Gut 1995; 37: 696-701

Clinical course of perianal fistulas in Crohn'sdisease

F Makowiec, E C Jehle, M Starlinger

AbstractThe clinical course of perianal fistulasand associated abscesses was evaluatedprospectively in 90 patients with Crohn'sdisease. Fistula type, rectal disease, faecaldiversion, and immunosuppression were

examined as prognostic indicators forfistula healing and recurrence. Medianfollow up was 22 months. The outcomewas evaluated with life table analysis.Prognostic factors were analysed bymultiple regression. Inactivation was

achieved in all patients. The risks ofrecurrent fistula activity were 48% at one

year and 59% at two years. Fistulas were

healed in 51% after two years butreopened in 44% within 18 months ofhealing. Faecal diversion and absence ofrectal disease decreased recurrence rates(p=0-01910.04) and increased healing rates(p=0.005I0.017). The outcome in patientswith trans-sphincteric fistulas was betterthan that in those with ischiorectal fistulasbut worse than in patients with subcuta-neous fistulas (p=0.015 for healing; p=0.007for recurrent fistula activity). After initialtreatment about 20% of the patients were

symptomatic and about 10% had painfulevents per six month period. Incontinencewas rare and did not increase during thestudy period. Perianal fistulas and associ-ated abscesses can be controlled safelyby simple drainage of pus coliections.Frequent reinfection and re-opening afterhealing of fistulas are characteristic.Fistula type, rectal disease, and stoolcontamination influence the clinicalcourse. Only a few patients, however,have continuous symptoms from perianalfistulas.(Gut 1995; 37: 696-701)

Keywords: Crohn's disease, perianal disease, analfistula, disease activity.

Department ofSurgery, Eberhard-Karls-University,Tubingen, GermanyF MakowiecE C JehleM Starlinger

Correspondence to:ProfessorM Starlinger,Chirurgische Universitatsklinik,Hoppe-Seyler-Strasse 3,D-72076 Tubingen, Germany.Accepted for publication23 March 1995

Perianal fistulas occur in up to 50/o of patientswith Crohn's disease. Perianal fistulas andabscesses are sometimes the first manifestationof the disease. Like intestinal disease, perianalCrohn's fistulas may take a chronic recurrentcourse. Once perianal fistulas have occurredmany patients will have symptoms for a longtime, frequently with recurrent fistula activityand repeated abscess formation, and thereforerequire repeated surgical treatment.A high percentage of healing of perianal

Crohn's fistulas has been reported in a fewretrospective studies where patients were seen

only several years after initial evaluation.' 2

Clinical risk factors for symptomatic recur-rence or factors favouring the healing of fistulashave not been clearly identified. It has beensuggested that medical treatment or surgicalremoval of actively diseased bowel may pro-mote healing of perianal fistulas.3 In addition,it is well recognised that the chance of healingis higher after the creation of a stool divertingstoma.4 5 However, from these retrospectivestudies it is difficult to differentiate whether themore rapid healing under ostomy protection isthe consequence of changes in rectal diseaseactivity or the lack of stool contamination.The objective of the prospective study

reported here was to describe the clinicalcourse of perianal fistulas in patients withCrohn's disease. We determined the timecourse of inactivation of perianal fistulas andassociated abscesses, the probabilities ofhealing and recurrent fistula activity, and therisk of re-opening of closed fistulas. Faecaldiversion, rectal disease, the type of fistula, andmedical treatment were evaluated as prog-nostic indicators of healing and symptomaticrecurrence of fistulas.

Methods

PATIENTSFrom May 1989 to October 1992, 106 con-secutive patients with perianal fistulas wereseen in a specialised Crohn's disease clinic inthe Department of Surgery at the University ofTubingen. Sixteen patients could not be con-sidered for the study. Nine patients were lost tofollow up after initial presentation and therapy.Three patients with a long history of perianalfistulas presented with complete anal inconti-nence after multiple operations: they under-went proctectomy early after first presentation.The remaining four patients were referredfrom other centres to receive a transanaladvancement flap in order to close the fistulas.The remaining 90 patients (53 women and 37men) were included in the study and werefollowed up for a median of 22 months (range6-44 months) until December 1993. Sixhundred and twenty one follow up examina-tions were documented (mean 6-9 per patient).The patients were also seen regularly in theDepartment of Medicine of our Universityhospital. The diagnosis of Crohn's disease(CD) was firmly established in all patients.The median duration of CD at the time of

inclusion in the study was 8 years (range 0-22years). Nine patients had ileal disease, 31 hadCrohn's colitis, and 50 had ileocolitis. Of the90 patients in our study, 55 (61%) had rectaldisease. The mean (SD) duration of perianal

696

on February 1, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.5.696 on 1 N

ovember 1995. D

ownloaded from

Clinical course ofperianalfistulas in Crohn 's disease

disease was 4.5 (4.8) years. Four of the 90patients presented without a history of perianaldisease. In two of these patients Crohn'sdisease had not been diagnosed before. Fortyone (47%) patients had had previous surgeryfor intestinal disease, and 69 (77%) patientsfor perianal fistulas or abscesses.

THERAPY OF PERIANAL FISTULAS ANDABSCESSESThe treatment policy during the study periodwas observation for asymptomatic fistulas anddrainage of symptomatic fistulas and associ-ated abscesses. During the later period selectedpatients without rectal disease were offeredoperative fistula closure by means of a rectaladvancement flap repair (n=20). The resultsafter this therapy will be reported in a separatepaper.6 Perianal abscesses were drained byincision (80 abscesses in 45 patients) and, inpatients with deep pus collection (n=25), byinsertion of a mushroom catheter untilmacroscopically visible purulent discharge hadstopped.

Active fistulas with purulent discharge andcommunication with the anal canal weredrained by seton (47 setons in 39 patients) fora mean duration of 5.3 weeks, until purulentdischarge or local pain had disappeared.

In seven patients a stool diverting stoma hadbeen performed before their first presentationat our clinic. In another seven patients a divert-ing stoma was performed for severe perianaldisease early during the study period. Twoother patients received a stool diverting stomaat first symptomatic recurrence.

THERAPY OF ACTIVE INTESTINAL DISEASEAll patients in the study were followed forintestinal disease in the Department ofGastroenterology. During the study period, 44episodes of Crohn's disease were diagnosed in29 patients by a Crohn's disease activity index(CDAI) >150 and typical clinical symptoms.The standardised treatment of these episodesconsisted of high dose corticosteroid therapy(1 mg prednisone/kg body weight/d). The drugregimens were then tapered off and, whenpossible, stopped. Thirty six patients weregiven prednisolone 10 mgfd, and six of thesealso received azathioprine (100 mgld). Twopatients received azathioprine withoutsteroids. For the evaluation of healing andrecurrence rates these 38 patients were classi-fied as a group receiving immunosuppressivetherapy.Nine patients with active fistulas were given

TABLE I Anatomical classification of analfistulas

No (%) ofpatients with

Fistula type No (%o) complex fistulas Group

Trans-sphincteric (including AV) 50 (56) 6 (12) Trans-sphincteric fistulas(15 AV) (n=50)

Subcutaneous 24 (26.7) 2 (8) Low fistulasIntersphinctenc 4 (4.4) 0 (n=28)Ischiorectal 11 (12-2) 6 (55) Ischiorectal fistulasSuprasphincteric 1 (1.1) 0 (n= 12)

AV=anovaginal fistula.

oral metronidazol (for 2 weeks-3 months).Increased rectal disease activity was treated byenemas (steroids or 5-ASA) in 12 patients.Intestinal resections were performed in 12patients during the study period forcomplications of intestinal disease.

METHODS OF FOLLOW UPPatients with active perianal fistulas orassociated abscesses were followed up at shortintervals (weekly) until active disease stopped.Afterwards, they were seen at least every sixmonths. Clinical and proctological examina-tion (including proctoscopy) were performedat each visit. When possible (no pain or analstenosis), endorectal ultrasound (EUS) wasperformed at each examination (hard cap 7.5MHz bifocal multiplane rectal transducer,Kretz Technik, Germany).7 When patientsunderwent surgery for active perianal fistulasor abscesses, EUS was used to locate abscessesbefore the drainage procedure.

DOCUMENTATIONAt study entry, demographic data and thehistory of previous perianal fistulas or abscesseswere carefully recorded. Activity of fistulas, peri-anal abscesses, the anatomical type of fistula,and rectal disease activity were classified asdescribed below. The following additional para-meters were documented at each presentation:CDAI, medical therapy, and history ofany acuteepisodes of Crohn's disease since the previousvisit, presence or absence of a stoma, and surgi-cal therapy of perianal and intestinal disease.Stool incontinence (for air, liquid, or formedstool), pain at rest, and pain at defecation werealso recorded. The data were recorded onstandardised questionnaires.

CLASSIFICATION OF FISTULA ACTIVITY, FISTULATYPE AND RECTAL DISEASE

PerianalfistulasThese were classified as being active if at leastone ofthe following parameters was present afterthe exclusion of other painful conditions (procti-tis, fissures, haemorrhoids: perianal abscess,fistula with purulent discharge, or severe localpain suggestive of active inflammation.

Fistula typeAnatomical classification of fistula type wasperformed as described previously8 andreassessed at each follow up visit by procto-logical and EUS examination. Explorationunder anaesthesia was performed only if anoperation was necessary for the placement ofsetons or drainage catheters. In addition, blindfistula tracts into the ischiorectal space withoutdemonstrable internal opening in the analcanal or rectum were defined as ischiorectalfistulas8 (Table I). At least one magneticresonance imaging (MRI) scan of the pelviswas performed in 62 patients. MRI furnishedimportant information on deep abscesses

697

on February 1, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.5.696 on 1 N

ovember 1995. D

ownloaded from

Makowiec, JYehle, Starlinger

(ischiorectal fossa, supralevator space) or theanatomical type of fistulas that were not clearlyidentified by proctological or EUS examina-tion.9 To characterise further the complexity ofsome of the fistulas (frequency of secondaryfistula tracts or multiple ramifications) we

evaluated the maximal number of externalfistula openings in each patient and classifiedpatients with more than two external openingsas having complex fistulas (Table I). A more

accurate description was not possible sinceoperative exploration was performed only ifnecessary for the treatment of active disease,and MRI was not routinely done in all patients.

Rectal disease activityDisease activity, as documented by procto-scopy, was classified as highly active in the caseof ulceration; moderately active if mucosalfriability, mucosal erythema, or a cobblestonepattern were present; or inactive if none of theabove mentioned criteria was detectable. Analstenosis that required bouginage was present innine patients and ulcerating anal disease in 10patients at study entry.

DEFINITIONSInactivation of perianal fistulas and associatedabscesses was defined as cessation of purulentdischarge from the fistulas and disappearanceof perianal pain. Healing was defined as

complete closure of the fistula without any signof activity or pain for at least one month.Re-opening of fistulas was defined as reappear-ance of perianal fistulas after prior healing.Symptomatic recurrence was the reappearanceof active perianal fistulas or associatedabscesses after prior inactivation or healing.

STATISTICAL ANALYSIS

Inactivation of perianal fistulas and abscesses,healing, re-opening, and symptomatic recur-

rence rates were analysed using Kaplan-Meiersurvival estimates. Patients were considered atrisk until the event occurred (inactivation, heal-ing, recurrence) or until the last follow upexamination. Factors influencing healing or

symptomatic recurrence were analysed by logrank and Wilcoxon rank tests (univariate analy-sis). The data underwent further independentanalysis using multiple regression according tothe proportional hazard model (Cox regressionanalysis).

Results

INACTIVATION OF ACTIVE PERIANAL FISTULASAND ASSOCIATED ABSCESSES

Seventy one of the 90 patients entered thestudy with active perianal fistulas or associatedabscesses. In all 71, inactivation was achievedbetween two and 81 weeks later. Median timeto inactivation was 6-5 weeks. In 25% ofpatients inactivation occurred after four weeks,in 50% after 6-5 weeks, and in 75% after 13weeks.

TABLE II Symptomatic recurrence andfistula healing

Symptomaticrecurrence Healing

No No (%) No No (%.)

All 80 42 (53) 90 42 (47)Faecal diversion:Yes 14 2 (14) 16 10 (63)No 66 34 (52) 74 32 (43)

Fistula type:Low 25 7 (28) 28 18 (64)Trans sphincteric 44 22 (50) 50 20 (40)Ischiorectal 11 7 (64) 12 4 (33)

Rectal disease:Yes 49 26 (53) 55 24 (44)No 31 10 (32) 35 18 (51)

Immunosuppression:Yes 35 18 (51) 38 16 (42)No 45 18 (40) 52 26 (50)

RECURRENCE OF FISTULA ACTIVITYTen patients experienced no further active peri-anal disease during the entire study period. Inaddition to the 71 patients who entered thestudy with active fistulas as described above,nine further patients with initially inactivedisease developed active disease early duringthe study. Recurrence of fistula activity afterinactivation was studied in these 80 patients.Forty two had at least one recurrence of fistulaactivity, at a median time interval of 3.8 monthsafter inactivation. The probability of sympto-matic recurrence was 35% after six months,48% after one year and 59% after two years.The crude data on symptomatic recurrence aregiven in Table II. Only two of 14 (14%)patients with faecal diversion but 34 of 66(52%) patients without ostomy developedsymptomatic recurrence. Using independentfactor analysis (Cox regression) fistula type(p=0.007), faecal diversion (p=0019), andrectal disease (p=0.041) appeared as indepen-dent prognostic factors: ischiorectal and trans-sphincteric fistulas recurred more frequentlythan low fistulas. Patients with faecal diversionhad a significantly lower recurrence rate thanpatients without ostomy. The presence of rectaldisease was a risk factor for symptomatic recur-rence whereas immunosuppressive therapy didnot influence this recurrence rate (p=048).

SYMPTOMATIC FISTULA RECURRENCE,INTESTINAL AND RECTAL DISEASE ACTIVITYIntestinal disease activity (CDAI) and rectalactivity at the time of first symptomatic recur-rence (n=36) were compared with the datadocumented at the examination before recur-rence. The mean (SD) time interval betweenthese two examinations was 3.8 (3.6) months.Rectal disease activity and CDAI before and atthe time of recurrence are given in Table III.

TABLE III Crohn's disease activity index (CDAI), rectaldisease activity, and number of liquid stools at the lastfollow up before recurrence and at time of recurrence

Before At time ofrecurrence recurrence p

CDAI (mean (SD)) 88 (62) 105 (85) 0.33Rectal disease activity:Absent 23 21Medium 6 7 0-68High 3 4

Mean (SD) no of liquidstools per wk 5-8 (13-4) 9-0 (14) 0.30

698

on February 1, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.5.696 on 1 N

ovember 1995. D

ownloaded from

Clinical course ofperianalfistulas in Crohn's disease

Neither the mean or median CDAI values northe frequency of rectal disease activityincreased. In addition, the number of liquidstools did not increase significantly whenfistula activity recurred.

FISTULA ACTIVITY AND FLARE UP OF CROHN'SDISEASEAt first presentation, 11 of 71 patients withactive fistulas had a flare up of Crohn's disease.Four other patients with active fistulas had aflare up 6 to 12 weeks before entering ourstudy and were still taking acute phase therapyat study entry. After the initial inactivation ofCrohn's disease, 19 patients with inactivefistulas had 25 'flare ups' of intestinal disease.Five intestinal disease recurrences wereaccompanied by symptomatic perianal fistularecurrence. In four of these, fistula activityrecurred within 12 weeks of the flare up (twopatients at four weeks, one at six weeks and,one patient at 12 weeks). The 16 other clinicalrecurrences were not followed by fistulaactivity within six months.

HEALING OF FISTULASDuring the study period perianal fistulashealed in 42 of 90 patients (crude data, TableII). The median time to healing was 11.4months. The cumulative change ofhealing was15% after six months, 300/0 after one year, 51%after two years, and 69% after three years.Multiple regression analysis showed that thepresence of a stoma was the most powerfulfactor influencing healing (p=0O005). As withsymptomatic recurrence, low fistulas had abetter prognosis (higher healing rate) thantrans-sphincteric or ischiorectal fistulas(p=0-015). Fistulas healed better in patientswithout than in those with rectal disease(p=0O017). Medical treatment (immuno-suppression or not) had no influence onhealing rate (p=0 65).

RE-OPENING OF FISTULAS AFTER HEALINGAfter healing, perianal fistulas recurred in 16 ofthe 36 patients available for this evaluationafter a median period of five months. Thecumulative risk of re-opening was 29% at sixmonths, 36% at 12 months, and 44% at 18months after healing. The number of includedpatients and observation times were too smallto evaluate risk factors for fistula re-opening.

FREQUENCY OF PERIANAL SYMPTOMSEighty per cent of the patients entered thestudy with active fistulas, explaining the highfrequency of symptomatic disease (84%) in thefirst six month period. After this initial activity,20-30% of the patients had symptomatic peri-anal fistulas during each of the following halfyear periods. Between 6% and 18% only com-plained at least once of local pain per sixmonth period. Pain at defecation was docu-mented in 71% of patients with a perianalabscess, in 34% of active fistulas, and in 5.7%

when fistulas were inactive. Pain at rest wasreported in 81% of patients with an abscess, in30% with active fistulas, but in only 4.4% ofthose with inactive fistulas.

STOOL INCONTINENCEAt study entry, 11 patients complained ofincontinence, eight for air and three for liquidstool. Three of these patients received anostomy for active perianal fistulas or abscesses.Only one of the remaining eight patients stillcomplained of incontinence at the last observa-tion (after 35 months). During the studyperiod only two patients developed partialincontinence (for air). Proctectomy was per-formed in two patients, but not for inconti-nence. Incontinence was more frequent inpatients with a perianal abscess (11%) than inpatients with inactive perianal fistulas (3.6%).

DiscussionBy using a standardised treatment and followup protocol we evaluated the clinical course ofperianal fistula disease. Active fistulas and peri-anal sepsis (abscesses) could initially be con-trolled effectively in all our patients withsimple drainage procedures. However, thesymptomatic recurrence rate was high (59%within two years). Except in a few patients,intestinal or rectal disease activity was notincreased when fistula activity recurred.At the last follow up visit, 46% of the

patients had healed fistulas. The chance ofspontaneous fistula healing was 50% after twoyears. However, healing was followed by fistulare-opening or by the development of a newfistula in a comparably high percentage ofpatients (44% at 18 months after healing). Theanatomical type of fistula turned out to be animportant prognostic factor influencing bothsymptomatic recurrence and fistula healing.

Subcutaneous fistulas had the highestchance of healing and the lowest recurrencerate. In contrast, blind ischiorectal fistula tractswithout a demonstrable internal opening in theanal canal or the rectum were characterised bylow healing and high recurrence rates. Thefrequent absence of an internal opening inischiorectal fistulas has previously beenidentified as an adverse prognostic indicator.'0

Proximal stool diversion significantlyimproved the course of the perianal fistulas(lower recurrence rate, higher chance of heal-ing). Rectal involvement of Crohn's diseasebut not the macroscopic grading of inflamma-tion of the rectal mucosa negatively influencedthe course of perianal fistulas. Since thenumber of liquid stools did not correlate withsymptomatic fistula recurrence, the presenceof stool per se in the rectum or the passageof stool through the anal canal seem to belocal factors that influence fistula activityindependently of mucosal inflammation.

Immunosuppressive treatment did not influ-ence the clinical course of perianal fistulas inour patients. Only eight of our patientsreceived azathioprine. A favourable influenceof 6-mercaptopurine/azathioprine on perianal

699

on February 1, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.5.696 on 1 N

ovember 1995. D

ownloaded from

Makowiec, _Jehle, Starlinger

fistulas has been suggested by others.1' Thisquestion can only be solved in a controlled trialtesting the effect of putative beneficial agents(azathioprine, metothrexate, antibiotics) onthe natural course of perianal fistulas inCrohn's disease.Most patients entered the study with active

perianal fistulas or associated abscessesexplaining the high frequency of perianal painin the first six month period. However, thefrequencies of perianal symptoms later in thestudy were low (20%) and correlated withrecurrent fistula activity. When perianalfistulas were inactive, local pain was reportedat only one of 20 follow up visits.

Stool incontinence is one of the main indica-tions for proctectomy in patients with perianalCrohn's disease. Incontinence was relativelyrare in our study population. It was transitoryand related to active perianal fistulas orabscesses, or related to stool passage throughfistula tracts in most cases. Stool incontinencedid not worsen during the study period underthe conservative local therapy. This confirmsthe observation of others'2 that most cases ofirreversible stool incontinence in perianalCrohn's disease are not a consequence of thelocal disease itself but of multiple operationsleading to partial or total anal sphincterdestruction.

In contrast to the increasing number ofprospective study data published on theclinical course of intestinal Crohn's disease,few data have been available regarding perianalfistulas in Crohn's disease. The few retrospec-tive studies on perianal Crohn's disease lookedat the final outcome (healing/closure of thefistulas or need for proctectomy) only and lackuniformity as some include anal fissures andskin tags. Buchmann et al 1 have suggested thatthe long term course of perianal fistulas inpatients with Crohn's disease is relativelybenign, resulting in asymptomatic or healedfistulas in most patients followed for 10 years.However, these conclusions are based on asmall number of patients, and the 10 yearstatus of only 65% of the initial study groupcould be evaluated. Another paper describedthe long term course in 184 patients with amedian follow up of 9 4 years.2 Forty per centof these patients had to undergo proctectomy,but spontaneous healing was observed in 47%followed by a two year recurrence rate of 35%.Eighty five per cent of the patients withoutproctectomy were healed at the end of followup but only 4°/0 of them had rectal disease.Therefore, patients without rectal involvementhad an excellent prognosis: most fistulashealed or were asymptomatic. A similar con-clusion was reached in a recent study describ-ing the results of surgical treatment in 27patients followed for a mean of 16 years.13Patients with ileocaecal disease had an 80%healing rate after surgical treatment, whereasonly four of 15 patients with Crohn's procto-colitis had a favourable long term outcome and11 patients eventually underwent proctectomy.The only paper dealing with the long term

perspective using lifetable analysis reported arelatively low proctectomy rate. 14 However,

the frequency of rectal involvement in thesepatients is not given.

Proctectomy may be the ultimate therapeu-tical possibility in rectal or perianal disease. Asmentioned above proctectomy rates varywidely in the published reports. The medianfollow up in our study is too short to estimatethe long term proctectomy rate in our patientswith perianal fistulas. Proctectomy was neces-sary in two of our patients - one with severeanal stenosis and the other with intractablehigh grade proctitis. However, this low numbermay not reflect the overall proctectomy rate inpatients with distal Crohn's colitis or perianaldisease, which is probably higher.2 15-17The indications for a more aggressive

surgical approach to treat perianal fistulas is amajor point of discussion. With varioussurgical procedures (fistulotomy, partialfistulotomy, drainage), high healing rates wereachieved (up to 85%) in selected groups ofpatients.10 13 16 18-24 Some of these papersreported only operations on active fistulas,22others only in patients where an aggressivesurgical approach seemed advisable (low lyingfistulas and absence of rectal disease asevidenced by rates of rectal involvement in thestudied patients as low as 30%,23 as opposed to61% rectal involvement in our patients).Furthermore, accurate rates of fistula recur-rence are frequently not available mostlybecause of incomplete data on the length offollow up. Actuarial methods of calculatingsuch end points as healing or recurrence havenot been used so far.We recently began to use excisional surgery

with fistula closure (endorectal advancementflap repair) in selected patients with trans-sphincteric and anovaginal fistulas.6 Therecurrence rate after flap repair was high, com-parable to recurrence rates in the presentstudy. Anovaginal fistulas seem to recur morefrequently after local repair than other trans-sphincteric fistulas.18 26 Since uncontrolledfaecal drainage from the vagina is bothuncomfortable and embarrassing for thepatient, however, local repair of these fistulasmay be worthwhile in these cases.24The results of our prospective follow up

indicate that the prognosis of perianal fistulasin Crohn's disease with regard to healing andrecurrence may be worse than reported in pre-vious studies. However, active perianal fistulasand associated abscesses could be controlledeffectively without aggressive local surgeryendangering the anal sphincter. After initialtherapy of active fistulas, only a minority ofpatients had symptoms related to the perianaldisease. The proctectomy rate was low in ourpatients, but this fact may change with a longerfollow up period. The clinical course wasimproved by faecal diversion protecting therectum and anal canal from contamination.Other risk factors like fistula type and rectaldisease give information on prognosis and helpplanning treatment in these patients. Intestinalor rectal disease activity appeared to influencethe activity of perianal disease in a smallnumber of patients only. A special treatmentpolicy to control perianal disease by treating

700

on February 1, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.5.696 on 1 N

ovember 1995. D

ownloaded from

Clinical course ofperianalfistulas in Crohn 's disease 701

intestinal disease activity is not warrantedtherefore unless rectal disease activity itselfgives rise to symptoms such as increased stoolfrequency or tenesmus.Parts of this paper have been published in abstract form at the2nd United European Gastroenterology week 1993, Barcelona,Spain and at the annual meeting of the AmericanGastroenterological Association 1992, New Orleans, USA.

1 Buchmann P, Keighley MRB, Allan RN, Thompson H,Alexander-Williams J. Natural history of perianal Crohn'sdisease. Ten year follow-up: a plea for conservatism. Am JSurg 1980; 140: 642-4.

2 Hellers G, Bergstrand 0, Ewerth S, Holmstr6m B.Occurrence and outcome after primary treatment of analfistulae in Crohn's disease. Gut 1980; 21: 525-7.

3 Heuman R, Bolin T, Sj6dahl R, Tagesson C. The incidenceand course of perianal complications and arthralgia afterintestinal resection with restoration of continuity forCrohn's disease. BrJ Surg 1981; 68: 528-30.

4 Harper PH, Kettlewell MGW, Lee ECG. The effect of splitileostomy on perianal Crohn's disease. BrJ Surg 1982; 69:608-10.

5 Sher ME, Bauer JJ, Gorphine S, Gelernt I. Low Hartmann'sprocedure for severe anorectal Crohn's disease. Dis ColonRectum 1992; 35: 975-80.

6 Makowiec F, Jehle EC, Becker HD, Starlinger M. Clinicalcourse following transanal advancement flap repair of peri-anal fistulas in Crohn's disease. BrJ Surg 1995; 82: 603-6.

7 Van Outryve MJ, Pelckmans PA, Michielsen PP, VanMaercke YMV. Value of transrectal ultrasonography inCrohn's disease. Gastroenterology 1991; 101: 1171-7.

8 Parks AG, Gordon PH, Hardcastle JD. A classification offistula-in-ano. BrJ Surg 1976; 63: 1-12.

9 Koelbel G, Schmiedt U, Majer MC, Weber P, Jenss H,Kueper K, Hess CF. Diagnosis of fistulae and sinus tractsin patients with Crohn's disease: Value of MR imaging.AJYR 1989; 152: 999-1003.

10 Levien DH, Surrell J, Mazier WP. Surgical treatment ofanorectal fistula in patients with Crohn's disease. SurgGynecol Obstet 1989; 169: 133-6.

11 Korelitz BI, Present DH. Favorable effect of 6-

Mercaptopurine on fistulae of Crohn's disease. Dig Dis Sci1985; 30: 58-64.

12 Alexander-Williams J, Buchmann P. Perianal Crohn'sdisease. WorldJ Surg 1980; 4: 203-8.

13 Nordgren S, Fasth S, Hulten L. Anal fistulas in Crohn'sdisease: incidence and outcome of surgical treatment.Int3c Colorect Dis 1992; 7: 214-8.

14 Wolff BG, Culp CE, Beart RW Jr, Ilstrup DM, Ready RL.Anorectal Crohn's disease. A long-term perspective.Dis Colon Rectum 1985; 28: 709-11.

15 Keighley MRB, Allan RN. Current status and influence ofoperation on perianal Crohn's disease. Int J Colorect Dis1986; 1: 104-7.

16 Van Dongen LM, Lubbers EJ. Perianal fistulas in patientswith Crohn's disease. Arch Surg 1986; 121: 1187-90.

17 Williams JG, Hughes LE. Abdominoperineal resection forsevere perianal Crohn's disease. Dis Colon Rectum 1990;33: 402-7.

18 Fry RD, Shemesh EI, Kodner IJ, Timmcke A. Techniquesand results in the management of anal and perianalCrohn's disease. Surg Gynecol Obstet 1989; 168: 42-8.

19 Sohn N, Korelitz BI, Weinstein MA. Anorectal Crohn'sdisease: definitive surgery for fistulas and recurrentabscesses. Am J Surg 1980; 139: 394-7.

20 Francois Y, Descos L, Vignal J. Conservative treatment oflow rectovaginal fistulas in Crohn's disease. Int _7 ColorectDis 1990; 5: 12-4.

21 Morrison JG, Gathright JB, Ray JE, Ferrari BT, Hicks TC,Timmcke AE. Surgical management of anorectal fistulasin Crohn's disease. Dis Colon Rectum 1989; 32: 492-6.

22 Pritchard TJ, Schoetz DJ, Roberts PL, Murray JJ, Coller JA,Veidenheimer MC. Perirectal abscess in Crohn's disease.Dis Colon Rectum 1990; 33: 933-7.

23 Williams JG, Rothenberger DA, Nemer FD, Goldberg SM.Fistula-in-ano in Crohn's disease. Dis Colon Rectum 1991;34: 378-84.

24 Scott NA, Nair A, Hughes LE. Anovaginal and rectovaginalfistula in patients with Crohn's disease. Br J Surg 1992;79: 1379-80.

25 Bayer I, Gordon PH. Selected operative management offistula-in-ano in Crohn's disease. Dis Colon Rectum 1994;37: 760-5.

26 Jones IT, Fazio VW, Jagelman DG. The use of transanalrectal advancement flap in the management of fistulasinvolving the anorectum. Dis Colon Rectum 1987; 30:919-23.

on February 1, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.37.5.696 on 1 N

ovember 1995. D

ownloaded from