Canad. Encephalopathy and Fatty of the Viscera in Hospital

522 Norman: Encephalopathy and Fatty Liver Canad. Med. Ass. J.Sept. 21, 1968, vol. 99

Encephalopathy and Fatty Degeneration of the Viscerain Childhood:

I. Review of Cases at The Hospital for Sick Children, Toronto (1954-1966)M. G. NORMAN, M.D.,* Toronto

IN 1963, Reye, Morgan and Baral1 of Aus-tralia reported a syndrome of encephalopathy

and fatty liver occurring in children. This as¬

sociation had been previously described,2^ andsubsequently cases have been reported fromNew Zealand,6 Czechoslovakia,7'8 the UnitedKingdom,912 South Africa13 and the UnitedStates of America.14"20 The admission to TheHospital for Sick Children, Toronto, of twosiblings of a child with this condition led to theexamination of the records of previous casesseen at this hospital during the past 15 years.

Cases were selected from the autopsy indexby reviewing all those coded under "fatty liver","encephalopathy of unknown origin" and "en-cephalitis, probably viral". Criteria for finalselection were a typical clinical course (see be¬low) and, at necropsy, a diffusely fatty liverwith no associated histological evidence of en-

cephalitis. A total of 21 cases satisfied thesecriteria and are included in the present study.Clinical CourseThe history and clinical course were fairly

uniform in the 21 cases collected. A prodromalphase, usually of about four days, but occa-

sionally lasting up to a week, was followed bya catastrophic event leading to unconsciousnessand death in about one day. This prodromalphase often consisted of a mild upper respira¬tory or other poorly defined febrile illness.Vomiting was frequent in this period but neitherthe child's physician nor the parents thoughtthat the child was seriously ill. The patientsusually appeared to recover for a day or so butwere brought to hospital because of focal or

generalized convulsions, delirium or a quietdrift into coma.These children usually were hyperventilating,

suggesting salicylate intoxication and promptingan estimation of blood salicylates. As death ap-proached, respirations became irregular andfinally failed. Alterations in reflexes were vari-able. Muscle tone fluctuated between hyper-tonic and hypotonic, and many of the childrenwhen touched or stimulated assumed a decere-

Fr©m the Department of Pathology, The Hospital forSick Children, Toronto, Ontario.?Chief Resident in Pathology.Reprint requests to: Dr. J. A. Lowden, Research Insti¬tute, The Hospital for Sick Children, 555 UniversityAvenue, Toronto 2, Ontario.

brate posture. No child showed signs of menin-geal irritation. Nine children were delirious.Nine had generalized convulsions and three hadfocal seizures. In 15 cases the liver was palpablebefore death. Vomiting of coffee-ground materialoccurred in six cases.

Treatment, which included varying combina-tions of intravenous fluid infusions, antibiotics,steroids, vasopressors, sedatives, anticonvulsants,hypothermia and assisted ventilation, was sup-portive and ultimately futile.Over 50% of the children were dead within

24 hours of admission to hospital. Of the sixwho survived longer than 30 hours, all but one

required assisted ventilation.

Laboratory DataHypoglycemia, metabolic acidosis and res¬

piratory alkalosis were frequently noted in thesechildren. In 14 cases blood glucose estimationswere available; eight of these were less than 50mg. per 100 ml. and levels as low as 5 and 11mg. per 100 ml. were recorded. In five casesthe blood glucose was within the normal range.One child had a high level (227 mg. per 100ml.), but before this determination her clinicalcondition had suggested hypoglycemia and shehad received 50% glucose in distilled water in-travenously. Cerebrospinal fluid glucose levelswere measured in eight cases; three of thesewere less than 50 mg. per 100 ml. but mostothers were measured after the intravenous ad¬ministration of glucose.Ten of 17 children had a blood pH less than

7.4, the lowest being 7.30. The C02 content was

always less than normal, and was less than 15mEq. per litre in eight cases.The white blood cell count was elevated in

15 of 16 cases, being above 20,000 cells perc.mm. in eight children, and ranging as high as

49,700 cells per c.mm. Differential cell countshowed a neutrophilia. Urinary ketones werefound in eight of nine cases tested.

Cerebrospinal fluid protein was not elevatedin the seven cases where it was measured, butin two cases 20 to 40 lymphocytes per cmm.and in three others between 5 and 10 cells werefound.

Liver function tests were rarely performed,but in two children cephalin-cholesterol floccu-lation and thymol turbity were positive and

Canad. Med. Ass. J.Sept. 21, 1968, vol. 99 Norman : Encephalopathy and Fatty Liver 523

their serum glutamic-oxaloacetic transaminaselevels were 1400 and 700 units, respectively. Theelectroencephalographic records in seven cases

were markedly and diffusely abnormal, but no

specific diagnostic patterns were noted.

Postmortem Findings

Histological sections of brain showed edema.Very occasional neurones had anoxic changes.In a few cases small perivascular hemorrhageswere present. In each of three cases, one sectionshowed a single vessel surrounded by a cuff oflymphocytes. These were insuflicient to warranta morphologic diagnosis of viral encephalitis.Fat stains of frozen sections of brain showed oc¬casional capillary endothelial cells containingoil-red-O-positive material but no fat emboli.

In all cases the liver was inereased in weight,although generally not more than 10 to 15%above normal; the organ showed various shadesof yellow. On microscopic examination diffusefatty changes were seen throughout the lobule.The shape of the liver cells was maintained, thenucleus remaining centrally placed with the fatdistributed in tiny droplets around it. Fat stainsof frozen sections confirmed the intense fattychange in the liver and the distribution of thefat in multiple tiny droplets around the nucleus.Kupffer cells appeared more prominent thannormal, and contained smudgy eosinophilic ma¬

terial and occasionally fat droplets. There was a

slight inflammatory cell infiltrate in the portaltracts in three cases, and two of these showedoccasional single necrotic liver cells adjacent tothe portal tracts. The "piecemeaF necrosis ofinfectious hepatitis was never seen. A section ofliver from one of the youngest children (5y2months) showed a few multinucleated giantcells resembling those seen in neonatal hepatitis.In another case (associated with chickenpox,where assisted ventilation and profound hypo-tension had lasted for eight days) there was bilestasis in the canaliculi.

In less than half the cases fat stains of frozensections of the kidney were available. Theseshowed tiny fat droplets in the basal portion oftubular epithelial cells, especially in the proxi-mal convoluted tubules. Hematoxylin and eosinstains of kidney showed only pallor and mini-mal fine vacuolation of epithelial cells of theproximal convoluted tubules. In the few cases

where fat stains of muscle and heart were avail¬able, mild fatty change of these tissues was

present. Adrenal cortices were depleted of lipidin all but one child who had a recent infarct ofthe pituitary gland attributed to a long periodof hypotension (this was the same patient whoshowed bile stasis in the liver).

Bronchopneumonia or pulmonary edema andfocal hemorrhage were present in 10 cases.

Small esophageal erosions or acute peptic ulcersof stomach or duodenum were found in sixcases, and blood was present in the gastroin-testinal tract in ten. Sections of pancreas were

normal in all but two cases, in each of which a

single duct was surrounded by polymorpho-nuclear leukocytes.General Factors

Retrospective studies have many sources oferror, and few conclusions may be drawn froma small series like this one. The disease affectsboth sexes (8 females and 13 males). Eight ofthe 21 children were aged 5 to 12 months, threewere between 1 and 4 years, and the remaining10 were 4 to 9y2 years old. The first case ofwhich we have record occurred in 1954 and theremaining 20 from 1958 to 1966, not more thanfour occurring in any one year. The majorityof the cases occurred in the fall and wintermonths: seven in January to March, four inApril to June, two in July to September, andeight in October to December. No geographicclustering of cases was evident.

Since 10 of the 13 children over 1 year of agehad previously responded normally to one or

more viral illnesses (rubella, mumps, chickenpoxor measles) it is unlikely that they had an in-born defect in their ability to tolerate such in¬fections. One 2%-year-old child had not hadany of these diseases, and in two cases no in¬formation was available. Little can be inferredfrom the children under 1 year of age (sevenof whom had had none of these diseases, whileone had survived measles). None of the childrenhad a history of being "sickly" or having an in-ordinate number of infections.One child had a history of rheumatic fever

seven months before death. Another 4y2-yeax-old girl had suffered from the nephrotic syn¬drome and had been treated with corticoster-oids, although none had been given for a yearbefore her death. A third had scarlet fever twoyears before death, and a fourth had bronchitisand a questionable allergy to wool and cereals.All four of these children had had a normalresponse to one or more of the common viraldiseases of childhood.

Association with Viral DiseaseThere was a background of viral illness or a

suggestive history in three-quarters of the cases.

An older brother and sister of a e^-year-oldgirl were admitted with infectious hepatitis 14and 25 days, respectively, after her death (this

524 Norman: Encephalopathy and Fatty Liver Canad. Med. Ass. J.Sept. 21, 1968, vol. 99

family prompted the present review). Viral cul-tures and virus antibody titres in this famlywere not remarkable, but during his illness thebrother showed a tenfold increase in the cox-sackie A9 titre. The condition of both siblingsimproved and they were discharged to con-valesce at home.Three children were admitted four to five

days after first developing a chicken-pox rash.One of these, a 3-year-old boy, was dead on ad¬mission to hospital. His 10-month-old brotherwas admitted on the same day with an illnesswhich started with fever, vomiting, diarrhea andconvulsions. For the next two to three days hewas semicomatose and had a palpable liver.Liver function tests were abnormal. He gradu-ally improved and was discharged 10 days later.An older and a younger brother of a 5-year-

old boy were sufficiently ill with a febrile con¬dition to warrant admission to hospital withina day of his death.Members of the family of a further five chil¬

dren had a "cold" or "upper respiratory infec¬tion". Two children who had not had mumpshad a history of contact with mumps.

Viral antibody studies were done in few ofthe patients. Titres were always low for theparticular antibodies tested and were generallynot helpful because paired sera were notstudied. Virus was not recovered from culturesof cerebrospinal fluid samples and postmortemtissues taken from six children, from cerebro¬spinal fluid alone from three children, or frompostmortem tissue alone from an additionalthree children.

Association with ToxinsBecause of the rapid downhill clinical course

and the evidence of severe liver disease, thepossibility of ingestion of toxic substances hasbeen considered. The parents of seven childrenin this series were specifically questioned butdenied the possibility of ingestion of poisons. Ahigh level of inorganic phosphorus was reportedin one child by another laboratory. No historyof exposure to phosphorus could be obtained,however, and clinical opinion was that the childwas not suffering from phosphorus poisoning.Only three of the 21 patients were in the agerange of 1 to 4 years, when accidental poison-ings are most common. This is probably a for-tuitous age distribution in this series.Blood salicylate levels, measured in 10 chil¬

dren, were all less than 25 mg. per 100 ml. Inseven cases the level was less than 20 mg., al¬though in many instances the children had re¬

ceived salicylates to control fever for some hoursbefore admission.

It is unlikely that the etiology of the fattyliver was iatrogenic, as no common single drugwas administered to every child. Seventeen chil¬dren received penicillin in some form, and nineof these were also given chloramphenicol. Onlyone child received tetracycline. Eleven were

treated with adrenal corticosteroids after hos-pitalization. Paraldehyde was administered to 11children to control seizures. A variety of otherdrugs was also used, but each drug was given toless than one-quarter of the patients.

Fifteen cases had negative lung and bloodcultures at autopsy, and these together with thedemonstration of negative blood cultures dur¬ing life, would appear to exclude bacterialtoxins.

DiscussionThe etiology of the syndrome is not clear.

The most favoured theories postulate an over-

whelming virus infection, a toxin, or a combina-tion of factors which terminate in the same

clinical and pathological picture.The syndrome has been described in associa¬

tion with chickenpox1416,21«22 and infectioushepatitis,4, 8 but five children in this series hadsurvived chickenpox before their fatal illness.Virus isolations have been infrequent. Utian,Wagner and Sichel13 reported isolating coxsackieA and a reovirus, while Golden and Duffell19found an echovirus, and Becroft6 reportedherpes simplex. Dvorackova, Vortel and Hroch8isolated adenovirus Type 3 from various tissues,including liver, at postmortem examination. Inanother case at this hospital (reported sep-arately in Part II) influenza B virus was re¬

covered from a liver biopsy and from varioustissues at autopsy.The significance of the history of contact with

other viral diseases such as mumps, or associa¬tion with viral illness in a family, is hard to as-

sess. Although a virus may be involved, most ofthe cases in this series occurred in winter whenmild viral illnesses are prevalent and probablya random group of children would show thesame number of contacts. The rarity of virusisolations from these children makes it difficultto attribute the syndrome to a viral infection.The demonstration of an increasing viral anti¬body titre in surviving members of the familywould strengthen the epidemiological evidence,but this was not tested in most of our cases.

Hepatitis-eneephalitis associated with reoviruscan occur in humans,23 but the changes des¬cribed are not those seen in this syndrome.Chang, Geyer and Andrus24 have reported iso-lation of a virus producing a lipogenic toxin.

Canad. Med. Ass. J.Sept. 21, 1968, vol. 99 Norman : Encephalopathy and Fatty Ltver 525

Fatty changes in the liver and kidney havesuggested a toxin, but only three positive toxi-cological studies of similar patients have beenreported. Curry, Guttman and Price2 foundpteridines in the urine and Randolph et al.17demonstrated trace amounts of isopropyl alcohol.

Recently, Glasgow and Ferris25 reported a

study of a 4-year-old girl with the clinical pic-ture described who died. Fatty liver was foundat necropsy. Because the child lived in thevicinity of a car-spraying establishment, a de¬tailed toxicological examination was performed,and isopropyl alcohol was found in the gastro-jejunal contents by gas-liquid chromatography.

Reye, Morgan and Baral1 pointed out the cli¬nical similarity to, and the pathological dif¬ferences of this syndrome from, the "vomitingsickness of Jamaica". That disorder results inhypoglycemia attributed to hypoglycins A andB, which are contained in the Ackes nut.26'27Becroft6 had drawn attention to aflatoxins.prod-ucts of Aspergillus flavus, which are are known,experimentally at least, to produce fatty liver.The search for a toxic substance may have tomove from the well-known and obvious poisons,like phosphorus and carbon tetrachloride, tofungal and vegetable poisons existing in food-stuffs, or to the complex chemicals such as in-secticides now in domestic use.

A third possibility is that a viral infection maybe required to trigger the syndrome in a childsuffering from intoxication with (for example) a

fungal or vegetable toxin. Finally, individualsuseeptibility or idiosyncrasy may be the cause.

Perhaps the syndrome can be produced byseveral different agents. Whatever is postulatedas the etiologic agent, it must explain not onlywhy isolated deaths occur in a family (wherepresumably all the children are exposed to moreor less the same environment), but also therelationship between the brain damage, hypo¬glycemia, acid-base imbalance and the fattyliver.

Usually the children had not eaten for severaldays before admission and were vomiting andketotie, both of which tend to lower the bloodsugar.28 Mortimer and Lepow16 have attributedthe hypoglycemia occurring in their cases ofvaricella to the ingestion of salicylates. Limbecket al.2Q have described several cases of hypo¬glycemia apparently resulting from extremesensitivity to salicylates. If salicylates were thecause of the hypoglycemia, it is difficult tounderstand why this syndrome was not recog-nized and reported in greater numbers earlierthan the 1960's. In this series blood glucoselevels of 11 mg. per 100 ml. were associated withblood salicylate levels of 10.2 mg. and 7.6 mg.

per 100 ml., respectively. Severe liver diseasecan cause hypoglycemia, but the degree of dis-turbance of liver function tests and the lack ofnecrosis in that organ make it difficult, short ofa specific block in an enzyme system, to attributethe hypoglycemia to liver disease alone.The brain damage could also result from hy¬

poglycemia, but it must be emphasized that lowblood sugars were not recorded in every case.The children were acidotic and hyperventilat-

ing at a rate which would seem excessive forthe degree of metabolic acidosis. Simpson9 hassuggested that the hyperventilation may be ofcentral origin.

In the absence of liver biopsy and liver func¬tion tests, retrospective diagnosis in survivingcases is difficult and speculative. Survivors havebeen reported,1 some with permanent cerebraldamage.13' 15>1Q

The etiology and pathogenesis are not clear,and therefore the most advantageous immediatecourse must be to pursue intensive investigation.Because of the rapidly fatal nature of the con¬dition, a pre-established protocol for investiga¬tion is called for. Becroft6 has suggested thefollowing:

1. Examination of the patient's environmentfor toxins, immediately rather than retrospec-tively, with particular attention to the possibilityof toxins in apparently innocuous foods andmedicines.

2. Further toxicological investigations, includ-ing analysis for fungal products, and examina¬tion of the urine for the fluorescent materialdescribed by Curry, Guttman and Price.2

3. Further viral studies by all available tech-niques.

4. Chemical analyses of the liver and serum

lipids, including estimations of serum lipo-proteins.

5. Studies of liver enzymes, particularly thoseconcerned with carbohydrate metabolism, pref-erably of biopsy rather than necropsy material.

6. Estimations during the acute stage of serumlevels of adrenal corticoids, insulin and growthhormone.To this, could be added:

7. Serial estimations of liver function, bloodglucose and blood insulin studies, done on thesame sample of blood and the results interpretedin relation to the amount of glucose being givenintravenously.

8. Estimation of serum proteins, with frac-tionation of the immunoglobulins.

9. Determination of interferon levels.10. Family studies with respect to associated

illness, liver function and change in viral anti¬body titres, with follow-up to discover illnesses

526 NORMAN: ENGEPHALOPATHY AND FATTY LIVER Canad. Med. Ass. J.99

manifested after the death of the index case ina family.

11. Inquiry into any dietary likes or dislikesof the children in the hope of explaining the iso-lated occurrence of these cases in a family.

12. Recording of the socioeconomic status,dietary habits and chemicals in use in the house-hold in question.

At necropsy:13. Assay of brain, liver and peripheral

muscle for lipids and glycogen.14. Study of pancreas for a and /3 cells.15. Virus studies of individual organs.

Summary The clinical and pathological featuresof encephalopathy and fatty liver oc-

curring in 21 children have been reviewed. The cli-nical picture was of a mild prodromal illness, fre-quently an upper respiratory infection lasting sev-eral days, followed by vomiting. The children thenrapidly became unconscious and died, three-quaitersof them in 24 to 30 hours. Laboratory investigationrevealed hypoglycemia, mild metabolic acidosis,respiratory alkalosis and disturbed liver function.Necropsy showed cerebral edema and anoxic neu-ronal changes. There was marked diffuse fattychange in the liver and fat was present in the kidneytubules. The clinical, biochemical and pathologicfeatures were not specific.The association of the syndrome with chickenpox

in three cases, with infectious hepatitis in otherfamily members in one case, and with mild viralillness in the family in a quarter of the cases, wasnoted. Although there was an epidemiological as-sociation with viral disease, vims could not be iso-lated from these patients, making it difficult toestablish the connection.

Lack of a history of exposure to toxins and lackof toxicological determinations did not exclude thepossible etiological role of toxins. The possibilitythat several factors combine to cause the syndromeis suggested. A plan for investigation of future casesis outlined.

Because hypoglycemia alone could cause thebrain damage, it should, when present, be treatedvigorously in children suffering from any acute en-cephalopathy.

R.8um. L'auteur passe en revue les cas de 21enfants ayant souffert d'enc6phalo-

pathie et de d6g6n6rescence graisseuse du foie eten analyse les caract6ristiques ciniques et patho-logiques. Le tableau dinique comportait d'aborddes prodromes b6nins, se manifestant souvent parune infection des voies respiratoires sup&ieuresd'une dur.e de quelques jours, puis par du vomis-sement. A ce moment, les enfants devenaient rapide-ment inconscients et mouraient, les trois-quarts densun d6lai de 24 . 30 heures. Les analyses de labora-toire r6v6iaient de rhypoglyc6mie, une l6g.re aci-dose m6tabolique, une alcalose respiratoire et des

fonctions h6patiques troubl6es. A la n6cropsie, onnotait un 6d.me c6r6bral et la pr6sence de lipidesdans les tubes urinaires. Les caract6ristiques cli-niques, biochimiques et pathologiques n'6taient passp6cffiques.On a constat6 que le syndrome coincidait, dans

trois cas, avec Ia varicelle, dans un autre cas, avecune h6patite infectieuse ayant atteint d'autres mein-bres de la famille et, dans 25% des cas, avec unemaladlie virale b6nigne frappant la farnille. Bienqu'il alt 6t6 incontestable qu'il existait tine associa-tion 6pid6miologique avec une maladie virale, onn'a Pu isoler de virus chez ces malades, ce qul arendu difficile l'identification d'un facteur 6tiolo-gique.

L'absence de renseignements sur les contacts avecdes toxines et le manque de d6terminations toxi-cologiques n'ont pas permis d'exclure le r6le .tio-logique possible des toxines. L'auteur estime, poursa part, qu'il faudrait incriminer plusieurs factetirspour la gen.se du syndrome. II propose tin pland'6tudes des futurs cas.

Etant donn6 que la seule hypoglyc6mie risque deprovoquer une l6sion c6r6brale, cette anomalie6ventuelle devrait .tre trait6e 6nergiquement cheztout enfant souffrant d'une enc6phalopathie aigu..

The author wishes to 'thank Dr. H. W. Barn for ac-cess to the clinical records of the patients; Dr. W. L.Donohue for access to the autopsy records; Drs. W. L.Donohue, M. J. G. Lynch, M. E. Delisle and J. A. Low-den for helpful comments and criticisms of the manu-script; Dr. W. S. Hartroft for helpful suggestions, andDr. E. G. Murphy for reviewing the electroencephalo-grams.

REFERENCES

1. REYE, R. D. K., MORGAN, G. AND BARAL, J.: Lancet,2: 749, 1963.

2. CuRRY A. S., GUTTMAN, H. A. N. AND PRICE, D. E.:Ibid., 1: 885, 1962.

3. BRAIN W. R., HUNTER, D. AND TURNBULL, H. M.:Ibsci., 1: 221. 1929.

4. ToMLINsoN, B. E.: Ibid., 1: 1300, 1955.5. ANDERSON R M Med. .f. Aust., 1: 573, 1963.6. BECROFT, b. M. 0.: Brit. Med. J'., 2: 135, 1966.7. STEJSKAL, J. AND KLUSKA, V.: Lancet, 1: 615, 1964.8. DVORACKOVA, I., VoRTEL, V. AND HROCH, M.: Arch.

Path. (Chicago), 81: 240. 1966.9. SIMPSON, H.: Lan,cet, 2: 1274, 1966.

10. SHAKESPEARE, W. G.: Bmt. Med. J., 2: 642. 1966.11. ELLIOTT, R. I. K., MANN, T. P. AND NASH, F. W.:

Lancet, 2: 882, 1963.12. MALONEY, A. F. J.: Ibid., 2: 1122, 1963.13. UTIAN H. L., WAGNER, J. M. AND SICHEL, R. J. S.:

Ibid., 2: 1043, 1964.14. JENKINS, R., DVORAK, A. AND PATRICK, J.: Pediatrws,

39: 769. 1967.15. Case Records of the Massachusetts General Hospital,

Weekly Clinicopathological Exercises, Case 1-1967:New Rag. .T. Med., 276: 47. 1967.

16. MORTIMER, E. A., JR. AND LEPOW, M. L.: Amer. J.Dis. Child., 103: 583. 1962.

17. RANDOLPH, M. et aZ.: Ibid., 110: 95. 1965.18. JASBOUR, J. T.. HOWARD, P. H.. JR. AND JAQUES, W.

E.: J. A. M. A., 194: 1245. 1965.19. GOLDEN, G. S. AND DTJFFELL, D.: Pediatrics, 36: 67,

1965.20. WINOGRAD, H. L.: ,.T. A. M. A., 194: 1247, 1965.21. BREEN, G. E. AND EDMOND, R. T. D.: Med. Press,

232: 251, 1954.22. ABRUZZI, W. A.: New York .1'. Med., 61: 3912, 1961.23. JOSKE, R. A. et al.: Arch. Intern. Med. (Chicago),113: 811, 1964.24. CHANG, R. S., GEYER, R. P. AND ANDRUS, S. B.: J.

Rxp. Med., 115: 959, 1962.25. GLASGOW, J. F. T. AND FERRIS, J. A. J.: Lancet, 1:

451. 1968.26. MARKS, V. AND ROSE, F. C.: Hypoglycaemia, Black-

well Scientific Publications Ltd., Oxford, 1965, p.168.

27. Idem: Ibid., p. 265.28. MADISON, L. L. et al.: J. Clin. Invest., 43: 408, 1964.29. LIMBECK, G. A. et al.: Amer. J. Dis. Child., 109: 165,

1965.