Subclinical hepatic encephalopathy in cirrhotic patients : Prevalence and relationship to liver function

'} POMIER-LAYRARGUES, NH NGUYEN, C FAUCHER, J-F GIGUERE, RF BUTIERWORTH. Subclinica1 hepatic encephalopathy in cirrhotic patients: Prevalence and relationship to liver function. Can J Gastroenterol 1991 ;5 (4 ): 121-125. Neuropsychological tests were adminisrere<l to 22 alcoholic cirrhotic patients, 20 nonalcoho lic cirrhotic pa tients and 42 control suhjects matche<l for age and educational background. Liver funct ion estimated by Pugh score was identical in a lcoholic and nonalcoholic c irrhotics. Thirty-two c irrhotic patients (76%) failed one or more psycho logical test. Performance was worse in c irrhotics than in controls in all but one test: the degree of neurological impairment was similar in alcoholics and nonalcoholics and corre lation with the severity of I iver dysfunction was only marginal. A combination of the five more sensitive tests was able to identify 31 of 32 cirrhotic pat ients with subclinical encephalopathy; this 'mini battery' of tests can be performed within 10 mins. S uhclinical hepatic encephalopathy is frequent in c irrhotic patients and can a<lversely affect the ir psychosocial behaviour.

H EPATIC ENCEPIIALOPATIIY IS usua lly clm,si(ie<l into four stage:, accord ing tll the scverit y of neurological symptoms.However, it is now well recognized rhat a significa nt rrorortit111 of c irrhotic patients show suhdc neurological dcfic ns that are not Jerecrnhlc vi,1 rout inc neurological exam ination but ca n be demonstrated vu1 eit her ncuropsychological test~ or visuall y evoked roten rn1 ls ( 1 ).T his entiry is generall y referred to as subclinical her atic encephalopathy.Its prevalence is va riable depend ing on the ratient population and on the battery of tests used to detect it.It is now estahl ished that subclinical hepatic cncephalorarhy may occur in both alcoholic and nonalcoholic cirrhotic patients (2)(3)(4)(5)(6).However, the relationship between neuropsycho logical imra irment and li ver function is still controversial (2, ' 3 ).

PATIENTS AND METHODS
S u bjec ts: Fo rty-two patients with h1opsy-pniven I 1ver L 1rrhosis were evaluateJ as 111-m outpatients: 22 had c 1rrhosb secondnry to chronic alcoholism (daily ingesrion of I 00 g ethan11I for more than 10 years) and 20 had nonalcohul1c cirrhusis (four, prima ry hiliary cirrhos is; seven, hepatitis Fl viral c irrhrn,is; three, hepatitis C ,•1rnl c 1rrhos1s; anJ six, c ryptogenlc c irrhosis).All of the alcoho lics had heen abstinent for at least twtl months.Three patients had had previous e nd -t<H,1c.lepo nacaval ~hunt  There was no significant difference between the two groups for any of me above parameters 122 None of d1L' patients had ,Nenx1s and handwrning wHs normal.
Amhmetic: The pat ienr is asked lll perform mentally eight L,1ku lat ions of 111 creasing difficulty.Months lmckwarcl: The pa ti e nt is asked t() name the month~ of the year 111 reverse order starting with Oecemher.D1gi1 span for ~uard and hackward: The digit span tests the numher lif digits the patient 1s ,1bk to recall immedunely from memnry from a previously given list.Al/)habet: The time needed tu enumerate the letters of the alphahet is measu red.
Visuos/)atial ability.The patient is asked to make a drawing of a c lt1ck 111d1caung a fixed tune ( 10 mms past 11 o'clock).
Psychomowr s(Jeecl.Digit symbol: ThL• subject must assoc1.1teone ser of symbol:, wit h another set of numbe rs.Reitan trnil makmg tests: The subject must connec t in order a series of e ithe r numbers {part A) nr numbers a nd letters (part B) ra ndo ml y presented.

STATISTICS
The resu lts of individual te.sts were con sidered ahnorma l 1f they fell outside t he range of two swmla rd dev1atiom above rhe mean value 111 control patients.Dara were a b o compared be t wc:en controls ,md urrhot1cs, ,me.I bet ween alcoholic ,md nonalcoholic cirrh otics using Student's L test fo r unpaired data.Corrd,1rnms were evaluated with the Spearman ran k corre lation test.

DISCUSSION
The present stud y reveals ~uhtlc impairment in ncuropsycholog1cal performance in a large percentage of cirrhotic patients in the absence of overt neurol()gica l abnorma litie~.Scvenry-six per cent of cirrhotic patients had subclinical hepatic encephalopmhy demonstrated by abnormalities in one or more rests.This prevalence was comparable with that previously reported ( l ).
Comparison of the cirrhotic group with a control group strictly matched fi.ir age and educatio nal background was performed in order to avoid any influence of these parameters on the performance of the tests used.A battery of tests was chosen which had previously been reported to be semitive and which was easi ly and quickly ad - or Bused alone min cnmhmamm with the block design tests were repeatedly reported to be must scn~itive (2-6, I 0-12); some authors suggested me,1surement of continuou, reactil,n time ( 13).

P0MIER-LAYRARULIE'i et al
It has abo been suggesceJ LhaL visual evokeJ potentials or somarosensory evoked pmentials coulJ be useful in detecting subclinical hepatic encephalopathy ( 14-17).However, their specificity is qucstionahlc anJ their efficacy b probably nor better than thaL l)( simrlc ncuropsyc hn mclrlc tests (1 5).The a uthors observeJ that th e most se nsitive rests we re Lh e Reitan trail making test part B fo lloweJ by alphabet, momhs hac kwarJ , arithmetic anJ the Reitan trai l making test part A.
This stud y also Jemonstratcs that subclinical hepatic encephalopa thy occurs in alcoho lics as well as nonalcoholics, whic h confirms results obtained by other groups in recen t years (2)(3)(4)(5)(6).This finding demonstrates that liver Jisease per se may induce alterations in n europsychological performance.The question then arises as to whether the severity of neurological abno rmalicics is corre lated with the J egrce of liver failure per se.Previous stuJies have generall y suggested that there is no relationship hetween these two pa rameters (3, I 8).However, one paper re porteJ a corre la t ion between abnormal performance in tests and serum albumin (2).In the present study, scrum albumin and Pugh score demonst rated a small hut significant correlation with one test, the Reitan trail making test part B. This finding is not surprising as the l)Ccurrence of suhclinica l hepatic encephalopathy is probably related not only LO liver function hut also to t he extent of spontaneous or surgically inJuceJ portal systemic shunting.Unfortunately thb latter parameter could not be evaluated in the present patients.
The fact that a majorit y of cirrhotic patients without overt hepatic encephalopathy display impairment in psychomotor performance has several practical implications.First, these deficits may affect the way of life of many patients, particularly if thei r work necessitates intellectual activity or precise mechanical manipulat ions.The problem of impairment in driving capacity was also addressed several years ago ( 19)

TABLE 3
Results of neuropsychological tests in control subjects and cirrhotic patients (mean ± standard deviation)