J Clin Pathol 1987;40:601-607

Restorative proctocolectomy with ileal reservoir:pathological and histochemical study of mucosalbiopsy specimens

N A SHEPHERD,*§ J R JASS,*t§ I DUVAL,T R L MOSKOWITZ,t R J NICHOLLS,ttB C MORSON*

From the Departments of*Pathology, tSurgery, and the tImperial Cancer Research Fund Colorectal CancerUnit, St Mark's Hospital, London, and the §Department ofHistopathology, St Bartholomew's Hospital, London

SUMMARY Mucosal biopsy specimens from the ileal reservoirs of 92 patients who had undergonerestorative proctocolectomy (12 with familial adenomatous polyposis, 78 with ulcerative colitis, andtwo with functional bowel disease) were studied. Chronic inflammation was found in almost all, aswas villous atrophy of varying severity. Other changes included pyloric metaplasia and mucosalprolapse. Acute inflammatory changes and ulceration were less common but, when present, corre-sponded to the clinical condition of "pouchitis". A grading system was devised to score acute andchronic inflammatory changes. There was a significant increase in acute inflammatory scores inulcerative colitis compared with those in familial adenomatous polyposis, and pouchitis was presentonly in patients who had had ulcerative colitis; the morphological features of pouchitis are similarto those seen in the colorectal mucosa in ulcerative colitis. Histochemical studies of mucin in thereservoirs of mucosa showed that there may have been a change from small intestinal mucin tocolonic mucin.

In severe diffuse mucosal disease of the colon and rec-tum the entire large bowel may have to be removed.Several operations have been tried to preservesphincter function and restore continuity of thebowel. Of these, restorative proctocolectomy withileal reservoir is the most recent.' The operation com-prises total proctocolectomy, construction of a reser-voir from the terminal ileum, and ileoanal anas-tomosis. The colon and about three quarters of therectum are removed, and the mucosa stripped fromthe remaining rectal stump down to the dentate line.The reservoir of small intestine is then brought downthrough the denuded muscle tube and anastomosedto the anal canal at the dentate line. The length ofrectal cuff left was about 8 cm in the early cases, but isnow about 4cm in some of the more recent cases.Various reservoir designs have been described.2 3 Theprocedure is covered by a temporary ileostomy that isusually closed about two months later, provided thatall suture lines have healed.

Diffuse mucosal disease, in the form of ulcerative

Accepted for publication 29 January 1987

colitis or familial adenomatous polyposis, accountedfor all but two of the 160 patients who have had proc-tocolectomy with ileal reservoir at St Mark's Hospi-tal. Inflammation often occurs in the reservoir, and ina recent publication the incidence and severity ofinflammation in 92 of these patients were describedand the role of various physiological and pathologicalfactors discussed.4 In this paper we ptesent anaccount of the histopathological changes occurring inthe mucosa of reservoirs, with particular reference tofeatures seen on histochemical staining for mucins.

Material and methods

Mucosal biopsy specimens from the reservoirs of 92patients who had undergone both stages of the oper-ation were studied. In 78 cases the original indicationwas ulcerative colitis, and in 12, familial adenomatouspolyposis. Two operations were performed for func-tion bowel disease, one a case of Hirschsprung's dis-ease and the other a young adult with an ileostomywho had had necrotising enterocolitis as a child.Biopsy specimens were taken through a sigmoido-scope from each patient at intervals of six months to

601

on May 20, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.6.601 on 1 June 1987. Dow

nloaded from

Shepherd, Jass, Duval, Moskowitz, Nicholls, Morson

-i..

Fig 1 Biopsy ofnormal small intestinal mucosafromreservoir.

Fig 2 Mucosal biopsy specimenfrom reservoir showingpartial villous atropby and diffuse chronic inflammation oflamina propria. (Haematoxylin and eosin.)

A- '1

Fig 3 Mucosal biopsy specimen from reservoir showing pouchitis, with subtotal villous atrophy, pronounced crypt hyperplasia,diffuse chronic inflammation oflamina propria and pyloric metaplasia. There is active inflammation ofsuperficial epithelium, andcrypt abscess is shown (arrowed). (Haematoxylin and eosin.)

602

-: . -.:, 1-7 - "..- -11- - -.

.1 - t., " .

:., -. - -.,e -, -. q- ,.,

on May 20, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.6.601 on 1 June 1987. Dow

nloaded from

Restorative proc tocolectomy with ileal reservoirF * s~~~~~~~~~~~~~~.PFig4 Mucosal biopsy specimenfrom reservoir with subtotalvillous atrophy and pronounced crypt hyperplasia withoutnoticeable inflammation. (Haematoxylin and eosin.)

Fig 5 Mucosal biopsy specimenfrom reservoir with diffusechronic inflammation, subtotal villous atrophy, crypthyperplasia and crypt distortion. (Haematoxylin and eosin.)

603

a year after closure of the ileostomy and were fixed in10% buffered formol saline. The biopsy specimenswere processed routinely in paraffin wax and stainedwith haematoxylin and eosin.A subgroup of 32 patients, who had had at least

three sequential mucosal biopsies of the reservoir,were further studied. In 25, paraffin embedded blocksof ileum, which had been obtained at the time of theoriginal proctocolectomy, were available. In additionto haematoxylin and eosin, high iron diamine-alcianblue (HID-AB) and periodate borohydride-potas-sium hydroxide-periodic acid Schiff (PB-KOH-PAS)stains were also performed. A modification of themethod of Spicer,5 with ferric chloride at 60% con-centration was used for HID-AB staining. All sec-tions were cut and stained by the same person usingthe same technique. Each slide was graded by oneobserver (NAS) for blue predominance, a mixedpattern, or brown predominance. Assessment byPB-KOH-PAS was less satisfactory as normal smallintestinal mucosa usually stains well, but a grading ofnone, mild, moderate or strong staining was used.

Results

Normal small intestinal mucosa was not usuallyfound in mucosal biopsy specimens from the reser-voirs; most showed chronic inflammatory changes. Inonly 12 (13%) did the first mucosal biopsy specimenafter ileostomy closure lack chronic inflammatorychanges and villous architectural abnormality (fig 1).The chronic inflammatory cell infiltrate in the laminapropria included all cell types associated withchronic inflammation-lymphocytes, plasma cells,eosinophils and histiocytes-although often eosino-philic polymorphs were the first to accumulate. Thechronic inflammatory changes were usually associ-ated with some degree of villous atrophy (fig 2), whichvaried from mild and partial to almost total villousatrophy, common in cases with acute inflammation(fig 3). The degree of villous atrophy varied both inthe sequential biopsy specimens and within the samespecimen. When considerable acute inflammationwas present, however, it was usually associated withextensive subtotal villous atrophy. Villous atrophywas always accompanied by pronounced crypt hyper-plasia, producing an appearance similar to that ofcolonic mucosa (fig 4), and when this was associatedwith extensive chronic inflammation, the histologicalappearances were reminiscent of chronic ulcerativecolitis in remission (fig 5).

In mucosal biopsy specimens from reservoirs whichshowed subtotal villous atrophy and crypt hyper-plasia the histological appearances were like those of

on May 20, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.6.601 on 1 June 1987. Dow

nloaded from

604

untreated coeliac disease (fig 2). There was, however,a notable difference in lymphocyte counts in the epi-thelium between biopsy specimens of small intestinalmucosa in coeliac disease and those of reservoirmucosa. Though intraepithelial lymphocytes are aprominent feature of untreated coeliac disease andare important in the assessment of response to agluten free diet, the intraepithelial lymphocyte countin mucosal biopsy specimens from reservoirs wasalways less than 5% of epithelial cells. The lympho-cyte count remained low even in the presence of activeinflammation, although numerous polymorphs wereusually present in the surface epithelium.

Other chronic changes seen in mucosal biopsyspecimens from reservoirs were pyloric metaplasiaand the histological features of mucosal prolapse.Pyloric metaplasia is a non-specific change thatoccurs in both small and large intestinal mucosa whenit is affected by chronic inflammatory bowel disease;in these cases it was a rare feature of activeinflammation in the pouch. It was seen in biopsyspecimens in three patients in this study, all of whomhad previously had acute inflammation (fig 3). In twoother patients there was clinical evidence of mucosalprolapse within the reservoir. Histological exam-ination of the mucosal biopsy specimens showedfibromuscular obliteration of the laminia propria,

5X'-'-'''

F j;> *do

Fig 6 Mucosal biopsy specimenfrom reservoir in patientwith clinical mucosal prolapse showing villous atrophy,fibromuscular obliteration oflamina propria, and cryptepithelial hyperplasia. (Haematoxylin and eosin.)

Shepherd, Jass, Duval, Moskowitz, Nicholls, Morsonvillous atrophy, and pronounced crypt basalhyperplasia (fig 6). Similar features are describedin mucosal prolapse of the rectum, the solitary ulcersyndrome.6

Acute inflammatory changes and ulceration wereseen less often. A patchy, superficial, acuteinflammatory cell infiltrate in the epithelium waspresent in at least one biopsy specimen from a func-tioning reservoir in 50 cases of ulcerative colitis (64%)and in six from cases of familial adenomatous poly-posis (50%). More extensive acute inflammatory cellinfiltration, with crypt abscesses and ulceration, wasseen in six cases (7%), all of whom had had ulcerativecolitis (fig 3). These histological changes, which wereaccompanied by pronounced chronic inflammationwith villous atrophy, were always associated withpouchitis, manifest by systemic symptoms, the pas-sage of at least six liquid motions per day, and macro-scopic inflammation on sigmoidoscopy.4 Extensiveactive inflammation was associated with reactive epi-thelial hyperplasia (fig 3) in these cases in which ulcer-ative colitis had been diagnosed previously: there wasno evidence of dysplasia in the reservoir. Two mildlydysplastic tubular adenomas were found in themucosa in two cases of familial adenomatous poly-posis.Using the HID-AB technique we showed a consid-

erable change from blue staining mucin of small intes-tinal goblet cell type to brown staining mucin ofcolorectal type in 16 of the 32 patients (fig 7). Thesepatients comprised 13 who had had ulcerative colitisand three with familial adenomatous polyposis. Theywere all associated with histological evidence ofchronic inflammation and villous atrophy. In three ofthe six with clinical pouchitis there was a change inthe reservoir goblet cell mucin to a predominantlybrown reacting product with HID-AB. With the PB-KOH-PAS staining sequence, there was a definiteincrease in PAS staining between ileal control and res-ervoir mucosa in only six patients, all of whom hadhad ulcerative colitis.The mucosa of the terminal ileum at the time of the

original proctocolectomy was histologically normalin 24 of the 25 patients for whom tissue was available.There was no evidence of villous atrophy, crypthyperplasia, or inflammation. In only one case, a 35year old man with ulcerative colitis, was there clin-ically important inflammation in the terminal ilealmucosa at the time of proctocolectomy. The HID-ABtechnique showed a predominantly brown reaction inthe goblet cells of the portion of terminal ileumincluded in the proctocolectomy specimen. The firstmucosal biopsy specimen of the reservoir showed apredominantly blue reaction, but later specimensshowed a reversion to a brown staining mucin of gob-let cell type, indicating an increased sulphate content.

on May 20, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.6.601 on 1 June 1987. Dow

nloaded from

Restorative proctocolectomy with dleal reservoirCt SW~~ ~~~~~~~~~~~~~~~~~~j:~

it. t~~~~~~~~~~~~~~~W

F-IgL 7 MAut/-I hitiowhemiszrvl of terminal i/ta! mucusa andreservoir muwosati trou Paientli ituft/ Prevlonis lcrtiveWt't colitis.a MiAttosea of tern Iinatl i/cantil re tovedat ptrocotoculeetoulv

S/t I-Itin a on/vI greyI-b/ae .taunnitg I IunctIn: ;1?nncosaf /'iojpstfrntrSerto0h(ir,one yea after closure, of ilcoistomit s/hottingtisLedl grevi-/Jbluc anid black -1-rottn'l stat.if.iiit suipliated muclhdein

iC ) lIlttco./tI biop)sy re(Serv,1o1ir specimienfriiut11 tWO tearlS aftertlSUVai ofi IosONtM I x

Discussion

Little attention has been paid to the histopathological

changes that occur in the small intestinal mucosa in

reservoirs after restorative proctocolectomy. As the

operation becomes more common pathologists will be

increasingly confronted by mucosal biopsy specimensfrom the reservoirs, and the findings of this prospec-

tive study may help them in their assessments. We

have shown that histological changes, which are

almost always present, are predominantly chronic,

and include both inflammation and villous atrophy.In cases with extensive acute and chronic

inflammatory changes there is such pronounced vil-

lous atrophy and crypt hyperplasia and distortion,

that the mucosa resembles that of the large intestine

in chronic ulcerative colitis. In our study only one of

25 cases showed histological abnormality in the ter-

minal ileal mucosa at the time of proctocolectomy,and therefore the changes seen in the reservoir

mucosa seem to be the direct result of creation of the

reservoir.

In a previous study, in which clinical and sig-

moidoscopic features were correlated with histologyof the reservoir mucosa, we introduced a histological

grading system for the assessment of severity of

inflammation and this is reproduced in the table.4

Scores for acute inflammation correlated with the

degree of macroscopic inflammation seen on sig-

moidoscopy, and with the frequency of defaecation,

and were lower in those patients with familial ade-

nomatous polyposis compared with those who had

ulcerative colitis. There was, however, no significantdifference in the score for chronic inflammation in

patients with ulcerative colitis, compared with those

with familial adenomatous polyposis.4 A histological

605

At

-S..

S1

C

on May 20, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.6.601 on 1 June 1987. Dow

nloaded from

606

Table Scoring system for histopathological changes inreservoir muscosa4

Histological feature Score

AcutePolymorph infiltration:None 0Mild and patchy infiltrate in the surface epithelium IModerate with crypt abscesses 2Severe with crypt abscesses 3

Ulceration:None 0Mild superficial IModerate 2Extensive 3

ChronicChronic inflammatory cell infiltration:None 0Mild and patchy IModerate 2Severe 3

Villous atrophy:None 0Minor abnormality of villous architecture IPartial villous atrophy 2Subtotal villous atrophy 3

grading for acute changes of four or more indicatedactive pouchitis, and correlated with the clinical crite-ria of this condition.4We have seen pouchitis only in those reservoirs cre-

ated for ulcerative colitis, but as these comprise over

80% of the total it is difficult to draw definite conclu-sions about the apparent absence of pouchitis fromthose reservoirs created for familial adenomatouspolyposis. More recent figures in our series show 10patients with pouchitis, all of whom had had ulcer-ative colitis, out of a total of 152 cases (126 ulcerativecolitis, 24 familial adenomatous polyposis, two mis-cellaneous). Furthermore, we have shown that thereis considerably more acute inflammatory change inthose reservoirs constructed from ulcerative colitis.4In the Kock continent ileostomy, where an ileal reser-voir is constructed with an abdominal stoma, definiteacute inflammation is also seen only in cases of ulcer-ative colitis.7 Current evidence suggests, therefore,that severe acute inflammation, including pouchitis, isa disease confined to reservoirs constructed for ulcer-ative colitis and that it does not occur in ileal reser-voirs constructed for familial adenomatous polyposis.

Although we have seen hyperplastic epithelialchanges in cases with active inflammation, we havenot seen dysplasia in the mucosa of those reservoirsconstructed for ulcerative colitis. None of the reser-

voirs, however, has been in situ for more than 10years. Two mildly dysplastic tubular adenomas were

found in two reservoirs constructed for familial ade-nomatous polyposis. The apparent colonic meta-plasia in the reservoir mucosa may lead to an

Shepherd, Jass, Duval, Moskowitz, Nicholls, Morson

increased propensity for adenomas to develop in thereservoir, and there is a case report of multiple ade-nomas (about 30) occurring in a Kock continent ile-ostomy.8 Adenomas may, however, also be seen inthe otherwise normal small intestinal mucosa infamilial adenomatous polyposis.9

Histochemical studies using mucin staining wereperformed on a subgroup of 32 patients, 26 withulcerative colitis and six with familial adenomatouspolyposis. The goblet cells of the terminal ileum stainmainly blue with HID-AB."0 This indicates the pres-ence of sialic acid with little or no sulphate. We havefound that there is some sulphomucin present in thenormal small intestine, especially in the epitheliumoverlying the lymphoid aggregates of Peyer's patches,and this emphasises the importance of using controltissues from the same patient. Our histochemicalstudies using mucin staining have shown a definitechange from small intestinal mucin to colorectalmucin rich in sulphate in 50% of the cases. Thechange was seen in patients with both ulcerative col-itis and familial adenomatous polyposis. This mucinchange seems to be independent of the original diag-nosis and probably occurs as a result of chronic,rather than acute, inflammatory change. It is, how-ever, known to be unstable. A change from highlysulphated colonic mucin to less sulphated small intes-tinal mucin (brown to blue with HID-AB) occurs incolonic mucosa in acute ulcerative colitis and in dys-plasia in ulcerative colitis.' In some of our casesthere may have been a change to colonic mucin, butthat the onset of active inflammation reversed thisprocess in a similar way to that seen in colonicmucosa in active ulcerative colitis. This may haveaccounted for the lack of obvious change in stainingin three of the cases with clinical and histological evi-dence of pouchitis.

Previous studies have shown that the onset of acuteinflammatory changes in the reservoir mucosa doesnot correlate with the type of reservoir, its emptyingcharacteristics, or its volume.4 12 This suggests thatstasis alone is not a factor in the development ofpouchitis. Recent bacteriological studies have notfound a consistent correlation between aerobic oranaerobic bacterial counts and inflammation, 13although one study showed some correlation betweenconcentrations of aerobic organisms andinflammation, 4 and another study showed asignificant inverse correlation between volatile fattyacids (which reflect anaerobic bacterial activity) andvillous atrophy in the reservoir mucosa. 15 Thisfinding suggests that high counts of anaerobic bacte-ria may protect the reservoir mucosa. Some patientswith pouchitis do, however, benefit frommetronidazole'3 which contradicts this observation.The creation of a neorectum, which becomes fur-

on May 20, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.6.601 on 1 June 1987. Dow

nloaded from

Restorative proctocolectomy with ileal reservoir 607

ther modified by a process of colonic metaplasia, maycreate an environment that favours the developmentof a condition similar to ulcerative colitis in the ilealreservoir. This hypothesis is supported by mor-phological changes in the pouch, which bear a strik-ing resemblance to ulcerative colitis, and the fact thatan inflammatory process does occur in small intes-tinal mucosa in ulcerative colitis. Currently there is noevidence to suggest that bacteriological or physicalfactors are implicated in the aetiology of pouchitis.

We thank Ms Jill Maybee for photographic help.

References

l Parks AG, Nicholls RJ. Proctocolectomy without ileostomy forulcerative colitis. Br Med J 1978;2:85-8.

2 Utsunomiya J, Iwama T, Imago M, etal. Total colectomy,mucosal proctectomy and ileoanal anastomosis. Dis ColonRectum 1980;23:459-66.

3 Nicholls RJ, Pezim ME. Restorative proctocolectomy with ilealreservoir for ulcerative colitis and familial adenomatous poly-posis: a comparison of three reservoir designs. Br J Surg1985;72:470-4.

4 Moskowitz RL, Shepherd NA, Nicholls RJ. An assessment ofinflammation in the reservoir after restorative proctocolectomywith ileoanal ileal reservoir. International Journal of ColorectalDisease 1986;1: 167-74.

5 Spicer SS. The use of various cationic reagents in histochemicaldifferentiation of mucopolysaccharides. Am J Clin Pathol1961;5:393-407.

6 Madigan MR, Morson BC. Solitary ulcer of the rectum. Gut1969;10:871-81.

7 Hulten L, Svaninger G. Facts about the Kock continent ile-ostomy. Dis Colon Rectum 1984;27:553-7.

8 Beart RW Jr, Fleming CR, Banks PM. Tubulovillous adenomasin a continent ileostomy after proctocolectomy for familialpolyposis. Dig Dis Sci 1982;27:553-6.

9 Hamilton SR, Bussey HJR, Mendelsohn G, et al. Ileal adenomasafter colectomy in nine patients with adenomatous polyposiscoli/Gardner's syndrome. Gastroenterology 1979;77:1252-7.

10 Filipe Ml. Mucins in the human gastrointestinal tract: a review.Invest Cell Pathol 1979;2:195-216.

11 Jass JR, England J, Miller K. Value of mucin histochemistry infollow up surveillance of patients with long standing ulcerativecolitis. J Clin Pathol 1986;39:393-8.

12 Dozois RR. Pathophysiology. Symposium; restorative procto-colectomy with ileal reservoir. International Journal of Col-orectal Disease 1986;1: 14-5.

13 Moskowitz RL. Pathophysiology. Symposium; restorativeproctocolectomy with ileal reservoir. International Journal ofColorectal Disease 1986;1:l16-7.

14 Nicholls RJ, Belliveau P, Neill M, Wilks M, Tabaqchali S.Restorative proctocolectomy with ileal reservoir: a patho-physiological assessment. Gut 1981;22:462-8.

15 Nasmyth DG, Godwin PGR, Dixon MF, Williams NS, JohnstonD. The relationship between mucosal struciure and intestinalflora in ileal reservoirs. Br J Surg 1985;72:S 129.

Requests for reprints to: Dr NA Shepherd, Department ofPathology, St Mark's Hospital, City Road, LondonEC IV 2PS, England.

on May 20, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.40.6.601 on 1 June 1987. Dow

nloaded from