403
Discussion Tests which could identify an individual’s specific diet-
Of the 20 patients who completed the trial, 14 res-
ponded more favourably to the antigen-avoidance diet
than to the control diet, whereas only 1 responded more
favourably to the control diet than to the trial diet. We
conclude that this antigen-avoidance diet will induce a
clinical improvement in many children with atopic
eczema. Of those who did not complete the trial twice as
many lapsed or withdrew during or after the control
period than did so during or after the trial diet period,
perhaps because of the lack of benefit experienced.
Analysis of the results of this trial revealed a rro-
nounced order effect; patients experienced more benefit
during the first diet period than the second, whatever
the nature of the diet. The mean severity of the eczema
(calculated by initial area score) of the children having
the trial diet first was very similar to that of those hav-
ing the control diet first. This order effect is difficult to
explain, and was observed in both the clinical and the
symptomatic scores. Nevertheless, the actual diet
effect-i.e., the effect which depended upon the actual
content of the diet-was always greater than the order
effect, except in the case of sleeplessness.
3 patients stopped all symptomatic treatments after
eczema virtually cleared during the antigen-avoidance
period. These 3 patients had had relatively mild eczema.
There may be several explanations for the incomplete-
ness of the response shown by the patients who im-
proved without complete clearance of eczema. A limited
range of allergenic foods was excluded, and many of the
patients may have been allergic to other foods. Incom-
plete adherence may have reduced the effectiveness of
the diet; patients who admitted more than five lapses
were excluded, but, presumably, there were unrecog-
nised and unadmitted lapses. A third possible explana-
tion may be the short period for which the diet was
given. Further experience suggests that six to twelve
weeks may be necessary for initial assessment of the
diet’s efficacy, though four weeks is adequate in most pa-
tients.
To increase the effectiveness of dietary treatment, the
antigens which are important in individual patients will
have to be identified. Of the 3 patients who gave a his-
tory of allergy (including exacerbation of eczema) to
eggs or cows’ milk, 2 benefited from antigen avoidance
in our study but 1 did not. Patients benefited from the
diet whether or not they had a history of overt allergy
to the excluded foods. Positive reactions to selected prick
test antigens confirm the atopic status of a child, but
since all our subjects reacted, we cannot comment on
whether this test is useful in selecting patients for diet-
ary treatment. The general impression that prick tests as
presently performed do not give information of thera-
peutic relevance in atopic eczema4is supported by our
finding of a lack of correlation between positive prick
tests to egg and cows’ milk antigens and response to the
diet. This lack of correlation may be explained by the
possibility that the antigens used for prick testing may
not be the antigens responsible for the disease. Food
antigens may be altered by digestion, which leads to the
production of relevant antigens. Furthermore, although
IgE-mediated responses occur in patients with atopic
eczema, we do not know whether they play a part in the
pathogenesis of eczema.
ary allergies would be of immense value in planning
more effective treatment. With this aim in mind, we are
currently investigating antibody and cell-mediated re-
sponses to food antigens in our patients, to determine
the criteria which would identify those benefiting most.
The results of this study suggest that the value of diet-
ary treatment in atopic eczema is greater than has been
widely appreciated. Similarly, appraisal is
critical
needed for avoidance of other dietary antigens in pa-
tients responding incompletely to cows’ milk and egg
avoidance alone.
We thank Cow & Gate Baby Foods for the soya and cows’ milk
powders. The late Dr R. Mendoza performed the prick tests on
children taking part in the trial. D.J.A. was in receipt of a Research
Training Fellowship from the National Fund for Research into Crip-
pling Diseases.
REFERENCES
1. Meara, R. H. Br. J. Derm. 1955, 67, 60.
2. Matthew, D. J., Taylor, B., Norman, A. P., Turner, M. W., Soothill, J. F.
Lancet, 1977, i, 321.
3. Smart, J. V. Elements of Medical Statistics; p. 108. London, 1970.
4. Parish, W. E., Champion, R. H. in Recent Advances in Dermatology (edited
by A. Rook); p. 193. Edinburgh, 1973.
MAINTENANCE TREATMENT OF
RECURRENT PEPTIC ULCER BY
CIMETIDINE
A. WALAN
G. BODEMAR
Departments of Clinical Pharmacology and Internal Medicine,
University Hospital, 581 85 Linköping, Sweden
68 patients with chronic peptic ulcer
Summary
took part in a double-blind trial in which
32 received the histamine H2-receptor antagonist cime-
tidine, (400 mg twice daily) and 36 received a placebo,
for a year. 6 of those on cimetidine had an endoscopi-
cally proven recurrent ulcer within a mean of 7 months,
while 30 of those on placebo had one after a mean of 4
months (P<0&middot;0005). 1 patient on cimetidine had two
recurrences compared with 12 patients on placebo
(P<0&middot;0005). No patient in the cimetidine group had
complications compared to 4 patients in the placebo
group, 2 of whom had mel&aelig;na and h&aelig;matemesis and 2
mel&aelig;na alone (P<0&middot;05). 15 patients on placebo and 1 on
cimetidine were referred for surgery during the trial
because of symptoms from their recurrent ulcers. Day
and night ulcer pain and antacid consumption were sig-
nificantly less, and general wellbeing better, in the cime-
tidine group. The only probable side-effect of cimetidine
was reversible drug-induced liver damage of hypersensi-
tivity type in 1 patient.
Introduction
CIMETIDINE, histamine H2-receptor antagonist, pro-
a
motes healing peptic ulcers and reduces ulcer pain
of
and antacid consumption when compared with a pla-
cebo,I-4 but its effect does not differ statistically from
that of frequent, large doses of antacids.1,6 It has been
404
suggested that, after cimetidine has been stopped, there
is a risk of of peptic ulcerbut there was
early relapse
no evidence for this when preliminary data from several
trials were pooled;8 there was also no increase in acid
secretion 48 hours after 3 months’ open treatment.9
ment.9
The aims of our 1-year double-blind trial were to
assess: the effect of continuous maintenance treatment
with cimetidine on the frequency of relapse of active
ulceration compared with the effect of a placebo; acid
secretion before, during, and after long-term treatment;
and the safety of long-term treatment with cimetidine.
Patients and Methods
Patients
68 adults with proven recurrent ulcers (i.e., those having at
least one previous ulcer demonstrated either on X-ray or
endoscopically, and an endoscopically proven ulcer in the 6
months before the start of the trial) were included. Ulcers were
classified according to whether they were in the duodenal bulb
(D.u.), in the prepyloric region (r.u.), or along the lesser curva-
ture proximal to the gastric angle (G.u.). Biopsies were done on
stomach ulcers to exclude malignancy.
65 patients had been treated with 1 g cimetidine daily until
their endoscopically proven ulcers had healed (56 patients im-
mediately before and 9 patients in the six months before enter-
ing the double-blind trial); 3 patients had been treated with
other regimens until their ulcers had healed. Patients entered
the trial only when endoscopy revealed no ulcers; they were
randomly allocated to maintenance treatment for up to 1 year
with either placebo or with 400 mg cimetidine twice a day.
The patients were followed-up monthly while on main-
tenance treatment and fortnightly while on open treatment
with therapeutic doses of cimetidine. Data recorded included:
the number of trial tablets taken; antacids taken; severity of
day and night pain (mild or severe); other symptoms; and the
patients’ assessment of their general wellbeing (very good,
good, fair, bad, or very bad); and the number of days off work
because of their ulcers. Attacks of pain and antacid consump-
tion have been summated for each individual and are pre-
sented as a pain-score and as antacid-consumption per month
of maintenance treatment. Endoscopy was repeated after 7
months and at the end of the trial or when the patients had
symptoms suggesting recurrence.
If an ulcer developed, open treatment with cimetidine 1 g
daily was given until further endoscopies showed complete
healing, when patients returned to the maintenance trial to
receive the same treatment they had originally been allocated.
If a second relapse occurred during the trial, patients either
received open treatment with cimetidine or, since most patients
had previously been considered for surgery because of their
recurrent illness, they were referred for selective proximal
vagotomy. 2 patients who had severe ulcer symptoms and
endoscopically proven severe duodenitis (plus a severe gastritis
in 1) but who had no ulcers were counted as relapses and
received open treatment with cimetidine.
Gastric Secretion
A standard pentagastrin test was done: before the trial; dur-
ing the 5th week of the first open course of treatment for recur-
rence ; after 7 months on maintenance treatment; just before
the end of the trial; and 2 days and 3 imonths after the com-
pletion of the long-term treatment. Basal acid-output (B.A.o.)
was measured for 1 hour before a subcutaneous injection of
pentagastrin (6-0 g/kg bodyweight) was given; aspiration of
acid continued for a further hour to measure maximum acid-
output and peak acid-output (B.A.O. measurements started
1.5-2 hours after a trial tablet had been taken). Samples for
blood-levels of cimetidine were taken 1.5, 2, and 2.5hours
after a tablet had been taken. The blood-level of cimetidine
was measured by high-pressure liquid chromatography and
ultraviolet-detection.
Laboratory Investigations
Measurements made before the trial and at every follow-up
were: haemoglobin level, red and white cell indices, platelet-
counts, plasma-electrolytes, serum-transaminases, serum-bili-
rubin, serum-alkaline-phosphatase, serum-creatinine; blood-
urea-nitrogen. Urine was examined for deposits, ketones, pro-
teins, and glucose.
Statistical Methods
Results were analysed by: the x2 test to assess ulcer-relapse
and complication-rate; Student’s t-test for acid secretion data
and analysis of changes in laboratory values; and Wilcoxon’s
rank sum test for analyses of pain-score, antacid consumption,
general wellbeing, and days off work.
Results
68 patients (42 D.u., 17 P.u., and 9 G.u.) with recur-
rent peptic ulcers entered the trial (table I). 32
patients--4 women and 28 men with mean age of 50.7
years (range 32-67 years)-received cimetidine 400 mg
TABLE I-CLINICAL OUTCOME OF TRIAL
n=no. of patients in the group.
twice daily; and 36 patients-6 women and 30 men with
mean age of 50-1years (range 30-64 years)-received
placebo twice daily. The mean duration of illness was
13-6 years in the cimetidine group and 10 7 years in the
placebo group. 17 patients in the cimetidine group had
had complications (1 patient had had a perforation,5
had bled once and 11 twice or more). In the placebo
group 11 patients had had complications (1 had a per-
forated ulcer, 4 had bled once and 6 twice or more). 88%
of the patients in the cimetidine group were smokers
compared with 75% in the placebo group.
Patients in the cimetidine group received maintenance
treatment for a mean of 10-5 months (range 3.5-13
months). Because 15 patients in the placebo group were
sent for surgery during the trial, placebo patients
received treatment for a slightly shorter period (mean
9-5months, range 3.5-13 months).
Endoscopy Findings (Table I)
6 out of 32 patients on cimetidine (3 D.u. and 3 P.u.)
had endoscopically proven recurrent ulcers after a mean
of 6.8months (range 4.5-12 months) and 30 out of 36
patients on placebo (18 D.u., 8 P.u., and 4 G.u.) after a
mean of 3-9 months (range 0. 5-11 months) (X2 96’1,
=
P<0.0005).
2 of the 6 patients with recurrences in the cimetidine

group and 7 of the 30 patients with recurrences in the
placebo group had no or mild symptoms although an
ulcer was seen at the routine endoscopies at 7 months or
at the end of the trial. All except 1 patient had their
recurrent ulcers located in the same region as their orig-
inal ulcer. 1 patient (D.U.) on cimetidine and 12 patients
(7 D.u., 3 P.U., and 2 G.u.) on placebo had a second
recurrence during the trial (x2 = 13 1; P<0-000.5).
No patient in the cimetidine group but 4 (2 D.u., 1
P.u., and 1 G.u.) in the placebo group had complications
during the trial (x,2=4.00; p<0.05) 2 had melaena and 2
both melxna and haematemesis, and 3 of these 4 were
the only patients who required hospital admission dur-
ing the trial.
1 patient on cimetidine and 15 patients on placebo
were operated on during the trial either because they
had two recurrences or because of severe symptoms at
the first recurrence (=19.6, p<0.00051
45% of the placebo group, but only 20% of the cimeti-
dine group, still had moderate or severe macroscopic
duodenitis and gastritis and/or more than one erosion at
the routine endoscopies after 7 months and 12 months
treatment although there were no ulcers.
Symptoms
Since D.u., P.u., and G.u. patients all behaved simi-
larly with respect to relief of pain, antacid consumption,
general wellbeing, and days off work, the data are pre-
sented for treatment or placebo groups as a whole.
Fig. 1-Eect of treatment on day pain.
Unshaded area mild pain; shaded area severe pain
= =
405
Figs. 1 and 2 show that the cimetidine group had a
significant reduction in the incidence of both day and
night pain when compared with the placebo group
(p<0.001). 20 out of 32 patients receiving cimetidine
did not have any ulcer pain throughout the whole trial
compared with 6 out of 36 patients receiving placebo.
Antacid consumption per month was also significantly
lower in the cimetidine group (p<0.001) (fig. 3). The pa-
tients’ overall assessment of their general wellbeing was
significantly better in the cimetidine-treated group (lst
month, p<0.02; 2nd month, p<0.02; 3rd month
P<0.001; 4th month, r<001). Comparisons were not
carried out beyond the 4th month because of the large
number of patients in the placebo group who had recur-
rences.
1 of the 32 patients on cimetidine was off work for 79
days (mean 2.8 days per patient per year) because of
symptoms compared with 23 of the 36 patients on pla-
cebo, who were off work for a total of 1405 days (mean
49-3days per patient per year) (P<0-001).
Acid Secretion and Blood Concentrations of Cimetidine
There was no significant difference in gastric-acid sec-
retion between the two groups before treatment (table n)
nor in degree of inhibition of acid secretion at various
times over the year’s treatment (table III). The inhibition
of acid secretion (percentage reduction in mean acid-
output compared with pretreatment values) by a 200 mg
dose of cimetidine was 80-90% for B.A.o. and 44-46%
for P.A.O. (43% in D.u. and P.u. and 73% in G.u.).
Blood levels of cimetidine remained steady at 5 weeks,
Fig. 2-Effect of treatment on night pain.
Unshaded area mild pain; shaded area severe pain.
= =
406

at 7 months, and at one year treatment. There was no TABLE II-PRETRIAL BASAL ACID OUTPUT (B.A.O.) AND
rebound effect on acid secretion either 2 days or 3.55 t<BMT4fmTCTM-<TTMTn ATED DTTTPTIT (P. A.O.)

months after the end of a year’s course of cimetidine in

the group as a whole. Differences were however seen
between D.u. and P.u. patients (table in). Patients on
placebo had a 50% rise inB.A.0. (P = 0 O1) after 7 months
and a 33% rise (N.s.) towards the end of the trial.
Acid secretion before treatment in cimetidine treated
patients who had recurrences did not differ from that in
patients with no recurrences.

Open Treatment with Cimetidine

Of the 56 patients who received cimetidine 1 g daily
immediately before entry to the trial, 47 (84%) had Results are given as mean &plusmn;s.E.
endoscopically proven healing after a mean of 7 weeks. n=no. of patients in group.
In the placebo group 28 patients were given a second
course of cimetidine 1 g daily for recurrent ulcers; in 23
TABLE III-BASAL ACID OUTPUT (B.A.O.) AND
(83%) healing took place after a mean of 7 weeks. 10 pa-
PENTAGASTRIN-STIMULATED ACID SECRETION DURING (P.A.O.)
AND AFTER LONG-TERM TREATMENT WITH CIMETIDINE COMPARED
tients in the placebo group were given a third course of WITH VALUES IN THE SAME PATIENTS BEFORE TREATMENT
cimetidine 1 g daily for a second recurrence; all had
healed ulcers after a mean of 6 weeks. All patients in the
placebo group who required therapeutic doses of cimeti-
dine during the trial had healed ulcers with prolonged
treatment.
patients in the cimetidine group who had
The few a

relapse during the trial had their ulcers healed by 6

weeks of an increase in dose from 800 mg cimetidine
daily to 1 g cimetidine daily (200 mg three times during
the day and 400 mg at bedtime).
When patients with severe symptoms had to be given
therapeutic doses of cimetidine during the trial, symp-

Results are given as mean &plusmn; s.E.

*p=O.05. P=0.01. P=0.001.

toms disappeared within a week in 50% having their

first course, in 57% having their second course, and in
50% having their third course of 1 g cimetidine daily.
Only in 6 of 94 open treatment courses, did symptoms
remain after 2 weeks on cimetidine 1 g daily.

Fig. 3-Effect of treatment on antacid-consumption.
Adverse Effects
1 patient in the placebo group stopped treatment after
9 months because of mild urticaria; 1 had thyrotoxico-
sis ; a third patient who had had a previous myocardial
infarction, had another myocardial infarction on the
second day of a second course of open treatment with
cimetidine (he recovered during continuous open treat-
ment with cimetidine); and 2 patients had several
attacks of vertigo.
In the treatment group a 65-year-old man had cardiac
failure which responded to digitalis and diuretics while
cimetidine therapy was continued; and 4 patients had
vertigo, 3 had single attacks, and 1 had vertigo for 2
weeks. No allergic reaction, drowsiness, tenderness in
the breasts, or galactorrhcea was observed in patients on
cimetidine.

Changes in Laboratory Variables
A threefold rise in serum-alanine-aminotransferase
(A.L.A.T.) which returned to normal after treatment had
stopped was seen in 4 patients taking cimetidine (2 on
maintenance doses and 2 on therapeutic doses) and in 1
patient on placebo. Liver biopsies from 2 of the patients
on cimetidine revealed slight to moderate hepatocellular
centrilobular necrosis compatible with drug-induced
liver damage of the hypersensitivity type in one and only
fatty infiltration in the other. A twofold rise in serum-
A.L.A.T. was seen in 1 patient on cimetidine and 1 on
placebo.There were no statistically significant changes
between the groups in mean serum-A.L.A.T. or serum-
aspartate-aminotransferase (A.S.A.T.).
A small, but significant rise in serum-creatinine was
seen during the first 2 weeks on either therapeutic or
maintenance doses of cimetidine (p<0.005) although
levels remained within the normal range. After 4 weeks
on treatment, values were still significantly higher than
on day 0 (P<0-01)
but the difference was smaller; after
6 weeks no significant difference was seen, and levels
remained unchanged thereafter. Changes in serum-crea-
tinine were not associated with changes in blood-urea-
nitrogen or urine microscopy.
The only change in hxmatological variables was a fall
in haemoglobin in 4 patients who bled while on placebo.
Discussion
This trial shows that a year’s maintenance treatment
with cimetidine prevents both recurrence of and com-
plications of peptic ulcers. Cimetidine significantly
reduced absenteeism, ulcer pain, and frequency of other
symptoms.
The degree of inhibition of acid-secretion and blood-
levels of cimetidine remained steady during the year’s
treatment- i.e., there was no evidence of tolerance to
cimetidine.
The mean blood-level of cimetidine which inhibits pen-
tagastrin-stimulated acid secretion by 50% ("C-50) has
been reported to be about 0.55 .g/ml in healthy men. 10
However, we have found the I.c.o to be about 1.0 p.g/ml
in 103 acid-secretion tests carried out in patients with
peptic ulcers in this and earlier trials.11 The blood-levels
of cimetidine achieved in this trial correspond well with
data from 37 other patients in whom the blood-concen-
tration curve of cimetidine was measured at close inter-
vals for 9 hours after a single dose.12
The cimetidine group as a whole showed no rebound
gastric-acid secretion either 2 days or 3.5months after
stopping long-term treatment. None of the patients
treated with cimetidine have had any complications of
their illness in the 6 months that have passed since the
end of the trial. Of the 4 patients in the placebo group
who had complications during the trial, 1 patient had
never previously been treated with cimetidine; 1 had
been treated 10 months, 1 patient 3 months, and 1 pa-
tient 1.5 months before haemorrhage from their recur-
rent ulcers.
Findings from
trial in which treatment was mostly
a prenaline with and without prior propranolol adminis-
open have suggested an increased risk of early, and in
some cases severe, relapse after cimetidine had been
stopped and that subsequent courses of open treatment
with cimetidine are not as effective as the first course.7
Our findings do not support either of these suggestions.
407
There have been isolated reports of mental confusion,
muscle twitches in a patient with renal-failure, and ex-
acerbation of ileus in burns cases possibly related to
treatment with cimetidine.13-16 Gynaecomastia and
galactorrhcea can develop during treatment with cimeti-
dine. 17-20 The only probable and serious side-effect of
cimetidine in our patients has been a reversible liver
damage of the hypersensitive type in 1 patient. The rise
(to within normal limits) of serum-creatinine at the start
of treatment with cimetidine still needs to be explained.
There were no abnormal haematological values among
our patients on long-term treatment with cimetidine and
cimetidine seems still to be reasonably safe.
We thank Mrs Ingrid Pettersson for technical assistance; Dr Bjorn
Norlander, Mrs Lillemor Fransson, and Mrs Eva Lilja for cimetidine
assays; and Smith, Kline, and French Laboratories for the cimetidine.
The study was partly supported by the Swedish Medical Research
Council.
REFERENCES
1. Bodemar, G., Walan, A. Lancet, 1976, ii, 161.
2. Bank, S., Barbezat, G. O., Novis, B. H., et al. S. Afr. med. J. 1976, 50, 1781.
3. Gray, G. R., McKenzie, I., Smith, I. S., Crean, G. P. et al. Lancet, 1977,
i, 4.
4. Frost, F., Rahbek, I., Rune, S. J., et al. Br. med. J. 1977, ii, 795.
5. Ippoloti, A. 1977. Personal communication.
6. Freston, J. W. 1977. Personal communication.
7. Saunders, J. H. B., Wormsley, K. G. Lancet, 1977, i, 766.
8. Bardhan, K. D., Blum, A., Gillespie, G., et al. ibid. p. 900.
9. Spence, R. W., Celestin, L. R., McCormick, D. A., Owens, C. J. in Cimeti-
dine (edited by W. L. Burland and M. A. Simkins); p. 163. Amsterdam,
1977.
10. Burland, W. L., Duncan, W. A. M., Hesselbo, T., Mills, J. G., Sharpe,
P. C. Br. J. clin. Pharmac. 1975, 2, 481.
11. Bodemar, G., Norlander, B., and Walan, A. in Cimetidine (edited by W. L.
Burland and M. A. Simkins); p. 224. Amsterdam, 1977.
12. Bodemar, G., Norlander, B., Walan, A. Scand. J. Gastroent. 1977, 12, suppl.
45, 10.
13. Grimson, T. A. Lancet, 1977, i, 858.
14. Robinson, T. J., Mulligan, T. O. ibid. 1977, ii, 719.
15. Grave, W., Nadorp, J. H. S. M., Rutten, J. J. M. H. ibid. p. 719.
16. Watson, W. C., Kutty, P. K., Colcleugh, R. G. ibid. p. 720.
17. Hall, W. H. New Engl. J. Med. 1976, 295, 841.
18. Delle Fave, G. F., Tamburrano, G., de Magistris, L., et al. Lancet, 1977, i,
1319.
19. Bateson, M. C., Browning, M. C. K., Maconnachie, A. ibid. 1977, ii, 247.
20. Carlson, H. E., Ippoliti, A. F. J. clin. Endocr. Metab. 1977, 45, 367.
BETA-BLOCKADE AND BLOOD-LEVELS
AFTER LOW-DOSE ORAL PROPRANOLOL:
THE HEPATIC "FIRST-PASS" THRESHOLD
REVISITED
ROY DAVIES* THOMAS G. PICKERING
ALBERTO MORGANTI GABRIO BIANCHETTI
PAULO L. MORSELLI JOHN ROMANKIEWICZ
JOHN H. LARAGH
Hypertension and Cardiovascular Center, New York Hospital
Cornell Medical Center, New York, N.Y. 10021, U.S.A., and
Department of Clinical Research Synthelabo&mdash;L.E.R.S.,
Paris, France
Summary Heart-rate, arterial pressure, and plasma-
renin activity were determined in six
normal subjects at rest and after an injection of 8 &micro;g iso-
tered orally in a dose of 5 mg 8-hourly for a total of five
doses. After propranolol, resting heart-rate, systolic
pressure, and plasma-renin activity all fell significantly
*Present address: Middlesex Hospital, London W1N 8AA.