OUP user menu

Journal of Crohn's and Colitis: 10 (8)


Laurence J. Egan, Ireland

Associate Editors

Shomron Ben-Horin, IsraelSilvio Danese, ItalyPeter Lakatos, HungaryMiles Parkes, UKJesús Rivera-Nieves, USABritta Siegmund, GermanyGijs van den Brink, NLSéverine Vermeire, Belgium


Published on behalf of

Surgical repair and biological therapy for fecal incontinence in Crohn's disease involving both sphincter defects and complex fistulas

J.A. Álvarez, F. Bermejo, A. Algaba, M.P. Hernandez, M. Grau
DOI: http://dx.doi.org/10.1016/j.crohns.2011.06.004 598-607 First published online: 1 December 2011


Background and aims: Surgeons have traditionally tried to avoid any complex surgical procedures in Crohn's patients with complex perianal diseases because of the fear of complications, worsening the patient's condition and risking an eventual proctectomy. The introduction of biological therapy has changed the management of Crohn's disease. This study assesses the long-term success of addressing defects in anal sphincter and complex fistula when patients receive anti-TNF-α antibodies.

Methods: Ten consecutive patients were prospectively scheduled for induction therapy with 5 mg/kg Infliximab at week 0, 2 and 6 and maintenance every 8 weeks associated with azathioprine. Elective surgery was performed conducting a simultaneous approach to the sphincter defect and fistula tracts. Outcomes were long-term continence, complications which were assessed by a Wexner's score along with a complementary questionnaire. Statistical analysis was performed using general linear model of repeated measures.

Results: Three patients had complications related to surgery: two abscesses and low intersphincteric fistula and one case of rectal stenosis causing fecal urgency. There was no suture dehiscence. Wexner's score improved at 12 months (10.0 ± 2.4 vs. 18.0 ± 2.6; p = 0.003) and over time (48 month 9.5 ± 2.8; p = 0.001). These scores were significantly worse when patients had urgency before treatment (12.8 ± 1.2 vs. 9.5 ± 2.8; p = 0.03) but not when the urgency appeared later. No patient remained incontinent to solid stools. Three patients had occasional incontinence to liquid stools associated to disease reactivation.

Conclusion: Surgical repair and immunomodulator therapy with infliximab could be an option in incontinent patients with Crohn's disease involving both a sphincter defect and severe or refractory fistulas.

Research highlights: ► Complex surgical procedures have traditionally been avoided in severe Crohn’s disease. ► Infliximab could change expectations achievable with surgical treatment. ► Sphincter defects and fistulas can be simultaneously addressed in incontinence. ► Combined biological therapy and surgery improved the Wexner and solved incontinence. ► Benefits of sphincteroplasty with antiTNF were stable after 48 months.

  • Crohn's disease
  • Infliximab
  • Sphincter defect
  • Sphincter repair
  • Fistula
  • Incontinence
  • Wexner

1 Introduction

Perianal complications are a common feature of Crohn's disease that afflict approximately one third of patients with a combination of fissure, fistula, stricture and abscess.1,2 As a consequence anal lesions are the main indication for surgery in 23% of cases and a contributing factor in an additional 43%.3 The cumulative risk for fistula increases with time.4,5 Results from surgical approaches are somewhat disappointing and poor outcomes can be anticipated.6,7 Fistulas due to Crohn's disease are usually complex, with curved and multiple tracts. Recurrent or persistent fistula with repeated inflammation, abscesses and pain lead to the deterioration of the patient's quality of life.8 In these circumstances both disease and surgical intervention play a role in the deterioration of sphincter function and integrity, and fecal incontinence arises as the main complaint. As the number of surgical procedures increases so does the likelihood that the patient must face diversion or a permanent stoma as the last therapeutic option which has crucial effects on their future.

The introduction of biological therapy has changed the management of Crohn's disease. Anti-tumor necrosis factor-α (TNFα) antibodies have improved the outcome of medical treatment both in the percentage of induction of remission as well as in the maintenance of the response.9,10 Nevertheless, Infliximab itself is not the final cure for every patient with perianal Crohn's disease because fistulas mostly recur when the antibody is discontinued and, in any case, response declines with time. Actually complete fistula healing without addressing the fistula surgically results in a low percentage of success.11 Likewise Infliximab can change expectations achievable with surgical treatment allowing more aggressive approaches to severe perianal disease.

Sphincter repair has already been performed in selected patients with sphincter disruption, but there are no clear guidelines concerning optimal therapeutic approaches in the case of coexisting complex fistulas and external sphincter defect.3 This study assesses the long-term success of simultaneously addressing defects in the external sphincter and complex fistulas in patients with Crohn's disease, whose complaint was gross incontinence not related to diarrhea, while they are receiving anti-tumor necrosis factor-α (TNFα) antibodies.

2 Material and methods

2.1 Patients

Between June 2004 and March 2006, 180 cases of Crohn's disease were treated from a population of 250,000 inhabitants. Thirty-eight patients (21.1%) underwent surgical treatment, twenty-five with perianal disease and thirteen with intestinal stricture. This is a consecutive series of ten patients with incontinence to solid stools as a consequence of sphincter defect who also suffered from a complex fistula. In all cases daily activities were severely impaired by the physical consequences of incontinence, feeling of uncleanliness, disturbing psychosocial functioning and diminishing the patient's image of the body. All patients had a proven diagnosis of Crohn's disease.12 Phenotypes were classified according to the Montreal Classification.13 The patients' symptoms with treatment were followed as well as disease-related complaints, continence status, and restrictions in their everyday life. The study was approved by the Ethics Committee of the steering centre and subjects gave informed consent to the work.

2.2 Diagnostic methods

Anorectal defect was investigated by physical examination, Endoanal Ultrasound Scanning (EUB-5500 HV, Hitachi Medical Systems Europe) and Magnetic Resonance Imaging in the coronal and axial planes with a body coil. Fistula-in-ano was classified in accordance with the guidelines of The American Society of Colon and Rectal Surgeons14 and the American Gastroenterological Association,15 and considered as complex when the tract crossed more than 50% of the external sphincter (high transsphincteric, suprasphincteric, and extrasphincteric).

2.3 Study variables

Outcomes were long-term continence, complications and impact on lifestyle from the patients' perspective. Grouping variables were age, gender, age at onset, location and amplitude in degrees of the sphincter defect, number of bowel movements before and after procedure and fecal urgency. Fecal incontinence was evaluated by the Wexner Continence Grading Scale by determining scores at enrollment, day of surgery and every 12 months over a period of 4 years. This classification can be used to demonstrate improvement in continence scores after treatment and as with other summary scales has been shown to correlate with quality of life instruments by evaluating lifestyle scores.16 Continence for solid stools was the main outcome. The Wexner's score was supplemented by three additional questions. Whether in the past month they had experienced bowel movements whose content had been fluid, to discover the liquid incontinence that occurs when diarrhea is not controlled. Secondly, in the case of liquid stools, if there had been accidental leakage and finally, the percentage of bowel movements that caused incontinence. Continence status was also considered taking into account fecal urgency and the ability to defer defecation. Patients' incontinence was a result of defects of the anal sphincter. Multiple previous operations to address the fistula and drainage of perianal sepsis had caused a breakage in the sphincter. Finally, patients had an anatomic defect that motivated permanent soiling and also one or more fistulas. All patients considered themselves as incontinent to solid stools and soiling forced them to wear pads on a permanent basis. Wexner Continence Grading Scale assesses continence by asking the patient about fecal incontinence for solid, liquid or gas, but does not ask for the discharge resulting from the fistula. Actually, pads can be worn to deal with soiling or pus discharge. However, when patients responded to the questionnaire and the three additional questions, they did so from the perspective of its main concern, fecal incontinence, and not by the production of pus. Finally, patients were asked to evaluate treatment outcomes, improvement achieved in continence and if they would opt for the same treatment.

All patients were operated on by the same surgeon. The monitor of the patient was conducted jointly in the Surgery and the Gastroenterology Departments. Complications were evaluated by a second surgeon and the physician of gastroenterology independently from the surgeon who had operated on patients. The assessment of continence was performed for each patient at each visit and included in the patient's history.

2.4 Preoperative management

The management plan started with induction therapy with Infliximab (Remicade®) at doses of 5 mg/kg at weeks 0, 2, and 6 using a standard protocol. The surgical repair was then performed and subsequently, maintenance dosing every 8 weeks. All patients took azathioprine at doses of 2.5 mg kg day for more than 6 months. Surgery was performed in an elective setting and delayed if there was an abscess or an acute flare of Crohn's disease, in particular if proctitis was present. This was the case of two patients with proctitis. In these cases, optimizing medical treatment and drainage of purulent material was a priority. Any surgical attempt to heal Crohn's anal fistulas was only undertaken once the patient was in a stable condition, with minimal residual perianal sepsis and good nutritional status. The signs of proctitis disappeared before treatment. No patient was excluded. All patients received metronidazole at 1.5 g/day or ciprofloxacin at 1 g/day preoperatively and postoperatively for 7 days. None of them underwent mechanical bowel preparation nor was there a diverting stoma.

2.5 Sphincter repair technique

Surgical procedure was performed under general anesthesia in the prone position. Sphincter repair and fistula treatments were simultaneously performed without a seton when mobilization of both ends of the defect in sphincters included the complex fistula. A fistulotomy was performed. Fibromuscular margins of the external sphincter and puborectalis mobilized cranially to the anorectal ring and overlapped with 2–0 Vicryl ™ (Ethicon) to avoid excessive tension on the repair. The first row of sutures connected rectum, puborectalis sling and external sphincter on both sides, and then a second row connected parts of the internal fibers, the submucosa and the mucosa. Suture of the internal sphincter occasionally may be performed separately but this is often difficult due to fragmentation of the muscle. A proper skin coverage over the defect was established with full-thickness skin either advanced or rotated into position, leaving part of the wound open for healing by secondary intention. At completion of surgery, the anal canal should admit the little finger snugly.

When the complex tract ran in a different quadrant from the sphincter defect, a transposition of the fistula tract was performed leaving a seton across the puborectalis sling (Vesseloop™, Medica Europe, The Netherlands). A fistulectomy or coring out was performed through the external opening to the intersphincteric plane, the extrasphincteric portion of the tract re-routed into an inter-sphincteric submucous position and curettage performed in the remaining tract above the sphincter. Seton was removed after the second postoperative dose of infliximab as an outpatient procedure. The presence of any additional secondary extensions was also identified, cored-out or thoroughly curettage and laid open (Fig. 1).

Figure 1

Treatment of complex fistula. Sagittal view shows the complex fistula in place on the right and transposition of the fistula tract, repair of the external sphincter, curettage and seton in the intersphincteric and submucous position on the left.

2.6 Statistical analysis

Analysis of variance was carried out to test for significant differences between means layering by grouping variables. In essence, we wanted to monitor the improvement of patients' Wexner scores over a 48 month period of follow-up. A standardized questionnaire was administered before treatment (level 1 of the repeated measures factor), and a comparable test was administered every 12 months. Thus, the repeated measures factor (Time) had 5 levels and the difference of the mean score on T1 from 0 indicated the improvement (or deterioration) of scores across the 5 levels of Time. Statistical analysis was performed using the General linear modeling (GLM). The GLM Repeated Measures procedure provides analysis of variance (ANOVA) when the same measurement is repeatedly made over a period of time with the same subjects or under varying circumstances. When we wanted to compare two groups, we used the t test for independent samples; when we wanted to compare two variables given the same subjects, we used the t test for dependent samples. This distinction – dependent and independent samples – is important for ANOVA as well. Basically, if we have repeated measurements of the same variable at different points in time on the same subjects, then the factor is a repeated measures factor (also called a within-subjects factor because by estimating its significance we compute the within-subjects sums of squares). If we compare different groups of subjects (e.g., procedure performed, fecal urgency or any of the study variables), we refer to the factor as a between-groups factor. The computations of significance tests are different for both factors; however, the logic of computations and interpretations is the same. Interactions between factors can also be investigated along with the effects of individual factors. In this case, each multivariate statistic is transformed into a test statistic with an approximate or exact F distribution. Analysis was performed using IBM® PASW® Statistics version 18 (IBM Corporation, NY). Statistical analysis of results was done by masking the names of patients'.

3 Results

Ten consecutive patients with perianal Crohn's disease with incontinence and complex fistulas underwent sphincter repair between June 2005 and March 2006. Follow-up was complete at 48 months in all cases. No patients were excluded. There were seven men and three women with a mean age of 37 (range, 21–53). The mean age at onset of symptoms was 24 (range, 19–38). The mean time between the diagnosis of Crohn's disease and the need for sphincter repair was 12 years (range, 2–20). There were no statistically significant differences layering by gender, age and age at onset of Crohn's disease (p = 0.51). Six patients had chronically active abdominal symptoms and four had a course that fluctuated between active and inactive disease. Patients' characteristics are summarized in Table 1. All patients had symptomatically active and complex perianal disease. Anal fistulas caused pain and discharge of pus from the external opening. They had undergone previous surgery for perianal fistula — mean 3 (range 1–5) or abscesses — mean 3 (range 1–10). All patients had complex fistulas, often with some additional secondary extensions, horseshoe tracts and abscesses, which were situated in the same quadrant or far from the sphincter damage (cases 2, 4, 7). They also showed scar and fibrosis associated with previous surgical interventions. The etiology of sphincter defect was the result of perianal septic episodes, fistulas and surgical treatments performed. None of the 3 women had previous obstetric trauma as the cause of sphincter defect. Mean defect in external sphincter was 114 grades (range, 75–160). There were no statistically significant differences in relation to the anterior or posterior location of the defect (92.5 ± 28.7 vs. 115.0 ± 27.3; p = 0.24). A keyhole deformity was obvious in one case and two additional patients showed intense disfigurements because of horseshoe tracts and fistula discharge. All patients considered themselves as incontinent to solid and liquid stools and soiling forced them to wear pads on a permanent basis. Two patients had proctitis and three patients had fecal urgency before surgery. Summary of previous surgical procedures and continence status before sphincter repair are shown in Table 2.

View this table:
Table 1

Clinical status and medical treatment at the time of enrollment.

CaseAgeGenderAge at onsetPhenotypeaDisease behaviorTreatmentb
137Male19A2 L3 B1pFluctuating diseaseAZA
221Male19A2 L3 B1pFluctuating diseaseAZA, CPF
353Male33A2 L2 B1pChronically active symptomsAZA
427Male17A2 L3 B1pChronically active symptomsAZA, CPF
532Female19A2 L3 B1pChronically active symptomsAZA, GCD
641Female26A2 L2 B1pFluctuating diseaseAZA
748Male38A2 L2 B1pChronically active symptomsAZA
834Male21A2 L3 B3pFluctuating diseaseAZA
933Female21A3 L3 B2pChronically active symptomsAZA
1053Male32A2 L3 B1pChronically active symptomsAZA, CPF
  • a Phenotypes categories according to the Montreal Classification of Crohn's disease.

  • b Treatment at the time of enrollment. AZA, Azathioprine 2.5 mg/kg/day. CPF, Ciprofloxacin 1 g/day. GCD, Glucocorticoids prednisone 0.8 mg/kg/day.

View this table:
Table 2

Summary of surgical history and continence status before treatment with Infliximab and surgical repair.

CasePrevious proceduresClinical status before treatmentProcedure performed
15 Fistula unspecified surgeriesAnterior 135° sphincter defect with extrasphincteric fistula, multiple tracts and keyhole deformity.Sphincter repair
3 Abscess drainage
21 FistulotomyPosterior 120° sphincter defect, horseshoe-shaped. Suprasphincteric fistula with secondary openings. Multiple simple fistulas, Left isquiorectal and levator abscess. Proctitis.Sphincter repair
1 Seton
2 Abscess drainage with associated fistulotomyFistula transposition with seton
31 FistulotomyPosterior 90° sphincter injury. Suprasphincteric fistula with secondary openings. Proctitis.Sphincter repair
6 Abscess drainage
41 Advancement flapLateral 160° sphincter defects. Right posterior extrasphincteric fistula to rectum, secondary openings, posterior horseshoe tract. Fecal urgency.Sphincter repair.
2 Fistulectomies,Fistula transposition with seton
1 Fistulotomy
1 Seton
51 Advancement flapPosterior 130° sphincter defect with a keyhole deformity. Multiple simple fistulas, trans-sphincteric fistula. Fecal urgency.Sphincter repair
1 Fistulectomy
1 Fistulotomy
2 Abscess drainage
61 FistulotomyAnterior 75° sphincter damage. High transphincteric fistula with multiple secondary tracts and rectovaginal involvement.Sphincter repair
1 Abscess drainage
72 FistulotomyPosterior 90° sphincter injury. Two extrasphincteric fistulas, two trans-sphincteric fistulas, mucosal prolapse.Sphincter repair.
2 SetonFistula transposition with seton
8 Abscess drainage
81 Advancement flapAnterior 85° sphincter damage. High transphincteric fistula with multiple secondary tracts.Sphincter repair.
1 Fistulectomy
1 Fistulotomy
2 Abscess drainage
91 FistulotomyAnterior 75° sphincter damage, horseshoe-shaped. High transphincteric fistula with rectovaginal involvement.Sphincter repair
3 Abscess drainage
101 FistulotomyPosterior 100° sphincter defect with a keyhole deformity. Two posterior trans-sphincteric fistula. Fecal urgency.Sphincter repair
Multiple abscess drainage

Three cases had been exposed to infliximab more than 12 months before enrollment. There were no statistical differences in patients who had already been treated with infliximab prior to enrollment in the study and the others who were started on biological therapy in line with to the study planning (18.1 ± 2.8 vs. 17.5 ± 2.5, p = 0.71). Evaluation of the impact of the introduction of infliximab on the patient's continence before surgical repair showed no differences at follow-up (10.5 ± 1.6 vs. 9.5 ± 1.8, p = 0.23).

3.1 Disease behavior

Case 3 suffered an outbreak of the underlying ileocecal disease that caused 10 liquid stools per day. It surged at the fifth month after the operation and took almost a year to achieve a complete recovery. The underlying disease in case 5 followed an aggressive course with severe swelling that forced an ileocecal resection. Infliximab was discontinued in case 6 because of urticarial reaction after initial 3 doses which repeated with certolizumab pegol. Three patients had loss in response that required dose escalation to 10 mg/kg in two cases and a switch to adalimumab 40 mg every other week in one case.

3.2 Complications

Three patients had postoperative complications (Table 3). Two patients developed an abscess at the second and third months after the wound appeared to be clinically cured. Surgical drainage was required in both patients. A recurrence of the fistula was detected immediately after the drainage in one case and 2 years later in the other. Both patients had a low intersphincteric fistula, whose treatment by fibrin-glue was ineffective. They remained occasionally symptomatic with intermittent discharge of pus in case 1 and serous in case 7. Perineal wound healing was delayed in one patient and needed 5 months to achieve complete healing. An endoscopy detected a rectal stricture with pseudodiverticula that caused occasional soiling problems for the patient. One case of related fecal urgency was detected early at the postoperative stage, but it only lasted 1 month improving as proctitis did. No patient had proctitis before enrollment in this study and only one patient had proctitis in the postoperative follow-up. Seven of ten patients had an uneventful postoperative course. There was no dehiscence in the sphincter repair evaluated by EUS and MRI.

View this table:
Table 3

Primary outcomes. Postoperative complications and continence.

CasePostoperative complicationsContinence
TypeOnsetTime to resolutionOutcomeTo solidTo liquid
1Perianal abscess3rd monthDaysCompletePartial
Low intersphincteric fistula24th monthLong-livedLittle symptomatic Purulent drainage
4Delayed healing time5 monthsCompleteComplete
Rectal stricture12th monthPersistentSymptomaticComplete
7Perianal abscess2nd monthDays
Low intersphincteric fistula2nd monthLong-livedLittle symptomatic serous drainageCompletePartial
10Proctitis1st month1 monthAsymptomaticCompletePartial

3.3 Continence

Wexner's score improved soon after the procedure (18.0 ± 2.6 vs. 10.0 ± 2.4, p = 0.003) and over the course of time (score at 48 month 9.3 ± 2.8; p = 0.001) (Fig. 2). The profile plot for the model-estimated means for Wexner's score at each of the four years of follow-up showed that there was no statistically significant differences for the between-subjects factors location of the sphincter defect, added transposition of the fistula and postoperative complications, i.e. they did not contribute significantly to the model (p = 0.45). Repeated contrasts were chosen, so each contrast compares the Wexner score at one time period to the score at the following period. The contrasts for time have significance value less than 0.05 only for the first time period — T1: before treatment vs. T2: Wexner at 12th month. The main effect on the Wexner's score is observed soon after surgical repair and remained stable during follow-up. The therapeutics' effect on continence appeared to be approximately the same for both male and female patients (p = 0.20).

Figure 2

The profile plot shows the model-estimated means for Wexner's score for each of the four years of follow-up. Wexner's score range from 0 to 20 — the higher the score the higher the alteration in continence. The model-estimated Wexner's score are plotted on the vertical axis. The follow-up is plotted on the horizontal axis. Separate lines are produced for sphincter repair vs. more complex sphincter repair with fistula transposition. GLM calculated for the effect on Wexner's score showed no statistically significant differences (p = 0.28).

Wexner improving was similar for all patients except for those who suffered from urgency before treatment (Fig. 3). These scores were significantly worse when patients had urgency before treatment (12.8 ± 1.2 vs. 9.5 ± 2.8; p = 0.03) but not when the urgency appeared at some other time (10.6 ± 1.8; p = 0.4). However, patients experienced more frequently incontinence with liquid stools on those occasions they had fecal urgency (30 and 40% of times they had diarrhea) and they considered themselves incontinent for liquid stools. Patients were not incontinent after sphincter repair. The two cases that had a fistula in the postoperative period clearly appreciated that they were continent but staining caused by the fistula.

Figure 3

Univariate test for the within-subjects factor fecal urgency. Wexner's score range from 0 to 20 — the higher the score the higher the alteration in continence. The model-estimated Wexner's score are plotted on the vertical axis and follow-up on the horizontal axis. Separate lines are produced for the presence or absence of fecal urgency. Results showed a statistically significant difference in Wexner's score when fecal urgency was a complaint (12.8 ± 1.2 vs. 9.5 ± 2.8; p = 0.03).

Univariate tests for the rest of the within-subjects factors showed significant values greater than 0.10, so it can be conclude that the effects of these components are not significant.

There were no statistical differences between those patients who have been treated previously with infliximab and the other ones that were started on biological therapy according to the study planning (18.1 ± 2.8 vs. 17.5 ± 2.5; p = 0.71). Evaluation of the impact of the discontinuation of infliximab on the patient's continence during follow-up showed no differences at 48 months (10.5 ± 1.6 vs. 9.5 ± 1.8, p = 0.23).

3.4 Patient's perspective

Evolution in continence status and perception of the results is depicted in Table 4. Symptoms of anal incontinence were significantly improved after surgical treatment from patients' perspective. Six patients considered themselves as full continent and one incontinent for flatus. Three patients were incontinent for liquid stools in one third of the occasions they had diarrhea. This incontinence emerged when the patients had acute flare or decompensation of Crohn's disease. Bowel movements before and after treatment remained unchanged (p = 0.32). All patients reported an improvement in their perianal symptoms and none of them has gross incontinence now.

View this table:
Table 4

Patients' perception of the outcome at follow-up.

CaseNumber of daily bowel movementsComplementary questionnaire to Wexner scorePatients' perspective
Before surgeryAfter surgeryFecal urgencyLiquid stools in the past monthAccidental bowel leakage of liquidBowel movements causing incontinenceLifestyle alterationaContinence status
13 liquid2 normalNoNoYes40%NoLiquid stools incontinence
221 (2)NoNoNo0%NoFull continence
34 (5)2 (4)YesYesNo10%NoFull continence
411NoYesNo10%YesFull continence
53–4 liquid2–3 liquidNoNoNo10%NoFlatus incontinence
622NoNoNo0%NoFull continence
72 (5)2 (4)YesYesYes30%YesLiquid stools incontinence
822NoNoNo10%NoFull continence
93 (5)3YesNoNo0%NoFull continence
103–4 liquid1–3 liquidNoYesYes30%YesLiquid stools incontinence
  • a Patients' answer to the question whether daylife limitations persisted after being treated.

Three patients felt that their lives had not changed and remained altered. Two of them had incontinence for liquids but the third felt that his life was limited despite stating that he was continent to solid and liquid. In contrast, one patient had incontinence up to 40% of the time in which he had diarrhea but felt he had no limitations in his lifestyle. All patients would repeat the same treatment.

4 Discussion

The surgical approach to perianal Crohn's disease has been characterized so far in limited actions.5,17,18 The causes can be found in the nature of the disease itself and because of the fear of complications, worsening the patient's condition and risking an eventual proctectomy. That is why surgeons have traditionally tried to avoid any complex surgical procedures. Occurrence of complications in patients undergoing resection associated with the use of immunosuppressive medication has recently been discarded.1921 Clinical outcomes after treatments based on infliximab made us wonder whether in the era of biological agents the sphincter repair may be associated with a fistula repair. That is, if it was possible to standardize the approach to patients amenable to treatment with infliximab and with one major flaw of continence, patients in whom there was no place for more conservative approaches. Desirable outcomes in perianal Crohn's disease should not focus exclusively on reduction in the number of septic events. In evaluating the results the age of the patients should be taken into account especially for these young people and because of the serious constraints imposed by incontinence in their daily activities. In all cases daily activities were deeply impaired by the physical consequences of incontinence, feeling of uncleanliness, disturbing psychosocial functioning and the diminishment of the patient's image of the body.

Antibiotics and azathioprine/mercaptopurine should be used as the first choice of therapy for complex perianal Crohn's disease in combination with surgical therapy, in spite of a lack of clinical trials. Infliximab or adalimumab should be used as a second line medical treatment, but most experts advocate combining anti-TNF and surgical therapy as first line therapy.22 Reparative surgery (e.g. mucosal flap or fistula plug) during infliximab therapy may improve longterm healing rates.23

Only a few cases treated by sphincter repair have been reported, although in these cases damage was a direct result of a previous perianal surgery.3 In our series, the procedure of repair was also conditioned by complex fistulas and especially by its location and relationship to the sphincter defect. When the complex route of the fistula crossed the area without sphincter the mobilization of both ends of the sphincter allowed a fistulotomy and simple sphincter overlapping. A more challenging situation arose when the complex fistula was far from the sphincter defect because it involved injury of the anorectal ring in two separate places which might have led to excessive fibrosis and stiffness. Therefore we performed the fistula approach by re-routing the extrasphincteric portion of the tract into an intersphincteric-submucous position with immediate repair of the external sphincter. Limited seton usage should permit the dissected sphincter musculature overlying the fistulous tract to heal by secondary intention with minimal functional disturbance. The point was that if it were not healed, at least the complex fistula would become simpler, less symptomatic and could be addressed later with a less aggressive procedure. This was the case for the two patients in whom fistula recurred. These are the only long-term symptomatic patients, so the results can be considered reasonably satisfactory in terms of fistula closure. Fistulas are now less symptomatic with occasional discharge without any impact on daily activities. Besides these two cases, perineal wound healing was delayed in one patient and took 5 months to achieve a complete closure of the perineal wound. There was no suture line dehiscence in our series, which is considered the most feared complication because these alterations may also be behind the decline of continence after an initially successful repair.24 Here is where immunomodulators could have an important role.25,26 Subsequently, in the absence of septic complications, medical treatment must be able to control the perianal disease with an increased likelihood of achieving and maintaining the fistula sealed. The complication rate is considered acceptable not because the overall percentage but as the importance of them. The low rate of complication after extensive perianal procedures in this high-risk population is interesting (3 out 10). We do not use protective colostomy in anorectal reconstruction in patients without Crohn's disease. Similarly, we believe that if the disease is controlled it is not necessary in these patients.

Severity was evaluated by the Wexner Continence Grading Scale which can be used to demonstrate improvement in continence scores after treatment, and as with other summary scales has been shown to correlate with quality of life instruments, by evaluating lifestyle scores.13 Combined biological therapy and surgery provided a clinically significant reduction in the Wexner's score, remarkable at 12 months after surgical repair and maintained throughout the follow-up. The profile plot for estimated marginal means for Wexner's score was similar regardless of the surgical technique. Therefore, it seems possible to make a simultaneous approach of the sphincter defects and contralateral complex fistula with no difference in the Wexner score and without added morbidity to the procedure.

Patients having proctitis may respond to Infliximab and so will likely see their symptoms of incontinence improve by a reduction of loose stool and urgency. Thus it may be difficult to differentiate from the effect of their sphincter repair. However, the etiology of incontinence was primarily an anatomic defect in the anal canal as a consequence of septic episodes perianal, fistulas and surgical treatments. In any case, differences in the Wexner's score at baseline were not significant. We have not noticed any differences in patients who were already treated with infliximab prior to enrollment in the study and the others who were started on biological therapy in accordance with the study planning.

A crucial issue for this therapeutic strategy was evaluating the stability of the long-term results. With a 48 month follow up it is possible to draw conclusions. The first outcome was to evaluate the continence for solids and all patients remained continent at follow-up without differences between those who were treated on biological therapy according to the study planning and the others whose treatment was discontinuous. A second outcome was to know what happened in the presence of liquid stools. The results showed a lesser control in continence for liquids when the patient had fecal urgency. In fact, in this subgroup of patients incontinence was present in one third of cases in which the stools were liquid. This symptom may be a between-subjects factor that limits the results of this procedure. Three patients still experience occasional incontinence when they have diarrhea. In concordance, they considered themselves as incontinent for liquid stools.

This case series includes a small number of patients because characteristics of these patients make it very difficult to design randomized or controlled trials. The benefits of sphincteroplasty with antiTNF were stable and the long term results can be considered promising with a major sepsis control or minimizing proctectomy rates.

From the standpoint of the patient, all of them reported an improvement in their perianal symptoms; the long-term outcome was considered satisfactory and would opt again for this strategy. Being a selected group of patients, improvement of continence and general status may be more favorably perceived, but at the same time the impact of this strategy in the control of the perianal symptoms should also encourage us to consider it as a valid alternative to proctectomy with good functional results.

In conclusion, surgical repair and immunomodulator therapy with infliximab could be an option in incontinent patients with Crohn's disease involving both a sphincter defect and severe or refractory fistulas. The effect of the therapeutic on continence appears to be stable although shows different effects for patients with fecal urgency. A larger study may be warranted to determine more precisely the role of urgency and proctitis on continence for liquid stools. In any case, before the widespread use of this therapeutic approach the results should be confirmed in a larger study group.

Competing interests

The authors do not have commercial associations that might pose in conflict of interest in connection with the submitted article.


No financial support was available for this project.


The text has been completely edited by Thomas Ian Ure.


The authors of this paper are part of a multidisciplinary group interested in inflammatory bowel disease and developed their work in the medical (FB), surgical (JA, PH) and research (AA, MG) areas. All have participated practically in the work to take public responsibility for the content of the article, have read and approved the final manuscript. All authors have made substantial contributions to the conception and design of the study (JA, FB), surgical treatment (JA, PH), medical treatment (FB), and acquisition and interpretation of data (AA, MG).


  • ☆☆ All authors have seen, approved and are fully conversant with the contents of the manuscript.

  • Conference presentation. The paper is based on a preliminary communication to the 4th Congress of the European Crohn's and Colitis Organization. February 5–7, 2009. Hamburg, Germany.


View Abstract