Gut, 1976, 17, 700-708

Hypergastrinaemia in cirrhosis of liverSHIU KUM LAM

From the Department ofMedicine, University ofHong Kong, Queen Mary Hospital, Hong Kong

SUMMARY The basal acid output (BAO), post-pentagastrin acid output (MAO), fasting and post-prandial gastrin levels in 40 patients with proven cirrhosis of the liver were compared with those in20 normal controls. The mean BAO and MAO were significantly lower than normal, the meanfasting gastrin level was significantly higher than normal, and the postprandial gastrin responsewas significantly increased and prolonged. These differences were still significant even when thepatients were divided into cryptogenic and alcoholic subgroups. A significant inverse relationshipbetween MAO and the integrated gastrin response to meal was observed both in the normal controlsand in the cirrhotic patients. The MAO and integrated gastrin response of the cirrhotic patients didnot correlate with the degree of liver function impairment. In five cirrhotic patients fasting andpostprandial gastrin levels were unchanged after splenorenal shunt operation. A more consistentabnormality of the gastric mucosa as assessed by endoscopy and biopsies appeared to be mucosalcongestion with occasional atrophic gastritis. The severity of mucosal abnormality, however, was

unrelated to the degree of hypoacidity. These results indicate, firstly, that the hypergastrinaemia incirrhotic patients is a reflection of gastric hypoacidity and bears no direct relationship to hepaticdysfunction. Secondly, the gastric hypoacidity does not accrue solely from mucosal abnormality.It is suggested that this hypoacidity may result from the presence of excessive amounts of cir-culating acid-inhibiting intestinal peptides, which the diseased liver fails to metabolise.

An increased incidence of pepticulceration inpatientswith cirrhosis of the liver had been observed bymany authors (Schnitker and Hass, 1934; Welbourn,1952; Palmer and Brick, 1953; Swisher et al., 1955;Fainer and Halsted, 1955; Hallen and Kroorc, 1963;Patek, 1963; Tabaqchali and Dawson, 1964;Bergman and van der Linden, 1965; Jackson et al.,1965; Orloff et al., 1969; Resnick et al., 1969;Jackson et al., 1971; Phillips et a!., 1975), butdisputed by some (Ratnoff and Patek, 1942; Lippand Lipsitz, 1952; Doll, 1952; Sullivan et al., 1954).The pathogenesis for the apparent increase infrequency is unknown. It is unrelated to gastrichyperacidity, since acid secretion in cirrhoticpatients has been shown to be abnormally low(Scobie and Summerskill, 1954; Ostrow et al., 1960;Schmidt and Martini, 1969) or to be normal (Tabaq-chali and Dawson, 1964).

It is now generally accepted (Grossman, 1974)that the relationship between pentagastrin orhistamine stimulated acid output and food stimu-'This study is supported by research grant no 158-295 of theUniversity of Hong Kong.Received for publication 4 June 1976

lated serum gastrin response is such that, in duodenalulceration, both secretions are on average higherthan normal, whereas, in gastric ulceration, therelationship is an inverse one, there being a highfasting serum gastrin reflecting a low acid output.

In this study, the relationship of acid and gastrinrelease in cirrhotic patients was examined, and anattempt was made to correlate the findings with theseverity of disturbance of liver function and themorphology of the gastric mucosa. In addition,the effect of a portosystemic shunting operation onacid and gastrin release was also studied.

Methods

PATIENTSForty patients, 35 males and five females, with agesranging from 24 to 65 years (mean 52-1 years) werestudied. The presence of cirrhosis was confirmed byneedle biopsies of the liver in 35 patients and atnecropsy in the remaining five patients who diedbetween one and seven months after the study. Inthe latter five patients needle biopsy of the liver wasnot performed at the time of the study because of

700

on Decem

ber 24, 2019 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.17.9.700 on 1 Septem

ber 1976. Dow

nloaded from

Hypergastrinaemia in cirrhosis of liver

persistent coagulation abnormalities. In 27 of thecirrhotic patients, no cause of cirrhosis could bedetermined and, for the purpose of this study, theywere designated cryptogenic cirrhotics. The morpho-logical features were those of post-necrotic scarring,as had been previously described (Cook et al., 1963).Among these, 16 were considered by the radiologiststo have a small-sized liver and nine to have a normal-sized liver on hepatic scintiscanrning. In the remain-ing 13 patients with cirrhosis a clear history ofexcessive alcohol consumption was present. Usuallymore than half a catty double distilled rice wine aday or its equivalent (about 100 g ethanol) wasconsumed, and these patients were considered tohave alcoholic cirrhosis. Ten of the patients in thisgroup had an abnormally large liver and three anormal-sized liver on hepatic scintiscanning. Thepatients were not allowed to take any alcohol for atleast seven days before the study. No peptic ulcerin the stomach or duodenum was demonstrated inany one patient by barium study, endoscopy, atsubsequent surgery and/or at necropsy. Blood ureaand creatinine were below 7 mmol/l and 125 ,umol/lrespectively at the time of the study in all patients.Gastrointestinal bleeding occurred in 12 patients,and for these a period of at least six weeks wasallowed before performing the study. When asciteswas present, this was ameliorated as far as possibleusing diuretics until the attending physician wassatisfied that no further improvement would belikely.

LIVER FUNCTION SCOREThe liver function of each patient was scored,using a modification of Child's classification (Child,1964). Five points were given for previous stage Iand II encephalopathy, 10 points if the encephalo-pathy had been stage III or IV (staging by Adamsand Foley, 1955). Five points were given whenascites were present, 10 points if this was poorlycontrolled. Two and a half points were given whenserum bilirubin was between 25-7-51.3 mmol/l,5 points if this was over 51-3 mmol/l. Two and a halfpoints were given when serum albumin was between25 and 30 g/l, 5 points when this was less than25 g/l. A patient could thus in theory score a maxi-mum of 30 points. Each patient was then graded ashaving mild (score 0-10), moderate (score 11-20),and severe (score 21-30) liver function impairment.

CONTROLSThese were carefully selected from among healthymedical and nursing staffs and patients in the generalmedical wards who had recovered from an unrelatedcondition and who were otherwise healthy. Therewere 20 such subjects, 16 males and four females,

aged from 19 to 65 years (mean = 46-2 years).Informed consent was obtained from all.

GASTRIC ACID OUTPUTThe basal acid output (BAO) and the maximumacid output (MAO) after an intramuscular injectionof pentagastrin (6 ,ug/kg body weight) was measuredin the conventional manner in all the 60 studysubjects (Baron, 1973). The acid content in thegastric juice was estimated by an automatic titrator(Radiometer, Copenhagen, type TTT II).

SERUM GASTRINOn a separate day after an overnight fast, a 19 gbutterfly needle was inserted into an antecubitalvein of the study subject and kept patent by a slowinfusion of normal saline. After two fasting bloodsamples were taken at - 15 minutes and zero time,each subject ingested a standard protein mealcomposed of 50 g protein, 40 g fat, and 40 g carbo-hydrate. Additional blood samples were taken at15, 30, 45, 60, 90, and 120 minutes. In 19 cirrhoticpatients and 10 normal controls, additional bloodsamples were collected at 150, 180, 210, and 240minutes after the meal. Each sample was storedimmediately at 4°C. The serum from all sampleswere extracted by centrifugation immediately at theend of the test and stored at - 20°C for subsequentradioimmunoassay of gastrin.

RADIOIMMUNOASSAY OF SERUM GASTRINThe details of the assay have been previouslyreported (Byrnes et al., 1976; Lam and Lai, 1976).The antiserum used had been raised in rabbit againstgastrin I and If extracted from pig antra using themethod of Gregory and Tracy (1964) and conjugatedto egg albumin. The association constant of theantiserum was 5-2 and 1011 I/mol. Sensitivity of theassay was down to 2-4 pmol/l (5 pg/ml) of serumusing synthetic human gastrin I (Imperial ChemicalIndustries) as standard. Cross-reaction by chole-cystokinin was 2 x 10-4 on a molar basis. Withinand between assay coefficients of variation were6-2% and 13% respectively. In each individual theintegrated gastrin response from zero time to 120minutes after the meal (£ GO-120) was calculatedfrom the area under the gastrin response curvewithin the two-hour period. Similarly the integratedgastrin response between 120 minutes and 240minutes after the meal (£ G120-240) was calculatedin those subjects who had blood taken up to fourhours after the meal.

GASTRIC MUCOSAL MORPHOLOGYTwenty-eight of the cirrhotic patients (20 crypto-genic and eight alcoholic) consented to the carrying

701

on Decem

ber 24, 2019 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.17.9.700 on 1 Septem

ber 1976. Dow

nloaded from

Table 1 Acid output and meal stimulated gastrin response in cirrhosis and controls (mean ± SEM)

No. Basal acid output Maximum acid output X GO-120 E G120-240(mmol/h) (nmollmin/l) (nmol/min/l)

(mmol/h) (mmol/h/kg)Normal 20 2-3 ± 0-3 13-6 1 1 0-27 0-02 54 + 04 5 3 ± 0-8

(10 subjects)CirrhoFsi

Cryptogenic 27 1-0 ± 0 4 6-7 i 15 0-13 0-03 8-2 0-8Alcoholic 13 0 7 ± 0-4 6-7 + 1-8 0-13 ± 003 8-7 1-0Total 40 0 9 ± 0 3 6-7 1-2 013 0-02 8-4 07 8-4 ± 0 9

(19 subjects)

p < 0 01 for all comparisons between normal and cirrhosis.

out of upper gastrointestinal endoscopy, for whichan end-viewing endoscope was used (Olympus,GIF-D2). This was performed by the author withoutprior knowledge of the patient's acid secretion. Atleast five biopsies were randomly taken from theantrum (one from the lesser curve, one from thegreater curve), and corpus of the stomach (two fromthe lesser curve, high and low, and one from thegreater curve) when no abnormality was obviousendoscopically. When abnormalities were detected,further biopsies were taken from these sites.

PATIENTS WITH SPLENORENAL SHUNTIn five of the patients with cryptogenic cirrhosis, atherapeutic splenorenal shunt was subsequentlyperformed. The acid output after pentagastrinstimulation and the serum gastrin response to thestandard protein meal was measured approximatelyone month after surgery. Serum samples for com-parison of gastrin content before and after theshunting operation were determined within the onegastrin assay. No attempt was made to establish thepatency of the shunt in each case.

pressing MAO in relation to body weight has theadvantage of eliminating the difference in acidsecretion in the two sexes (Lam and Sircus, 1975) aswell as in different ethnic groups (Lam, 1974). Therewas no difference in BAO or MAO between thecryptogenic and alcoholic cirrhotic patients.

SERUM GASTRIN (Fig. 1 and Table 1)The mean fasting serum gastrin (at zero time beforethe meal) of the cirrhotic patients as a group(41'6 ± SE 3-1 pmol/l) was significantly higher(p < 0-025) than that of the controls (26'3 ± SE3.4 pmol/l). This holds true even when the patientswere divided into cryptogenic (41'1 ± SE 4 0pmol/l) and alcoholic (42-7 ± SE 5-1 pmol/l) groups(p < 0-025 and 005 respectively).

SERUMGASTRINpmoVL

100n

80-

STATISTICAL ANALYSISStudent's t test was used to compare unpaired data.For comparison of the postprandial gastrin responsebetween groups of subjects, analysis of varianceusing a two-way classification was used, and theresidual variance obtained was used for the calcula-tion of standard error and for carrying out t testwhen needed.

Results

GASTRIC ACIDITY (Table 1)The BAO of the cirrhotic patients, whether as agroup or subdivided into cryptogenic and alcoholic,was significantly (p < 0'01) lower than that of thecontrols. Likewise, the MAO, whether this wasexpressed as mmol/hour or in relation to total bodyweight-namely, mmol/hour/kg TBW-was signifi-cantly (p < 001) lower in the cirrhotic patients thanin the controls. It has been established that ex-

60-

40-

20-

0

P.M

- ALCOHOLIC

o.... CRYPTOGENIC

*-. NORMAL

F- -_ _'

.....

,i

MINUTES

0 30 60 90 120

Fig. 1 Mean serur gastrin levels (+ standard error) innormal controls (20) and in patients with cryptogenic (27)and alcoholic cirrhosis (13): fasting and in response to astandardprotein meal(PM). SEMas computedfromresidual variance = 3-4for values after stimulation.

The serum gastrin response in the two hours afterthe standard protein meal was significantly higherin the cirrhotic patients as a group than in the

702 Shniu Kum Lam

on Decem

ber 24, 2019 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.17.9.700 on 1 Septem

ber 1976. Dow

nloaded from

Hypergastrinaemia in cirrhosis ofliver

SERUMGASTRINpmol/L

100-1

CIRRHOSIS(19)a NORMAL(10)

Fig. 2 Gastrin response to astandardprotein meal (PM) meas-

uredfor four hours in 10 normalcontrols and 19 patients withcirrhosis. SEM as computed fromresidual variance = 5-6for valuesafter stimation.

controls (F = 49-27, p < 001). This is still truewhen the patients were divided into cryptogenicgroup (F = 42Q25, p < 0-01) and alcoholic group(F = 40-78, p < 0-01). The mean serum gastrinlevels at 15, 30, 45, 90 and 120 minutes after themeal were significantly higher in the cirrhoticpatients as a group than in the controls (p < 0 01,0-0125, 0-0025, 0.001, and 0001 respectively). Thisalso holds true for the patients with cryptogeniccirrhosis (p < 0 01, 0-025, 0 01, 0 01, 0 01 respec-tively) and those with alcoholic cirrhosis (p < 0 05,0-05, 0 005, 0 001, 0 001 respectively). A significantlyhigher than normal £ GO-120 was observed in thecirrhotic patients as a group, whether taken as a

whole or when broken down into cryptogenic andalcoholic subgroups (p < 0 01 in all instances).

Because the initial results showed an increasedand sustained release of gastrin in the two hoursafter the protein meal in the cirrhotic patients thisgastrin response was followed for a total of fourhours in the subsequent 19 patients and 10 normalcontrols studied (Fig. 2). The results indicated thatthe serum gastrin response declined in the thirdand fourth hour, but this response was still signi-ficantly higher than that of the normal controls(F = 21-84, p < 0-01). The mean serum gastrinlevels at 150, 180, 210, and 240 minutes after themeal were significantly higher than the correspon-

ding values in the controls (p < 0-001 for all pointsof time). In addition, the mean I G120-240 was

significantly higher in the cirrhotic patients than inthe controls (p < 0-01).

RELATIONSHIP BETWEEN ACID OUTPUT AND

GASTRIN RELEASE (Figs. 3 and 4)In the normal controls, a significant inverse cor-relation (r = 0-4929, p < 0-05) was observedbetween the MAO when this was expressed inmmol/hour/kg TBW and the GO-120. Therelationship between MAO as expressed in mmol/hour and Z GO-120 just failed to reach statisticalsignificance (r = 0-4423, p > 0-1).The MAO/kg TBW and Z GO-120 were correlated

in the cirrhotic subjects, both as a group and whendivided into cryptogenic and alcoholic subgroups.A significant inverse correlation was observed ineach case (r = 0-5237, p < 0-001; r = 0-4559,P < 0-05; r = 0-7286, p < 0-01).

In the 10 normal controls and 19 patients withcirrhosis who had the G120-240 measured, thiswas correlated with the MAO/kg TBW. No signi-ficant correlation was observed in the normalcontrols(p > 0- 1), but a significant inverse correlationwas present in the cirrhotic patients (r = 0-5032,p < 0-05),No significant relationship was demonstrable in

the normal controls or patients with cirrhosis betweenBAO and fasting (zero time) serum gastrin (r =

0-0998, r = 0-2968 respectively), as well as betweenBAO and GO-120 (r = 0-3843, r = 0-1885respectively). The corresponding coefficients for thecryptogenic cirrhotic group (r = 0-2882, r = 0-2138respectively) and alcoholic cirrhotic group (r =

0-3343, r = 0-3258 respectively) were also insigni-ficant.

PM. -k_

.r >s~~~~~~~~~~~~~~~~Ii,

80-

60-

40-

20-

0

MINUTESt

0 30 60 90 120 150 180 210 240

703

I.Z"

ii

0-

on Decem

ber 24, 2019 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.17.9.700 on 1 Septem

ber 1976. Dow

nloaded from

Shiu Kum Lam

I GO-120nmol.min/L

0251

r 0.52p (0.001

0*

oO

O a

0o

I

r:0.49 p 0.05 7

a ** I a

I-r0 0-2 0-4

MAXIMUM0-6

ACID

1G120-240nmolmin/L

1 0 r:0.50p <0.05

O

° 8\° \

iI I

r:0.22 p,0.1a

* U

I.I I0 0-2 0.4OUTPUT mmol/hrkg

Fig. 3 Relationship between maxi-mum acid output in mmol/h/kg totalbody weight (MAO/kg) and integratedgastrin response in thefirst two hours(2 GO-120) and in the third andfourthhours (_ G120-240) after ingestionofa standardprotein meal in normalcontrols and in patients with cirrhosis(open circles = cryptogenic cirrhosis,closed circles = alcoholic cirrhosis).

Table 2 Relationship ofliverfunction to acidandgastrin secretion (mean ± SEM)

Liverfunction score

0-10 11-20 21-30

Total cirrhosis Cryptogenic Total cirrhosis Cryptogenic Total cirrhosis CryptogenicNo. 23 13 10 8 7 6BAO (mmol/h) 0 9 + 0-40 0-5 ± 0.2d 0-6 t 0-2b 0 4 00-le 13 i 0.8c 1-6 0 9MAO (mmol/h) 6-2 ± 1-5a 57 ± 19d 5.9 ± 1-2b 6-2 1-4e 9.7 ± 4-6c 9-6 5-4fMAO (mmol/h/kg) 0-12 ± 0.030 0-12 ± 0.04d 0-12 ± 0.02b 0-12 ± 0.02e 0-18 ± 0.08c 0-18 + 0-IfrGO-120(nmol.min/1) 8-3 ± 1Os 8-3 + 1-6d 8-7 ± 1-2b 8-2 ± 0*9e 8 0 ± 1-2c 8-2 + 1-4t

aandbaandcbandc J

dande Inot significant d and f not significant

eandf J

RELATIONSHIP BETWEEN LIVER FUNCTIONAND ACID AND GASTRIN SECRETIONS (Table 2)No difference in mean BAO, MAO (in mmol/hourand mmol/hour/kg TBW) and £ GO-120 wasobserved between groups of cirrhotic patients havingdifferent degrees of liver function impairment as

assessed by the scoring system. There appeared to bea trend for patients with a score of 21-30 (severe

impairment) to have a higher basal and post-pentagastrin acid output than those with lower score(mild to moderate impairment). This trend becamemore obvious if the cryptogenic cirrhotic patientsalone were considered, but the difference was notstatistically significant. No attempt was made toanalyse the patients with alcoholic cirrhosis, as thegroups with moderate or severe disturbance in liver

704

4-

0

lmUV

-I

40z

20

15

10

5-

0

10 -

5-

o

.

0

0

a

s

on Decem

ber 24, 2019 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.17.9.700 on 1 Septem

ber 1976. Dow

nloaded from

Hypergastrinaemia in cirrhosis ofliver

NORMAL CRYPTOGENIC

r 038

a Ua a. * -

0

00fooo000

r:0-21

0

0

r:O33

r

I I 1 I I 1

r:o.io

.

taL

.a

.

%; a

I I

0 4 8

r:0o29

0o0

0 0

0 000 0O0o0 O0

00I I

4 8 0

r 033

r

Fig. 4 Relationshipbetween basal acidoutput (BAO) andfastingserumgastrin,andbetween BAOandintegratedgastrinresponse in the twohours after ingestionofa standardproteinmeal (Q GO-120) innormal controls andin patients withcirrhosis.

I I I4 8

BASAL ACID OUTPUT mmol/hour

SERUMGASTRINpmol/l

100'-

80-

60-

40-

20-

0

o----- BEFORE SHUNT

AFTER SHUNT

tPM

MINUTES

.II0 30 60 90 120

Fig. 5 Gastrin response to a standardprotein meal(PM)infivepatients with cryptogenic cirrhosis: before andonemonth after splenorenal shunt. Mean + SEM.

Table 3 Acidsecretion before andafter splenorenalshunt (mean + SEM)

BAO MAO MAO/kgBefore 1 1 ± 0.3* 6-7 ± 0 7* 0-15 ± 0-02*After 3 0 i 1-2t 8-8 ± 3 0t 0-21 i 0-08t

*andtp > 0-1

function were too small for statistical evaluation.

EFFECT OF SPLENORENAL SHUNT ON ACID

AND GASTRIN OUTPUT (Fig. 5 and Table 3)The mean BAO and MAO appeared higher afterthe operation in the five patients studied, but thedifferences were not statistically significant (Table 3).Figure 5 shows that the mean fasting and post-prandial gastrin levels after surgery were no differentfrom the ones before surgery.

MORPHOLOGY OF GASTRIC MUCOSA (Tables 4and 5)In general, the commonest endoscopic finding wasdiffuse congestion of the whole of the gastricmucosa, as evidenced by the presence of generalisedhyperaemia with or without the presence of oedema.This occurred in 15 of the 28 patients. The nextmost common finding was mucosal atrophy,suggested by the fact that the mucosa appeared thinso that the underlying vessels became clearly visible.This was recorded in eight patients and was alwaysobserved in the corporal and fundic regions. Activegastritis was diagnosed when patchy areas ofhyperaemia, granularity or nodularity, and/orundue friability were observed. This was found inthree patients. In seven patients no abnormalitywas detected.

20-

10-

mm

5' EU-0Coo0 E

ALCOHOLIC

705

0

.-

-I0

la.

20-

10-

0

01

.

.

.

on Decem

ber 24, 2019 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.17.9.700 on 1 Septem

ber 1976. Dow

nloaded from

Shiu Kum Lam

Table 4 Upperpanel: gastroscopicfindings in 28cirrhotic patients. Lowerpanel: number ofpatients withcorrelating histology

Normal Mucosal Mucosal Active atrophiccongestion atrophy gastritis

Endoscopy 7 15 8 3Histology 7 7 2 3

Normal Superficial Atrophic Atrophic gas-gastritis gastritis tritis severe

mild and activeScore 0 1 2 3

Table 5 Relationship ofgastric mucosal morphologywith acid secretory capacity (MAO), gastrin secretion( E GO-120) and liverfunction impairment

Mucosal morphology score

0 1-3 4-7No. 7 15 6MAO (mmol/h) 6-7 ± 3-7a 6 5 ± 2-5b 5.4 ± 0.7cMAO (mmol/h/kg) 0-12 + 0 07^ 0-12 ± 0-04b 0 12 ± 0-02cZGO-120(nmol/min/1) 99 2-8a 8-6 0-6b 7-8 ± 0.9

Liver function score 6-8 1iSa 9-0 1-7b 10 4 ± 2-1C

aandbba and c not significantb and c

The histological findings were categorised asnormal, superficial gastritis, mild atrophic gastritis,and severe active atrophic gastritis according toWhitehead (1973). In general, the findings correlatefairly well with the endoscopic findings of normalmucosa, mucosal congestion, and active atrophicgastritis, but only poorly so with the endoscopic diag-nosis of mucosal atrophy (Table 4). The commonestoccurrence in the histological sections was normalhistology, there being 14 such reports, for which theendoscopic diagnoses were normal (seven), mucosalcongestion (three), andmucosalatrophy (four). Super-ficial gastritis was reported in nine cases and was ob-served in both antral and corporal biopsies. Atrophicgastritis was seen in five cases and was found in allcases in the corporal and fundic biopsies, corres-ponding to the endoscopic diagnosis of mucosalatrophy and active atrophic gastritis in these regions.The endoscopic findings were scored, giving one

point if mucosal congestion was present, two pointsif mucosal atrophy was seen, and three points if thiswas associated with features of active gastritis. Amaximum of four points could thus be scored.Similarly, the histological findings were scored,giving one point if superficial gastritis (oedema andmild inflammation) was observed, two points ifmild atrophic gastritis was present, or three pointsif the gastritis was severe and active, so that a maxi-mum of three points could be scored.

The total score (endoscopy and histology) wascorrelated with the MAO, 2 GO-120 and liverfunction score. No significant relationship, however,was observed in each case. There appeared to be atrend, however, for those with more severe liverfunction impairment to have more severe mucosalabnormality (Table 5). Comparison of the mor-phology score between the cryptogenic (2-5 ± SE0-5) and alcoholic (1-8 ± SE 0-7) groups revealedno significant difference (p > 0-2).

Discussion

Mazzacca et al. (1974) observed higher than normalserum gastrin levels both in the basal state and afteran oral dose of glycine in patients with cirrhosis.The post-glycine hypergastrinaemia, however, isdifficult to interpret, as the patients were given aparenteral dose of atropine before the oral glycine.The present study confirms the presence of anincreased basal serum gastrin and demonstrates thatpatients with cirrhosis, be this cryptogenic oralcoholic in origin, do indeed have an abnormallyhigh and, in addition, an abnormally prolongedgastrin release in response to a physiological meal(Figs. 1 and 2). The capacity to secrete acid, whetherin the basal state or upon stimulation by pentagastrinis appreciably lower than normal (Table 1), afinding that agrees with most previous reports(Scobie and Summerskill, 1954; Ostrow et al.,1960; Schmidt and Martini, 1969) but differs fromothers (Tabaqchali and Dawson, 1964; Mazzaccaet al., 1974). Although there is no relationshipbetween BAO and fasting serum gastrin, a significantreciprocal relationship is present between the acidsecretory capacity (as expressed in MAO/kg TBW)and the post-prandial gastrin response (expressed asthe integrated gastrin response, 2 GO-120 and2 G120-240). This inverse relationship (betweenMAO/kg and 2 GO-120) is also seen in the normalcontrols, agreeing with previous observations(Byrnes et al., 1976). These findings thus suggest aninverse relationship between the functioning parietalcell mass and the functioning gastrin cell mass inthe cirrhotic subjects and also suggest that theincreased gastrin release is a result of gastric hypo-acidity and hence decreased feedback inhibition ofgastrin secretion. The significance of this in thepathogenesis of the alleged higher incidence ofpeptic ulcer in cirrhotic patients as reported by many(for references, see introductory section) is unknown.It is of interest to note that an inverse relationshipbetween gastrin release and gastric acidity has beendescribed in atrophic gastritis (Ganguli et al., 1971;Strickland et al., 1971) and in gastric ulceration(Trudeau and McGuigan, 1971; Korman et al.,

706

on Decem

ber 24, 2019 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.17.9.700 on 1 Septem

ber 1976. Dow

nloaded from

Hypergastrinaemia in cirrhosis of liver 707

1972), there being hypergastrinaemia and hypo-acidity in both situations.The hypergastrinaemia and hypoacidity in the

cirrhotic patients is unrelated to the degree of liverfailure as assessed by the liver function scoringsystem (Table 2). The same scoring system has beenfound to be of definite use as an index of the chanceof survival in cirrhotic patients undergoing emer-gency ligation of bleeding oesophageal varices (Onget al., 1976). As far as the role of the liver in theinactivation of endogenous gastrin is concerned,studies so far (Temperley et al., 1971; Reeder et al.,1972; Dencker et al., 1973; Debas and Grossman,1974) indicate that this is unlikely to be of anyimportance. This is further supported by theobservation that a splenorenal shunting operationdoes not affect the serum gastrin level (Fig. 5).

Is the gastric hypoacidity in any way linked withthe gastric mucosal morphology? As assessedendoscopically and by multiple biopsies, a moreconsistent picture of abnormality of the mucosaappears to be mucosal congestion with occasionalatrophic gastritis. This picture, however, can at bestpartly explain the gastric hypoacidity, as there is norelationship between the mucosal morphology andgastric acidity (Table 5). As expected, no relationshipexists between the post-prandial gastrin responseand mucosal abnormality. It may be argued that themethod of histological assessment of the mucosalabnormality is incomplete, as this is based on biopsyspecimens, although these were multiple, endosco-pically guided, and, as far as possible, representativeof both the antrum and the corpus. A more likelyexplanation may lie in the observation that certainpeptides of intestinal origin inhibit acid secretion-namely, secretin (Johnson and Grossman, 1971),gastric inhibitory polypeptide (Pederson and Brown,1972), and vasoactive intestinal peptide (Barbezatand Grossman, 1971; Rayford et al., 1974). Vaso-active intestinal peptide has been observed to bepresent in the serum in abnormally high concen-tration in patients with cirrhosis (Said et al., 1974)and not to inhibit gastrin release except in unphy-siologically high doses (Rayford et al., 1974). Inaddition, oral ingestion of alcohol results in a largerelease of secretin (Straus et al., 1975). The role ofthese peptides in the genesis of gastric hypoacidityin cirrhosis is only speculative and remains to bedefined.

The author is particularly indebted to Professor D.Todd for his valuable criticisms and suggestionsconcerning this study. He wishes to thank SisterIvy Cheng and members of her staff for performingthe acid tests, and Mr E. Ng for his technicalassistance in the gastrin assay.

References

Adams, R. D., and Foley, J. M. (1953). The neuroIogicaldisorder associated with liver disease. Proceedings of theAssociation for Research in Nervous and Mental Disease,32, 198-237.

Barbezat, G. O., and Grossman, M. I. (1971). Intestinalsecretion: stimulation by peptides. Science, 174, 422-424

Baron, J. H. (1973). The clinical application of gastricsecretion measurements. Clinics in Gastroenterology, 2,293-314.

Bergman, F., and van der Linden, W. (1965). The associationof peptic ulcer with cirrhosis of the liver, an analysis of anautopsy series. Acta Pathologica et Microbiologica Scan-dinavica, 65, 161.

Byrnes, D. J., Lam, S. K., and Sircus, W. (1976). The relationbetween functioning parietal cell and gastrin cell masses intwo groups of duodenal ulcer patients. Clinical Scienceand Molecular Medicine, 50, 375-383.

Child, C. G., III (1964). The Liver and Portal Hypertension,p. 50. Saunders: Philadelphia.

Cook, J., McFadzean, A. J. S., and Todd, D. (1963).Splenectomy in cryptogenic splenomegaly. British MedicalJournal, 2, 337-344.

Debas, H. T., and Grossman, M. I. (1974). Hepatic in-activation of gastrointestinal hormones. Abstracts ofFifth World Congress of Gastroenterology, Mexico,Oct. 13-19, p. 494.

Dencker, H., HAkanson, R., Liedberg, G., Norryd, C.,Oscarson, J., Rehfeld, J. F., and Stadil, F. (1973). Gastrinin portal and peripheral venous blood after feeding in man.Gut, 14, 856-860.

Doll, R. (1952). Peptic ulcer. I. Epidemiology. In ModernTrends in Gastroenterology, pp. 361-379. Edited by F.Avery Jones. Butterworth: London.

Fainer, D. C., and Halsted, J. A. (1955). Sources of upperalimentary tract hemorrhage in cirrhosis of the liver.Journal of the American Medical Association, 157, 413-414.

Ganguli, P. C., Cullen, D. R., and Irvine, W. J. (1971).Radioimmunoassay of plasma-gastrin in perniciousanaemia, achlorhydria without pernicious anaemia,hypochlorhydria, and in controls. Lancet, 1, 155-158.

Gregory, R. A, and Tracy, H. J. (1964). The constitutionand properties of two gastrins extracted from hog antralmucosa. Part I. The isolation of two gastrins from hogantral mucosa. Gut, 5, 103-117.

Grossman, M. I. (1974). Pathophysiology of gastric andduodenal ulcers. Abstracts of Fifth World Congress ofGastroenterology, Mexico, Oct. 13-19, p. 9.

Hal.n, J., and Krook, H. (1963). Follow-up studies on anunselected ten-year material of 360 patients with livercirrhosis in one community. Acta Medica Scandinavica,173,479-493.

Jackson, F. C., Perrin, E. B., Felix, W. R., and Smith, A. G.(1971). A clinical investigation of the portacaval shunt:V. Survival analysis of the therapeutic operation. Annalsof Surgery, 174, 672-701.

Jackson, F. C., Perrin, E. B., Smith, A. G., Dagradi, A. E.,and Nadal, H. M. (1968). A clinical investigation of theportacaval shunt. II. Survival analysis of the prophylacticoperation. American Journal of Surgery, 115, 22-42.

Johnson, L. R., and Grossman, M. I. (1971). Intestinahormones as inhibitors of gastric secretion. Gastroenter-ology, 60, 120-144.

Korman, M. G., Soveny, C., and Hansky, J. (1972). Gastrinstudies in gastric ulcer Gut, 13, 166-169.

Lam, S. K. (1974). Clinical and Pathophysiological Studies owrDuodenal Ulcers: Evidence for the Existence of TwoPopulations. M.D. Thesis: University of Hong Kong.

Lam, S. K., and Lai, C. L. (1976). Inhibition of sulpiride oa

on Decem

ber 24, 2019 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.17.9.700 on 1 Septem

ber 1976. Dow

nloaded from

708 Shiu Kum Lam

the cephalic phase of gastric acid and gastrin secretion induodenal ulcer patients. Scandinavian Journal of Gastro-enterology, 11, 27-31.

Lam, S. K., and Sircus, W. (1975). Studies on duodenalulcer. I. The clinical evidence for the existence of twopopulations. Quarterly Journal of Medicine, 44, 369-387.

Lipp, W. F., and Lipsitz, M. H. (1952). The clinical signi-ficance of the co-existence of peptic ulcer and portalcirrhosis, with special reference to the problem of massivehemorrhage. Gastroenterology, 22, 181-189.

Mazzacca, G., Budillon, G., De Marco, F., and De Ritis, F.(1974). Serum gastrin in patients with cirrhosis of theliver. Digestion, 11, 232-239

Ong, G. B., Cheng, F. C. Y., Lam, W. K., and Lam, S. K.(1976). An experience in transabdominal ligation of bleed-ing gastro-oesophageal varices. Proceedings of the VthAsian-Pacific Congress in Gastroenterology, May 23-291976. (In press.)

Orloff, M. J., Chandler, J. G., Alderman, S. J., Keiter, J. E.,and Rosen, H. (1969). Gastric secretion and peptic ulcerfollowing portacaval shunt in man. Annals of Surgery, 170,515-527.

Ostrow, J. D., Timmerman, R. J., and Gray, S. J. (1960).Gastric secretion in human hepatic cirrhosis. Gastro-enterology, 38, 303-313.

Palmer, E. D., and Brick, I. B. (1953). Sources of uppergastrointestinal hemorrhage in cirrhotic patients withesophageal varices. New England Journal of Medicine, 248,1057-1058.

Patek, A. J., Jn. (1963). Portal cirrhosis (Laennec's cirrhosis).In Diseases of the Liver, 2nd edn. Edited by L. Schiff.Lippincott: Philadelphia.

Pederson, R. A., and Brown, J. C. (1972). Inhibition ofhistamine-, pentagastrin-, and insulin-stimulated caninegastric secretion by pure 'gastric inhibitory polypeptide'.Gastroenterology, 62, 393-400.

;Phillips, M. M., Ramsby, G. R., and Conn, H. 0. (1975).Portacaval anastomosis and peptic ulcer: a nonassociation.Gastroenterology, 68, 121-131.

"Ratnoff, 0. D., and Patek, A. J., Jr (1942). The naturalhistory of Laennec's cirrhosis of the liver: an analysis of386 cases. Medicine, 21, 207-268.

,Rayford, P. L., Villar, H. V., Reeder, D. D., and Fender, H. R.(1974). Effect of GIP and VIP on gastrin release and gastricsecretion. Physiologist, 17, 319.

iReeder, D. D., Brandt, E. N., Jr, Watson, L. C., Hjelmquist,U. B. E., and Thompson, J. C. (1972). Pre- and posthepatic.measurements of mass of endogenous gastrin. Surgery

72,34-41.Resnick, R. H., Chalmers, T. C., Ishihara, A. M., Garceau,A. J., Callow, A. D., Schimmel, E. M., O'Hara, E. T., andthe Boston Inter-Hospital Liver Group (1969). A con-trolled study of the prophylactic portacaval shunt. A finalreport. Annals of Internal Medicine, 70, 675-688.

Said, S. I., Faloona, G. R., Deon, H., Unger, R. H., andSiegel, S. R. (1974). Vasoactive intestinal polypeptide:elevated levels in patients with hepatic cirrhosis. ClinicalResearch, 22, 367A.

Schmidt, H. A., and Martini, G. A. (1969). Gastric secretionin chronic liver disease. German Medical Monthly, 14, 246-254.

Schnitker, M. A., and Hass, G. M. (1934). A histologic studyof the liver in patients affected with peptic ulcer. AmericanJournal of Digestive Disecses, 1, 537-543.

Scobie, B. A., and Summerskill, W. H. J. (1954). Reducedgastric acid output in cirrhosis: quantitation and relation-ships. Gut, 5,422-428.

Straus, E., Urbach, H. J., and Yalow, R. S. (1975). Alcohol-stimulated secretion of immunoreactive secretin. NewEngland Journal of Medicine, 293, 1031-1032.

Strickland, R. G., Bhathal, P. S., Korman, M. G., and Hansky,J. (1971). Serum gastrin and the antral mucosa in atrophicgastritis. British Medical Journal, 4, 451-453.

Sullivan, B. H., Jr, Lienhard, W. F., Jr, and Lukeman, J. M.(1954). Massive hemorrhage in patients with both eso-phageal varices and duodenal ulcer. Gastroenterology, 26,868-877.

Swisher, W. P., Baker, L. A., and Bennett, H. D. (1955).Peptic ulcer in Laennec's cirrhosis. American Journal ofDigestive Diseases, 22, 291-294.

Tabaqchali, S., and Dawson, A. M. (1964). Peptic ulcer andgastric secretion in patients with liver disease. Gut, 5,417-421.

Temperley, J. M., Stagg, B. H., and Wyllie, J. H. (1971).Disappearance of gastrin and pentagastrin in the portalcirculation. Gut, 12, 372-376.

Trudeau, W. L., and McGuigan, J. E. (1971). Relationsbetween serum gastrin levels and rates of gastric hydro-chloric acid secretion. New England Journal of Medicine,282, 408-412.

Welbourn, R. B. (1952). Benign ulcers of the greater cur-vature associated with ulcerative colitis and cirrhosis ofthe liver. British Journal of Surgery, 39, 303-308.

Whitehead, 0. (1973). Simple (non-specific) Gastritis. InMucosal Biopsy of the Gastrointestinal Tract, pp. 17-32.Saunders: Philadelphia.

on Decem

ber 24, 2019 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.17.9.700 on 1 Septem

ber 1976. Dow

nloaded from