Col6i ee|al

Disease
9 Springer-Verlag 1989

Note to subscribers
Since the number of high-quality manuscripts submitted for
publication has been continually increasing and we would like to
guarantee rapid publication in the interests of both authors and
readers, it is now necessary to increase the number of pages per
volume. For this reason, it will also be necessary to increase the price
per volume.

~ SpringeInrternational

Col6i'eeial
Disease
9 Springer-Verlag 1989

Int J Colorect Dis (1989) 4:205 229

Workshop
Pouchitis
The following presentations were given at the
Workshop on Pouchitis which took place at St.
Mark's Hospital, London, on 27 January 1989. A
brief account of discussions which took place is
also included.

Evacuation and pouchitis

P. R. O'Connell

Illeal pouch motility

D. K u m a r and N. S. Williams

Presentations

The pathology of the ileal reservoir

Introduction

N. A. Shepherd

R. J. Nicholls

Pouchitis - Incidence and characteristics in the

continent ileostomy
L. Hult6n

The role of endoscopy in pouch monitoring and

pouchitis
G. N. J. Tytgat

Clinical diagnosis
R. J. Nicholls

Bacteriology I
D. G. Nasmyth

Bacteriology H
M. J. Hill and F. Fernandez

Inflammatory mediators in ulcerative colitis

D. J. Gertner and D. S. Rampton

Faecal bile acids in pouch and pouchitis patients

M. J. Hill and R. W. Owen

Pouchitis: defining an objective method of diagnosis

W. A. Kmiot and M. R. B. Keighley

Pouchitis is a problem which those of us who have

done pouch surgery will have come across and had
to do our best with, and we really know very little
about it. I suspect that we do not yet have a proper
definition of it and perhaps this day, which I hope
will comprise as much discussion as presentation,
will give us some basis, certainly in terms of a
definition which will allow us in the future to be
able to study it in different units talking the same
language and using the same definitions. This really
is the purpose of this meeting.
It is my great pleasure to welcome three people
from outside the United Kingdom as speakers, Professor Guido Tytgat from Amsterdam, Professor
Leif Hult6n from Gothenburg and Dr. R o n a n
O'Connell from Dublin. All three of them are very
involved with pouch work and Dr. O'Connell has
been at the Mayo Clinic so we can have some of his
impressions and information from the work he did
over there. In addition, I would like to welcome
Professor Brummelkamp.
I would also like to thank the companies who
have supported this meeting. Smith Kline &
French, Squibb Surgicare and Pharmacia have all
generously funded and supported the visitors from
abroad and also the refreshments during the day.

206

The pathology of the ileal reservoir

N. A. Shepherd
Restorative proctocolectomy with ileal reservoir is
a surgical procedure of increasing importance [1].
The operation is undertaken in patients with diffuse
mucosal disease of the large intestine: most patients
have ulcerative colitis whilst a small percentage of
operations are performed on patients with familial
adenomatous polyposis (FAP). The pathological
features discussed in this article are based on a
study of 160 patients who have undergone restorative proctocolectomy with ileal reservoir at St.
Mark's Hospital, London. All of these patients are
under rigorous follow-up and sigmoidoscopic biopsies of the reservoir mucosa are taken at least once
a year. Some of the findings described in this paper
relate to a smaller cohort of patients, 92 in number,
who were extensively investigated in a systematic
functional and pathological study [2, 3].
The creation of an ileal reservoir with ileo-anal
anastomosis is complicated by chronic inflammatory changes in the mucosa of the reservoir in the
majority of patients. The infiltrate in the lamina
propria includes all cell types associated with
chronic inflammation; lymphocytes, plasma cells,
eosinophils and histiocytes. These changes are usually associated with villous atrophy, varying from
minor irregularities of villous architecture to subtotal villous atrophy (Fig. 1). The inflammatory infiltrate and villous atrophy may vary in sequential
biopsies and even within the same biopsy. However
when significant acute inflammation is present,
there is usually extensive subtotal villous atrophy.
Villous atrophy is always accompanied by crypt
hyperplasia. The mucosa may resemble that in
coeliac disease: the pouch mucosa, however, lacks
the high intra-epithelial lymphocyte count of the
small intestinal mucosa in coeliac disease. Pyloric
metaplasia is present in a small percentage of reservoirs and usually occurs in pouches with previously
documented acute inflammation. It may therefore
be a useful marker of previous inflammation, particularly pouchitis. The histological changes of
mucosal prolapse are a not unexpected finding in a
small proportion of biopsies. In very occasional
biopsies, well formed epithelial cell granulomas are
seen. Granulomas may be seen as a consequence of
histocytic reaction to foreign material after superficial ulceration. It should be emphasised therefore
that granulomas in the pouch do not necessarily
indicate a diagnosis of Crohn's disease.
At St. Mark's Hospital a system for grading the
histological severity of both chronic and acute in-

flammation in the mucosa of ileal reservoirs has

been introduced [2]. Polymorph infiltrate and ulceration are used as parameters of acute inflammatory change: chronic inflammatory cell infiltrate
and villous atrophy are those for chronic change.
These changes are graded from 0 to 6. In our studies there are no significant differences in the chronic
inflammatory scores between reservoirs created for
ulcerative colitis and those in FAP patients, and
there are no detectable differences in inflammatory
scores for the different reservoir designs. A patchy
superficial acute inflammatory cell infiltrate in the
epithelium is not uncommon in pouches of both
ulcerative colitis and FAP patients: scores for acute
inflammation correlate with the degree of macroscopic inflammation seen on sigmoidoscopy and
with the frequency of defaecation. Extensive acute
inflammtory cell infiltration, with crypt abscesses
and ulceration, corresponds to the clinical condition of pouchitis (Fig. 2). Reservoirs in ulcerative
colitis patients show a highly significant increase in
acute inflammatory scores compared with those in
FAP patients. Furthermore pouchitis has only been
described in patients with a previous diagnosis of
ulcerative colitis [4].
In a high proportion of reservoirs, the combination of villous atrophy and inflammation creates a
histological appearance that is reminiscent of
chronic ulcerative colitis (Fig. 1). Mucin histochemical studies have demonstrated a change from
small intestinal type goblet cell mucin to colorectal
type mucin in about half of the reservoirs studied
[3]. This mucin change has been seen in both ulcerative colitis and FAP patients. Both the morphological and histochemical changes in the reservoir
mucosa appear to be the direct result of creation of
the reservoir. These features suggest that the pouch
is undergoing a form of colonic metaplasia. This
'colonisation' may have major implications: there
may be a propensity to dysplasia and carcinoma in
ulcerative colitis patients and to adenoma and carcinoma development in FAP.
The aetiology of pouchitis is still uncertain.
There is no evidence to implicate specific bacteriological changes or functional abnormalities such as
stasis. It has been suggested that pouchitis is the
result of chronic mucosal ischaemia in the reservoir
although many would dispute this theory. Clinical
and pathological studies have shown that acute inflammation is more marked in ulcerative colitis
reservoirs and that pouchitis only occurs in this
group [4]. A hypothesis for the pathogenesis of
pouchitis is that the creation of a neorectum which
becomes further modified by a process of colonic
metaplasia creates an environment which favours

207

Fig. 1. Pouch mucosa from an ulcerative colitis patient. There is

subtotal villous atrophy and a
moderate chronic inflammatory infiltrate in the lamina propria. The
appearances closely resemble those
of inactive chronic ulcerative colitis
in colonic mucosa. There is a
prominent lymphoid follicle left.
H&E xl00

Fig. 2. Pouch mucosa from a patient with pouchitis. There is subtotal villous atrophy, a dense
chronic inflammatory infiltrate in
the lamina and several incipient
crypt abscesses. There is a small
area of superficial ulceration (arrow). H & E x 100

the development of an ulcerative colitis-like condition in the ileal reservoir. If this thesis is correct, the
reservoir may well prove a useful human model for
studies of the aetiology and pathogenesis of ulcerative colitis.

Discussion
Professor Keighley (Q.): What about focal variation in biopsies?

Dr. Shepherd (A.)" We haven't specifically studied

this but there is no doubt that histological changes
can be focal from looking at multiple biopsies
taken at the same time.
Professor Keighley (Q.): H o w do the histological
appearances change with time?

Dr. Shepherd (A.): Inflammation seems to come

and go as judged by biopsies taken at intervals over
a period of time. This appears to be irrespective of
treatment.
Dr. Morson (Q.): Is it worthwhile for the pathologist to report whether he thinks the pouchitis is
active or inactive?
Dr. Shepherd (A.). Yes, we do this routinely. The
grading system gives an idea of activity. This relates
significantly with endoscopic appearances.
Dr. Morson (Q.): Does pouchitis not occur in polyposis?
Dr. Shepherd (A.)" As far as we can see, no. We
need a definition of pouchitis which involves clinical and histopathological features.
Mr. W. H. F. Thomson (Q.)." Has dysplasia occurred in the pouch?

208

Dr. Shepherd (A.): We have not seen a single case

of dysplasia in a colitic pouch. Some polyposis patients have adenomas but this has been regarded as
part of the disease.
Dr. O'Connell (Q.): Is it truly a pouchitis or could
it be a distal ileitis that these patients get? Are
changes confined to the pouch or do they go more
proximally?
Dr. Shepherd (A.): I do not know.

P o u c h i t i s - incidence and characteristics

in the continent i l e o s t o m y
L. Hult6n
With the introduction of the continent ileostomy in
1969, a new syndrome appeared caused by a mucosal inflammation restricted to the pouch or extending sometimes also into the adjacent intestine.
This disorder which has been called mucosal enteritis, pouch ileitis or simply pouchitis is characterized
by episodes of crampy pain, diarrhoea with liquid,
bloody foul-smelling faeces and general malaise.
When severe the pouchitis may be associated with
systemic manifestations such as weight loss, fever
and arthralgia.
Endoscopically the mucosa is reddened, granular, friable, with contact bleeding with pin point
superficial patchy or linear ulcerations. The histological picture is that of an unspecific acute inflammatory reaction varying from mild to moderate or
severe. Mucosal histology has been difficult to
grade in an appropriate way and in my experience
correlates poorly with the clinical picture. In fact,
varying degrees of inflammation may be seen in
asymptomatic patients undergoing pouch biopsy.
The need for frequent emptying is not due to
increased faecal volumes per se. Decrease of pouch
capacity or compliance in combination with an exaggerated motility are contributing factors as are
also lowered sensory thresholds for pouch filling or
urge.
Treatment of the pouchitis has not been systematically evaluated. Favourable responses have been
reported with antibacterial agents such as vibramycin, salazopyrine, metronidazole or a combination of these. Local and systemic steroids may
also be effective. Recurrent episodes are common
however. Local cholestyramine, sucralphate have
proved ineffective.
In severe therapy-resistant cases, continuous
drainage of the pouch with parenteral nutrition or
a defunctioning ileostomy may be needed. In other
severe unresponsive cases resection of the pouch
may be required.

The reported incidence of this complication

varies between 5 and 25%, a discrepancy that may
to a great extent be due to variability in definition.
Another reason for the great variation in the pouchitis figures is the different numbers of patients
investigated and the length of follow-up.
In my group we have defined pouchitis as
episodes of increased stool frequency with watery
and/or blood stained stools associated with urgency and an inflamed reservoir mucosa on endoscopy and with symptoms severe enough to require treatment. Today I am going to present
results based on 84 patients followed for between
3.5 and 10 years. By using actuarial methods which
compensate partly for the abovementioned errors
the cumulative probability of developing a first attack of pouchitis over a 10-year period would be
about 35% (Fig. 3).
Looking at the pattern of pouchitis in the 28
patients affected (Fig. 4) it can be seen that many
patients had one single episode responding
promptly on treatment, others had frequent episodes and a few continuous and/or severe pouchitis
requiring more active long lasting medical therapy
or a defunctioning ileostomy or resection.
More than half of the patients developed the
first episode within 6 months postoperatively but
there were a few patients who did not experience an
attack until 2 or 3 years had elapsed (Table 1).
There is little information on laboratory values
in pouchitis. In a smaller series of my patients it was
possible to compare routine laboratory tests during
pouchitis with pre-illness status. While haemoglobin, TIBC and serum-albumin were mostly
within normal there were gross abnormalities in
sedimentation rate and serum-iron reflecting the
severity of inflammation. Low serum-iron does not
mean microcytic anaemia in these patients but
rather infection. It is an acute phase reactant as
reliable and sensitive as is seromucoid or haptoglobin.
The intestinal leakage of plasma proteins,
known to occur over an inflamed mucosal surface,
was also studied by in vivo labelling by intravenous
injection of trace amounts of s 1CrC13 measuring its
faecal excretion. The median excretion was 3.5%
Table 1. Occurrence of first episode (n = 28 patients)

Within
Within
Within
Within

6
1
2
4

months
year
years
years

No. of pts

16
18
24
28

57
65
85
100

209
percent
100
90
80
7O
6O
5O
4O
. . . .

3O
2O
10

Patients

75
84

at risk

t"

10 Y e a r s
post,-op.

66
66

62

48

44

33

27

25

19

13

10

64

55

height and a 4-fold increase in mitotic activity indicating an increased cellular turnover, findings that
initiated speculations that the changes might represent a premalignant potential. Continued followup shows evidence, however, that the early changes
may be transient only and th~tt there is a tendency
towards normal with time. Even in patients followed-up for 16 to 20 years it has not been able to
demonstrate dysplasia. Whether the pouch mucosa
is more sensitive and apt to develop inflammation
is not known.

:j

-7

LuuP i ~

I
:
I =

I
:

|
:

mm

tuup

;
LOOP

12

1'8

24

3'0

3'6

42

Fig. 3. The cumulative incidence in pouchitis

in a group of 84 patients followed for up to 10
years

48 m o n t h s

Fig. 4. The frequency and duration of pouchitis in 28 patients

followed for 48 months. Loop =loop ileostomy

which is similar to that in conventional ileostomy

patients with prestomal recurrence of Crohn's disease.
The aetiology of pouchitis is still unknown. In
the search for that it may be important first to look
at the consequences that are peculiar to the continent ileostomy.

Long-term consequences
The early changes in mucosal morphology of the
Kock pouch involve a 50% reduction in villous

The bacterial flora in the continent ileostomy is

more colon-like than that in a conventional
ileostomy. More important is, perhaps, that the
flora is constantly in contact with the ileal mucosa.
It seems logical to ask whether the change to a
colonic flora with increasing numbers of anaerobes
bears a relationship to the development of pouchitis. Prompt response to antibiotics may speak in
favour of a bacteriological factor, but good response may also be achieved with local steroids.
Bacteriology is qualitatively and quantitatively the
same in patients with and without pouchitis and
bacterial cytotoxins have not been isolated. There
is no proven relationship between pouch stasis or
poor emptying and subsequent development of
pouchitis.
Intestinal absorption in patients with continent
ileostomy appears to be largely normal and similar
to that in conventional ileostomy patients but there
are two important differences in the pouch patients. Apart from defective B,z absorption there is
an increased excretion of bile acids. Whether the
bacterial contamination or the morphological

210
changes in the reservior contribute more to this
malabsorption is not clear. Whether and if so to
what extent B12 and bile acid malabsorption will be
more severe during episode of pouchitis is not clear.
Bile salts are known to be cytotoxic but whether they
are responsible for the development of pouchitis is
unknown.

Aetiology
There are reasons to suspect that neither bacterial
overgrowth nor other metabolites are causative factors. At least not alone. Prompt response to broadspectrum antibiotics, salazopyrine or metronidazole
may speak in favour of a bacteriological factor.
The mode of action of antibiotic treatment is not
clear, however. Salazopyrine is a composite of salicylic acid and a sulphonamide. Is it the antibacterial or the anti-inflammatory component that is the
prime beneficial factor? Metronidazole has also
been suggested to have effects on the immunological system. An intriguing observation is that pouchitis is seen predominantly in patients where proctocolectomy has been done for ulcerative colitis,
whereas patients with polyposis do not exhibit the
syndrome, at least not as severely. Could it be that
the local as well as systematic manifestations in
pouchitis may be the result of immune complex formation? Could it be after all that ulcerative colitis,
terminal ileitis, perhaps even backwash ileitis and
pouchitis, is a single disease entity? Hypothetically
the passage of faecal antigens across a permeable
mucosa and their access to subepithelial lymphoid
tissue, might reproduce the pathogenic mechanism
of inflammatory bowel disease.

Dr. Shepherd (Q.)." Do you think that they are

increased over what you would expect in normal
ileal mucosa?
Professor Hultbn (A.)." I don't know.
Mr. Nicholls (Q.): Were some of those pouchitis
patients ultimately identified as being cases of
Crohn's disease?
Professor Hultdn (A.): Not very often. We have
always taken special care to exclude Crohn's disease. The majority of patients are converted after
previous conventional proctocolectomy, so the entire specimen is available for pathological examination. Probably it is a safer way to get an accurate
diagnosis. Somebody asked whether pouchitis is
confined to the pouch. I have to add then that in
colitic patients with pouchitis, inflammation can
extend up into the efferent limb. It is very difficult
to get up with an endoscope and most examiners
will miss these lesions. Fifteen years ago our
pathologists were very sure in their diagnosis but
today they are often confused and consider many
of the changes nonspecific.
Dr. Morson (Q.): One of the things that I find
remarkable about the disease we call ulcerative colitis is that, despite a long history of severe chronic
inflammation, there is always a minimal fibroblastic response, and that also seems to be true of pouchitis. You don't get strictures, you don't get submucosal fibrosis, you don't get shrinkage of the
pouch? Do you get fibrosis?
Professor Hultdn (A.): I do not know exactly the
frequency but the case is typical, namely the occurrence of a stricture of the efferent limb where it goes
into the pouch. This seems to be a very isolated
short stricture, and often associated with inflammatory reaction in the mucosa, proximally.

Discussion
Mr. Mortensen (Q.): Have you had patients with
the Kock pouch who did not originally have U C
and who developed a clinical pouchitis syndrome?
Professor Hultdn (A.): We do not have very many
familial polyposis cases in our clinical series but
that among those 12 or 15 cases seen over a 10-15
years period we have not had this problem. A couple of years ago I wrote to units world wide with
experience of the Kock ileostomy. None had observed pouchitis in the polyposis patients.
Dr. Shepherd (Q.): Have you seen adenomas in
Kock pouches in patients with polyposis?
Professor Hultdn (A.): Yes, there have been a few
cases but again our experience is very limited as
regards familiar polyposis.

The role of endoscopy in pouch monitoring

and pouchitis
G. N. J. Tytgat
Role of endoscopy in pouch monitoring
Endoscopy has a dual role in pouch monitoring.
First of all, it is important in monitoring the healing process before closure of the protective
ileostomy. Initially, the area of the anastomosis is
severely inflamed and friable and some ulceration
and swelling is seen along the suture lines. Later on,
after two to three months, both the anastomotic

211

Fig. 5. Appearance of the pouch before ileostomy closure

Fig. 6. The differential diagnosis between severe pouchitis and
ischaemia may be difficult; this markedly congestive haemorrhagic appearance is more compatible with ischaemia

Fig. 11. Severe pouchitis with extensive deep mucosal ulceration

Fig. 12. a Pouchitis before medical therapy, b Apperance after
therapy with metronidazole. Note the marked improvements of
the endoscopic appearance

Fig. 7. Appearance of the pouch soon after ilestomy closure

Fig. 8. Mild pouchitis with mainly swelling and erythema
Fig. 9. Pouchitis with erythema, swelling and spread-out superficial mucosal destruction
Fig. 10. More severe pouchitis with extensive superficial mucosal destruction and copious amounts of mucopurulent exudate

line and the suture lines have completely healed

(Fig. 5). It has been our policy to close the
ileostomy only after full endoscopic healing of the
pouch. The normal small bowel mucosa has a
clearly visible vascular pattern upon distension.

212
The mucosa is smooth and shiny and non-friable.
Occasionally, tiny areas of highlighting may be visible, presumably corresponding to lymphoid follicles. In addition some rather opalescent milkylooking areas may be seen, again, presumably
corresponding to lymphoid aggregates.
In some patients abnormalities may be detected
which may delay re-anastomosis. One may see
small fistulous openings along suture lines. A few
weeks later, such fistulous openings usually have
completely healed. More common is a partial dehiscence of the anastomosis. An extensive dehiscence of the anastomotic line leading to quite deep
defects is rare. A very haemorrhagic aspect of the
mucosa is presumably due to ischaemic changes in
the pouch (Fig. 6). Ileostomy closure has to be delayed till full mucosal healing has occurred. We
wonder whether prolonged rather severe inflammatory changes close to the anastomosis with rather
slow healing of the suture lines may not be the
expression of impaired micro-circulation.
The aspect of the pouch always changes as soon
as the ileostomy is closed and faecal material enters
the pouch (Figs. 5, 7). There are always some mild
inflammatory changes both at the level of the anastomosis and in the pouch itself. The pouch mucosa
becomes slightly swollen and somewhat redder in
appearance. The Kerckring fold pattern becomes
poorly visible or disappears entirely.
If one examines pouches in patients who are
clinically completely well, there always are some
minor abnormalities detectable, especially in patients who have had pouchitis in the past.
Endoscopy is very important in the diagnosis of
pouchitis. Pouchitis remains a very significant
source of postoperative morbidity. It may be seen
in 1 0 - 2 0 % , even up to 25% of patients, especially
in those operated upon for chronic colitis and in
those exhibiting extra-intestinal manifestations of
their disease. These have led to speculation that
pouchitis may actually be a further manifestation
of inflammatory bowel disease. The pathogenesis
of pouchitis, despite all recent hypotheses, is still
poorly understood.
The criteria upon which the endoscopic diagnosis of pouchitis is based are the well known indicators of inflammation: swelling, erythema or redness, friability and petechial punctate haemorrhagic spots, excessive mucopurulent exudative areas and superficial erosive defects or larger ulcerative destruction of the mucosa. Mild pouchitis is
characterized by discrete swelling of the mucosa
around the anastomosis and in the pouch and some
mild friability (Fig. 8). In others, there is obvious
erythema, mucopurulent punctate exudate, some

small defects and friability (Fig. 9). Occasionally, in

slightly more severe forms spread-out small ulcerations, perhaps with some predilection for the mucosal folds may be visible. This appearance of the
mucosa may readily mimic the abnormalities seen
in Crohn's disease. We have the distinct feeling that
far too often the diagnosis of Crohn's disease is
reconsidered if surgical problems occur or if pouchitis develops. To some extent this is supported by
the pathologists who are indeed rather uncertain
and confused at the present time what to call ulcerative colitis and what to call Crohn's disease. Less
common is striking erythema, petechial haemorrhages, diffuse inflammation, friability and superficial defects.
The spectrum of the severe forms of pouchitis is
rather impressive and consists of striking diffuse
erythema, formation of copious exudate and extensive superficial necrosis (Fig. 10). In other patients there is extensive deep ulceration surrounded
by severe inflammatory changes of the mucosa
(Fig. 11). Such severe changes occasionally occur
only a few weeks after closure of the ileostomy.
Such abnormalities may also mimic to some extent
Crohn's disease. [n some patients it may be difficult
to distinguish pouchitis from ischaemic damage
(Fig. 6). Initially there is severe congestion of mucosa. Later, serpiginous coalescent ulceration develops as one may see in ischaemic damage of the
colon.
Endoscopy is also useful in monitoring the effects of therapy. Medical therapy includes metronidazole, drainage, if necessary irrigation and occasionally even steroids, systemically or in enema
form, together with 5-aminosalicylic acid (Fig. 12 a,
b). Some patients with very severe pouchitis do not
improve with metronidazole therapy only. Only after switching to corticosteroids and 5-aminosalicylic acid does gradual improvement in the endoscopic appearance occur.
In summary, pouchitis is a new disease, an intriguing medical problem for which endoscopy
plays an essential role in the proper diagnosis and
in monitoring the response to therapy.

Discussion

Professor Keighley (Q.): Well, we are absolutely

thrilled with the pictures and I would like to ask
how Professor Tytgat manages to get them. What
about bowel preparation on these patients?
Professor Tytgat (A.)" Just a saline enema 30 minutes before the examination.

213

Professor Hultbn (Q.)." The slides you showed with

the nice vascular pattern, were all of them taken
before you had closed the ileostomy?
Professor Tytgat (A.): Yes. Once the ileostomy is
closed you never see the vascular pattern any more.
Mr. Marks (Q.)." Have you had to ever give a
patient with a pouch an ileostomy again and what
happens to the pouch mucosa then? Do you have to
give them back an ileostomy? Does it get better?
Professor Tytgat (A.)." We have never performed an
ileostomy for pouchitis only.
Mr. Mortensen (Q.): Someone was asking a bit
earlier about the proximal ileum upstream of the
pouch.
Professor Tytgat (A.): Usually it is normal but we
have seen patients with severe pouchitis where there
definitely was extension quite a long distance in the
pre-pouch ileum.
Mr. Mortensen (Q.): Have you seen these sorts of
dramatic changes in polyposis patients?
Professor Tytgat (A.)." Never.
Dr. O'Connell (Q.)." I think your data show that we
can scotch the idea that a chronic ischaemia is responsible for much of the pouchitis as a hypothesis.
It must be due to stasis with something that is
causing an effect on the mucosa because of that
stasis. When it is defunctioned it is normal, we have
seen that, and you don't get the same changes. They
are similar but they are not the same with chronic
ischaemia. Would you agree with that?
Professor Tytgat (A.): Not entirely. I am not entirely convinced that we know after hooking up the
ileostomy, whether the vascularisation of the pouch
is optimal in these patients. I think there may be
patients who are on the borderline and then, depending upon distension and other factors, mucosal perfusion may decrease critically.
Mr. Nicholls (Q.): Do you think you see a pouchitis before the ileostomy closure?
Professor Tytgat (A.)." No.
Mr. Nicholls (Q.)." So the changes that we might
say were pouchitis before closure of the ileostomy
you would consider as likely to be due to ischaemia.
Professor Tytgat (A.): The changes that we have
seen so far before closure of the ileostomy we have
interpreted as compatible with ischaemia.
Dr. Morson (Q.)." Has there been a biopsy done on
these patients?
Professor Tytgat(A.): I do not know. I concentrated on endoscopy and did not look at all the
histology material.
Dr. Morson (Q.)." If you think it is ischaemia a
biopsy would be likely to show changes different
from what have been illustrated as pouchitis.

Professor Tytgat (A.)." It all depends at which time

you take the biopsies, in the very early days or after
the phase of sloughing.
Mr. Mortensen (Q.): How many of those completely resolved?
Professor Tytgat (A.)." They all did.
Mr. Mortensen (Q.): There is always a nice, pink,
healthy pouch in Amsterdam before closure of the
ileostomy.
Professor Tytgat (A.): We aim for that.
Mr. Kmiot (Q.): In the case with extremely gross
inflammatory changes there was no response to
metronidazole. Since the patient responded slowly
to steroids and salazopyrin, was this not occult
Crohn's you were looking at?
Professor Tytgat (A.): Why should I think so? This
patient has had bona fide ulcerative colitis with all
the standard classical criteria. Why should I now
reconsider and look for another disease because he
has severe pouchitis? The literature is filled with
this sort of reasoning.
Professor Keighley (Q.): It does occur. Would you
not agree?
Professor Tytgat (A.)." I have not seen one case. I
doubt it very much. It is very, very rare.
Professor Keighley (Comment): We have two very
well documented cases. There is no doubt that the
histopathological changes in the excised pouch indicated Crohn's disease, despite the fact that review
of all the previous sections, the colectomy, and the
subsequent rectal excision suggested ulcerative colitis.
Professor Tytgat (Q.)." Well, what did the pathology show on the excised pouch to be typical of
Crohn's disease?
Professor Keighley (A.): Lesions which were typical of Crohn's disease. Granulomas, transmural inflammation, fissures, etc.
Professor Tytgat (Q.): But any diseased mucosa,
depending on the intra-luminal pressure, can cause
ulceration cracks and fissuring. I doubt very much
that this is a true specific pathognomonic sign of
Crohn's disease.
Clinical diagnosis
R. J. Nicholls
The reported incidence of pouchitis has ranged
from less than 5% to over 50%. It seems very unlikely however that the criteria used by the various
authors have been the same. Some degree of inflammation in the reservoir is very common and
other circumstances, e.g. outlet obstruction, can

214
lead to frequency of defaecation. It is probable that
when both are present, some may make the diagnosis of pouchitis. It is also possible that most colitics
suffer a degree of low-grade pouchitis. Thus pouchitis has probably been over diagnosed.
Nevertheless certain patients manifest an obvious clinical syndrome which is associated with endoscopic and histopathological inflammation of a
severe degree. The symptoms include frequency of
defaecation, watery stool, sometimes malaise and
occasionally an activation of extra-alimentary disorders, e.g. arthropathy if previously present.
The histological grading system developed by
Neil Shepherd has enabled some objectivity to be
introduced. Richard Moskowitz personally examined 55 patients by sigmoidoscopy and recorded on
a scale 0 - 6 the degree of macroscopic inflammation. Two correlations emerged. Macroscopic inflammation score was firstly related significantly to
the frequency of defaecation (Fig. 13). Secondly it
was also related to the histological grade of acute
inflammation (also graded 0 - 6 ) in biopsies taken
at the same time as the sigmoidoscopic assessment.
In this last correlation there was a cluster of six
cases with both severe (Grade 4 - 6 ) macroscopic
and histological inflammation who on symptoms
alone had already been considered to have pouchitis (Fig. 14). This gives a prevalence of 11% in the
series but as with the Kock reservoir there is likely
to be a cumulative incidence with the passage of
time. These findings led us to feel that pouchitis
should be diagnosed only by a combination of clinical, endoscopic and histopathological criteria with
severity of inflammation being the essential feature.
There appears to be an important distinction
between acute and chronic inflammation, the former only being diagnostic of pouchitis. In a larger
group of 90 patients (including 77 with ulcerative
colitis and 13 with familial adenomatous polyposis), 78 (87%) showed some degree of chronic and
only 27 (30%) acute inflammation. While cases
with severe chronic inflammation tended to have
acute inflammation also, only those with acute
changes manifested symptoms.
Possible risk factors for pouchitis have been
studied at the Mayo Clinic, Leeds and St. Mark's
Hospital. These can be summarised as follows. No
mechanical factor, e.g. type of reservoir, emptying
properties or compliance, or the presence of quantitatively assessed bacterial species can be related to
the condition. Indeed the only positive association
is with the original diagnosis. Pouchitis is almost
unheard of in patients with familial adenomatous
polyposis, almost all cases occurring in those with
ulcerative colitis. Even in the absence of pouchitis,

10

9--q ~

;.

'.

m~N
Se

2 _~O

q~

i
1

I
2

I
3

I
4

I
5

I
6

Histopathological grade

Fig. 13. Frequency of defaecation and histological grade of inflammation in 55 patients examined by one clinician. (Correlation coefficient of linear regression r=0.65; p < 0.001)
6

8'

oo

~.

eo
oo
oo
1

o,0:
gee
ooo
ooo

ooo
0

Macroscopic

{sigmoidoscopic)

grade

Fig. 14. Histological grade of inflammation and macroscopic

grade on sigmoidoscopy in 55 patients examined by one clinician. (Correlation coefficient of linear regression r=0.77;
p < 0.001) (Figs. 13 and 14 taken from Moskowitz et al., Int J
Colorect Dis 1: 167-174)

colitics tend to have higher grades of histological

inflammation in pouch biopsies. We also have evidence to suggest that patients who are most at risk
are those originally having total rather than left
sided colitis. In a study of 83 patients with ulcera-

215
tive colitis, there were 14 cases with pouchitis. All
occurred in the 68 patients with total colitis and
none in the 15 with left sided colitis. No correlation
between pouchitis and backwash ileitis has been
found.
Is pouchitis ulcerative colitis in the small intestine? There are the diagnostic correlations and similar histological appearances. Furthermore pouchitis seems to take the forms of a persisting chronic
disorder or one that demonstrates exacerbations
and remissions, similar to ulcerative colitis.
With regard to treatment the picture is confusing. There is no doubt that metronidazole can,
sometimes dramatically, induce a remission. We do
not however know how often it is effective or which
type of patient is likely to respond. Are the failures
of metronidazole treatment more likely to respond
to conventional ulcerative colitis treatment? Again
we do not know.
Our knowledge so far can be summarised as
follows. Pouchitis should be diagnosed on the basis
of clinical endoscopic and histological features, a
histological grading system being strongly recommended. We are ignorant of its cause but a combination of disease-related susceptibility (UC), possibly with a bacteriological factor (response to
metronidazole) should give useful clues. Treatment
needs to be assessed through properly designed
clinical trials.

Discussion
Professor Tytgat (Q.): Did you notice any correlation as to whether the patients had after elective or
emergency colectomy developed pouchitis? What
about the influence of postoperative pelvic sepsis?
Mr. Nicholls (A.): I cannot answer either question.
Mr. Marks (Q.): In the treatment of metronidazole, if it does not help what else have you tried?
Mr. Nicholls (A.): One tries steroids either locally
or systemically but we do not know objectively how
effective they are.
Mr. Mortensen (Q.)." Can we ask Professor Tytgat
about the newer kinds of salazopyrin substitutes? Is
there any evidence they work better?
Professor Tytgat (A.): It does something. Primarily
in the sense in passive therapy, it does something.
Mr. Mortensen (Q.)." Especially when used in the
form in which you can leave it in the pouch?
Professor Tytgat (A.)." It is impossible to say that it
is better, or equivalent to oral medication.
Dr. O'Connell (Comment): One of the patients I
have come across responded to erythromycin, not
to metronidazole.

Mr. Nicholls (A.)." I think we desperately do need

a proper clinical trial of several different agents in
this kind of situation. You have got to have the
patient base to start, though.
Mr. Mortensen (Comment)." Getting a definition
everybody agrees about, too. You have got to get
that group of patients and everybody agrees about
who they are.
Mr. Everett (Comment)." I find augmentin as effective as metronidazole. I have several patients who
need to take antibiotics more or less continuously,
and they can get peripheral neuritis from metronidazole as you know, and so I have put them on
augmentin. They take one tablet three times a day
for a month and then reduce it, but they usually
stay on just one tablet and that seems to suit them
all right.
Mr. Nicholls (A.): There are some patients who,
when you give them metronidazole get better and
you stop it. Months later perhaps they get another
attack. Other patients as indicated respond to the
metronidazole agent but need to maintain a low
dose therapy. There are other patients you try it on
and nothing happens at all.
Professor Tytgat (Q.): Do you know of any study
where lactobacilli have been used to try and modify
the faecal flora?
Mr. Nicholls (A.): No.
Mr. Mortensen (Comment): Anecdotally, we have
tried yoghurt every now and then.
Professor Tytgat (Q.)." What sort of yoghurt?
Mr. Mortensen (A.): Active yoghurt.
Professor Tytgat (Q.): There are no viable bacilli.
It is only certain brands of yoghurt which have live
bacilli.
Mr. Mortensen (A.): It would be live.
Professor Tytgat (Q.)." But there are now good
preparations for the first time and you can truly
modify the faecal flora with these newer products.
Mr. Nasmyth (Q.): In active ulcerative colitis there
is a change in the pH of the stool. Has anybody
looked at the pH of the stool in patients with pouchitis, and perhaps when they have recovered to see
if there is any change there?
Mr. Nicholls (A.): Not that I know of.
Professor Keighley (Q.): Has anybody seen
Clostridium diJficile in pouches because we certainly have, and I am a little bit worried about
longterm antibiotic therapy for that reason. It is
interesting that the mucosa of these pouches does
become colonified, and I suspect it is therefore an
organ that can become affected by Clostridium difficile toxins.
Mr. Nicholls (A.)." We do look for it. We have had
one case with Clostridium difficile. He was quite ill

216

and finally lost the pouch. So we have had one case.

We may well have missed others but we do look
for it.
Professor Tytgat (Q.)." In pouchitis defunctioning
should cause it to heal up unless there is faecal
contamination. If it does not then you should
worry that something else is going on, for example
chronic ischaemia or Crohn's disease. But normally
a severely inflamed pouch, if you bypass stool and
rinse it well, should improve.

Table 2. Concentration of organisms in faeces from patients

with an ileal reservoir and permanent ileostomy

Median and range

Pouch

Ileostomy

Anaerobes/aerobes *

100
(18 621 1)
9.8
(0-11.9)
8.7
(0-11.9)
6
(0-10.4)
2.1
(0.4 29.2)

4
(120 0.1)
5.7
(0 10.1)
0
(0-9.6)
0.4
(0-10.7)
0.5
(0.1 15.6)

Bacteroides **
Bifidobacteria *
Propionate *
Butyrate **

log 1~ cfu
logl o cfu
mmol/kg
mmol/kg

Bacteriology (I)
* p<0.05;

** p < 0 . 0 1

D. G. Nasmyth
In the normal individual bacterial counts increase
along the small intestine. In the ileum the ratio of
anaerobes to aerobes remains approximately equal,
although the type of anaerobic flora changes and
there are increased numbers of Coliforms, Bacteroides and Clostridia [5]. The effect of further
changes in the bacterial flora on the structure of
ileal mucosa was investigated in patients who have
undergone proctocolectomy for ulcerative colitis
(conventional ileostomy n = 12, pouch-anal anastomosis n = 11).
Fourteen genera of faecal bacteria were identified and quantitated. Anaerobic bacterial fermentation of carbohydrate in the colon or ileum results
in the production of volatile fatty acids (VFA). To
determine whether there is a association between
VFA and pouch mucosal change, faecal VFA from
pouches and ileostomies were measured by gas liquid chromatography [6].
It might be expected that incomplete emptying
of the pouch would be associated with an increased
predominance of anaerobes and their metabolites
in the pouch flora. This hypothesis was investigated
by measuring the completeness of pouch evacuation with a synthetic stool labelled with 51Cr
sodium chromate [7].
Biopsies of ileal mucosa were taken at the same
time as faecal samples. Four components of mucosal inflammation were assessed by scoring each
on a scale from 0 (absent) to 3 (severe). Villous
atrophy was assessed by measuring the ratio of
mucosal surface length (MS) to the area of the
lamina propria (LP) using the Kontron IBAS 1
image analyser [6].
When compared with faecal samples from
ileostomies, those from pouches showed greater
anaerobe predominance, and greater numbers of
Bacteroides and Bifidobacteria. There were also

greater concentrations of faecal propionate and butyrate in the pouches (Table 2).
There was no significant correlation between
the score for mucosal inflammation and the number of bacteria isolated, or the concentration of
faecal VFA. However, the greater the number of
Bacteroides the more severe was the villous atrophy
( r s = - 0 . 9 3 , p<0.01). Conversely, the higher the
concentration of faecal butyrate the less severe was
the villous atrophy (rs = 0.68, p < 0.05). No association was demonstrated between the completeness
of pouch emptying and faecal bacteria or mucosal
change but the concentration of faecal propionate
was higher in those pouches in which there was
greater retention of 51Cr (rs=0.82 , p < 0 . 0 1 ) [6].
These associations do not demonstrate cause
and effect, but they do suggest mechanisms by
which bacteria may affect mucosal structure. Villous atrophy in the pouch has recently been shown
to be associated with an increased crypt cell production rate. This study has shown that the higher
the concentration of butyrate the less severe the
villous atrophy; in other words a higher concentration of butyrate is associated with a lower turnover
of epithelial cells. This fits in with studies in vitro in
which butyrate has been associated with an arrest
of cell proliferation [8]. It is also consistent with
findings in the colon in which colonic neoplasia has
been associated with low concentrations of butyrate [9]. It might seem paradoxical that the more
severe villous atrophy should be associated with
increased numbers of Bacteroides yet one of the
products of anaerobic bacterial fermentation, butyrate, is associated with less severe villous atrophy.
There is no paradox because very few species of
Bacteroides produce significant amounts of butyrate, and it might be speculated that in vivo butyrate suppresses the growth of Bacteroides. The

217
association between pouch emptying and faecal
propionate suggests that when the transit time of
pouch contents is increased by incomplete emptying there is increased anaerobic fermentation, especially by bacteria that produce propionate.

Bacteriology (II)
M. J. Hill and F. Fernandez
Chronic inflammation in the reservoir occurs in
almost every case; acute inflammatory changes, often severe and associated with diarrhoea (pouchitis) occur in some colitic patients but not in those
with polyposis. The clinical features of pouchitis
respond to metronidazole, suggesting a bacterial
aetiology. We therefore decided to study the bacterial flora of pouch effluent in pouchitis patients
compared with colitic and polyposis patients without pouchitis.
Samples of faeces were collected from 6 colitic
patients with pouchitis, 20 colitics without pouchitis and 7 polyposis patients. The faecal samples
were emulsified in a cryoprotective transport broth,
frozen at 40 ~ and transported to the laboratory
for analysis. The bacteriological investigation took
two forms, first a search for putative pathogens
such as Clostridium difficile, Yersinia spp., Campylobacter spp., Salmonella, Shigella, Vibrio and enterotoxigenic Esch. coli; and secondly a study of the
bacterial flora profile. To achieve these two aims
the samples were plated on non-selective media for
total counts and on a range of selective or diagnostic media for the counts of specific organisms.
We were unable to detect any of the pathogens
sought and conclude that pouchitis is unlikely to be
caused by such organisms or their toxins.
In general the bacterial counts in pouch patients
were only a little higher than those seen in
ileostomy effluent and were 4 logarithm (base 10
units) lower than those normally seen in faeces.
This suggests that either the period of stasis in the
pouch is low or that conditions are unfavourable
for bacterial growth. Nevertheless the relative
proportions of the major genera in pouch faeces
were more similar to those seen in "normal" faeces
than in ileostomy effluent (e.g. Bifidobacterium spp.
were as numerous as Bacteroides spp.; lactobacilli
were a major component).
Trends in the composition of the flora from
colitic pouches with and without pouchitis suggested that both aerobic and anaerobic organisms
were more numerous in pouchitis patients than in
healthy pouches, but there was no difference in the

relative proportions of anaerobes and aerobes.

None of the trends in differences in individual genera were statistically significant but this could have
been simply due to the small number of pouchitis
patients.
It is unlikely that a known intestinal pathogen
is the cause of pouchitis, but it is possible that the
disease is caused by an at present unidentified abnormality in the floral profile or in the immune
response to that profile.

Discussion

Mr. Mortensen (Q.)." H o w happy are you with

grading of histological biopsies? H o w easy is it to
do it reliably and efficiently, and if you gave the
same biopsy to 25 histologists were they all graded
equally?
Mr. Nasmyth (A.): Well, I have to confess this was
only done with one histologist. Obviously, one
took the advice of the histopathologist on how to
do it and this was the grading system that they
recommended. All sections were submitted to the
pathologist blind. Probably, on occasions it was
fairly obvious that I was showing him sections from
an ileostomy as opposed to a pouch, simply because of the technique of taking the biopsy, but he
had no relevant clinical information when he
scored them.
Dr. Hill (Q.): Have you looked at how much variation in bacteriological counts there is from day to
day?
Mr. Nasmyth (A.): We did duplicate samples.
Quantitative bacteriology is quite expensive in
terms of the numbers of plates you need to do, so
we only repeated a few. Qualitatively, the results
were fairly similar on follow-up. Quantitatively,
there were occasionally quite big differences.
Dr. Hill (Q.): And the same applies to your volatile
fatty acids?
Mr. Nasmyth (A.): Yes.
Dr. Shepherd (Q.)" Why did you use lamina propria fibrosis as a parameter and how often did you
see it?
Mr. Nasmyth (A.): We saw it very rarely. There
were one or two where this change was obvious but
only one or two.
Dr. Shepherd (Q.): What do you think this represents?
Mr. Nasmyth (A.): Presumably the late stage of
chronic inflammation and more than that I cannot
say.
Dr. Shepherd (Q.): Or ischaemia?
Mr. Nasmyth (A.): Possibly.

218

Mr. Nicholls (Q.)." The bacteriological differences

between an ileostomy and pouch are significant.
Agreed?
Dr. Hill (A.): Yes.
Mr. Nicholls (Q.)." In which case the only difference that immediately comes to mind is the fact
that one is freely evacuated and the other one is
not. So there is an element of stasis is there not?
Dr. Hill (A.): Yes. I wouldn't argue with that at all
and it is interesting that the changes that we have
seen in our study were entirely manifested in terms
of anaerobic flora, so there was obviously quite a
lot of fermentation going on which would tie in well
with the fatty acid pictures that were obtained. But
is is not the sort of longterm festering stasis that
causes the production of large amounts of toxic
substances that might well induce pouchitis. On the
other hand, if you wanted to propose a protective
substance then you might well be able to argue that.
Dr. O'Connell (Q.)." Were any of the patients in
your series early "S" pouches from the St. Mark's
group who have stasis and intubate? If so, do you
know whether that influenced the bacterial appearance in the pouches?
Dr. Hill (A.): There was a mixed bag of pouches,
so to speak and there was no pouch-related difference in the bacteriology.
Mr. Mortensen (Q.)." There was a paper from the
Mayo Clinic suggesting that a proportion of patients did have a syndrome related to proximal
small bowel overgrowth and I wonder if there is
any information on that.
Dr. O'Connell (A.)." Yes. We identified a group of
people who had jejeunal bacterial overgrowth and
this was related to a high stool volume and a poor
result, and it exactly mirrored what Kelly had
found in a group of Kock pouch patients some 5
years earlier, also with pouch dysfunction.
Professor HultOn (Q.)." In the Blind Loop Syndrome you have a bacterial contamination and a
high flow from the intestine, but you rarely see
mucosal inflammation as seen in pouchitis. How
about that?
Dr. Shepherd (A.): I think the major change you
see, as George Nasmyth suggested, was villous atrophy. You do see chronic inflammation. You don't
tend to see acute inflammation unless you have got
an additional mechanical factor which will cause
local ulceration, such as in sigmoid diverticulosis
with a faecolith or something like that. The changes
tend to be just villous atrophy and chronic inflammation.
Mr. Nasmyth (Comment): There are two points
here which are quite important. One is perhaps the
length of time for which the pouch has been con-

structed and what influence that has, and the second is to classify what one is looking at. In our
study we were not looking at patients with pouchitis. Some, at least two, had had documented
episodes of pouchitis but at the time they were
studied they did not have it. I expected when I
started this that we would find a significant correlation between our index of mucosal inflammation
and villous atrophy. We didn't. Maybe we didn't
have enough numbers. In the small bowel the
change is predominantly villous atrophy and not
inflammation. From the observations on the Kock
pouch 10 years out, marked infiltration of the lamina propria with both acute and chronic inflammatory cells occurs early on. Later this levels off and
remains stable so that the difference at 5 years and
10 years is minimal. However there might be quite
a marked difference between one year and 5 years.
In our flush of enthusiasm to study these patients
we looked at them fairly early after the pouch had
been constructed and it may well be that some of
the conclusions we have drawn might have to be
re-interpreted in the light of long-term follow-up.
Mr. W. H. F. Thomson (Q.): Did the patient who
had a clinical pouchitis treated with metronidazole
successfully have the faeces scrutinised bacteriologically before and after?
Mr. Nasmyth (A.)." We did not study this, it is
obviously something that needs to be done.
Mr. Nicholls (Q.): Have the Mayo done it?
Dr. O'Connell (A.). Not that I know of.
Mr. Kmiot (Q.)." Do you think we should be looking at adherent mucosal flora?
Mr. Nasmyth (A.): Looking at the literature the
bacteria isolated from the mucosa do not differ
significantly from those in the lumen.
Dr. Hill (Comment): I agree that, qualitatively,
they are very similar. The relative proportion is
however very different; there are many fewer anaerobic flora. We looked bacteriologically at biopsy
specimens from colitics with and without pouchitis
and could not see any differences between them.
Mr. Nicholls (Q.)." Having not established very
much so far from a lot of work, do you feel there
is still a role or place for study of bacteriology,
perhaps in a more functional way?
Dr. Nasmyth (A.)." I think part of the problem is
that we don't understand the aetiology of colitis
and that represents the main difficulty. The pouch
offers a system which might be more amenable to
study than the colon, particularly in looking at
products of bacterial metabolism, if they are important. Perhaps the chance to look at the effects of
treatment of pouchitis, particularly by antibiotics
and to follow changes in the flora, might give us a

219
clue as to which bacteria we should look at in relation to the aetiology of colitis.
Dr. Hill (A.)." I am sure we should continue to look
at it. It is a problem knowing how to go about it.
We have been doing a study on colitis with a group
of clinicans in Salisbury, Swindon and Wolverhampton looking at the initial stages of colitis. So
we were looking at the flora in patients first diagnosed with colitis, then patients in remission. This
has yielded a bacterial library of patients in remission and we are just watching and waiting for them
to go into relapse and see what changes we can see.
Mr. Nasmyth (Comment)." Concerning the action
of metronidazole, it is fairly well established that it
has a genuine immunosuppressive action and undoubtedly affects T-cell function. It may therefore
be that it is the mucosal response to the bacterial
flora that is significant rather than the change in the
flora itself.
Dr. Hill (A.): Yes, it may well be. Metronidazole
seems to have an increasing profile of activities and
its effect on the immune system is just one of them.
However the antibacterial action of metronidazole
is rapid but when used as an anti-tumour agent it is
very much more slowly active. I do not know what
I would expect as far as its anti-immune reaction
would be.
Professor Keighley (Q.): If the butyrate hypothesis
on the aetiology of ulcerative colitis is right, one
would expect that those individuals who cannot
handle butyrate are going to be those who developed pouchitis.
Dr. Hill (A.)." The colonic mucosa has an absolute
requirement for luminal butyrate as a nutrient. It is
its main energy source. Most organisms are very
good butyrate producers and will be very sensitive
to metronidazole, so when you treat with metronidazole there will be a massive decrease in the
amount of butyrate. Therefore I do not think it is
a matter of patients not being able to handle butyrate.
Professor Tytgat (Q.): But if butyrate is the main
energy source for colonic mucosa and Bacteroides
suppresses the butyrate producers then levels will
be low.
Mr. Nasmyth (A.)." Well, that was pure speculation. I am just hypothesising.
Professor Tytgat (Q.): You said something like
that.
Mr. Nasmyth (A.)." No, what I said was that the
situation in which the Bacteroides proliferate might
well be one in which the organisms which are producing butyrate do not.
Dr. Hill (Q.)." Yes, there is a limited amount of
nutrient and so if you have a person whose flora is

dominated by bacteroides it must be at the expense

of something else. So it must be at the expense of
well-known butyrate producers. If you have got
very large numbers of Bacteroides then maybe you
have decreased numbers of the others. So that
would fit to some extent.
Professor Tytgat (Q.): But it would not explain
why pouchitis improves if you bypass faecal flora
from the inflamed pouch.
Dr. Hill (A.)." No, that would not fit at all.
Professor Tytgat (Q.)." Because butyrate production decreases dramatically as soon as you bypass?
Dr. Hill (A.): Yes, that wouldn't fit, nor would the
effect of metronidazole.

Inflammatory mediators in ulcerative colitis

D. J. Gertner and D. S. Rampton
Eicosanoids, 20-carbon fatty acids derived from
arachidonic acid, have been implicated as mediators in several inflammatory conditions. Since
Gould first described increased prostaglandin excretion in stools in patients with ulcerative colitis,
there have been numerous reports of increased eicosanoid synthesis and release in the inflamed colon [10].
What physiological role do eicosanoids play in
normals and might they play a role in the pathogenesis of ulcerative colitis or pouchitis? Oral and parenteral PGE 1 and PGF2c~ provoke watery diarrhoea in humans, probably by effects on intestinal
water and electrolyte transport and gut motility.
Splanchnic blood flow may also be altered by
prostaglandins. Prostaglandins stimulate mucus
production and, in the stomach, exert a "cytoprotective" effect; whether this occurs in the colon is
not clear. Leukotrienes increase vascular permeability and are highly chemoattractant for leucocytes. Eicosanoids could therefore explain many of
the features of ulcerative colitis [10].
The clinical problem in colitis is twofold. Firstly
the cause of colitis is unknown and secondly, curative therapy is lacking. Moreover the mechanism of
action of drugs currently used in colitis is not well
understood.
Our experiments have been designed, firstly, to
examine those factors which stimulate the inflammatory response and thereby improve our
understanding of the pathogenesis of colitis, and
secondly, to attempt to inhibit the inflammatory
response and thus develop novel therapeutic
agents. We have used an in vitro method in which

220
rectal biopsies are incubated in oxygenated Tyrode's medium which is changed every 20 min. The
release of P G E 2 and LTB 4 into the medium, before
and after the addition of potential trigger factors or
drugs, is measured by radioimmunoassay. We have
shown that eicosanoid release by colitic mucosa is
increased compared with normal and correlates
with disease activity. Exogenous arachidonic acid
amplifies LTB 4 and P G E 2 release indicating enhanced 5-1ipoxygenase and cyclo-oxygenase activity in inflamed mucosa [11] (Fig. 15).
Why are these two enzyme systems switched
on? Of the potential initiating factors currently under investigation (e.g. F M L P , a bacterial derived
peptide, bacterial endotoxin, bile acids and PAF),
two ( F M L P and bile acids) are known to be capable of producing colitis in experimental animals [12,

duced as are thromboxane and leukotriene synthesis; these agents have also been shown to act as
free-radical scavengers [10].
Pouchitis appears to occur only in patients in
whom the operation was performed for ulcerative
colitis. The histological abnormalities are similar to
those found in ulcerative colitis. The beneficial effect of metronidazole in pouchitis contrasts with
that in ulcerative colitis and might suggest a bacterial aetiology perhaps mediated by F M L P or endotoxin. Data on inflammatory mediators in pouchitis are currently lacking. Studies on eicosanoid
synthesis and release in this iatrogenic condition
may provide further clues to the aetiology and therapy not only of pouchitis but also of ulcerative
colitis.

13].
Attempts to inhibit these enzyme systems with
specific pharmacological agents have not yet
proven effective in colitis [10]. Indomethacin and
flurbiprofen, potent cyclo-oxygenase inhibitors, do
not improve colitis and may worsen it. 5-1ipoxygenase inhibition with benoxaprofen was ineffective;
newer, more potent 5-1ipoxygenase inhibitors a r e
currently being evaluated. Dietary therapy could
also be efficacious. A fish-oil diet rich in polyunsaturated (w-3) fatty acids inhibits 5-1ipoxygenase and
results in the synthesis of P G E 3 and LTBs, which
a r e thought to be less pro-inflammatory than PGE 2
and LTB 4. Conventional treatment may or may
not act by modification of arachidonic acid metabolism. Experimentally, corticosteroids inhibit the
release of membrane bound arachidonic acid and
therefore limit the availability of free substrate for
these two enzymes. Sulphasalazine and its metabolites exert diverse and dose-dependent effects:
prostaglandin synthesis and degradation are re-

700
600

PGE2 release(ng/g w.w./20min8

LT84 release(ng/g w.w./20mins

?0

500

i60
i
!50

400

40

300

30

200

20

100
0

~- N87

F p(0*01~

.,a
Normal
(n-7)
1

r p'0"01 "7

UC
(n-43)
Control

Normal
~

(n-7)
AA (101Jg/ml)

r - pr

10
UC

(n-43)

Mean * 8EM
Wilcoxon paired and unpaired tests

Fig. 15. Effect o f exogenous arachidonic acid on eicosanoid re-

lease by normal and colitic mucosa

Discussion
Professor Tytgat (Q.): We have just finished a
study measuring the PAF content in biopsies from
normals, patients with ulcerative colitis, pouch patients with and without pouchitis, and the actual
content of PAF in all these four categories was
exactly the same. There was certainly no increase in
PAF in pouchitis mucosa. That does not mean that
the generation capacity may be different but the
actual content was indistinguishable.
Dr. Gertner (A.)." We have actually tried to look at
the PAF release and production in colitis and it is
quite a difficult chemical to get out.
Professor Tytgat (Q.): But that was generation capacity and that is what you expect in a mucosa with
so many inflammatory cells. That doesn't mean a
thing, and the actual content I think is much more
important.
Dr. Gertner (A.)." One of the problems with this is
that this could all be an epiphenomenon. We aren't
sure if it is perhaps just an aftereffect of the inflammatory infiltrate switching on all these enzyme
mechanisms. What we don't know is what actually
switches it on in the first place and this is an attempt to get a handle on that really.
Mr. Nicholls (Q.): Where do these reactions occur?
Dr. Gertner (A.)." Membrane bound arachidonic
acid is released by phospholipase into the cytosol,
and these mediators are produced within the cytosol and then released.
Professor Hultdn (Q.): The neurohumoral regulation of blood flow differs very much in the stomach, small intestine and colon. It is only in the
stomach and colon that we have vaso-dilatator
fibres. You focus your interest on prostaglandins.
What about the bradykinin mechanism?

221

Dr. Gertner (A.)." I could show you a slide with 10

other cascades, all of which involve different pathways such as the kinins, for example. They may all
be important in inflammatory mechanisms but we
do see a particular correlation between these proinflammatory mediators, the leucotrienes, and the
disease condition in colitis. That is not to say that
these other cascades are not also important. We just
do not know.

Faecal bile acids in pouch and pouchitis

patients
M. J. Hill and R. W. Owen
In previous studies we have reported on the faecal
bile acids in patients with chronic ulcerative colitis
and in familial adenomatous polyposis (FAP) patients. In the former group of patients there was a
relationship between the faecal bile acid concentration and the severity of epithelial dysplasia and the
risk of colorectal cancer. The latter group were
characterised by a lack of bacterial metabolites despite the bacterial flora having a normal composition and having normal bile acid degradative activity in vitro. These results suggested that in FAP
there was an inhibitor of bacterial metabolism present in the colon and secreted either by the colonic
mucosa or in some small intestinal secretion (e.g.
bile).
It has been demonstrated that bile acids are
toxic to the intestinal mucosa and are responsible
for cholerrheic diarrhoea. It was therefore decided
to study the faecal bile acids in pouch and pouchitis
patients to determine whether the diarrhoea and
severe inflammation in pouchitis could be due to an
abnormally high intestinal bile acid concentration.
In addition, by comparing the faecal steroids of
pouch patients treated for FAP with those treated
for UC, evidence of the cause of the lack of steroid
metabolism in FAP patients might be obtained.
The faecal steroids in the 20 healthy UC-Pouch
patients, the 7 healthy FAP-Pouch patients and the
6 UC-Pouch patients with pouchitis were assayed
by the method of Owen et al. [14].
In contrast to ileostomy patients, the faecal
steroids were mainly deconjugated (68% in UCPouch and 81% in FAP-Pouch). The hydrolase enzyme produced by the gut flora is highly active and
evidently even the brief transit time in the pouch
patients was sufficient to facilitate the reaction.
Similarly, whereas in ileostomy patients the bile
acids excreted have usually undergone only slight
dehydroxylation, in the pouch patients 23% of the

Table 3. Mean bile acids (mg/g faeces)

UC
n=20
Totalbile acids(TBA) 7.68
Total conjugates
2.68(35%)**

UC
pouchitis
n=6

FAP
n-7

5.65
0.52(9%)**

7.15
1.56(22%)

0.17 *
0.35
5.14
2.87
0.24
1.86
0.18

0.46
1.10
5.60
2.35
0.90
1.73
0.54

(% TBA)
Tauro-conj
Glyco-conj
Total free bile acids
CA
DCA
CDCA
LCA

0.44 *
2.24
5.0
1.94
0.51
2.04
0.51

UC, ulcerative colitis; FAP, familial adenomatous polyposis;

* Significantly different, p < 0.025; ** significantly different,
p < 0.05

free bile acids had undergone dehydroxylation. The

extent of dehydroxylation was as high in FAPPouch patients as in UC-patients.
There is no evidence from this study that the
faecal bile acid concentration is high in pouchitis;
in fact the reverse appears to be true since the faecal
bile acid concentration in pouchitis patients was
13.98 m M compared to 17.27 and 18.42 m M respectively in the FAP-Pouch and the UC-Pouch
patients. It can be seen in Table 3 that the pouchitis
patients had significantly lower levels of glycoconjugates in the faeces than colitics without pouchitis.
The bile acids do not appear to have a cholerrheic role in the causation of pouchitis. Interestingly, the results on the pouch patients with no
pouchitis provide useful evidence that the lack of
metabolic activity in the colonic flora of FAP patients is dependent on the presence of the colon and
may be caused by a colonic secretion.

Discussion
Professor Hultdn (Q.)." Which of the different bile
acids are the most cytotoxic?

Dr. Hill (A.): Lithocholic and deoxycholic are

both highly cytotoxic but lithocholic seems to be
marginally more so.
Professor Tytgat (Q.): If there is less fermentation
of fibre there should be less dilution of bile salts.
Dr. Hill (A.)." The food residues are of such low
molecular weight that they don't have much osmotic effect, so they won't really affect the degree
of diet hydration.
Professor Tytgat (Q.): Therefore, there should be
less dilution.
Dr. Hill (A.): Yes, but I am not talking about
aqueous dilution, I am talking about solid dilution.

222

Professor Tytgat (Q.)." I was talking about

aqueous. Could it be that some of the bile salts are
sequestered by the fibres?
Dr. Hill (A.)." Yes.
Professor Tytgat (Q.): And you extract it in your
extraction procedure?
Dr. Hill (A.): Yes. We are happy that we can extract any bile acids that would have absorbed to
solid surfaces.
Mr. Kmiot (Q.)" Over how many days did you
collect the faecal sample? There has been some recent work which suggests that to get a valid result
you need at least 7 days and some people say more
than 7 days.
Dr. Hill (A.): That is if you are going to look at an
individual not if you want to look at populations.
If you are looking at populations, then you can use
single samples. If you are looking at individuals
then you must have at least a 3-day collection.
Mr. Kmiot (Q.): Could a single sample have picked
out a person when they were just secreting a low
amount of bile acid at that time point of sampling?
Dr. Hill (A.). Well, we have never seen tremendous
diurnal variation even from day to day. If you look
at the concentrations, we haven't seen the variations that others have reported. We found an error
of plus or minus 10% by taking a single sample
rather than a 3-day collection.
Professor Keighley (Q.)." We know that transit is
grossly abnormal in these pouch patients, and solid
lumps of food are there. This must be a factor. Did
you standardise diet in this study? They were not all
starved? Because it makes such a difference to what
you collect?
Dr. Hill (A.)." We reckoned that since these patients
didn't have to be starved to get pouchitis we ought
to look at their steroids under normal dietary circumstances.
Professor Keighley (Q.)" But because eating food
has such an impact on the faecal composition
surely you should standardise it somehow?
Dr. Hill (A.): I think that in a perfect study we
would have had diet standardised.
Professor Keighley (Q.): Have you studied
ileostomy patients who have had a large segment of
terminal ileum resected?
Dr. Hill (A.): No. We haven't looked at that.
Mr. Nicholls (Q.): Is there any difference in bile
salt concentrations in polyposis patients with an
intact colon and those that have had a colectomy
with ileorectal anastomosis?
Dr. Hill (A.)." If they have had a colectomy they
behave from the metabolic point of view as if they
have an ileostomy. There is no metabolism at all.
But there is no metabolism in a colitis patient who

has had a colectomy either, neither is there in a

polyposis patient with a complete colon. We have
explained the lack of metabolism in the ileorectal
ones as being due to very short transit. When we
analysed these patients by type of pouch, we had
much greater metabolism in the ones who needed
to catheterise. So it is a bit of a blow to hear that
those with the artificial emptying didn't have any
longer stasis than the others. But for some reason
or other they do have more metabolism.
Professor Williams (Q.): How much can you rely
on the fact that you have only got six subjects? We
heard the suggestion that diet is important. What
about age, length of time since surgery, length of
bowel remaining? How reproducible are these measurements?
Dr. Hill (A.): We could have improved the study a
great deal by standardising the diet first of all and
obviously by getting greater numbers. We could
improve it by taking longer collections instead of
single samples. We would possibly get different results in patients with a pouch that had been done
many years previously rather than a few months.
There must be some period of stabilisation.
Professor Williams (Q.)." Is there any difference in
body weight, age and sex?
Dr. Hill (A.)." Yes. The faecal bile acid output increases with increasing body weight. It is higher in
women than in men and it increases with age. There
is a relatively small increase with age.

Pouchitis: defining an objective method

of diagnosis
W. A. Kmiot and M. R. B. Keighley
The reported incidence of pouchitis varies from
7 - 4 3 % in patients following restorative proctocolectomy, although it is not clear if different series
are describing the same condition. Recent work has
documented an incidence of 11% on the basis of
specifically defined clinical, sigmoidoscopic and
histological criteria of acute inflammation. Despite
this, the definition ofpouchitis remains unclear and
the diagnosis difficult due to the relapsing and remitting nature of the disorder and its patchiness at
a macroscopic and microscopic level.
To define a more objective diagnostic method
we have used the specific inflammatory cell seen in
pouchitis, i.e. the neutrophil, and performed isotope scans after labelling it with 111Indium.
Twenty-two patients have had the radiolabelled
neutrophil scans performed at a median of 22
(range 7-46) months following pouch formation.

223
The study groups were stratified into three; (I) pouchitis on clinical, microscopic and sigmoidoscopic
criteria (n=9); (2) poor functional result without
evidence of pouchitis (n=7); (3) good functional
result (n = 6).
Sixty ml. of venous blood was removed from
each patient and 15 ml layered over 7 ml of FicollHypaque in four separate tubes. Following centrifugation at 400 g for 60 min a distinct neutrophil
band was visible which was aspirated. After a single
wash in saline, the cells were labelled with 5 M B q of
Indium oxime and resuspended in platelet free
plasma. This was centrifuged for a final time, the
supernatant counted to determine labelling efficiency, and the cell pellet injected into each patient
following a further resuspension in plasma.
All patients had gamma camera scans 4 and
24 h following cell reinjection and a 4-day stool
collection counted isotopically, along with specific
clinical, haematological and sigmoidoscopic assessment. All nine subjects in the pouchitis group had
a l-month course of metronidazole. All patients
were re-studied I month later and those with a
positive scan initially were rescanned.
Two out of nine scans in the pouchitis group
were unsuccessful and of the remaining seven, four
were positive. After treatment, repeat scans showed
reduced pouch activity in each case. All other patients had a negative scan.
Faecal granulocyte excretion was positively
correlated with scan activity and was decreased in
each re-scanned case. The 24-h defaecation frequency in patients with a positive scan significantly
improved from a pre-treatment value of 8 (7-11)
times per 24 h (median and range) versus a posttreatment value of 6 (4-8), a difference not detected in the negative scan group.
Histological grade and sigmoidoscopic appearance improved following therapy only in those patients with a positive scan.
Haematological and biochemical indices of inflammation tested, namely haemoglobin, white cell
count, ESR, CRP, AGP and albumin were not significantly different between the three study groups.
Absolute values did not correlate with a positive
scan.
This pilot study demonstrates that 111Indium
granulocyte scanning is a sensitive method of assessing acute pouch inflammation. The absence of
positive scans in asymptomatic patients refutes the
existence of a syndrome of subclinical pouchitis.
Specificity however is less, with the three false negatives reflecting the inconsistent response of the
condition to metronidazole and its intermittently
active nature.

Therefore use of this technique in conjunction

with the other three diagnostic criteria may enable
improved detection of active pouchitis.

Discussion
Mr. Nicholls (Q.): There must be polymorphs in
the pouches, even in the negative ones. Is it a matter
of the method being too insensitive to pick them
up?
Mr. Kmiot (A.): The amount of neutrophils in each
patient in the pouchitis group was similar. The labelling in patients with positive and negative scans
in that group was similar as well. I think this may
be a question of neutrophil kinetics, perhaps the
cells are not entering the mucosa in a patient with
a negative scan. One presumes that this disease is in
fact in remission at the time.
Dr. Shepherd (Q.): Isn't it more likely to be a reflection of the presence of ulceration? If you look at
ulceration histopathologically there are thousands
of polymorphs.
Mr. Kmiot (A.)." Well, if that was the case you
would expect it to be picked up on the faecal excretion.
Mr. Bartolo (Q.): Did you employ any method for
counting the neutrophils before you put them back
into the patient?
Mr. Kmiot (A.)." Yes, they were all Coulter counted.
The median values were between 107 and l0 s .
Mr. Bartolo (Q.)." Did you assess how accurate
your labelling was before putting them back into
the patient?
Mr. Kmiot (A.): We did. We centrifuged the neutrophils after labelling and then measured the activity in the supernatant; the labelling efficiency was
around 66%.
Professor Williams (Q.)" I have patients who have
urgency of defaecation who I would classify as having pouchitis because on biopsy they have an acute
inflammatory reaction, but frequency is only four
times a day. Other patients are very well for 3 or 4
months and then they get an acute attack of what
I again would classify as pouchitis. So one wonders
about in your poor result group. Did any of these
go on and have this sort of cyclical type of problem
that some patients get, and indeed in the so-called
pouchitis group, were you looking at them during
an acute attack?
Mr. Kmiot (A.): Both the positive controls and the
study group had this history. They were selected
exactly for that reason. So in fact, you might suspect pouchitits in them, even though there was
nothing on biopsy or on the sigmoidoscope for you
to confirm your preliminary diagnosis.

224

Mr. Nicholls (Q.): In the group with poor result

but no pouchitis what do you think the dysfunction
was due to?
Mr. Kmiot (A.)." Two of the seven patients were
found to have jejeunal bacteriological overgrowth.
Of the other five one had thyrotoxicosis and the
other four were labelled as functional diarrhoea.
Mr. 0 'Connell (Q.)." We had this morning a virtual
agreement that pouchitis occurs only when you
have had an antecedent history of ulcerative colitis.
It seems to occur chiefly in patients with pancolitis,
and I suggest that it also seems to occur mainly in
patients who have had extracolonic manifestations
of their disease, because those who have recurrent
pouchitis are those who complain of arthralgia,
uveitis, general malaise and a high ESR. We have
such a spectrum of the disease that perhaps if we
can define those about whom none of us would
disagree, that would be a starting point.
Mr. Bartolo (Q.)." When we diagnose someone
with ulcerative colitis we use clinical and endoscopic features and histology. Pouchitis could well
be ulcerative colitis occurring again in colonic
metaplasia in the small bowel. Why can't we just
use those same features that we use to diagnose
ulcerative colitis?
Mr. Kmiot (A.)." On the basis that, in our patients
for which we used those criteria, three did not have
any clinical response to metronidazole.
Mr. Nicholls (Q.). We do not know anything like
enough about drug treatment in this condition to
be able to accept that as a diagnostic criterion. It
has been suggested that there are at least two different forms of pouchitis, that it is a heterogeneous
condition. There is a bacteriologically related one
which may respond to metronidazole and there is
one which does not respond which might do so to
steroids.
Mr. Hawley (Q.)." I would support that. For me
pouchitis is a clinical diagnosis, but not all cases
respond to metronidazole. I have got about six patients who have had to go back on steroids and they
are totally cured with a small dose of steroids.
Professor Williams (Comment): You should not
even close your mind to two causes of it. This is
probably a multi-factorial problem.

Evacuation and pouchitis

P. R. O'ConneU
Construction of an ileal pouch reservoir produces
stasis within the distal small bowel. This is desirable

in patients with an ileal pouch-anal anastomosis or

a Kock continent ileostomy, as post operative stool
frequency is directly related to the maximum capacity of the ileal reservoir [15, 16]. The corollary is
that the ileal reservoir must evacuate efficiently to
maximise benefits of a large reservoir and to minimise biological effects of stasis within the distal
ileum [17]. The importance of efficient pouch evacuation to clinical outcome has been clearly demonstrated in patients with ileal 'S' pouch anal anastomosis [4]. Impairment of ileal pouch evacuation
resulted in ileal pouch intubation at least once per
day in 50% of patients. Animal studies confirmed
that the 'S' pouch reservoir with a long efferent
limb empties poorly, perhaps by overflow [18].
Contrast radiology confirmed that a long ileal
pouch efferent limb was obstructive [19]. These observations led to changes in ileal pouch design with
shortening of the obstructive efferent limb in 'S'
pouches and to development of other non-obstructive ileal pouch designs.
Conventional contrast radiological studies of
the ileal pouch reservoir provide excellent detail
and are useful in establishing anastomotic integrity
before closure of a protective loop ileostomy. Evacuation proctography is also useful in investigation
of patients considered to have inefficient ileal
pouch evacuation. However neither technique provides a quantitative measurement of ileal pouch
emptying which is needed for comparative studies
of ileal pouch design and clinical outcome. Furthermore, the radiation dosage of conventional radiological methods make control and sequential studies difficult to justify.
Several non-radiological and scintigraphic
methods which quantify ileal pouch evacuation
have been developed. To simulate physiological
evacuation, the media employed must have a consistency which resembles semisolid ileostomy stool.
Methyl cellulose paste [20], porridge [2], aluminium
magnesium silicate gel [21, 22] and radiolabelled
egg albumin [23] have been used. Using silicate gel
containing the colourmetric marker phenolsulphonpthalein, Stryker found a direct relationship between the efficiency of ileal 'J' pouch evacuation and functional result [21]. Later scintigraphic
studies have shown that ileal pouch emptying is
similar to that of the healthy rectum and that little
difference exists in the efficiency of evacuation of
the various ileal pouch designs [22, 23].
Pouchitis occurs in approximately 14% of patients following ileal pouch-anal anastomosis [24].
An antecedent history of ulcerative colitis and stasis within an ileal reservoir are prerequisite to development of the condition. The aetiology is widely

225
believed to be due to stasis within the ileum leading
to bacterial overgrowth and mucosal inflammation. To study this hypothesis, we investigated 20
patients following ileal pouch-anal anastomosis.
Ileal and jejeunal bacterial growth and efficiency of
ileal pouch evacuation were correlated with clinical
outcome. Enteric bacteriology and ileal pouch
evacuation were not different in patients with recurrent pouchitis (n = 6) when compared to patients
with a good clinical outcome (n = 8). Of interest
was identification of jejeunal bacteriological overgrowth in patients with a poor clinical outcome
(n = 6) associated with a large stool volume [24].
Our observations regarding ileal pouch evacuation and its relationship to pouchitis are supported
by those of Moskowitz [2] and Heppell [23]. These
cast doubt on the theory that pouchitis is due simply to stasis with quantitative bacterial overgrowth
within the ileal pouch. Whether a qualitative
change in ileal pouch flora is responsible is as yet
uncertain. An attractive alternative is that, in certain individuals, intraluminal microbial antigens
produce an inflammatory response within ileal
pouch mucosa which closely resembles the antecedent colitis [I 7]. Investigation of this hypothesis
may provide answers to many questions regarding
the pathophysiology of ulcerative colitis.

Discussion
Mr. Bartolo (Q.)." Were your patients in the left
lateral position, and were you evacuating from the
neorectum or were you using a radio-opaque balloon?
Dr. 0 'Connell (A.): The material was inserted with
the patient was in the left lateral, but then we had
a plastic commode on which the patient sat. Lateral
views only were found to be useful. The balloon
was used to look at angles not for quantifying emptying.
Professor Tytgat (Q.): Do you instruct the patients
to press when they empty their pouches?
Dr. O'Connell (A.): We just left them beside the
gamma camera and went off to one side and said
"Evacuate just as you would normally when you
want". We left the machine on continuous acquisition for 3 minutes and when they said they were
finished we went back in.
Professor Hultdn (Q.)." Do you instruct your pelvic
patients to strain?
Dr. O'Connell (A.)." No, I don't give any specific
instructions. I would certainly think I would dissuade somebody from staining a lot because prolapse might occur.

Professor Keighley (Q.)." Have you looked at

pouch emptying in your jejeunal overgrowth syndrome? What about transit?
Dr. O'Connell (A.)." The transit work from the
Mayo Clinic was specifically aimed to see whether
the ileal pouch-anal anastomosis affects the ileal
break phenomenon, and it is still intact. They have
not yet looked at patients with jejeunal bacterial
overgrowth.
Professor Keighley (Q.)." What would you predict?
Dr. O'Connell (A.)." Proably there will be a slower
transit in those with overgrowth.
Professor HultOn (Q.): Is there a relationship between incomplete emptying and pouchitis? In the
continent ileostomy most patients become continent eventually. Many irrigate the pouch with saline after every evacuation but the patients who do
that have a similar risk of pouchitis as the others.
Dr. 0 "Connell (A.). Today has reinforced my opinion that pouchitis is not due to stasis per se. Stasis
may be part of the whole, it creates the whole
milieu, but it is not just stasis.
Mr. Nicholls (Q.): Have you studied the species of
bacteria in patients with jejeunal overgrowth?
Dr. O'Connell (A.): No. In the cost-conscious
Mayo they decided to do Gram stains only.
Mr. Nicholls (Q.)." Were they quantitative?
Dr. O'Connell (A.): No, we quantified the various
growths and sub-divided those on the basis of the
Gram strain. We could say whether they were
Gram negative bacilli, but we couldn't go into the
subspecies. It would cost $1500 a specimen to process in greater detail.
Professor HultOn (Q.)." Were there correlations between defaecation frequency and pouchitis?
Dr. O'Connell (A.): Some people with pouchitis
have a greater stool frequency than when they
don't have pouchitis.
Professor Hultbn (Q.): H o w did you define pouchitis in Rochester?
Dr. 0 'Connell (A.)." We defined it as a clinical syndrome.
Mr. Nicholls (Comment). We might leave that till
the end. I do not wish to interrupt but I think that
in fact that might be germane to the final 10 minutes of discussion, and we are all ears to know how
they define pouchitis in Rochester. But thanks very
much indeed.
Ileal pouch motility
D. Kumar and N. S. Williams
The diagnosis of pouchitis is made either on clinical
symptoms of excessive stool frequency and malaise

226
or macroscopic/microscopic inflammation of the
ileal reservoir. However, there is a significant overlap in pouch function or dysfunction between the
inflamed and normal pouches. Also our current
understanding of pouch motility is based on static
measurements of pouch pressure independent of
small bowel motility proximally and anal canal motor activity distally. In order to define patterns of
pouch motility and establish relationships between
motility of the ileal reservoir, the small bowel and
the anal canal on one hand and pouch function
morphology on the other, we have studied anal
manometry, reservoir capacity and compliance and
mucosal morphology in 15 patients [7] and ambulatory measurements of small bowel motility, pouch
motility and anal canal motility in 8 patients. Static
measurements (anal manometry, pouch capacity
and compliance) were made using a water filled
balloon connected to a pressure transducer and a
latex balloon attached to an 8 F G catheter respectively. Ambulant measurements of small bowel,
pouch and anal canal manometry were made by a
fine (OD 2 mm) catheter-mounted microtransducers. The signal was recorded on audio tape on a
portable tape-recorder. The patients were fully ambulant during the study.
Resting anal canal pressure was significantly
lower (p < 0.01) in pouch patients but the maximum
squeeze pressure was unaltered. Patients with minor leakage had significantly lower resting pressures (p<0.05) than those who were fully continent. Patients with urgency of defaecation had
significantly lower squeeze pressures (p < 0.05) than
those who could defer defaecation for more than
30 min. Similarly patients who could defer defaecation had greater pouch compliance and larger
pouch capacity. None of the pouches showed
macroscopic inflammation when examined endoscopically. However, on microscopic examination,
both fibrosis and moderate to severe infiltration of
the lamina propria with acute and chronic inflammatory cells was evident. There was no significant
correlation between the degree of mucosal inflammation and either frequency or urgency of defaecation.
Prolonged manometry of the pouch showed
two types of motor activity. First, there were large
( > 50 mm Hg) amplitude contractions which lasted
for 0.5_+0.1 min (mean_+SEM) duration and had
no relation to pouch filling. These giant contraction waves were significantly (p < 0.05) more frequent in pouches with a capacity of 300 ml (mean)
or less. They occurred with similar frequency during the day and at night. When three or more such
contractions appeared over a period of 5 min, the

patients felt an urge to empty the pouch. The second type of contraction occurred with a frequency
of 6-8/min. This type of activity was predominantly seen after meals. Giant contraction waves
were also seen in association with 6 - 8 per minute
activity in the postprandial period. The anal canal
showed slow contractile activity at a frequency of
2 3/min. The upper anal canal immediately adjacent to the anastomosis showed complete inhibition of motility in response to pouch distension; the
mid and lower anal canal were not affected in a
similar manner. On the other hand, small bowel
motility in patients with ileal reservoirs was characterised by a significantly (p<0.05) shorter cycle
length (42.5_+11.4min) when compared with
healthy controls. The contractile frequency, phase
I, and phase II activity in the small bowel in ileal
pouch patients was similar to that seen in healthy
controls. Whether this is the effect of colectomy or
pouch formation on small bowel motor activity has
yet to be determined.
Prolonged monitoring of ileal reservoirs has
provided us with a pouch motility marker (giant
contraction waves) which can be easily measured to
give us an indication of pouch function, i.e. giant
contractions occur more frequently in smaller capacity pouches and are related to the frequency of
bowel action. Although these types of contractions
do not reflect the aetiology of pouch dysfunction,
they may provide a useful guide to its severity.

Discussion
Professor Hultkn (Q.): H o w about motility in the
pouch with inflammation?
Mr. Kumar (A.)." I wish I knew how to classify
them into the two separate groups of no pouchitis
and definite pouchitis. However, the patients with
inflammation would tend to have a frequency of
more than 6 per minute, whereas those who have a
good result tend to show a lower frequency of these
contractions.
Mr. Kmiot (Q.)." We still don't really know why
patients with pouchitis evacuate so frequently. Is it
increased motility or is it excessive secretion?
Mr. Kumar (A.)." I am not sure what you mean
when you say "pouchitis" and what Leif Hult6n
means when he says "pouchitis". There are one or
two patients in my practice who get absolute classic
colitis-type symptoms and have everything else
there, but I also have patients who are asymptomatic for periods of time and are fine, then suddenly they have an attack of increased frequency.
Furthermore, I have patients who have frequency

227

of bowel action but they don't have much in the

way of inflammation.
Mr. Nicholls (Q.)- All pouch dysfunction is not
pouchitis, and indeed your relationship between
sphincter relaxation, sphincter performance and
urgency suggests a separate functional disorder,
not pouchitis.
Professor Williams (A.)." Well, just to be controversial, with the Kock ileostomy you are dealing with
a pouch which has an egress to the surface. There
is a sphincter of sorts but it is not a real sphincter
and also it is not in the pelvis. It is simplistic to rule
- out other factors and to say just because there is
inflammation and poor function, then the patient
has pouchitis. I suspect that most such patients
have not been fully evaluated in motility terms.
Professor Hultdn (Q.): I think you are making it
more complicated than it is in practice. If you have
got a pouch patient with a steady frequency of
about 3 to 5 per day, who then develops increased
frequency with bloody stools and malaise, there is
no difficulty on seeing a red bloody mucosa there to
say that patient has got pouchitis.
Professor Keighley (Q.): We know there is a
tremendously variable response to modalities of
medical treatment and we are probably talking
about a spectrum of disordered pouch function.
One of the components is something that is associated with a lot of diarrhoea, it may be sudden or
insidious, and you see inflammation when you put
the sigmoidoscope in. The histopathologist sees a
lot of acute inflammatory cells as opposed to the
normal villous atrophy that is almost normal in
these patients. Those taken together with a clinical
symdrome ought to be the criteria for the diagnosis.
Mr. Nicholls (Q.): So we have gone full circle back
to this morning. We are looking at a clinical disorder which, like all diseases, can be defined by appearances, both macro and micro, plus the clinical
features.
Professor Williams (Q.)." We are looking at a syndrome, aren't we? We are not looking at a disease
entity but a pouchitis syndrome for which there
may be a variety of causes.
Dr. Kamm (Q.): The problem is similar to others in
inflammatory bowel disease in terms of nature and
definition. Using the Crohn's disease activity index,
especially in following therapy, you cannot include
treatment as one of the parameters, but you can use
a complex of symptoms, signs and laboratory data
and come up with a final figure.
Mr. Nicholls (Q.): It seems that we are talking
about a syndrome of clinical, histological and
macroscopic features, the cause of which is obscure
because the cause of ulcerative colitis is obscure

because we have so far not unearthed any aetiological factors that can be related to the condition.
Mr. Mann (Q.): If you tried to define ulcerative
colitis as a syndrome in terms of bowel frequency,
urgency and incontinence and everything else, it
would be extremely difficult. The only objective
data that are relatively independent of subjective
influences are the histological criteria and I would
make a plea that we do observe this as very much
part of our appreciation of what we mean by pouchitis.

References
1. Parks AG, Nicholls RJ (1978) Proctocolectomy without
ileostomy for ulcerative colitis. Br Med J 2:85-88
2. Moskowitz RL, Shepherd NA, Nicholls RJ (1986) An assessment of inflammation in the reservoir after restorative
proctocolectomy with ileoanal ileal reservoir. Int J Colorect
Dis 1:167 174
3. Shepherd NA, Jass JR, Duval I, Moskowitz RL, Nicholls
RJ, Morson BC (1987) Restorative proctocolectomy with
ileal reservoir the histopathology of mucosal biopsies. J
Clin Pathol 40:601-607
4. Moskowitz RL (1986) Pathophysiology. Symposium:
Restorative proctocolectomy with ileal reservoir. Int J Colorect Dis 1:2-19
5. Gorbach SL, Plaut AG, Nahas L, Weinstein L, Spanknebel
G, Levitan R (1967) Studies of intestinal microflora II. Micro-organisms of the small intestine and their relation to
oral and faecal flora. Gastroenterology 53:856-867
6. Nasmyth DG, Godwin PGR, Dixon MF, Williams NS,
Johnston D (1989) Ileal ecology after pouch-anal anastomosis or ileostomy. A study of mucosal morphology, fecal
bacteriology, fecal volatile fatty acids and their interrelationship. Gastroenterology 96:817 824
7. Nasmyth DG, Johnston D, Godwin PGR, Dixon MF,
Smith A, Williams NS (1986) Factors influencing bowel
function after ileal pouch-anal anastomosis. Br J Surg
73:469 473
8. Kruh J (1982) Effects of sodium butyrate, a new pharmacological agent, on cells in culture. Mol Cell Biochem 42:6582
9. Weaver GA, Krause JA, Miller TL, Wolin MJ (1988) Short
chain fatty acid distributions of enema samples from a sigmoidoscopy population: an association of high acetate and
low butyrate ratios with adenomatous polyps and colon
cancer. Gut 29:1539-1543
10. Rampton DS, Hawkey CJ (1984) Prostaglandins and ulcerative colitis. Gut 25:1399-1413
11. Gertner D J, de Nucci G, Rampton DS, Cynk E, LennardJones JE (in press) Exogenous arachidonic acid potentiates
LTB 4 and PGE 2 synthesis by colitic mucosa in vitro. Gut
12. LeDuc LE, Nast CC, Patterson JB, Zipser R D (1987)
Chemotactic peptides induce colon inflammation: a new
model and mechanism of colitis. Gastroenterology 92:1496
13. Graven PA, Pfanstiel J, Saito R, DeRubertis FR (1987)
Actions of sulfasalazine and 5-aminosalicylic acid as reactive oxygen scavengers in the suppression of bile acid induced increases in colonic epithelial cell loss and proliferative activity. Gastroenterology 92:1998-2008
14. Owen RW, Thompson MH, Hill MJ (1984) Analysis of
metabolic profiles of steroids in faeces of healthy subjects

228

15.

16.
17.
18.
19.
20.

21.
22.
23.

24.

undergoing chenodeoxycholic acid treatment by liquid-gel

chromatography and gas-liquid chromatography-mass
spectrometry. J Steroid Biochem 5:593-600
Heppell J, Kelly KA, Phillips SF, Beast RW, Telander RL,
Perrault J (1982) Physiologic aspects of continence after
colectomy, mucosal proctectomy and endorectal ileo-anal
anastomosis. Ann Surg 195:435 443
O'Connell PR, Pemberton JH, Kelly KA (1987) Determinants of stool frequency after ileal pouch-anal anastomosis.
Am J Surg 153:157-164
Phillips SF (1987) Biological effects of a reservoir at the end
of the small bowel. World J Surg 11:763 768
Cranley B, McKelvey STD (1982) The pelvic ileal reservoir:
an experimental assessment of its function compared with
that of the normal rectum. Br J Surg 69:465 469
Pescatori M, Manhire A, Bartram CI (1983) Evacuation
pouchography in the evaluation of ileoanal reservoir function. Dis Colon Rectum 26:365-368
Nasmyth DG, Williams NS, Johnston D (1986) Comparison of the function of triplicated ileal reservoirs after mucosal proctectomy and ileoanat anastomosis for ulcerative
colitis and polyposis coll. Br J Surg 73:361-366
Stryker SJ, Kelly KA, Phillips SF, Dozois RR, Beart RW
(1986) Anal and neorectal function after proctocolectomy
and ileal pouch-anal anastomosis. Ann Surg 203:55-61
O'Connell PR, Kelly KA, Brown ML (1986) Scintigraphic
assessment of neorectal motor function. J Nuc Med 27: 460464
Heppell J, Belliveau P, Taillefer R, Dube S, Derbekyan V
(1987) Quantitative assessment of pelvic ileal reservoir emptying with a semisolid radionuclide enema. A correlation
with clinical outcome. Dis Colon Rectum 30:81-85
O'Connell PR, Rankin DR, Weiland LH, Kelly KA (1986)
Enteric bacteriology, absorption, morphology and emptying after ileal pouch-anal anastomosis. Br J Surg 73:909914

Further bibliography
Becker JM, Raymond JL (1986) Ileal pouch-anal anastomosis.
A single surgeon's experience with 100 consecutive cases.
Ann Surg 204:375-383
Breuer NF, Rampton DS, Tammar A, Murphy GM, Dowling
RH (1983) Effect of colonic perfusion with sulfated and
non-sulfated bile acids on mucosal structure and function in
the rat. Gastroenterology 84:969-977
Coleman R (1987) Bile salts and biliary lipids. Biochem Soc
Trans 15:68S-80S
Dozois RR (1986) In: Symposium, Restorative proctocolectomy
with ileal reservoir. Int J Colorectal Dis 1:14 15
Fleshman JW, Cohen Z, McLeod RS, Stern H, Blair J (1988)
The ileal reservoir and ileo-anal anastomosis procedure.
Factors affecting technical and functional outcome. Dis Colon Rectum 31:10-16
Fonkalsrud EW (1987) Update on clinical experience with different surgical techniques of the endorectal pull-through operation for colitis and polyposis. Surg Gynec Obstet
165:309 316
Goldberg SM (1986) In: Symposium. Restorative proctocolectomy with ileal reservoir. Int J Colorectal Dis 1:15 16
Gorbach SL, Nahas L, Weinstein L (1967) Studies of intestinal
microflora. IV. The microflora of ileostomy effluent: a
unique microbiological ecology. Gastroenterology 53:874
880

Gustavsson S, Weiland LH, Kelly KA (1987) Relationship of

backwash ileitis to ileal pouchitis after ileal pouch-anal anastomosis. Dis Colon Rectum 30:25 28
Handelsman JC, Fishbein RH, Hoover HE, Smith GW, Haller
JA (1983) Endorectal pull-through operation in adults after
colectomy and excision of rectal mucosa. Surgery 93:247
253
Harig JM, Soergel KH, Komorowski RA, Wood CM (1989)
Treatment of diversion colitis with short chain fatty acid
irrigation. New Eng J Med 320:23-28
Hult6n L (1985) The continent ileostomy (Kock's pouch) versus
the restorative proctocolectomy (pelvic pouch). World J
Surg 9:952 959
Hult6n L, Svaninger G (1984) Facts about the Kock continent
ileostomy. Dis Colon Rectum 27:553-557
Jarvinen H J, Makitie A, Sivula A (1986) Long-term results of
continent ileostomy. Int J Colorectal Dis 1:40-43
Kelly DG, Phillips SF, Kelly KA, Weinstein WM, Gilchrist MJ
(1983) Dysfunction of the continent ileostomy: clinical features and bacteriology. Gut 24:193 201
Kock NG, Darle N, Hultdn L, Kewenter J, Myrvold H, Philipson B (1977) Ileostomy. Curr Probl Surg 14:36 38
Kock NG, Myrvold HE, Nillson LO, Philipson MB (1985) Continent ileostomy: the Swedish experience. In: Dozois RR (ed)
Alternatives to conventional ileostomy. Year Book Medical
Publishers, Chicago, pp 163 175
Lohmuller JL, Pemberton JH, Dozois RR, Ilstrup D (1989) The
relationship between pouchitis after ileal pouch-anal anastomosis and extra-intestinal manifestations of chronic ulcerative colitis. Paper presented to American Society of Colon
and Rectal Surgeons, June 1989
Luukonen P, Valtonen V, Sivonen A, Sipponen P, Jarvinen H
(1988) Fecal bacteriology and reservoir ileitis in patients
operated on for ulcerative colitis. Dis Colon Rectum
31:864-867
Luukonen P, Jarvinen H, Lehtola A, Sipponen P (1988) Mucosa
alterations in pelvic ileal reservoirs. A histological and ultrastructural evaluation in an experimental model. Ann Chir
Gynaecol 77:91-96
Luukonen P (1989) Restorative proctocolectomy for ulcerative
colitis. A clinical and experimental study. Academic Dissertation, University of Helsinki
Miglioli M, Barbara L, Di Febo G, Gozzetti G, Lauri A, Paganelli GM, Poggioli G, Santucci R (1989) Topical administration of 5-aminosalicylic acid: a therapeutic proposal for
the treatment of pouchitis. New Eng J Med 320:257 (letter)
Nasmyth DG, Godwin PGR, Dixon MF, Williams NS, Johnston D (1985) The relationship between mucosal structure
and intestinal flora in ileal reservoirs. Br J Surg 72:$129
Nicholls RJ, Belliveau P, Neill M, Wilks M, Tabaqchali S (1981)
Restorative proctocolectomy with ileal reservoir: a pathophysiological assessment. Gut 22:462 468
Nicholls RJ, Moskowitz RL, Shepherd NA (1985) Restorative
proctocolectomy with ileal reservoir. Br J Surg 72 [Suppl]:
$76-$79
Nicholls R J, Lubowski DZ, Hill M J, Moskowitz RL, Shepherd
NA, Owen RW, Fernandez F (1989) Pouchitis after restorative proctocolectomy: faecal bacteriological and bile salt
analysis in patients with ulcerative colitis and familial adenomatous polyposis. Paper presented to Tripartite Meeting,
Birmingham, June 1989
Nilsson LO, Kock NG, Lindgren I, Myrvold HE, Philipson
BM, Ahren C (1980) Morphological and histochemical
changes in the mucosa of the continent ileostomy 6-10 years
after its construction. Scand J Gastroenterol 15:737 747
Pemberton JH, Kelly KA, Beart RW, Dozois RR, Wolff BG,
Ilstrup DM (1987) Ileal pouch-anal anastomosis for chronic
ulcerative colitis. Long-term results. Ann Surg 206:504 513

229
Philipson B, Brandberg A, Jagenburg R, Kock N, Lager I,
Ahren C (1975) Mucosal morphology, bacteriology and absorption in Intra-abdominal ileostomy reservoir. Scand J
Gastroenterol 10:145- 1 5 3
Roediger WEW (1980) The colonic epithelium in ulcerative colitis an energy deficiency disease? Lancet 2:712 715
Silk DBA (1989) Fibre and enteral nutrition. Gut 30:246 264
Stelzner M, Fonkalsrud EW, Lichtenstein G (1988) Significance
of reservoir length in the endorectal ileal pullthrough with
ileal reservoir. Arch Surg 123:1265 1268
Taylor BA, Wolff BG, Dozois RR, Kelly KA, Pemberton JH,
Beart RW (1988) Ileal pouch-anal anastomosis for chronic
ulcerative colitis and familial polyposis coli complicated by
adenocarcinoma. Dis Colon Rectum 31:358-361
Tytgat GNJ, van Deventer SJH (1988) Pouchitis. Int J Colorectal Dis 3:226-228

List of speakers
Dr. D. Gertner, St. Mark's Hospital, City Road, London
EC1V 2PS, UK
Dr. M. J. Hill, Director, Bacterial Metabolism Research Laboratory, Public Health Laboratory Service, Porton Down, Salisbury, Wilts SP4 0JG, UK

Professor L. Hult6n, Department of Surgery II, Sahlgrenska

sjukhuset, S-413 45 Goteborg, Sweden
Professor M. R. B. Keighley, Department of Surgery, Queen
Elizabeth Hospital, Edgbaston, Birmingham B15 2TH, UK
Mr. W. Kmiot, Department of Surgery, Clinical Teaching
Block, The General Hospital, Steelhouse Lane, Birmingham
B4 6NH, UK
Mr. D. Kumar, Department of Surgery, Queen Elizabeth Hospital, Edgbaston, Birmingham BI 5 2TH, UK
Mr. G. Nasmyth, Senior Surgical Registrar, Department of
Surgery, Royal Liverpool Hospital, Prescot Street, Liverpool
L69 3BX, UK
Mr. R. J. Nicholls, St. Mark's Hospital, City Road, London
EC1V 2PS, UK
Dr. P. R. O'Connell, Senior Lecturer, Surgical Unit, The London Hospital, Whitechapel, London El 1BB, UK
Dr. N. A. Shepherd, Department of Histopathology, Gloucestershire Royal Hospital, Great Western Road, Gloucester
GLI 3NN, UK
Professor G. N. J. Tytgat, Department of Hepato-Gastroenterology C2, Academic Medical Centre, Meibergdreef 9, NL1105 AZ Amsterdam Zuidoost, The Netherlands
Professor N. S. Williams, The London Hospital Medical College, Surgical Unit, The London Hospital, Whitechapel, London E1 1BB, UK

Col6i'ectal
Disease

Int J Colorect Dis (1989) 4:230-233

9 Springer-Verlag 1989

Original articles
The role of CEA, TPA and CA 1 9 - 9 in the early detection
of recurrent colorectal cancer
P. Barillari, G. Ramacciato, R. de Angelis, P. Gozzo, P. Aurello, M. Indinnimeo, S. Valabrega,
F. D'Angelo and G. Fegiz
1st Surgical Clinic, University of Rome "La Sapienza", Rome, Italy

Abstract. Eighty-eight consecutive patients w h o

u n d e r w e n t curative resection for colorectal cancer
between 1983 and 1985 were studied prospectively
to evaluate the roles o f sequential C E A , T P A and
C A 19-9 d e t e r m i n a t i o n s a n d i n d e p e n d e n t clinical
e x a m i n a t i o n in the early diagnosis o f resectable recurrences. Twenty nine recurrences were detected
between 8 and 38 m o n t h s after p r i m a r y surgery.
C E A , T P A and C A 1%9 showed a sensitivity o f
7 2 % , 6 2 % a n d 38%, and a specificity o f 78%, 86%
and 9 7 % , respectively. O f eight recurrences in
which C E A was not raised, five induced a rise in
T P A a n d two a rise in C A 19-9. T h e rise in the
s e r u m c o n c e n t r a t i o n o f one o f the three m a r k e r s
was the first sign o f relapse in 23 (79%) patients.
Two s e c o n d - l o o k l a p a r o t o m i e s based solely on a
rise in s e r u m m a r k e r s were p e r f o r m e d . In one case
diffuse recurrent disease was found, a n d in the other a resectable solitary hepatic metastasis was
found.

claim that a C E A rise generally occurs late in the

n a t u r a l history o f the disease [8-11], and t h a t small
recurrences are n o t detected by serum C E A assay
[12]. Others have suggested the use o f other m a r k ers such as T P A (tissue p o l y p e p t i d e antigen) which
seems to be synthesized by tissue u n d e r g o i n g rapid
g r o w t h [13], a n d whose values m a y be elevated even
w h e n small a m o u n t s o f c a n c e r o u s tissue are present. A n o t h e r m a r k e r , C A 19-9, has been r e p o r t e d
to have a better sensitivity a n d specificity t h a n
C E A . It reacts specifically with a m o n o c l o n a l antib o d y designated 1116 NS 19-9 raised against intact
colon cancer cells [14-16].
The aim o f this p a p e r is to analyse the sensitivity and specificity o f C E A , T P A a n d C A 19-9, and
to evaluate their predictive value in detecting recurrences in patients already treated for cure for colorectal cancer. T h e effect o f this m o n i t o r i n g on the
survival was also studied.

Material and methods

Introduction
R e g u l a r follow-up in patients after surgery for colorectal cancer has b e c o m e a m a t t e r o f tradition.
M a n y tests are available t h a t m i g h t be used to detect recurrent colonic cancer. These include history
a n d physical e x a m i n a t i o n , the stool guaiac test,
b a r i u m enema, chest x-ray, liver snacs, c o m p u t e r ized t o m o g r a p h y , c o l o n o s c o p y , and even a s y m p t o m a t i c l a p a r o t o m y as p r o p o s e d by W a n g e n s t e e n
et al. [1, 2] in 1951.
F o r m a n y years use o f C E A m o n i t o r i n g in follow-up o f patients after curative surgery for colorectal cancer has been said to be o f value in detecting recurrences early e n o u g h to m a k e surgical rem o v a l possible [3-7]. H o w e v e r , s o m e a u t h o r s

Eighty-eight consecutive patients with colorectal cancer who

underwent radical surgery between January 1983 and December
1985 were followed up for a period ranging from 20 to 54
months (median 40 months). Patients undergoing palliative procedures or who were submitted to adjuvant treatments were
excluded.
All patients underwent preoperative and 10 days postoperative CEA, TPA and CA 19-9 assays. If there was no decrease
in the values of the three markers at the postoperative assay to
below the upper limit of normal, patients were excluded from
the follow-up programme. Patients were monitored with CEA,
TPA and CA 19-9 by 3-monthly assay. All surviving patients
had at least six consecutive assays of the three markers with a
maximum interval of 4 months between the assays. Seventy-two
patients had a complete follow-up with the three markers.
Fifty-three patients were male and 35 female with a mean
age of 65.3 years. Eight tumours were Dukes' stage A; 49 were
stage B and 31 were stage C. Thirty-two tumours were in the
rectum, 13 cases in the rectosigmoid region and 43 in the colon.

231
CEA assay. CEA was analysed using a direct radioimmunologic
method (CEA-PR, Sorin Biomedica). The upper limit of normal
for this assay was 3 ng/ml.

bones in 2 cases, in the brain in one case and in the

pelvis in 10 cases (8 from rectal cancer).
In 13 patients recurrence was found in multiple
organs. There were no recurrences in stage A tumours, 15 in stage B tumours (31%), 14 in stage C
tumours (45%). Twelve patients (41%) underwent
surgical treatment for the recurrent disease, and 6
of them (50%) had resectabte disease. Four patients are still living 9, 16, 21, and 31 months respectively after re-operation without evidence of neoplastic disease. Nineteen patients died of disease,
while six patients are still living with evidence of
recurrent disease.

TPA assay. TPA was analysed by the TPA-RIA system (AB

Sangtec Medical), with an upper limit of normal of 85 U/1. The
lowest detectable level was 11 U/1.
CA 19-9 assay. CA 19-9 was analysed using the Gica-RIA system (Sorin Biomedica) with an upper limit of normal of 25 U/1.
The lowest detectable level was 2 U/1.

As reported by Fucini et al. [17], defining the baseline as the

lowest marker value recorded after operation, any elevation of
one of the antigen levels greater than the limit defined by the
between-assay coefficient of variation (calculated on the basis of
two standard deviations) was defined as significant and the
assay was repeated after 10 days. If the marker rise was confirmed, an abdominal or total body CT scan, a chest x-ray
examination, a bone scan and clinical examination were performed. If recurrent neoplastic disease was confirmed by these
examinations the patient was evaluated for appropriate therapy;
otherwise he was reviewed at the next scheduled clinic visit. An
exploratory laparotomy was performed when all three markers
were elevated even if recurrence was not confirmed by total body
CT scan and clinical examinations. The follow-up schedule is
shown in Table 1. Data were analysed using the chi-square test.

CEA results

An increase in CEA levels in two consecutive serum

samples within 10 days was considered a positive
test and this occurred in 34 patients. Out of these,
21 had recurrence (positive predictive value 62%).
Eight patients with recurrence had normal CEA
levels (test sensitivity 72%). There were 13 false
positive CEA rises (test specificity 78%). In 18 patients (62%) who subsequently developed recurrent
disease the preoperative CEA level was raised.
Four (31%) of the 11 patients with a normal preoperative CEA did not develop a raised CEA when
recurrence occurred compared to 4 (22%) patients
of those with a high preoperative CEA (chi-square
0.68; NS). In 16 (55%) patients recurrence was heralded by an increase in CEA.

Results

No patient was lost to follow-up. Four patients

died during follow-up from non-neoplastic causes;
none had recurrence at autopsy.
Recurrent neoplastic disease was discovered in
29 patients (33%). Eight recurrences were local (all
from a rectal cancer treated with an anterior resection); 21 were metastatic, five of which also had
local recurrence. Recurrences were located in the
liver in 17 patients (7 from rectal cancer), in the
lung in 3 cases (2 from rectal cancer), in the peritoneum in 10 cases (3 from rectal cancer), in the

TPA results

An increase in TPA levels in two consecutive serum

samples was recorded in 25 patients. Of these 17
proved to have recurrent disease (positive predic-

Table 1. Follow-up schedule

Months

CEA; TPA; CA 19-9

Clinical examination
Abdominal scan
Chest x-ray
Barium enema
Colonoscopy
CT scan
Bone scan

12

15

18

21

24

30

33

36

39

42

48

54

60

+
+

+
+
+

+
+

+
+

+
+

+
+
+

+
+

+
+

+
+
+

+
+

+
+

+
+

+
+
+

+
+

+
+
+

+
+

+
+

+
+

+
+
+

+
+
+

+
+

+
+

+
+
+

+
+

+
+

+
+
+

232
tive value 68%). Twelve patients developed recurrence with normal TPA levels (test sensitivity 62%).
False positive TPA results occurred in eight patients (test specificity 86%). Twelve (75%) of 16
patients who developed recurrence had a raised
preoperative TPA, compared to 4 (31%) of 13 with
a normal preoperative TPA (chi-square 3.58; NS).
In 15 (52%) patients recurrence was heralded by an
increase in TPA.

cases) was the first sign of recurrence, and the diagnosis was established by clinical methods later. In
this group the median time for diagnosis of recurrence based on initial marker increase in comparison with routine clinical and instrumental followup was 2 months for liver metastases and 4 months
for disseminated metastases.

Discussion
CA 19-9 results

A CA 19-9 rise in two consecutive serum samples

occurred in 13 cases. Of these, 11 were shown to
have recurrence (positive predictive value 85%).
Eighteen patients who developed recurrence had
normal CA 19-9 levels (test sensitivity 38%). There
were two false positive CA19-9 results (test
specificity 97%). Five (71%) of seven patients with
raised preoperative CA 19-9 results developed recurrence compared to six (27%) of 22 with a normal preoperative CA 19-9 (chi-square 5.35;
P<_0.05). In eight (28%) patients the diagnosis of
recurrence was heralded by an increase in CA 19-9.

CEA, TPA, CA 19-9 and recurrences

Taking all three markers together, an increase in

the value of at least one of them, confirmed in two
serum samples, was recorded in 43 patients. Of
these, 28 were shown to have a recurrence (positive
predictive value 65%). Only one patient who presented with recurrence had no increase in any of the
three markers. This patient had a solitary brain
metastasis which was diagnosed when he developed
hemianopsia. In 15 patients without neoplastic recurrent disease there were 2 consecutive rises in one
of the 3 markers, but none of these patients showed
a simultaneous increase in all 3 markers. In two
cases a rise in the levels of all three markers was
seen as evidence of recurrent disease, and an exploratory laparotomy was performed. In one case
the operation revealed peritoneal carcinomatosis
while in the other a resectable subdiaphragmatic
hepatic recurrence was found.
Of eight recurrences in which CEA was normal,
five induced a rise in TPA value and two a rise in
the CA 19-9 level. These were in the majority of
cases (75%) with disseminated peritoneal metastases.
In 23 cases the rise in the value of CEA (16
cases) and/or TPA (15 cases), and/or CA 19-9 (8

Our results indicate that CEA is the most sensitive

marker available nowadays for the detection of
recurrent colorectal cancer. However, TPA and
CA 19-9 had significantly better specificity than
CEA. Furthermore, of eight patients with recurrence and normal CEA levels, five had a raised TPA
level, and two raised CA 19-9; the majority (75%)
of these recurrences were disseminated. Our results
are not completely consistent with those reported
by Fucini et al. [17] who found for CEA, TPA and
CA 19-9 sensitivities of 90%, 60% and 20%, and
specificities of 83%, 50% and 90%, respectively. In
our study TPA showed a better specificity (86%)
and sensitivity (62%), and CEA had the lowest
sensitivity (72%).
Some authors have suggested that periodic CT
scans lead to an increase in resectability rate for
colorectal recurrences. Deveney and Way [9],
adopting a follow-up which included a 6-months
interval CT scan, obtained a resectability rate of
26%. Sugarbaker et al. [18] followed 66 high risk
patients using review of symptoms and physical
examination, CEA assay, abdominal CT scan, full
lung tomography and liver-spleen scan every 3
months, and found that the serum CEA was positive in 81% of patients with recurrent disease, while
CT scan was abnormal in 17/33 (51%), with a sitespecific positive rate of 85% (17/20). Furthermore,
they found that a CEA rise was the first sign of
recurrence in 67% of patients, clinical examination
in 21%, and CT scan in 6%. In our study the
increase of one of the three markers was the first
sign of recurrence in 79% of patients, while clinical
examination, CT scan and chest x-ray were useful
in early diagnosis in three (10%), in two (7%), and
in one case (4%), respectively. Once recurrent disease was suspected on the basis of an increase in a
marker value, CT scan, full lung tomography, hepatic echotomography and bone scintigraphy were
helpful during work-up to localise recurrence.
In this study patients with low preoperative
TPA and CA 19-9 levels rarely developed a rise in
the marker values when recurrent disease occurred.

233

However, we believe that it is useful to monitor

patients with these markers because of the low cost
of the examination, and the increase in sensitivity
and specificity of the follow-up.
In our study, two patients developed a rise in all
three markers, leading to second-look laparotomy,
revealing one disseminated and one resectable recurrence. If CEA rise alone had been used as the
indicator for second-look surgery 13 unnecessary
laparotomies would have been performed. This
conforms with the experience of Fucini et al. [17]
who abandoned second looks in asymptomatic patients based solely on elevation of CEA after the
first five failures (four negative laparotomies and
one showing diffuse disease). However, Martin
et al. [4] found that of 146 asymptomatic patients
re-explored on the basis of a rise in CEA, 55% had
resectable recurrence. They performed seven (5%)
negative laparotomies, though six of these subsequently developed recurrent tumour. Minton et al.
[5] reported 43 patients who underwent secondlook surgery based solely on an elevated CEA level,
and found recurrences in 92%; they reported a
30% 5-year survival rate.
In conclusion, our study indicates that colorectal cancer follow-up based on the periodic evaluation of CEA, TPA and CA 19-9 may be useful in
early detection of recurrence and gives better results than CEA alone. Our data are not able to
demonstrate that second-look surgery based on rising marker levels improves resectability of recurrent disease or improves survival; a controlled
study is needed to answer these questions.

References
1. Wangensteen OH, Lewis FJ, Tongen LA (1951) The secondlook in cancer surgery. Lancet 71 (August):303-307
2. Wangensteen OH, Lewis FJ, Arthelger SW, Muller JJ,
Maclean LD (1954) An interim report upon the "second
look" procedure for cancer of the stomach, colon and rectum and for "limited intraperitoneal carcinosis". Surg Gynecol Obstet 99:257 267
3. Martin EW, Cooperman M, Carey TC, Minton JP (1980)
Sixty second look procedures indicated primarily by the rise
in serial carcino-embryonic antigen. J Surg Res 28:38 394
4. Martin EW, Minton JP, Care TC (1985) CEA-directed second-look surgery in the asymptomatic patients after primary resection of colorectal carcinoma. Ann Surg 202:310
317

5. Minton JP, Hoehn JL, Gerber DM, Horsley JS, Connolly

DP, Salwan F, Fletcher WS, Cruz AB Jr., Gatchell FG,
Oviedo M, Meyer KK, Leffall LD Jr., Berk RS, Stewart PA,
Kurucz SE (1985) Results of a 400-patients CEA secondlook colorectal cancer study. Cancer 55:1284-1290
6. Staab HJ, Anderer FH, Stumpf E, Hornung A, Fisher R,
Kleninger G (1985) Eighty-four potential second-look operations based on sequential CEA determinations and clinical
investigations in patients with recurrent gastrointestinal
cancer. Am J Surg 149:198-204
7. Attiyeh FF, Stearns MW (1981) Second-look operations
based on CEA elevations in colorectal cancer. Cancer
47:2119-2125
8. Beart RW, Metzger PP, O'Connell MJ, Schutt AJ (1981)
Postoperative screening of patients with carcinoma of the
colon. Dis Colon Rectum 24:585-589
9. Deveney KE, Way LW (1984) Follow-up of patients with
colorectal cancer. Am J Surg 148:717-722
10. Moertel CG, Schutt AJ, Go VL (1978) CEA test for recurrent colorectal carcinoma: inadequacy for early detection.
J A M A 239:1065-1066
11. Steele G, Zamcheck N, Wilson R (1980) Results of CEAinitiated second-look surgery for recurrent colorectal cancer. Am J Surg 139:544-548
12. Finlay IG, MacArdle CS (1983) Role of CEA in detection
of asymptomatic disseminated disease in colorectal carcinoma. Br Med J 286:1242-1244
13. Bjorklund B (1980) On the clinical use of TPA. Tumor
Diagnostik 1:9 20
14. Koprowski H, Herlyn M, Steplewski Z, Sears H F (1981)
Specific antigen in serum of patients with colon carcinoma.
Science 212:53 55
15. Sears HF, Herlyn M, Del Villano BC, Steplewski Z, Kapprowski H (1982) Monoclonal antibody detection of a circulating tumor-associated antigen. II. A longitudinal evaluation of patients with colorectal cancer. J Clin Immunol
2:141 149
16. Ritts RE, Del Villano BC, Go VL, Hebermann RB, Klug
TL, Zurawski VR (1984) Initial clinical evaluation of an
immunoradiometric assay for CA 19-9 using the NCI serum
bank. Int J Cancer 33:339-345
17. Fucini C, Tommasi SM, Rosi S, Maltantis G, Cardona G,
Panichi S, Bettini U (1987) Follow-up of colorectal cancer
resected for cure. Dis Colon Rectum 30:273-277
18. Sugarbaker PH, Gianola FJ, Dwyer A, Neuman N R (1987)
A simplified plan for follow-up of patients with colon and
rectal cancer supported by prospective studies of laboratory
and radiologic test results. Surgery 102:79-87

Accepted: 24 April 1989

Dr. P. Baritlari
I. Clinica Chirurgia
Universita "La Sapienza"
Viale del Poticlinico
1-00168 Rome
Italy

C oloJ'eetal
Disease

Int J Colorect Dis (1989) 4:234 243

9 Springer-Verlag 1989

Management of anal epidermoid carcinoma- an evaluation

of treatment results in two population-based series
S. Goldman 1, B. Glimelius 3, Ulla Glas 2, G. Lundell 5, L. P~hlman 4 and Elisabeth Stfihle 4
Departments of 1Surgery and 20ncology, S6dersjukhuset, Stockholm,
Departments of 30ncology and 4 Surgery, Akademiska sjukhuset, Uppsala and
Department of 50ncology, Karolinska sjukhuset, Stockholm, Sweden

Abstract. Between 1978 and 1984, two unselected

population-based groups of patients with anal epidermoid carcinoma were analysed: (1) a retrospective group (Stockholm region, 90 cases), where the
treatment varied considerably (partly radiation
therapy _+ chemotherapy _+ surgery, partly surgery
alone), and (2) a prospective group (Uppsala region, 51 cases) mainly treated by primary irradiation __ chemotherapy followed by surgery in some
cases. At diagnosis, 106 of the patients were free
from metastases. Two of these patients died before
treatment began. Of the remaining 104 patients,
77 received primary radiotherapy _+ chemotherapy,
44 to a dose of 3 0 - 4 0 Gy and 33 to a higher dose
level, 55-65 Gy. Radiotherapy was followed by
surgery in 28 cases. Twenty-seven patients were operated on primarily. The projected 5-year survival
rate was significantly higher in the Uppsala than in
the Stockholm region (all patients: 55% versus
43%; patients with no initial dissemination: 75%
versus 48%). The prognosis was better in patients
initially treated with radiotherapy than in those
initially treated with surgery. Long-term diseasefree survival was 88% in patients treated with radiation alone to an adequate (high) dose level. Multivariate analyses indicated that besides stage and
sex, initial treatment and region gave statistically
significant prognostic information. There was no
evidence that chemotherapy (Bleomycin) conferred
any additional benefit. It is concluded that the initial treatment in anal carcinoma should be radiotherapy (__ chemotherapy). In patients with no initial dissemination, this therapy seems to improve
5-year survival by 2 5 - 3 0 % compared with primary
surgery.

Introduction
Up to the last decade, surgery alone was the treatment of choice in patients with epidermoid carcinoma of the anus. For those patients (less than
1 0 - 1 5 % ) with a small, highly differentiated and
preferably distally located tumour, a local excision
was advocated, whereas for patients with a larger
tumour or a tumour higher up in the anal canal, an
abdomino-perineal resection was chosen [1, 2]. Radiotherapy was reserved for advanced or unresectable tumours.
During the past 10-15 years, a considerable
change of opinion in favour of non-surgical treatment has been evident in several centres. The results from these series indicate that radiation alone
or combined with chemotherapy can control local
disease and at the same time preserve anal function
[3-8]. On the basis of these observations, several
centres now recommend that surgery should be reserved for patients with residual or recurrent disease after radiation therapy or, in certain initially
large tumours, after preoperative radiotherapy.
However, the question as to whether primary
radiotherapy_+ chemotherapy alone or combined
with surgery has changed the prognosis has not
been settled. There are also several unanswered
questions concerning the way in which the non-surgical therapy should be performed. For example,
the most appropriate target volume, fractionation
scheme and final target dose are not yet fully established. Nor is it known whether it is of advantage
to combine radiation with chemotherapy.
In order to improve our knowledge about these
questions, two unselected population-based groups
of patients were analysed: (1) a retrospective group

235

(Stockholm region 1978-1984), where the treatment varied considerably (partly radiation therapy
chemotherapy surgery, partly surgery alone)
and (2) a prospective group (Uppsala region 19781984), where the treatment mainly consisted of primary irradiation chemotherapy, followed by surgery in certain patients.
Material and methods

Patients
In order to obtain unselected patients, data on all those with
diagnosed rectal and anal carcinoma (International Classification of Diseases, ICD 8, No. 154) who were residents of the
Swedish counties of Stockholm, Gotland, Uppsala, Kopparberg, Vfistmanland and Jfimtland between January 1978 and
December 1984 were gathered from: (1) the diagnostic files of all
surgical and oncotogical departments in the counties, and (2) the
Cancer Registry of the Stockholm and the Uppsala Health Care
Regions. In this way 141 patients with biopsy-proven primary
invasive epidermoid carcinoma of the anal canal were identified.
All microscopic material was re-evaluated by one experienced
pathologist [9]. No patient was lost to follow-up. The general
characteristics of the patients in the two population-based
groups are shown in Table 1.

1. Characteristics of the patients in the two populationbased series

Table

Stockholm
region

Uppsala
region

Total

Number of patients

90

51

141

Median age (range)

68 (42-93)

69 (34-92)

Female/male

67/23

42/9

Stage
T 1 2
T 3-4
T 1 4
T 1 4

NOM0
NOM0
N+M+
NXMX

32
38
18
2

14
22
13
2

46
60
31
4

Histological type
Squamous cell
Cloacogenic

33
57

17
34

50
91

Treatment groups"
(NOM0)
TI-2
RT+APR
5
"Low-dose" RT
10
"High-dose" RT
6
LE
4
APR
7
A P R + RT
No treatment
-

T3-4
18
9
3
7
1

TI-2
9
5
-

68 (34 93)
109/32

T3
5
1
11
2
2
1

4 T1-4
28
20
29
9
16
2
2

a For grouping and abbreviations, see text

Tumour staging
The tumour staging was based upon the tumour size, a tumour
less than 2 cm in diameter being designated T x, between 2 4 cm
T 2 and above 4 cm T 3. Tumours invading other organs except
the rectum and skin were designated T 4 irrespective of tumour
size. This classification follows in principle that proposed by
Papillon [5] for tumours in the anal canal. Lymph node involvement and metastatic disease were recorded as suggested by the
UICC [101.

Treatment." Stockholm region

During the period in question, the treatment varied considerably
in this region. In some hospitals the patients were operated upon
primarily, either by a local excision or by an abdomino-perineal
resection, and only occasionally were they sent for radiotherapy.
In other hospitals, all patients were given primary radiochemotherapy. No single protocol was followed. During the
first 4 5 years most patients received the radio-chemotherapy
as preoperative treatment to a dose of 30 to 40 Gy given by
either a single perineal portal or anterior/posterior portals, using
6~ or 6 - 1 6 MV photons from a linear accelerator. The daily
target dose varied, but the majority of the patients received 1.8
to 2.0 Gy 5 days a week. The radiation treatment was usually
combined with 5 mg Bleomycin i.m. either alone (daily or three
times a week; B), together with 0.5 mg Vincristine i.v. (twice a
week; BV) or together with 200 mg Cyclophosphamide and
250 mg 5-Fluorouracil i.v. (twice a week; BCF). These patients
were re-examined by the surgeon 3 - 4 weeks after completion of
the radiation therapy, and surgery was usually performed if any
palpable and/or biopsy-confirmed residual tumour was found.
If no palpable mass persisted after this "preoperative" dose,

most patients were considered disease-free by the surgeon and

no further treatment was given.
During the last 1 - 2 years, the policy was changed and
treatment in most patients was similar to that used in the Uppsala region (see below).

Treatment." Uppsala region

Patients referred to the Department of Oncology in Uppsala
were treated according to a standard protocol [6]. All radiation
therapy was given with an 8 or 16 MV linear accelerator. The
target volume included the primary tumour and all lymph nodes
in the lower dorsal part of the pelvic cavity. The inguinal and/or
the external iliac lymph nodes were included only if there was
primary perianal extension of the tumour and in cases with
known groin metastases. In the latter patients, treatment was
given by anterior/posterior portals, whereas the former patients
were treated in a prone position with one mid-dorsal and two
angled dorsal portals. The daily target dose was 2.0 Gy. After an
initial dose of 4 0 - 50 Gy ( _+Bleomycin), the radiation treatment
was individualised, using either the 3-portal technique or a single
perineal portal with electrons, depending upon the initial tumour burden. Chemotherapy (5 mg Bleomycin i.m.) was only
given to patients with carcinoma of the squamous cell type, and
was administered one hour before each of the first 15-18 radiation treatments, to a total dose of 7 5 - 9 0 mg.
The total target dose varied with the tumour stage. Patients
with a tumour in stage T 1 - T 2 or T~ were given 65 Gy (60 Gy
with Bleomycin) over 8 to 9 weeks with an interval of 2 - 3 weeks
after 40 Gy, corresponding to a cumulative radiation effect
(CRE) value of 17.5 - 18.0 (16.8 - 17.2 with Bleomycin). Patients
with a T 3 tumour received 5 0 - 5 5 Gy (40-45 Gy with

236
Bleomycin) followed by an interval of 3 weeks. If by then there
was no evidence of any residual tumour, the therapy was continued to 65 Gy (60 Gy + Bleomycin). If there was a residual tumour, an abdomino-perineal resection was performed. In patients with verified groin metastases, a groin dissection was carried out after the radiation treatment.
The patients not referred to Uppsala were treated either by
a local excision (mostly elderly patients) or by a standard abdomino-perineal resection.

1.0

".a o5
>.

o
t~_

Cumulative radiation effect

To compare the effects of different total radiation doses given by
different fractionation schemes and during different time periods, the CRE-value within the irradiated volume was calculated
as described by Kirk [11].

~05

Out of the 106 patients without initial metastases, two patients

died before treatment was begun. The remaining 104 patients
were assigned retrospectively to one of five treatment groups:

&

Statistical analysis
The survival curves were produced and the statistical significance tested (log rank test) according to the methods of Peto
et al. [12]. The Cox proportional hazards model was used for
multivariate analyses [I 3]. Differences between two proportions
were tested with X2 analysis. Unless otherwise indicated, the
statistical significance level used for testing the various hypotheses was p=0.05. The results are presented as cancer specific
which means that patients dying of intercurrent diseases are not
included in the population at risk after their death, provided
they were in complete clinical remission; if not, they were regarded as dying of anal carcinoma. All patients were followed up
until January 1988, which resulted in a median follow-up of
living patients of 75 months (range 30 119 months).

J4

3'6

12

24
36
/+8
Time [monthsl

#8

Time (rnonthsl

10

Treatment groups

1. Preoperative radiotherapy + cytostatics followed by an abdomino-perineal resection ( n = 2 8 , median age, 66 years)

(RT+APR).
2. Low-dose radiotherapy + cytostatics ( n = 2 0 , 71 years), no
surgery (low-dose RT, maximum CRE 14.9).
3. High-dose radiotherapy _+ cytostatics ( n - 2 9 , 62 years), no
surgery (high-dose RT, minimum CRE 15.0).
4. Local excision (n = 9, 78 years) (LE).
5. Abdomino-perineal resection _+ radiotherapy (n = 18, 69 years)
(APR + RT).

i2

6'0

Fig. 1. Probability of cancer specific survival in patients from

the Stockholm region ( o - - o ) and the Uppsala region (o--o).
The upper part illustrates the survival in the total group of
patients (Stockholm: n = 9 0 ; Uppsala: n = 5 1 ; log-rank test
p = 0.1) and the lower part the survival in patients without initial
metastases (Stockhohn: n = 7 0 , Uppsala: n = 36; p<0.01). The
figures in parenthesis indicate the number of patients at risk
after 60 months

initial dissemination: 75% versus 48%, p<0.01).

The proportion of patients with initially disseminated disease and those with large tumours (T3_4)
among those without initial dissemination was
somewhat higher in the Uppsala region, and the
proportion of men tended to be higher in Stockholm, but neither of the differences was statistically
significant. Other prognostic variables were evenly distributed between the two patient groups
(Table i).

Treatment results according to treatment group,

non-disseminated cases (Table 3)

Results
The treatment results for all patients without initial
metastases from the Stockholm and Uppsala regions are summarised in Table 2, and the cancer
specific survival for all patients are given in Fig. 1.
The projected 5-year survival rate was higher in the
Uppsala region than in the Stockholm region (all
patients: 55% versus 43%, p = 0.1; patients with no

1. R T + A P R . In this group of 28 patients, all but 4

received a dose of about 40 Gy (30-40 Gy), usually
(15 patients) with a perineal field only. The four
exceptional patients were given 50, 52, 55 and 68
Gy. Nineteen patients received Bleomycin (average
90 mg, range 35-115 mg), 5 VB and 3 BCF.
Five patients had small tumours (T 1 Tz). At
surgery, all specimens were free of viable cancer.

237
Table 2. Overall treatment results for patients with resectable carcinomas (T 1 - 4 NOM0) of the anal canal in the two populationbased series
Number

Alive and
well

Recurrences
Local
Inguinal
Distant
N u m b e r of patients (per cent)

Stockholm region
Uppsala region

69"
35"

29 (42)
23 (66)

16 (23)
4 (11)

10 (14)
5 (14)

11 (16)
2 (6)

Died of
intercurrent
disease

Died of
cancer

6 (9)
5 (14)

34 (49)
7 (20)

a One patient in each region not included, died before treatment

Table 3. Results of different treatment strategies for "Small" (T 1 2 NOM0) and "Large" (T 3 4 NOM0) carcinomas of the anal
canal
Treatment group

Number
cases

Alive and
well

Recurrences

Died of
intercurrent
disease

Local
Inguinal
Distant
Number of patients (per cent)
RT + A P R
T 1-2
T 3-4

5
23

5 (100)
9 (39)

3 (13)

.
6 (26)

"Low dose" RT
T 1-2
T 3-4

10
10

6 (60)
-

2 (20)
3 (30)

"High dose" RT
T1-2
T3-4

15
14

14 (93)
9 (64)

1 (11)
.

LE
T 1-2
T3-4
APR +_RT
T 1-2
T 3-4

7
11

4 (57)
4 (36)

.
3 (13)

11 (48)

2 (20)

1 (10)
5 (50)

2 (20)
-

2 (20)
10 (100)

6 (100)

1 (7)
1 (7)

1 (7)

1 (7)

1 (7)
3 (21)

2 (/4)

13 (93)
8 (89)

5 (56)

2 (22)
.

2 (22)

6 (67)

1 (100)

2 (29)
2 (18)

3 (27)

2 (29)
3 (27)

Alive
without
colostomy ~

3 (13)

Died of
cancer

.
3 (43)
7 (64)

a % of alive patients

All patients are alive with no evidence of disease

45-111 (median 62) months after treatment.
Large tumours (T3-T4) were found in 23 patients. In eight (35%) of these, including the four
who received more than 40 Gy, no residual carcinoma was found in the specimens. Nine (39%)
are alive and free from disease 48 to 89 (median 62)
months after treatment. Eleven patients died from
cancer, 3 of whom had local, 5 inguinal and 3 distant recurrences, 7 - 4 0 (median 12) months after
primary treatment. The remaining 3 patients died
with no signs of residual tumour after 5, 17 and 47
months.
2. Low-dose R T . In this group of 20 patients, the
majority (14 patients) were treated with a perineal
field with 6~
Fifteen of the patients received

Bleomycin to a dose of about 90 mg (40-150 mg),

4 VB and one BCF.
Six (60%) of the 10 patients with a small tum o u r ( T I - T 2 ) are alive, disease-free and with intact anal function, 36-106 (median 66) months after initial treatment. Three patients had a recurrence (two local and one distant) 4, 7 and 40
months after treatment. Two of these three patients
died from the disease, whereas the third was successfully retreated with an abdomino-perineal resection. One patient died of cardiovascular disease
2 months after the primary treatment without any
signs of residual tumour.
All 10 patients with large tumours (T3-T4) had
a recurrence 6 - 3 2 (median 12) months after treatment - 3 local and 2 inguinal recurrences and 5
distant metastases. All died from their disease.

238
3. High-dose RT. Twenty-nine patients were treated
to a dose of 5 5 - 6 5 Gy with an interval of about 3
weeks after 3 6 - 4 0 Gy. Most patients were treated
with either two or three portals. Bleomycin was
given to 19 patients to a dose of about 90 mg (range
40 155 mg), whereas 6 patients received no cytostatics. The remaining four patients received either
VB or BCF.
Of 15 patients with a small tumour (T~-T2), 14
(93%) are alive and disease-free and have been followed up for 30 to 106 (median 66) months. Only
one of them has had a local recurrence after 54
months successfully treated with an abdomino-perineal resection. Twenty months later, she has liver
metastases. One patient died from heart disease
after 6 months, with no evidence of residual disease
at autopsy. One of the patients had a palpable nodule one month after treatment, which was treated
by a limited excision. Microscopy revealed no residual carcinoma. This patient developed perineal
wound sepsis postoperatively, with secondary fibrosis and incontinence, and now has a colostomy.
Among the 14 patients with a large tumour, 9
(64%) are alive with no evidence of cancer after
2 8 - 8 9 (median 60) months. However, one of them
was found to have an inguinal node metastasis 28
months after treatment. After surgery she had further radiotherapy to the inguinal region, which had
not primarily been irradiated. She is disease-free 41
months later. In another 2 patients recurrence was
manifested after 23 and 58 months - one local and
one distant. Both patients died of their disease.
Three patients died without residual tumour after
16, 54 and 55 months.
4. Local excision ( L E ) . The patients of this group
were somewhat older (median 78 years) than those
of the other groups. All had small tumours (T~T2). Only one of the 9 patients is alive after 36
months, and she has intact anal function. One patient died after 13 months with no evidence of disease. Seven of the patients obtained a recurrence (5
local, 2 inguinal) after 7 - 3 0 (median 8) months, all
of them died ~1 54 (median 25) months after the
primary treatment.
5. A P R +_RT. Eighteen patients underwent A P R as
primary treatment. Two of them received postoperative radiotherapy, one to 40 Gy, and the other to
60 Gy.
Four (57%) of the 7 patients with a small tumour
(T1-T2) are alive and well after 4 2 - 6 6 (median 48) months. However, one of these patients
showed a local relapse after 3 months. She was then
given radiotherapy to a dose of 50 Gy and is now
disease-free 63 months later. The other 3 patients

displayed loco-regional recurrence 6, 27 and 36

months after treatment. They all died of their disease.
Four (36%) of 1 ] patients with a large tumour
are alive 3 6 - 8 6 (median 70) months after treatment. One of them received additional postoperative radiation to a dose of 60 Gy. One of the patients developed an inguinal recurrence, which was
succesfully removed. This patient is disease-free 73
months after recurrence. Seven patients died from
their disease, all of them with a recurrence (3 local,
one inguinal and 3 distant) after 6-41 (median 13)
months.

Treatment results, initially disseminated cases

(any T N+_ and/or M+_)
Thirty-five (25%) patients either presented with initial metastases, most of them with verified lymph
node metastases, or could not be properly assessed
regarding stage (Table 1). Sixteen (46%) of these
patients were treated to complete clinical remission
with primary radiation therapy (__ chemotherapy).
Four (25%) of the 16 patients are alive 3 0 - 4 2
months after treatment, but one of them has pulmonary metastases. One patient died of cardiovascular disease after 54 months with no signs of residual tumour and 11 died of cancer 2 - 3 9 (median 14)
months after treatment. Nineteen patients received
no or only palliative therapy because of advanced
disseminated disease and/or a poor general condition. They all died after 1-20 (median 2) months.

Treatment results in relation to radiation dose,

non-disseminated cases
There was a clear correlation between the radiation
dose and the outcome of treatment. This relationship concerned both patients treated with radiation
(_+chemotherapy) alone and those in which this
was followed by major surgery, as illustrated in
Fig. 2. For example, among patients treated with
radiation only, 4 (14%) of 29 cases with a dose
giving a CRE value above 15.0 relapsed, as against
5 (50%) of 10 in the group with a CRE value of
12.0-15.0, and 8 (80%) of 10 in the group with
C R E below 12.0. The differences are statistically
significant, e.g. above and below CRE 15.0,
p < 0.001. Among the 28 patients who had surgery
in addition, 5 (33%) of 15 with a CRE value above
12.0 had a recurrence, compared with 7 (54%) of 13
with CRE below 12.0. In the group of patients

239
20
160

18

.:

16
9 ~

&&
12- 9
10- ~II 9

Ao o

no

m
A

OO
O
O
OO

o9

100

OCOt~O

A 9

oo
0

00
O

Re
0

8
9

~0
/N/VX,

Time (months)

Fig. 2. Treatment results in 49 patients without initial metastases treated with radiotherapy _+ chemotherapy alone (left) and
in 28 patients treated with radiotherapy + chemotherapy followed by surgery (right), according to the radiation dose (CRE
value) and length of time after diagnosis. Alive with no evidence
of disease (o), dead with no evidence of disease (A), local recurrence (e), inguinal metastasis (m), and distant metastasis (A)

Survival according to treatment, non-disseminated

cases (Figs. 4 and 5; Tables 4 - 6 )

The retrospective classification of patients into

groups treated in different ways showed that cancer
specific survival was substantially better in the two
groups of patients who had received either highdose radiotherapy (5-year survival 88%) or radiotherapy followed by surgery (61%) than in the other groups (about 40%) (Fig. 4). Further, on classification of the patients into only two groups, primar-

12

I~

I; ' 'c~9~~
18
2~

Fig. 3. Treatment results in 40 patients without initial metastasis according to different doses of bleomycin and irradiation
(CRE-value) without additional surgery. Six of them received no
bleomycin. Alive with no evidence of disease (o), dead with no
evidence of disease (zx), and any recurrence (e)

1.0

E
,.~ 0.5

In order to investigate whether the additional cytostatic drug therapy influenced the treatment results,
the outcome for patients treated with radiation
without additional surgery according to different
doses of bleomycin was analysed (Fig. 3). As can be
seen in the figure, there was no evidence that
bleomycin had any beneficial effect. The number of
patients treated with BV or BCF was small, but
again there was no indication that this treatment
improved the treatment results (not illustrated).

10

ERE-vaLue

treated with primary surgery, the number of patients with recurrences was 19 (70%) out of 27.

Treatment results according to Bleomycin dose,

non-disseminated cases

&o

>,

(4~

'L

(4J

&

Time (months(
Fig. 4, Probability of cancer specific survival in patients without
initial metastasis treated according to different treatment
modal(ties: High-dose radiotherapy+chemotherapy (n--m,
n = 29), low-dose radiotherapy _+ chemotherapy ( e - - e , n = 20),
radiotherapy-t-chemotherapy followed by surgery ( o - - o ,
n=28), abdomino-perineal resection-t-radiotherapy (A--A,
n = 18) or local excision only (m--I, n = 9). Two patients who
died before therapy are excluded

ily non-surgical and primarily surgical treatment,

the results strongly favoured the primarily non-surgical approach (66% versus 38%, p<0.01, logrank test) (Fig. 5).
Since a number of factors other than the type of
treatment are of prognostic importance, univariate

240
Table 6. Stage, gender, region and radiation dose in anal car-

!t0t

cinoma treated with radiation ( _+chemotherapy) only - a multivariate analysis in 49 non-disseminated cases

s_._231

Estimation variable
Step Entering

Univariate
Z2

1.
2#
36
L~8
Time (months)

Z2

(values in the
final model)

{4)

12

Multivariate

2.

Radiation dose
(CRE>_/< 15.0)
Stage
(TI 2 v s T 3 4)
Gender
Region

14.6

14.8

p <0.001

9.9

20.5

p < 0.001

3.0
8.3

7.6
4.8

Fig. 5. Probability of cancer specific survival in patients without

initial metastasis primarily treated with radiotherapy +_chemotherapy (o o, n = 77) or primarily treated with surgery ( o - - o ,
n = 27)

3.
4.

Table 4. Stage, gender, region and initial treatment in primary

and multivariate analyses were performed in order

to investigate the relative importance of each factor. When all patients were considered, four factors
were found to be of statistically significant importance, viz. stage (T 1 2 better than T3_4), gender
(female better than male), region (Uppsala better
than Stockholm) and treatment (primary radiotherapy better than primary surgery) (Table 4). In a
multivariate analysis, all factors gave additional
prognostic information, although gender was favoured first (Table 4).
in the group of patients who received primary
radiotherapy, the radiation dose (CRE > / < 15) but
not the question as to whether surgery was performed or not gave statistically significant information. The radiation dose gave statistically significant information besides stage, gender and region,
although stage was included first (Table 5). In the
group of patients irradiated without any surgery,
the same factors were of independent prognostic
importance, although in this subset of patients, the
radiation dose was included first (Table 6).

anal carcinoma a multivariate analysis in 106 non-disseminated cases

Estimation variable
Step Entering

Univariate
){2

Multivariate
X2

(values in the
final model)
1.
2.
3.
4.

Gender
Region
(Uppsala vs Stockholm)
Stage
(T 1-2 vs T 3-4)
Treatment
(primary radiation vs
primary surgery)

8.7
7.8

7.7
7.8

p<0.01
p<0.01

8.4

11.0

p<0.001

5.1

5.3

p<0.05

p<0.01
p<0.05

Table 5. Stage, gender, region and treatment in primary anal

carcinoma treated with initial radiation - a multivariate analysis

in 77 non-disseminated cases
Estimation variable

Univariate

Multivariate

Step Entering

)~2

Z2
p
(values in the
final model)

1.
2.
3.
4.

16.0
13.6
9.4
9.3

18.9
10.7
17.9
4.2

Stage
Radiation dose
Gender
Region
(Uppsala vs Stockholm)

Not included, additional surgery, Z2= 1.3

p<0.001
p<0.01
p<0.001
p < 0.05

Discussion

It is evident from this study that patients treated

with different modalities have different prognoses.
For example, the proportion of patients who were
disease-free for a long period (tentatively cured)
was higher after a higher radiation dose (at least 50
Gy, C R E > 15.0) than after a lower one, even if the
radiation therapy was followed by surgery, or after
surgery alone. This difference in outcome is substantial and indicates that a primarily non-surgical
approach, i.e. radiotherapy +_ chemotherapy, improves the survival as compared with a primarily

241
surgical approach. The results also indicate that the
radiotherapy should be delivered at a sufficiently
high dose level.
It should be remembered that these conclusions
are not based upon a randomised comparison, a
comparison which in view of the rarity of the disease can be difficult to achieve. Also, the treatment
results after primary and optimal radiation appear
so much better (higher survival, more frequent anal
preservation) that a randomised comparison comparing primary surgery versus primary radiotherapy _+ chemotherapy is no longer justified. The
more frequent preservation of the anus after radiation therapy has been known for a long time [14,
15], but the improved survival has only been suggested from previous comparisons [4-8]. The present comparisons were made within two populationbased and thus entirely unselected patient groups.
In one of the regions, the Uppsala region, the treatment followed a standard protocol (high-dose radiation followed by surgery in selected cases)
throughout the period in question, whereas in the
other, the Stockholm region, this treatment protocol was only introduced in the last 2 years; before
that, no single protocol was followed. This allowed
several comparisons, all of which showed improved
survival after the primarily non-surgical approach.
Even if the above-mentioned treatment protocol
was not followed for all patients in the Uppsala
region and some patients in the Stockholm region
were treated in this way, the overall survival was
better in the Uppsala material. The figures indicate
that the difference in survival is substantial, the
5-year survival differing by 25 30%. The proportion of patients with disseminated disease was evenly distributed between the two groups, as also were
other variables of prognostic relevance.
The statistical analysis indicated that the initial
treatment was an important and independent prognostic factor. In addition, region was an independent prognostic factor. This latter finding could
imply that the mere fact that a prospective study is
going on (e.g. with a protocol to follow, and with
a special interest from certain investigators) is beneficial.
Although selection principles might explain
some of the marked differences in outcome between
the retrospectively collected treatment groups, such
an explanation appears unlikely. One of the more
important prognostic factors, clinical stage, was either evenly distributed between groups (e.g. highdose versus low-dose RT) or distributed in such a
way that it would decrease rather than increase the
differences (e.g. the local excision and primary
A P R groups had more favourable stages than the

group treated with radiotherapy followed by

surgery). The multivariate analysis confirmed these
assumptions.
The overall radiotherapy results in the present
study are not as good, however, as those reported
from other centres [6, 7, 16-23]. This difference
could be explained both by the fact that all patients
who had radiotherapy to any dose level were included in the analysis and by the fact that the cases
were unselected and population-based. It should be
emphasised that the 5-year survival in patients
without initial dissemination referred to Uppsala
exceeded 90% in spite of the fact that several of
these patients had very large tumours [6]. Thus,
even if primary radiotherapy is of great benefit to
patients with anal carcinoma, the most appropriate
protocol of irradiation and the additive effect of
chemotherapy has not yet been established. The
ideal would have been a randomised clinical trial.
In the absence of such trials, we believe, however,
that these unselected population-based groups,
even if partly of a retrospective nature, might serve
as a background to improve curative treatment
strategies for patients with anal carcinoma.
The addition of bleomycin to irradiation did
not, in this study, appear to be of advantage over
irradiation alone. Two randomised studies in
squamous cell carcinoma of the head and neck did
not provide such evidence [24, 25]. Therefore, in the
following discussion, all irradiated patients are
grouped together irrespective of whether they received chemotherapy or not.
The present results seem to indicate that a more
aggressive radiation dose is necessary to cure patients with an epidermoid carcinoma, irrespective
of the turnout size. Even if anal epidermoid carcinomas appear to be very radiosensitive, in comparison with other types of carcinoma, a dose
above 55 Gy (CRE 16.5) is required in order to
achieve prolonged local control with a high probability. For example, none of the patients with a
tumour in stage T 3 treated with a radiation dose of
about 40 Gy (CRE below 15.0) is alive after 3 years.
Nevertheless, in 6 out of 10 patients with a small
tumour, tumour control was achieved with such a
low dose. All these patients showed rapid and complete regression in response to the radiation therapy and may thus have represented a selected group
of patients with an "extremely" radiosensitive tumour. A further reason for concluding that these
patients were highly selected is the fact that viable
cancer was found in 15/19 specimens (79%) after a
preoperative dose giving a C R E of below 15.0.
On the other hand, reports from other centres
have shown that a radiation dose of about 40 Gy

242

given together with 5-fluorouracil (5-FU) and mitomycin-C (MTC) gives excellent treatment results
with high local control rates [7, 8, 21-23]. It has
been claimed from these studies that the cytostatics
potentiate the effects of irradiation and that lower
doses therefore can be given [3, 4, 7, 8]. Experimental studies have, however, shown extremely complex results and an improved therapeutic index has
not been convincingly demonstrated [26-28]. Another explanation for the excellent treatment results in studies using low irradiation doses (about
40 Gy) together with 5-FU and M T C is patient
selection with a predominance of small tumours [4,
7, 8, 21, 23].
An essential point to discuss is whether local
excision should be used in any but highly selected
patients, as there are considerable discrepancies between reported results, with a recurrence rate ranging from 0 to 75% [29-33]. According to a report
from the Mayo clinic, only one of 13 patients with
a small superficial tumour treated by local excision
had a recurrence [30]. This excellent result contrasts
with our own findings as well as those of others [29,
31, 33]. In our experience, it is difficult to evaluate
the question of tumour invasion into the internal
and external sphincter by digital examination
alone. Our results from local excision indicate that
the depth of tumour infiltration is often underestimated by the surgeon. If a local excision is used, a
more precise evaluation of tumour invasion is necessary. Preoperative ultrasound appears to be of
value for selecting those few very small superficially
located tumours [34].

Conclusions

Use of a primarily radiotherapeutic approach will

result in a larger number of long-term disease-free
survivors. In addition, more patients will retain
anal function. From the data reported in the present study, together with the excellent survival figures reported from hospital-based series, we conclude that primary surgery should no longer be the
initial treatment of epidermoid carcinoma of the
anus.
The most appropriate treatment schedule has
not yet been established. It seems that a dose of
about 5 5 - 6 0 G y (CRE value of 16.5) is high
enough to control about 85% of the tumours. The
possible benefit of bleomycin in addition to irradiation is questionable. Awaiting randomised clinical
trials, we hope that more well-controlled, preferably population-based series, using a consistent
and optimised technique, will be published.

References
1. Golden GT, Horsley JS (1976) Surgical management of
epidermoid carcinoma of the anus. Am J Surg 131:275-280
2. Beahrs OH, Wilson SM (1976) Carcinoma of the anus. Ann
Surg 184:422-428
3. Cummings BJ (1982) The place of radiation therapy in the
treatment of carcinoma of the anal canal. Cancer Treat Rev
9:125-147
4. Nigro ND (1984) An evaluation of combined therapy for
squamous cell cancer of the anal canal. Dis Colon Rectum
27:763 766
5. Papillon J, Mayer M, Montbarbon JF, Gerard J, Chassard
JL, Bailly C (1983) A new approach to the management of
epidermoid carcinoma of the anal canal. Cancer 51:18301837
6. Glimelius B, Pfihlman L (1987) Radiation therapy of anal
epidermoid carcinoma. Int J Radiat Oncol Biol Phys
13:305 312
7. Michaelson RA, Magill GB, Quan SHQ, Leaming RH,
Nikrui M, Sterns MW (1983) Preoperative treatment and
radiation therapy in the management of anal epidermoid
carcinoma. Cancer 51: 390- 395
8. Sischy B (1985) The use of radiation therapy combined with
chemotherapy in the management of squamous cell carcinoma of the anus and marginally resectable adenocarcinoma of the rectum. Int J Radiat Oncol Biol Phys
11:1587-1593
9. Goldman S, Glimelius B, PShlman L, St~hle E, Wilander E
(1988) Anal epidermoid carcinoma: a population-based
clinico-pathological study of 164 patients. Int J Colorect Dis
3:109 118
10. Spiessel B, Schiebe O, Wagner G (1982) U I C C - T N M Atlas.
Springer, Berlin Heidelberg New York
11. Kirk J, Gray WN, Watson ER (1971) Cumulative radiation
effect. Part I. Fractionated treatment regimes. Clin Radiol
22:145 155
12. Peto R, Pike MC, Armitage P, Breslow NE, Cox DR,
Howard SV, Mantel N, MacPherson K, Peto J, Smith PG
(1977) Design and analysis of randomized clinical trials requiring prolonged observations of each patient. II. Analysis
and examples. Br J Cancer 35:1-39
13. Cox D R (1972) Regression models and life tables. J R Stat
Soc B 26:103-110
14. Papillon J (1974) Radiation therapy in the management of
epidermoid carcinoma of the anal region. Dis Colon Rectum 17:184-187
15. Rotman M, Aziz H (1986) Carcinoma of the anus. Int J
Radiat Oncol Biol Phys 13:465-466
16. Adam YG, Efron E (1987) Current concepts and controversies concerning the etiology, pathogenesis, diagnosis and
treatment of malignant tumours of the anus. Surgery 101:
253-266
17. Cummings BJ (1987) Treatment of primary epidermoid carcinoma of the anal canal. Int J Colorect Dis 2:107-112
18. Nigro N (1987) Multidisciplinary management of cancer of
the anus. World J Surg 11:446 451
19. John M, Flare M, Lovalvo L, Mowry PA (1987) Feasibility
of nonsurgical definitive management of anal canal carcinoma. Int J Radiat Oncol Biol Phys 13:299-303
20. Papillon J, Montbaron JF (1987). Epidermoid carcinoma of
the anal canal. Dis Colon Rectum 30:324-333
21. Leichman L, Nigro N, Vaitkevicius VK, Considine B,
Burober T, Bradley G, Seydel HG, Olchowski S, Cummings
G, Leichman C, Baker L (1985) Cancer of the anal canal.
Am J Med 78:211-215

243
22. Shank B (1985) Treatment of anal carcinoma. Cancer
55:2156-2162
23. Meeker Jr WR, Sickle-Santanello BJ, Philpott G, Kenady
D, Bland KI, Hill GH, Popp MB (1986) Combined
chemotherapy, radiation and surgery for epithelial cancer of
the anal canal. Cancer 57:525-529
24. Vermund H, Kaalhus O, Winther F, Transj6 J, Thorud E,
Harang R (1985) Bleomycin and radiation therapy in
squamous cell carcinoma of the upper aero-digestive tract:
A Phase Ill Clinical trial. Int J Radiat Oncol Biol Phys
11:1877 1886
25. Eschwege F, Sancho-Gamier H, Gerard JP, Madelain M,
DeSaulty A, Jortay A, Cachin Y (1988) Ten-year results of
randomized trial comparing radiotherapy and concomitant
bleomycin to radiotherapy alone in epidermoid carcinomas
of the oropharynx: experience of the European Organization for Research and Treatment of Cancer. NCI Monogr
6:275 278
26. Byfield JE, Barone R, Mendelsohn J, Frankel S, Quinol L,
Sharp T, Seagren S (1980) Infusional 5-fluorouracil and X
ray therapy for non-resectable eosphageal cancer. Cancer
45:703 708
27. Rockwell S (1982) Cytotoxicities of mitomycin C and Xrays to aerobic and nypoxic cells in vitro. Int J Radiat Oncol
Biol Phys 8:1035 1039
28. Wu D-Z, Zhang Y-Q, Keng P, Sutherland RM, Lasagna L
(1985) The interaction between bleomycin and radiation on
cell survival and DNA damage in mammalian cell cultures.
Int J Radiat Oncol Biol Phys 11:2125-2131

29. A1-Jurf AS, Turnbull RB, Fazio VW (1979) Local treatment

of squamous cell carcinoma of the anus. Surg Gynecol Obstet 148:576-578
30. Boman BM, Moertel CG, O'Connell M J, Scott M, Weiland
LH, Beart RW, Gunderson LL, Spencer RJ (1984) Carcinoma of the anal canal: a clinical and pathologic study of
188 cases. Cancer 54:114-125
31. Lindkaer Jensen S, Hagen K, Harling H, Shokouh-Amiri
MH, Vagn Nielsen O (1988) Long-term prognosis after radical treatment for squamous-cell carcinoma of the anal
canal and anal margin. Dis Colon Rectum 31:273 278
32. Greenall MJ, Quan S, Stearns M, Urmacher C, DeCosse J
(1985) Epidermoid cancer of the anal margin. Am J Surg
149:95 100
33. Clark J, Petrelli N, Herrera L, Mittelman A (1986) Epidermoid carcinoma of the anal canal. Cancer 57:400-406
34. Goldman S, Glimelius B, Norming U, P~hlman L, Seligson
U (1988) Transanorectal ultrasonography in anal carcinoma. Acta Radiol 29: 337-341

Accepted: 8 May 1989

Dr. S. Goldman
Department of Surgery
S6dersjukhuset
S-100 64 Stockholm
Sweden

Announcemen t
3 - 6 October 1990 - Minneapolis]
Minnesota]USA

Principles of Colon and Rectal Surgery (Postgraduate Course)

For further information contact: Continuing Medical Education, University of Minnesota Medical School, Box 202 UMHC, 420 Delaware Street Southeast, Minneapolis,
MN 55455, USA. Phone: (612) 626-5525

243
22. Shank B (1985) Treatment of anal carcinoma. Cancer
55:2156-2162
23. Meeker Jr WR, Sickle-Santanello BJ, Philpott G, Kenady
D, Bland KI, Hill GH, Popp MB (1986) Combined
chemotherapy, radiation and surgery for epithelial cancer of
the anal canal. Cancer 57:525-529
24. Vermund H, Kaalhus O, Winther F, Transj6 J, Thorud E,
Harang R (1985) Bleomycin and radiation therapy in
squamous cell carcinoma of the upper aero-digestive tract:
A Phase Ill Clinical trial. Int J Radiat Oncol Biol Phys
11:1877 1886
25. Eschwege F, Sancho-Gamier H, Gerard JP, Madelain M,
DeSaulty A, Jortay A, Cachin Y (1988) Ten-year results of
randomized trial comparing radiotherapy and concomitant
bleomycin to radiotherapy alone in epidermoid carcinomas
of the oropharynx: experience of the European Organization for Research and Treatment of Cancer. NCI Monogr
6:275 278
26. Byfield JE, Barone R, Mendelsohn J, Frankel S, Quinol L,
Sharp T, Seagren S (1980) Infusional 5-fluorouracil and X
ray therapy for non-resectable eosphageal cancer. Cancer
45:703 708
27. Rockwell S (1982) Cytotoxicities of mitomycin C and Xrays to aerobic and nypoxic cells in vitro. Int J Radiat Oncol
Biol Phys 8:1035 1039
28. Wu D-Z, Zhang Y-Q, Keng P, Sutherland RM, Lasagna L
(1985) The interaction between bleomycin and radiation on
cell survival and DNA damage in mammalian cell cultures.
Int J Radiat Oncol Biol Phys 11:2125-2131

29. A1-Jurf AS, Turnbull RB, Fazio VW (1979) Local treatment

of squamous cell carcinoma of the anus. Surg Gynecol Obstet 148:576-578
30. Boman BM, Moertel CG, O'Connell M J, Scott M, Weiland
LH, Beart RW, Gunderson LL, Spencer RJ (1984) Carcinoma of the anal canal: a clinical and pathologic study of
188 cases. Cancer 54:114-125
31. Lindkaer Jensen S, Hagen K, Harling H, Shokouh-Amiri
MH, Vagn Nielsen O (1988) Long-term prognosis after radical treatment for squamous-cell carcinoma of the anal
canal and anal margin. Dis Colon Rectum 31:273 278
32. Greenall MJ, Quan S, Stearns M, Urmacher C, DeCosse J
(1985) Epidermoid cancer of the anal margin. Am J Surg
149:95 100
33. Clark J, Petrelli N, Herrera L, Mittelman A (1986) Epidermoid carcinoma of the anal canal. Cancer 57:400-406
34. Goldman S, Glimelius B, Norming U, P~hlman L, Seligson
U (1988) Transanorectal ultrasonography in anal carcinoma. Acta Radiol 29: 337-341

Accepted: 8 May 1989

Dr. S. Goldman
Department of Surgery
S6dersjukhuset
S-100 64 Stockholm
Sweden

Announcemen t
3 - 6 October 1990 - Minneapolis]
Minnesota]USA

Principles of Colon and Rectal Surgery (Postgraduate Course)

For further information contact: Continuing Medical Education, University of Minnesota Medical School, Box 202 UMHC, 420 Delaware Street Southeast, Minneapolis,
MN 55455, USA. Phone: (612) 626-5525

Col6i'ee/al
Disease

Int J Colorect Dis (1989) 4:244 246

9 Springer-Verlag 1989

Surgical repair of vulvar anus in adults

R. Rintala 1, p. Luukkonen 2 and H . J . Jiirvinen 2
1 Department of Pediatric Surgery and z Second Department of Surgery, Helsinki University Central Hospital, Helsinki, Finland

Abstract. Three adult women with previously unoperated vulvar anus underwent surgical treatment
for faecal incontinence which developed in adulthood. The obvious cause of incontinence in these
patients was the weakening of pelvic floor musculature by aging and pregnancy. Two of the patients
had several associated anomalies. The anus was
transposed to the normal perineal position through
the well-developed external sphincter found posterior to the vulvar anus. Faecal continence improved markedly in all patients.

Introduction

Vulvar anus is rare type of female anorectal malformation. The anorectum opens into the posterior
vestibular area. Unlike the more common recto- or
anovestibular fistulae, no treatment has been advocated for vulvar anus. Instead, it has been assumed
that the orifice has normal function [1] even though
the anus opens, in fact, outside the external sphincter on the anterior side, and the puborectalis and
levator muscles alone are responsible for voluntary
anal control. The exact description for this condition should be vulvar ectopic anal canal or ectopic
bowel outlet [2] instead of vulvar anus, used here
because of simplicity. We report the anatomical
surgical repair of vulvar anus in three adult patients
who gradually developed anal incontinence after
childhood. Two of these patients had associated
malformations compatible with the VATER association [3].
Case reports
Case 1
A 40-year old woman was admitted for investigations in 1986
because of total anal incontinence, which had developed gradu-

ally during the previous year. As a child she had undergone

repair of a cleft palate. She had had three pregnancies with
vaginal deliveries. A hysterectomy was performed at the age of
34 years for duplication of the uterus. Radiologicatly, congenital
anomalies were found in the lower lumbar vertebrae and dysmorphism was detected in the right kidney.
A typical vulvar anus was found on clinical examination.
There was an anal dimple with clearly demonstrable voluntary
sphincter activity about 3 cm posterior to the vulvar anus. On
manometric evaluation the anal pressure was found to be low
(15 cm H 2 0 ) and no voluntary squeeze was observed. The patient underwent anal transposition. At operation a strong circular external sphincter was identified with a neurostimulator at
the site of the anal dimple. The sphincter complex formed by the
external sphincter, levator ani and puborectalis muscles was also
well developed. The anal canal was transplanted through the
circular external sphincter, and the anterior wall of the voluntary sphincter funnel was reconstructed from the anterior tails of
puborectalis muscle (Fig. 1). Great care was taken to preserve
the distal bowel end and the internal sphincter during the dissection and transposition.
At follow-up (24 months) faecal continence had improved
significantly. The patient does not have to use any perineal
protective aids. The anal basal pressure is the same as preoperatively, but the patient can contract the sphincters voluntarily
up to 40 cm H20.

Case 2
This 37-year-old woman was treated at the age of 8 months in
a paediatric department because of a urinary tract infection and
slightly stenotic vulvar anus. The anus was dilated without any
difficulties at that time. It was noted that the anus was situated
just behind the vagina and there was virtually no perineal body.
There were also abnormalities in the thumb and hands. When
the patient was 10 years old she underwent left nephrectomy for
renal hypoplasia and an ectopic left ureter. The patient has had
two pregnancies with deliveries by caesarean section. For the
last 10 years the patient had experienced gradually worsening,
but still partial, urinary and faecal incontinence. She underwent
teflon injections in the bladder neck with some improvement in
urinary incontinence. During clinical examination in 1987 the
patient was found to have a patulous vulvar anus. External
sphincter contraction was observed a few centimetres posterior
to the anal orifice where a typical anal dimple was identified.
Measured manometrically, the basal anal canal pressure was

245
Table 1. Manometric findings before and after the rerouting
procedure (open perfused continuous pullthrough)
Case no.

MABP

MASP

HPZL

1. Before
6 months after
2. Before
6 months after
3. Before
2 months after

15
28
32
16
19
45

30
40
38
35
25
78

13
20
20
40
12
22

MABP: maximal anal basal pressure (cm/H20); MASP: maximal anal squeeze pressure (cm/H20); HPZL: length of anal high
pressure zone (mm)

""~, !.
i
9

.~.
:.

....

' .

:':,.

32 cm H20, and the sphincter squeeze was weak (max. pressure

38 cm H 2 0 ). The patient underwent anal transposition using a
technique similar to that used in case 1. In this case also there
was a well developed circular external sphincter. Postoperatively
the faecal soiling disappeared, and at the last follow-up, six
months after repair, the patient was actually slightly constipated
and was prescribed bulk laxatives. However, there was no improvement in basal pressure and maximal squeeze pressure.

Case 3

I.

A 33-year-old woman developed total anal incontinence after

her second vaginal delivery. During the delivery there occurred
a minor perineal tear, which was sutured. The first delivery, 5
years earlier, was aided by an episiotomy.
The patient came for proctological consultation a year after
the second delivery with a suspicion of an anal sphincter tear. A
typical vulvar anus was found. The external sphincter was situated posterior to the anal orifice where there was an anal dimple.
No other anomalies were detected. On anal manometry, both the
basal (19 cm H 2 0 ) and squeeze (25 cm H 2 0 ) pressures were
found to be low. After anal transposition through the external
sphincter complex, this patient also regained full continence
subjectively; within 2 months anal manometry showed clear
improvement.
The results of pre- and postoperative manometry in all cases
are summarised in Table 1. The lengthening of the anal high
pressure zone is shown in all cases.

Discussion

Fig. 1. A Distal bowel end mobilised and the external sphincter

ring identified and marked with tapes. B The rectum rerouted
through the external sphincter. C Anal skin sutured to the bowel
end and a rectal tube inserted; perineal body repair not completely finished. V: posterior wall of the vagina; E: anterior
border of the external anal sphincter; P: perineat body; R: distal
rectum

A n o r e c t a l m a l f o r m a t i o n s are detected neonatally

in the great majority o f cases. Surgical t r e a t m e n t is
necessary in m o s t patients to establish the flow o f
bowel contents. In females the m o s t c o m m o n
a n o m a l y is the anovestibular fistula, where the rectum opens into the posterior fourchette o f the
vestibule with a stenotic orifice or tract. Anal transposition is usually a d v o c a t e d for its t r e a t m e n t [4].
Vulvar anus is a rare a n o m a l y where the rectum
terminates in the vestibular area with a n o r m a l or
nearly n o r m a l orifice. Usually n o t r e a t m e n t is necessary during c h i l d h o o d because in most cases the

246

anus functions normally. Faecal continence is

maintained by the puborectalis part of the voluntary sphincter complex and the internal anal
sphincter [1]. An external sphincter does not contribute to faecal continence in these cases. However, as demonstrated in the patients described above,
the external sphincter apparatus is well developed
and situated in its normal perineal position.
Anal incontinence developed in our patients after childhood. The first patient had had three childbirths by vaginal delivery and the third had had
two. Vaginal delivery is a known risk factor for
anorectal incontinence [5]. In anorectal malformations the nerve supply of the voluntary sphincter
complex takes a more medial course than that in
normal subjects [6], which makes the nerve supply
even more vulnerable to stretching and damage. If
the puborectalis is the main voluntary factor in
faecal continence, even minor damage to its integrity may cause faecal incontinence. Patient two had
delivered her two children by caesarean section but
urinary and faecal continence had worsened in this
patient gradually during the last 10 years. The
weakening of the pelvic floor by aging and pregnancies is the probable cause for faecal incontinence in this case.
In our patients faecal continence improved significantly after anal transposition. Although in two
patients there was no major improvement in the
manometrically measured anal pressures, it is assumed that lengthening of the muscular anal canal
and normalisation of the anorectal angle improved
the barrier against faecal flow. Clinically, a voluntary external sphincter contraction was observed in
all patients postoperatively.
Similar adult cases of vulvar anus presenting
with incontinence seem to be rare. Katz et al. [7]
reported one case identified at the age of 25 years
and satisfactorily corrected at the age of 55 years.
In an inquiry for similar cases by these authors, 29
other cases from the United States were revealed.
Recently, five more adult patients undergoing anal
transpositions have been described, four of these
patients having undergone previous reconstructions of their anorectat malformations during
childhood [8, 9]. Because within 2 years we have
observed this anomaly in an incontinent woman

aged 62 years but not yet treated, in addition to the

three cases presented here, its prevalence may be
higher than has been previously reported. This kind
of anomaly is not generally known as a cause of
incontinence by general surgeons and may be easily
overlooked, as was the case in two of our patients,
who had undergone several previous examinations.
The present cases demonstrate that clinical improvement can be achieved by restoring the anorectal anatomy even during later years. The detailed
operative technique, used also in the present cases,
has been described recently by Keighley [9].

References
1. Stephens FD, Smith ED (1971) Ano-rectal malformations in
children. Year Book Medical Publishers, Chicago London,
pp 107-108
2. Penninckx F, Kerremans R (1986) Internal sphincter saving
in imperforate anus with or without fistula. A manometric
study, lnt J Colorect Dis 1:28-32
3. Rintala R, Lindahl H, Louhimo I (1986) VATER association
and anorectal malformations. Z Kinderchir 41:22-26
4. Tempteton JM, O'Neill JA (1986) Anorectal malformations.
In: Welch K J, Randolph JG, Ravitch MM, O'Neill JA, Rowe
MI (eds) Pediatric surgery, 4th edn. Year Book Medical Publishers, Chicago London, pp 1027-1029
5. Snooks SJ, Swash M, Henry MM, Setchell M (1986) Risk
factors in childbirth causing damage to the pelvic floor innervation. Int J Colorect Dis 1:20-24
6. Scott JES, Swenson O (1959) Imperforate anus, results in 63
cases and some anatomical considerations. Ann Surg 150:477
7. Katz LD, Zinkin LD, Stonesifer GL, Rosin JD (1978) Imperforate anus and ectopic orifices in adult patients. Dis Col
Rect 21:633-635
8. Pescatori M, Vulpio C, Castiglioni GC (1986) Congenital
anovulvar fistula in a 30-year-old woman. Br J Surg 73:161
9. Keighley MRB (1986) Re-routing procedures for ectopic
anus in the adult. Br J Surg 73:974-977

Accepted: 26 April 1989

Dr. R. Rintala
Department of Pediatric Surgery
Children's Hospital
Helsinki University Central Hospital
Stenbackinkatu 11
SF-00290 Helsinki
Finland

Col0i'eclal
Disease

lnt J Colorect Dis (1989) 4:247 250

9 Springer-Verlag 1989

Can the external anal sphincter be preserved in the treatment

of trans-sphincteric fistula-in-ano?
J. P. S. Thomson and A. H. McL. Ross *
St. Mark's Hospital, London, U K

Abstract. A method of preserving the external anal

sphincter in the treatment of complex trans-sphincteric fistula-in-ano is described. Cure without division of the external sphincter was possible in 44%
of cases. Disturbances of continence, common after
conventional fistula surgery, appeared to be reduced in those patients whose extenal sphincter remains intact.

Introduction

Although a number of methods of treatment of

trans-sphincteric fistula-in-ano have been described, laying open of the primary and secondary
tracks remains the standard management [1]. When
accompanied by the presence of circumferential
spread of infection the healing of fistula wounds is
prolonged [2]. Furthermore much external and internal anal sphincter muscle may need to be divided, resulting in varying degrees of clinical and
physiological anal sphincter insufficiency [3, 4]. In
general documentation of the functional results of
such surgery is scanty.
In the light of these considerations and previous
findings on the aetiology of anal sepsis and fistula
[5, 6], Parks described techniques for the management of complex trans- or supra-sphincteric fistula
aimed to preserve as much external anal sphincter
as possible [7]. This paper describes and assesses a
modification of Parks' method which aims to conserve the external sphincter completely whilst treating complex trans-sphincteric fistulae.

Materials and methods (male 29, female 5)

Thirty-four consecutive patients treated by a single surgeon between 1977 and 1984 have been evaluated retrospectively by
scrutiny of clinical records, questionnaire and where possible
clinical evaluation. All had a trans-sphincteric fistula-in-ano [8]
with the primary track passing through the external sphincter at
or above the level of the anal valves (Table 1). The relationship
of the primary track to the upper border of the puborectalis was
difficult to define in two patients and it is possible that they had
supra-sphincteric fistulae. Three additional patients were excluded from the series. One died at home of a myocardial infarction 3 months after her operation. The second refused to attend
for outpatient review and the third had records insufficient for
analysis.

Operative details
At operation the primary and secondary tracks were defined and
the latter laid open curetting away chronic granulation tissue.
The inter-sphincteric space was laid open and drained by internal sphincterotomy from the anal verge to the level of the primary track. The primary track through the external anal sphincter was loosely encircled by a monofilament nylon (0-gauge)
seton to act: first as a drain for the primary track, secondly to
stimulate fibrosis around it, and thirdly to act as a marker of the
track.
All wounds were dressed once or twice a day with lay-in
hypochlorite dressings and reviewed regularly under general

1. Level oftheprimary track, presence of secondary tracks,

and incidence of previous fistula surgery

Table

Group 1
Internal
opening

At dentate line
Above dentate line
Through puborectalis

Group 2

4/15 (27%) 4/19 (21%)

8/15 (53%) 12/19 (63%)
2/15 (13%) 2/19 (11%)

Secondary Supralevator extension 8/15 (53%) 11/19 (58%)

tracks
Horseshoe
10/15 (67%) 9/19 (47%)
Neither
3/15 (20%) 6/19 (32%)
*

Present address: Surgical Unit, Broomfield Hospital, Chelmsford, Essex CM1 5ET, U K

Previous surgery for fistula

2/15 (13%)

6/19 (32%)

248
anaesthesia when required. The seton was removed when adequate wound healing around it had occurred. The external
sphincter was only divided if the primary track failed to heal
after seton removal or because of persistent local sepsis around
the seton.
Twenty-five patients had circumferential or hemicircumferential spread of sepsis (19) or a supralevator extension (19).
Eight patients had previously been operated on for fistula and
22 for simple drainage of perianal sepsis prior to referral
(Table 1).
Patients were divided into two groups. Group 1 included
those in whom fistula healing occurred after removal of the
seton thus successfully preserving the external sphincter. Group
2 included those in whom the external sphincter was divided
following the failure of the seton technique to heal the fistula.
The Wilcoxon Rank sum test and the Chi-squared test with
Yates correction were used for statistical comparisons.

Results

Complete healing without the need for division of

the external sphincter after removal of the seton
occurred in 15 patients (Group 1). The remaining
19 patients required external sphincter division at
or after the time of seton removal (Group 2). Thus
the technique succeeded in curing the fistula whilst
preserving the sphincter in 44% of treated patients.

Duration of seton use

There was no difference in the time the seton was
left in situ between Group 1 and 2 (median and
range: Group 1, 38 days, 1-126; Group 2, 133
days, 7-394). The seton was left in situ for 6 weeks
or less in 16 patients (Group 1, 9 patients; Group 2,
7 patients).

Fig. 1. The postoperative appearance at 6 months in a patient in

G r o u p 1. Preservation of the external sphincter has allowed
healing to result in a normal looking anus and anal canal (white
arrow) with no distortion or guttering. The scar of the healed
circumferential secondary track is indicated by the black arrows

Table 2. Number of procedures under general anaesthetic during treatment

Group

(n)

Number of procedures
1

15

19

Median

2.7
3.2

Wound healing
Assessment of the time taken for complete wound
healing from the initial operation was possible in 31
patients (Group 1, 13; Group 2, 18). Healing was
complete in all patients (median and range: Group
1,153 days, 35-1232; Group 2, 212 days, 91-689).
Healing time was protracted in four patients in
Group 1 (though in three, assessment based on
clinic notes is probably excessive). There was no
significant difference in healing time between the
two groups. The normality of the anal appearance
when the seton method was successful in preserving
the external sphincter is seen in Fig. 1.

longed in many patients as a consequence of the

complexity of their fistulae and the need to prevent
premature closure of the laid-open tracks. There
was no significant difference between the two
groups.

Number of procedures requiring general anaesthesia

There was no significant difference between the two
groups in the number of procedures carried out
under general anaesthesia (Table 2).

In-patient stay

Follow-up

In-patient stay (median and range: Group 1, 30

days, 19-97; Group 2, 46 days, 14-91) was pro-

All patients were followed up until complete healing of the fistula wound had occurred. Thirty one

249
Table 3. Continence following fistula surgery

Group

Patients
assessed

Fully
continent

(,0

1
2

12
16

10 (83%)*
5(32%)*

Incontinent
Liquid
stool or
flatus

Solid Pad
stool use

Sphincter
repair

2
10

0
9

0
3

1
5

* p<O.05

patients were followed up after complete healing of

their wound had occurred by a mean (and range)
for Group 1 of 18.8 months, (1-75) and for Group
2 of 15.8 months, (1-47). Thirty patients answered
the questionnaire (mean and range: Group 1, 54.8
months, 20-159; Group 2, 38.6 months, 21-90).

Continence

Satisfactory assessment of continence was possible

in 28 patients (Table 3). A paraplegic patient in
Group 1 whose pre-operative incontinence was unaffected by surgery has not been included. Significantly more patients whose external sphincter was
preserved during treatment (Group 1) reported full
continence than among those whose external
sphincter had been divided. Conversely a greater
proportion of patients in Group 2 reported impaired control of flatus, liquid stool, or solid stool
and three in Group 2 have required an overlap
sphincter repair. Only one patient in Group 1 required regularly to use a perineal pad to control
soiling. Of 13 patients who reported defects in anal
control only 7 regarded this as abnormal.

Discussion

The primary consideration in the treatment of fistula-in-ano is the eradication of sepsis and any new
technique must not compromise this. The second
important object is to achieve this with the maxim u m preservation of anal function.
Conventional treatment of trans-sphincteric fistula-in-ano results in division of the external
sphincter to a variable degree [1]. It causes shortening of the anal canal and lowers anal sphincter
pressures, effects which may result in significant
functional disturbance [3]. When questioned
closely, up to 36% of all patients treated for fistula
will note a disturbance of anal continence with a
higher incidence if only patients with complex fistu-

lae are considered [9]. In view of this and in an

attempt to preserve external sphincter muscle, Parks
described a series of conservative procedures for
trans- and supra-sphincteric fistulae. These involved
partial division (V3 V2) of the external sphincter
below the track and the use of a loose seton to act
as a primary track drain as well as to promote
fibrosis around it [7]. Using this technique approximately 45% of his cases subsequently required external sphincter division to cure the fistula. The
majority of these fistulae were classified as suprasphincteric.
A variety of other methods aimed at preservation of the external sphincter have been described.
These range from coring out the fistula track, partial external sphincterotomy and fistulectomy [7],
to re-routing of the track into the intersphincteric
space [10]. We consider that coring out or re-routing the track would have been impossible or impractical in the fistulae treated in this series. Closure of the internal opening by mucosal flap
advancement after coring out the primary track has
been described with excellent results in 151 patients. Cure was achieved in 98% between 8 and 84
months. However, the fistulae treated were unclassified and the prevalence of horseshoe and supralevator extensions not recorded [I1].
Given that considerable differences exist between the patients and assessments in this and
Parks' series, in our selected group of patients with
complex fistulae, a smaller proportion have been
cured using the seton technique. In contrast, the
external sphincter was entirely preserved (in 44%)
when the method proved successful. The slightly
higher failure rate may relate to the suggestion that
premature healing of the postanal cavity is more
likely to occur when the distal external sphincter
muscle is not partly divided [7]. In addition, the
protracted healing times and number of surgical
procedures required in this series may relate to the
high incidence of horseshoe fistulae and supralevator extensions of the sepsis, confirming other authors' experience of the difficulties in treating such
complex fistulae [1, 2, 7, 9].
Accurate assessment of continence is difficult.
Up to 10% of a normal population will admit to
minor symptoms of anal leakage [9], though the
patient's reaction to these will be greatly influenced
by personality [12]. Co-existent disorders, particularly irritable bowel syndrome, may play an important contributory role. Comparison with other
studies is complicated by differing classifications of
fistulae and by poor data. Bennett found that a
high proportion (55%) of patients had difficulty
with anal control after treatment of horseshoe fis-

250
tulae [9], and a similarly high prevalence (31 - 3 3 %)
is quoted in other series in which complex fistulae
have been treated [4, 7]. Surprisingly, some authors
claim no problems with post-operative incontinence [13, 14]. In our study a high prevalence of
disturbed control of flatus and liquid stool occurred when the external sphincter was divided,
and in three such cases a sphincter repair was eventually required. In contrast, continence was the
norm when the external sphincter muscle was preserved and significantly more common than when
the external sphincter muscle had been divided. It is
of interest that only 7 of the 13 patients who reported defects in their anal control regarded these
as unsatisfactory, suggesting that a significant
number will accept such defects of control as a
reasonable price to pay for the cure of their troublesome fistula symptoms.
There has been much disagreement over the
function of the seton in fistula surgery. Parks and
others viewed it as a drain for the primary track as
well as a promoter of fibrosis which might subsequently prevent muscle retraction when the external sphincter was divided [15, 16]. Other authors
doubt whether the seton acts as more than a marker
of the primary track [1].Some use a tight seton to
cut slowly through the encircled muscle and thus
minimise muscle retraction [13, 17]. We view its
primary use as a drain with its secondary function
to mark the primary track. The incidence of disturbed control in the failed seton group (Group 2)
suggests that any stimulation of fibrosis that does
occur does not protect against incontinence when
the sphincter is divided. It also raises the possibility
that disordered control may, in addition to the consequences of muscle division, have a further aetiological component such as damage to the nerve
supply to the puborectalis and the external sphincter caused by surgical dissection or as a result of
sepsis [9]. Preliminary evidence from this hospital
suggests such damage does occur in a proportion of
these patients.
The seton method described offers the prospect
of cure for a significant proportion of patients with
complex trans-sphincteric fistulae, without division
of any external anal sphincter muscle. The cardinal
principle of eradication of sepsis is fulfilled. Healing time is prolonged particularly when the method

fails but this is so whatever method is employed.

Disorders of continence are less common when the
method succeeds. The method should be considered when treating complex fistulae and especially
when the patient is at particular risk of incontinence after division of the sphincters.

References
1. Hawley PR (1975) Anorectal fistula. Clin Gastroenterol
4:635-649
2. Thomson H R (1962) The orthodox conception offistula-inano and its treatment. Proc Roy Soc Med 55:754-756
3. Belliveau P, Thomson JPS, Parks AG (1983) Fistula-in-ano:
a manometric study. Dis Col Rect 26:152 154
4. Marks CG, Ritchie JK (1977) Anal fistulas at St Mark's
Hospital. Br J Surg 64:84-91
5. Parks AG (1961) Pathogenesis and treatment of fistula-inano. Br Med J i:463-459
6. Eisenhammer S (1956) The internal anal sphincter and the
anorectal abscess. Surg Gynaecol Obstet 103:501-506
7. Parks AG, Stitz RM (1976) The treatment of high fistula-inano. Dis Col Rect 19:487 499
8. Parks AG, Gordon PH, Hardcastle JD (1976) A classification of fistula-in-ano. Br J Surg 63:1 12
9. Bennett RC (1962) A review of the results of orthodox
treatment for anal fistulae. Proc Roy Soc Med 55:756 757
10. Mann CV, Clifton M (1985) Rerouting of the track for the
treatment of high anal and anorectal fistulae. Br J Surg
72:134 137
11. Vasilevsky CA, Gordon PH (1985) Results of treatment of
fistula-in-ano. Dis Col Rect 28:225-231
12. Hill JR (1967) Fistulae and fistulous abscesses in the
anorectal region. Personal experiences in management. Dis
Col Rect 10:421-434
13. Ewarth S, Ahlberg J, Collste G, Holmstrom B (1978) Fistula-in-ano. A six year follow-up study of 143 operated patients. Acta Chir Scand [Suppl] 482:53-55
14. Hanley PH, Ray JE, Pennington EE, Grabowsky OM
(1976) Fistula-in-ano: a ten year follow-up of horseshoe-abscess fistula-in-ano. Dis Col Rect 19:507-515
15. Gabriel WB (1963) The principles and practice of rectal
surgery, 5th edn. Thomas, Springfield
16. Kuypers HC (1984) Use of the seton in the treatment of
extrasphincteric anal fistula. Dis Col Rect 27:109 110
17. Hanley PH (1978) Rubber band seton in the management of
abscess-anal fistula. Ann Surg 187:435-437
Accepted: 4 July 1989
Mr. A. H. McL. Ross
Surgical Unit
Broomfield Hospital
Chelmsford
Essex CMI 5ET
UK

Col6i eeial
Disease

Int J Colorect Dis (1989) 4:251 254

9 Springer-Verlag 1989

A comparative study of polydioxanone (PDS |

and polyglactin 910 (Vicryl | in colonic anastomoses in rats *
E. Andersen 1, K. Sondenaa 1 and J. Holter 2
Departments of 1 Surgery and 2 Pathology, Rogaland County Central Hospital, Stavanger, Norway

Abstract. The use of polyglactin 910 (Vicryl | in

colonic anastomoses is theoretically undesirable
because its short dissolution time and multifilament structure may lead to local sepsis and anastomotic leakage. We have compared Vicryl with a
newly introduced monofilament absorbable suture
which has a longer dissolution time than Vicryl. In
a study of 98 rat colonic anastomoses no difference
was found in complication rates or cellular reaction
to the suture material between Vicryl and polydioxanone (PDS|
The use of monofilament suture
with longer dissolution time does not necessarily
imply added security for colonic anastomoses.

tains its strength longer than Vicryl [3, 4]. In this

research project a comparison was made with the
aim of revealing any disparity in the tissue reaction
to the sutures and possible differences in postoperative complication rates.
Material and methods
Forty-nine female Wistar rats (Mollegaard Breeding Center
Ltd., Denmark), average weight 198 g, were operated on under
aseptic conditions using general anaesthesia (fentanyl-fluanisone 0.1 ml/100 g subcutaneously) after being fed only tap
water during the previous 24 h. Through a midline incision both
the right and left colon were divided in turn and reanastomosed
using a single row of 12 14 interrupted, inverting, all-coats
sutures. In one half of the animals 6 - 0 Vicryl was used in the
right colon while 6 - 0 PDS was used in the left colon and con-

Introduction

The synthetic absorbable braided suture Vicryl has

been available for the past 15-20 years and has
wide application in general surgery. Due to the
rigidity of this polymer it may be used as a braided
suture only. Some authors [1, 2] have stressed the
disadvantage of using a braided suture with short
dissolution time in colonic anastomoses as the
rapid decline in tensile strength may jeopardise t h e
integrity of the anastomosis. Furthermore, the multifilament structure combined with bacterial contamination may cause micro abscesses and anastomotic leakage.
The use of a non-capillary suture has been advocated [1, 2]. PDS is an absorbable monofilament
suture with a slower dissolution time, which main* This study was supported by a grant from the Board of Research, Rogaland Central Hospital and has previously been
presented at the annual meeting of the Norwegian Surgical
Society in autumn 1987. An abstract was printed in the meeting catalogue (Vitenskapelige Forhandlinger)

t Segment
for
histological

evaluation

Fig. 1. Two end-to-end anastomoses were performed in each

animal, one in the ascending colon and one in the descending
colon

252

Fig. 2 a, b. The colonic anastomosis on the 28th postoperative day.

The suture materials are surrounded by some fibroblasts and polymorphonuclear leucocytes. The PDS
suture (a) has kept its integrity
while the Vicryl suture (b) shows
cellular infiltration (arrow) and
commencing disintegration. Hematoxylin-eosin stain; magnification
x 25
versely in the other half (Fig. 1). A standard surgical knot was
applied for both materials. The abdominal wall was closed with
a running suture of 4 0 polypropylene (Prolene | through all
layers. After the operation 5 ml of 0.9% saline was given subcutaneously. The animals were housed in pairs in plastic cages with
steel barred bottoms and were allowed tap water ad lib from the
operation and food pellets (Ewos-AB, Sweden) from the first
postoperative day.
The rats were sacrificed after 7, 14, 28 and 56 days using the
same anaesthesia and aorta sectioning. The anastomotic segments of the colon were excised, cut open, pinned to a cork
board, fixed and prepared for histological examination. From
each anastomosis three specimens were examined by a pathologist who was not told which suture had been used. The presence
of suture material was recorded and infiltration of fibroblasts,
macrophages and polymorphonuclear leucocytes around the suture was evaluated semiquantitatively as absent/few, abundant
or very abundant on a scale ranging from one to three.

Results
O n e a n i m a l d i e d o n the f o u r t h p o s t o p e r a t i v e d a y
after i n c i s i o n a l dehiscence a n d evisceration, a n d was
e x c l u d e d f r o m the project. C h r o n i c o b s t r u c t i o n d u e
to s t r i c t u r e o f the p r o x i m a l a n a s t o m o s i s (Vicryl)
was o b s e r v e d i n o n e a n i m a l a t sacrifice o n the 56th
day. A p a r t f r o m t h a t n o c o m p l i c a t i o n s were o b served, p a r t i c u l a r l y n o a b d o m i n a l or p e r i c o l i c a b scesses. S u t u r e m a t e r i a l c o u l d be verified i n 85 o f
the 96 a n a s t o m o s e s . T h e r e s u l t s o f the h i s t o l o g i c a l
e x a m i n a t i o n are p r e s e n t e d i n T a b l e 1.
M o s t o f the s p e c i m e n s s h o w e d m o d e r a t e l e u c o cyte a n d m a c r o p h a g e i n f i l t r a t i o n o n d a y 7 w h e r e a s
f i b r o b l a s t s were very a b u n d a n t a r o u n d 11 o f 12

253
Table 1. Histological reaction a to the suture material in 85 anastomoses

Days
postop,

Suture
material

Leucocytes

Macrophages

Fibroblasts

1 2 3

1 2 3

123

Vicryl
PDS

1
2

10
10

1
0

0
0

11
10

1
2

0
0

1
6

11
6

14

Vicryl
PDS

4
8

8
4

0
0

1
3

10
9

1
0

0
1

6
8

6
3

28

Vicryl
PDS (11)

9
9

3
2

0
0

6
8

5
2

1
1

2
4

9
6

1
1

56

Vicryl (6) b

6b 0
7
1

0
0

5b 1
6
2

0
0

3
4

38

0
1

PDS

(8)

a Semiquantitative classification: 1 = absent/few; 2 = abundant;

3 = very a b u n d a n t
b One animal with stricture of the proximal anastomosis

Vicryl suture specimens compared with 6 of 12 PDS

sutures. Up to day 28 the inflammatory cell reaction decreased gradually, as did the number of fibroblasts. At the 28th day disintegration of the
Vicryl suture had started with fragmentation and
cellular infiltration among the strands (Fig. 2),
while PDS retained its integrity throughout the observation period. On the 56th day only a slight
cellular reaction was observed around the sutures.

Discussion

Braided sutures evoke a stronger inflammatory reaction than monofilament suture in contaminated
tissue [5]. This phenomenon is related to the capillarity of the suture, its capacity for fluid absorption
and its ability to shield bacteria within its structure
[6, 7]. Theoretically a braided suture placed in contaminated tissue may contribute to suture abscesses
giving rise to anastomotic failure. Paterson-Brown
et al. did not, however, find more infectious complications with multifilament suture compared with
monofilament when tested in wounds contaminated with Escherichia coli and Bacteroides fragilis
[8]. Durdey and Bucknall found in their study [2]
that absorbable sutures, with the exception of PDS,
lose strength rapidly and produce prolonged tissue
reaction; they concluded that Vicryl is unsuitable
when used alone in colonic anastomoses. They recommended a monofilament suture like PDS. Lord
et al. [9] showed that braided sutures cause more
damage to submucosal collagen than the smooth
surface of monofilament sutures. In the experimental work of Houdant and coworkers, PDS produced only slight tissue reaction in the colon [10].

The same claim for PDS was made in a review of

colonic anastomoses [1].
In the study presented, more fibroblast reaction
was observed around Vicryl during the first week.
This may indicate increased tissue trauma caused
by the rougher surface of Vicryl but the difference
did not reach statistical significance (Fischer exact
probability being 0.069). However, no differences
in the overall foreign body inflammatory reaction
and no micro abscesses were observed associated
with either material (Table 1). Therefore we conclude that in our study the braided suture did not
cause significantly more tissue reaction than the
monofilament suture. We recorded only one long
term complication, namely a stricture in the right
colon associated with the use of Vicryl. This may
well have been caused by a technical error. The
rapid dissolution of Vicryl did not cause any suture
line breakdown, confirming the view of Fontaine
and Dudley [11] who pointed out that suture
durability is of less importance in securing an anastomosis. This is also supported by clinical data [12].
Experimental data indicate that qualities other
than the physical configuration of the suture are
important. The chemical composition may play an
essential part in the development of early infection
[13, 14]. There are also indications that glycolic
acid, the degradation product of polyglactin 910,
may act as an antibacterial agent [13, 15].
Our results do not justify any claim that Vicryl
is unsafe in colon anastomoses. N o advantage of
PDS was revealed under the present circumstances.
References
1. Khoury GA, Waxman BP (1983) Large bowel anastomosis.
The healing process and sutured anastomosis. A review. Br
J Surg 70:61 63
2. Durdey P, Bucknall TE (1984)Assessment of sutures for use
in colonic surgery, an experimental study. J R Soc Med
77:472-477
3. Ray JA, Doddi N, Regula D, Williams JA, Melveger A
(1981) Polydioxanone (PDS), a novel monofilament synthetic absorbable suture. Surg Gynecol Obstet 153:497- 507
4. Craig PH, Williams JA, Davis KW, Magoun AD, Levy A J,
Bogdansky S, Jones JP (1975) A biologic comparison of
polyglactin 910 and polyglycolic acid synthetic absorbable
sutures. Surgery 141:1-10
5. Blomstedt B, Osterberg B (1978) Suture materials and
wound infection. Acta Chir Scand 144:269-274
6. Blomstedt B, Osterberg B (1977) Fluid absorption and capillarity of suture materials. Acta Chir Scand 143:197-204
7. Osterberg B, Blomstedt B (1979) Effect of suture materials
on bacterial survival in infected wounds. Acta Chir Scand
145:431 434
8. Paterson-Brown S, Cheslyn-Curtis S, Biglin L Dye J, Easmon CSF, Dudley H A F (1987) Suture materials in contaminated wounds: a detailed comparison of a new suture with
those currently in use. Br J Surg 74:734-735

254
9. Lord MG, Broughton AC, Williams HTG (1978) A morphologic study on the effect of suturing the submucosa of
the large intestine. Surg Gynecol Obstet 146:211 216
10. Houdart R, Lavergne A, Valleur P, Hautefeuille P (1986)
Polydioxanone in digestive surgery. An experimental study.
Am J Surg 152:268-271
11. Fontaine C J, Dudley H A F (1978) Assessment of suture materials for intestinal use by an extramucosal implant technique and a quantitative histological evaluation. Br J Surg
65:288-290
12. Blomstedt B, Jacobsson SJ (1977) Experiences with
polyglactin 910 (Vicryl) in general surgery. Acta Chir Scand
143:259-263
13. Edlich RF, Panek PH, Rodeheaven GT, Turnbull VG,
Kuntz LD, Edgerton MT (1973) Physical and chemical configuration of sutures in the development of surgical infection. Ann Surg 177:679 687

14. Ching-Chang Chu, Williams D F (1984) Effects of physical

configuration and chemical structure of suture materials on
bacterial adhesion. A possible link to wound infection. Am
J Surg 147:197-204
15. McGeehan D, Hunt D, Chaudhuri A, Rutter P (1980) An
experimental study of the relationship between synergistic
wound sepsis and suture materials. Br J Surg 67:636-638
Accepted: 29 March 1989

Dr. Egil Andersen

Department of Surgery
Rogaland Central Hospital
N-4011 Stavanger
Norway

Book review
D. Beyer, U. M6dder: Diagnostic Imaging of the Acute Abdomen.
Berlin, Heidelberg, New York: Springer 1988. XV, 453 pp, 250
figures, containing 680 separate ills., DM 180.00, Hardcover,
ISBN 3-540-17520-2.
"Diagnostic imaging of the acute abdomen: a clinico-radiologic approach" is written by Dr. Dieter Beyer from Cologne
and Dr. Ulrich M6dder from Dfisseldorf with contributions by
other German colleagues. This is an English translation, the
original German edition was also published in 1988.
The authors state that their aim is to outline the applications
of imaging techniques in patients with an acute abdomen. It is
designed as a quick reference to show the radiological and imaging techniques that are currently in use. The main emphasis is on
plain radiographs. The roles of ultrasound and computer tomography (CT) are described. The use of gastrointestinal contrast studies, angiography and radionuclides is also included.
Acute disorders of the gastrointestinal tract, hepatobiliary system, pancreas, spleen, renal tract, peritoneal cavity, retroperitoneum and pelvis are described. The text is short and contains
useful lists of differential diagnoses. The book is comprehensively illustrated and it is well referenced, many of the references
are from the German literature.
The book achieves the main aim of the authors. The text is
informative and the illustrations are of a high standard. I would
disagree however with the authors on a number of technical
aspects. In the section on plain radiographic technique, I fully

agree that the upright abdominal view is of little value and the
supine view should be supplemented by the left lateral decubitus
view instead. However, an upright chest postero-anterior (PA)
radiograph should be routinely taken in patients being investigated for acute abdominal disorders; there is no reference in the
text and no illustration of the role of the upright chest in the
detection of pneumoperitoneum. Most authorities, including
those referenced by the authors, emphasise the importance of
upright chest views. The authors advocate the use of water-soluble contrast medium in the investigation of upper gastrointestinal and small intestinal obstruction, whereas most centres now
consider barium to be a superior contrast medium. The radiological appearances of most acute gastrointestinal disorders are
described in detail but closed-loop obstruction received little
attention and the relationship of appendicoliths to a high incidence of appendicitis complications is not mentioned. The figure
legends would be less confusing if each figure referred to one
patient.
The book has translated well with only the occasional
spelling mistake.
Apart from these minor criticisms this is an excellent book.
Radiologists, and particularly those in training, will find it to be
most informative, while clinicians who deal with acute abdominal emergencies will find it a useful source of reference.
D. J. Nolan (Oxford)

254
9. Lord MG, Broughton AC, Williams HTG (1978) A morphologic study on the effect of suturing the submucosa of
the large intestine. Surg Gynecol Obstet 146:211 216
10. Houdart R, Lavergne A, Valleur P, Hautefeuille P (1986)
Polydioxanone in digestive surgery. An experimental study.
Am J Surg 152:268-271
11. Fontaine C J, Dudley H A F (1978) Assessment of suture materials for intestinal use by an extramucosal implant technique and a quantitative histological evaluation. Br J Surg
65:288-290
12. Blomstedt B, Jacobsson SJ (1977) Experiences with
polyglactin 910 (Vicryl) in general surgery. Acta Chir Scand
143:259-263
13. Edlich RF, Panek PH, Rodeheaven GT, Turnbull VG,
Kuntz LD, Edgerton MT (1973) Physical and chemical configuration of sutures in the development of surgical infection. Ann Surg 177:679 687

14. Ching-Chang Chu, Williams D F (1984) Effects of physical

configuration and chemical structure of suture materials on
bacterial adhesion. A possible link to wound infection. Am
J Surg 147:197-204
15. McGeehan D, Hunt D, Chaudhuri A, Rutter P (1980) An
experimental study of the relationship between synergistic
wound sepsis and suture materials. Br J Surg 67:636-638
Accepted: 29 March 1989

Dr. Egil Andersen

Department of Surgery
Rogaland Central Hospital
N-4011 Stavanger
Norway

Book review
D. Beyer, U. M6dder: Diagnostic Imaging of the Acute Abdomen.
Berlin, Heidelberg, New York: Springer 1988. XV, 453 pp, 250
figures, containing 680 separate ills., DM 180.00, Hardcover,
ISBN 3-540-17520-2.
"Diagnostic imaging of the acute abdomen: a clinico-radiologic approach" is written by Dr. Dieter Beyer from Cologne
and Dr. Ulrich M6dder from Dfisseldorf with contributions by
other German colleagues. This is an English translation, the
original German edition was also published in 1988.
The authors state that their aim is to outline the applications
of imaging techniques in patients with an acute abdomen. It is
designed as a quick reference to show the radiological and imaging techniques that are currently in use. The main emphasis is on
plain radiographs. The roles of ultrasound and computer tomography (CT) are described. The use of gastrointestinal contrast studies, angiography and radionuclides is also included.
Acute disorders of the gastrointestinal tract, hepatobiliary system, pancreas, spleen, renal tract, peritoneal cavity, retroperitoneum and pelvis are described. The text is short and contains
useful lists of differential diagnoses. The book is comprehensively illustrated and it is well referenced, many of the references
are from the German literature.
The book achieves the main aim of the authors. The text is
informative and the illustrations are of a high standard. I would
disagree however with the authors on a number of technical
aspects. In the section on plain radiographic technique, I fully

agree that the upright abdominal view is of little value and the
supine view should be supplemented by the left lateral decubitus
view instead. However, an upright chest postero-anterior (PA)
radiograph should be routinely taken in patients being investigated for acute abdominal disorders; there is no reference in the
text and no illustration of the role of the upright chest in the
detection of pneumoperitoneum. Most authorities, including
those referenced by the authors, emphasise the importance of
upright chest views. The authors advocate the use of water-soluble contrast medium in the investigation of upper gastrointestinal and small intestinal obstruction, whereas most centres now
consider barium to be a superior contrast medium. The radiological appearances of most acute gastrointestinal disorders are
described in detail but closed-loop obstruction received little
attention and the relationship of appendicoliths to a high incidence of appendicitis complications is not mentioned. The figure
legends would be less confusing if each figure referred to one
patient.
The book has translated well with only the occasional
spelling mistake.
Apart from these minor criticisms this is an excellent book.
Radiologists, and particularly those in training, will find it to be
most informative, while clinicians who deal with acute abdominal emergencies will find it a useful source of reference.
D. J. Nolan (Oxford)

Int J Colorect Dis (1989) 4:255

The patient, a 39-year-old female, was referred

from abroad with a two-year history of recurrent,
quite severe, right sided abdominal pain and vomiting. She also suffered from constipation and required to take laxatives. During the two years she
had lost 10 kilograms in weight. Clinical examination and laboratory tests were normal. Two barium
enemas, performed before the patient was referred,
were reported to be normal. A further barium enema (Fig. 1) showed a dilated splenic flexure with
two short strictures (arrows), one in the transverse
colon and the other in the descending colon. There
was also spasm of the sigmoid colon during the
examination. Colonoscopy was subsequently performed and the two areas of narrowing were confirmed but no mucosal abnormality could be identified. It was considered likely that the strictures
were causing the patient's episodes of intermittent
obstruction.

-uoganalsqo jo sapos!da
lUa.t.mao.t gmsne3 aanla!~ls Ims!p aaom oql Rq
pasnea s
se~ oanxa U a!ualds aql jo no!l
-elN!p a q l "aanxou o!uolds aql jo so!l!tuallx0 aql le
paleao I sa~nla91s o,m aaaA~0.I0ql ase3 s!ql uI "mma~3
-qas! Jo apos!da sno~Ao.ld olq~3s!u~oao.t e Jo aauos
-qe oql u~. luosold ol aanxau 3!ualds 0ql jo uotffaa
oql u! so.mlo.uls o.~tuoeqos! Ioj iensnun lou s! 1I
opos!da atmae
-qas! paz!u~oaaiun Xlsno!aaad e Jo llnsaa 0ql uooq

OAeq O1 potunso.~d SeA~ ~UtA~O:I.~eU poleIaosse pue

s!so/qu oR& "puo qaeo le s.~so-Nu poz!IeaOl jo sluotu
-~os pmeposse o,~1 ql!Ar uoumN!p paz,lJ'em po~oqs
uomtaods pmaasoJ oql jo uo!leu!mexo Iea!~oloqled
9s!sotuolseue 3:~etm.ld e ql.tAr pmaosaI o.lo~ 'sluotu
-ffos po~ao:ueu o,m oql ffu!pnpui 'o~nxo U a!uolds
jo sa.nam.nua3 ,~luoAsL "UOlOajo luom~as ~U!UOA:tm
-u.t oql jo uo!l~ltq!p qlt.~ solnlo.t~ls oql s'e uo.tl~oo[
otues oql le punoj oaa~ ~u!uozla!ql Iesoantu jo se~.~e
o~al pue pomJojJod s ~ s
V :aotasug

Acknowledgement
Since its inception, the Journal has undergone several developments due largely to changes in the
numbers and type of submissions, and the Editorial
Board and Editors feel that this should be reflected
in the format of the Journal. One of the consequences of this is that the Quiz will no longer appear as a regular feature. The Editors would there-

fore like to express their appreciation to those who

have contributed to the interest of this feature during the last four years:
Volumes 1 to 3: A. Price, London and J. Virjee,
Bristol
Volume 4:
D . J . Nolan, Oxford and
I. C. Talbot, London.

Col6rec/al
Disease

Int J Colorect Dis (1989) 4:256

9 Springer-Verlag 1989

Acknowledgement to Referees
The Editors wish to express their thanks to the referees of papers published in this volume.
Their time and trouble is much appreciated.

S. Alexander, London
S. Arnott, London
D. C. C. Bartolo, Bristol
R. H. L. Begent, London
J. Beynon, Bristol
J. Christiansen, Copenhagen
Z. Cohen, Toronto
J. P. Cruse, London
M. F. Dixon, Leeds
P. Durdey, London
M. S. Elliot, Cape Town
W. G. Everett, Cambridge
L. P. Fielding, Waterbury, Conn.
P. Finan, Leeds
D. Galloway, Glasgow
S. M. Goldberg, Minneapolis, Minn.
P. H. Gordon, Montreal
B. D. Hancock, Manchester

M. M. Henry, London
P. Hermanek, Erlangen
U. Hildebrandt, Homburg
G. L. Hill, Auckland
M. A. Kamm, London
M. Kaufman, London
M. R. B. Keighley, Birmingham
M. G. W. Kettlewell, Oxford
R. H. S. Lane, Winchester
F. Lazorthes, Toulouse
D. J. Leaper, Bristol
R. Leicester, Gosport
D. Z. Lubowski, Sydney
J.-P. Mach, Epalenges
D. McGibbon, London
R. S. McLeod, Toronto
M.-C. Marti, Geneva
S. Nivatvongs, Rochester

B. Nordlinger, Paris
T. Oresland, G6teborg
J. Papillon, Lyons
J. H. Pemberton, Rochester, Minn.
F. Penninckx, Leuven
R. K. S. Phillips, London
A. V. Pollock, Scarborough
A. Price, London
N. W. Read, Sheffield
J.-C. Sarles, Marseilles
A. Sitges-Creus, Barcelona
A. N. Smith, Edinburgh
E. St/ihle, Uppsala
M. Swash, London
W. H. F. Thomson, Gloucester
R. C. N. Williamson, London