holics, the differences noted in the water and lipid profiles of the whitematter may be an indication of a subtle structural change which can not be identified by the usual subjective histological examination of human material. Thus there may be two different pathological changes in the white matter as a result of alcohol abuse: anirreversiblecomponent due to neu-ronal death and subsequent Wallerian degeneration and a reversible component as noted above. This research was supported by a grant from the Australian Associated Brewers and the N.H. & M.R.C. (No. 87/0564). The authors thank Ms Joy Gillieswhoprepared this manuscript.

References

I Ron MA.The alcoholic brain: CTscanandpsychologicalfinding Psvchol Med1983;monograph supplNo3:1-33.

2 LishmanWA,JacobsonRR,AckerC. Braindamagein alcohol-ism: current concepts Acia Med Scand 1987;222;suppl 717: 5-17.

3 Carlen PL,WortzmanG, Holgate RC,WilkinsonDA, Rankin

JG.Reversible cerebralatrophyinrecentlyabstinent chronic

alcoholicsmeasuredbycomputed tomographicscans.Science

1978;200:1076-8.

4Cala LA,Jones B, Mastaglia FL,WileyB. Brainatrophyand

intellectual impairmentinheavydrinkers: aclinical, psycho-metric andcomputerisedtomography study. AustNZ JMed 1978;8:147-53.

5 WalshKW.Alcohol related braindamage:Anhypothesis.Aust

Ale/DrugRev1983;2:1-84.

6 Harper CG.Neuropathologyof braindamagecausedbyalcohol. MedJ Aust 1982;2:277-82.

7 Harper CG,KrilJJ, HollowayRL. Brainshrinkageinchronic

alcoholics: apathologicalstudy.BrMedJ1985;290:501-4. 8 Harper C, Kril J. Pathological changes in alcoholic brain

shrinkage.Med J Aust1986;144:3-4.

9 HarperC, KrilJ,DalyJ. Arewedrinkingour neuronesaway?

BrMedJ1987;294:534-6.

10 Paton A, SaundersJB. ABC of alcohol. Br MedJ 1981;283: 1248-50.

11 Lelbach WK.(1974). OrganicPathologyrelatedtovolume and patterns ofalcoholuse.In:GibbinsRJ,etal(eds). Research

913

Advances in Drug and Alcohol Problems. Volume1.New York. JohnWileyand Sons.

12 AustralianBureau of Statistics. Alcohol and Tobacco

Consump-tion Patterns (Catalogue No 4312.0). Canberra. Australian

Bureauof Statistics, 1977.

13 PolsRG,Hawks DV. Is there asafelevel ofdailyconsumptionof alcohol for men and women? National Health & Medical Research Council Report for Health Care Committee, Canberra, 1986.

14 HayDA, Cummins JE.Geneticdependence of alcohol related disease: animal models. Aust NZ J Med1981;11:148-53.

15 HarperCG, Blumbergs PC. Brain weights in alcoholics. J Neurol Neurosurg Psychiatry 1982;45:838-40.

16 Harper C, Kril J, Raven D, Jones N. Intracranial cavity

volumes-a new method and its potential applications.

Neuro-patholAppINeurobiol1984;10:25-32.

17 Harper CG, Kril JJ, Daly JM. Brain shrinkage in alcoholics is notcausedby achangein hydration: a pathological study.

JNeurol Neurosurg Psychiatry 1988;51:124-7.

18 HarperCG,Kril JJ, Daly JM. The specific gravity of the brainsof alcoholic and control patients: a pathological study. Br J Addict1987;82:1349-54.

19 Phillips SC,HarperCG,Kril J. A quantitative histologicalstudy

of the cerebellar vermis in alcoholic patients. Brain1987;110:

301-14.

20 Harper C, Kril J. Brain atrophy in chronic alcoholic patients: a quantitative pathological study. J Neurol Neurosurg

Ps'vhiatrs}

1985;45:211-7.

21 YatesAJ, ThelmoW, PappiusHM.Postmortem changesin the

chemistryandhistologyof normal and edematous brains. Am JPathol1975;79:555-64.

22 ShawDM, Frizel D, CampsFE,WhiteS. Brainelectrolytes in depressive and alcoholic suicides. Br J Psychiatris 1969;

115:69-79.

23 FolchJ,Lees M,Stanley GHS.Asimple methodfortheisolation andpurification oftotallipidsfrom animal tissue. J Biol Chem

1957;226:497-509.

24 Marmarou A, Poll W,ShulmanK.Bhabavon H. A simple gravi-metric technique for measurement of cerebral oedema. JNeurosurg 1978;49:530-7.

25 GornallAG,BardawillCJ,David MM. Determinationofserum

proteinsbytheBiuretreaction. J BiolChem1949;177:751.

26 AherneWA,Dunnill MS.(1982)MorphometrY.London.Edward Arnold.

27O'Brien JS, Sampson EL. Lipid composition of the normal human brain: grey matter, white matter andmyelin. JLipid

Res1965;6:537-44.

28 Davison AN,Wajda M. Cerebral lipids in MultipleSclerosis.

JNeurochem1962;9:427-32.

29Cumings JN. (1965) Cerebral lipid biochemistry in the de-myelinations. In: Cumings JN, Kremer MFA. (eds) "Bio-chemical Aspectsof Neurological Disease"Philadelphia,Davis Co. 229-51.

30 Lesch P, Schmidt E, Schmidt FW. Effects of chronic alcohol abuse onthe structurallipidsinthe human brain. Z Klin Chem KlinBiochem 1972;10:410-5.

31 Chin JH,Goldstein DB. Membrane-disordering actionof eth-anol. Variation withmembrane cholesterol content anddepth