Introduction

Crohn’s disease (CD) is a chronic, relapsing, incurable inflammatory condition that can affect any gastrointestinal site from the oral cavity to the anus. Approximately 20–30% of patients suffer from colonic CD [1, 2]. The frequency of perianal complications ranges from 17 to 43% of patients [3,4,5,6,7,8]. Both of these disease phenotypes lead to significant symptomatology, complications, and impairment in quality of life.

Although medical management is the mainstay of treatment for CD, up to 70% of patients require surgery during their lifetime [7, 8]. Surgery typically involves removal of the diseased portion of intestine with anastomosis between healthy segments. However, when CD affects the colon, rectum, and anus, reconstructive options that maintain gastrointestinal continuity are limited. Thus, not surprisingly, previous studies have demonstrated that 31% of patient with perianal CD [9] and 28–31% of patients with Crohn’s colitis require a permanent stoma [10, 11]. Indications for stoma creation in perianal CD include perianal sepsis, complex anorectal fistula/abscess, rectovaginal fistula, anorectal stricture, and persistent anal ulcer [9, 10, 12, 13]. In colonic CD, fecal diversion is used to manage medically refractory colitis, fulminant or toxic colitis, stricture, and spontaneous or iatrogenic colonic perforation [14, 15].

Luminal antigens are thought to be involved in the chronic inflammation seen in CD [16], and accordingly, studies looking at fecal diversion without resection have found high rates of subsequent clinical remission [17, 18]. Historically, fecal diversion was adopted to induce symptom control and avoid urgent surgery, while allowing for time to optimize medical treatments. However, in mostly retrospective studies, disappointingly high relapse rates after stoma takedown were noted [13, 17]. Defunctioning stoma creation without resection for either colonic or perianal CD did result in acute remission, defined as reduction or cessation of gastrointestinal symptoms, decrease in steroids or other medication, and improvement in general well-being in up to 86% of patients [13]. In another more recent retrospective study, 55% of patients who underwent temporary fecal diversion for perianal or colorectal CD experienced either partial (41%) or complete (14%) remission [17]. Fecal diversion has also been found to be associated with improvement in inflammatory serologic marker derangements (increase in albumin and hemoglobin and decrease in erythrocyte sedimentation rate), decreased need for steroid medication, and improvement in patient quality of life [15, 19]. However, permanent stomas are poorly accepted by patients and, further, carry risks of complications, including peristomal skin irritation, high output, hernia, prolapse, retraction, ischemia, and stenosis [12, 20].

Previously published reports evaluating fecal diversion in CD show low rates of stoma reversal. However, most studies did not address post-diversion medical management and the impact of treatment optimization on outcomes, namely stoma reversal. In a meta-analysis of 16 studies evaluating fecal diversion for the management of perianal CD, restoration of bowel continuity was attempted in 34.5% of patients with success in only 17% [18].

The aim of this study was to determine whether temporary fecal diversion for the indication of severe perianal and/or colonic CD can lead to clinical remission and restoration of intestinal continuity after optimization of medical therapy in patients treated at a tertiary referral center.

Methods

Study Design and Patient Identification

We conducted a retrospective study of adult patients with CD evaluated at the University of Maryland Inflammatory Bowel Disease (UM IBD) Program from 2004 to 2014. All patients had temporary fecal diversion for the indications of: (1) medically refractory/severe colonic CD, (2) medically refractory/severe perianal CD, or (3) medically refractory/severe colonic AND perianal CD. Fecal diversion was intended as “temporary” with the future prospect of stoma reversal. The decision to undergo diversion was based upon IBD provider discretion. Exclusion criteria included preexisting diagnosis of ulcerative or indeterminate colitis, disease confined to the ileum, and age less than 18 years old.

Study Variables

Demographic information was obtained from a prospectively maintained, IRB-approved clinical database that contains clinical information on approximately 1500 patients that are followed at our center. Chart review was utilized to ascertain detailed medical and surgical therapeutic history for identified patients.

Demographic data included gender, race, current age, age at diagnosis, smoking status, CD duration, family history of IBD, and Montreal classification of disease [21]. The diagnosis of CD was confirmed by standard clinical, histopathologic, and endoscopic criteria [22]. Colonic CD behavior was classified as non-stricturing, non-penetrating, stricturing, or penetrating at the time of surgery using the Montreal classification. Perianal disease was categorized as simple or complex according to the AGA classification system using magnetic resonance imaging (MRI) or endoscopic ultrasound (EUS). A simple fistula is one that is superficial, inter-sphincteric, or low trans-sphincteric, has only one opening, and is not associated with an abscess. A complex fistula involves the anal sphincters, has multiple openings or horseshoeing, or is associated with a perianal abscess. Connection to an extraluminal organ also qualifies fistulae as complex [23].

Medical and surgical history was obtained at baseline and post-fecal diversion. Prior history of immunomodulator (6MP/azathioprine/methotrexate) and biologic (infliximab, adalimumab, certolizumab pegol, natalizumab, and ustekinumab) use was extracted. We also evaluated chronic narcotic use. History of prior intestinal and perianal surgeries, including the number and type of perianal operations (fistulotomy, advancement flap, incision and drainage of abscess, and fistula plug), was recorded.

Outcome measures included rates of colectomy and restoration of intestinal continuity (i.e., stoma reversal) at multiple time points over the 2-year following fecal diversion.

Statistical Analysis

We chose to mainly use categorical data for the purpose of this study, and all categorical variables were described in the form of proportions and compared using the Chi-square test. The continuous variables were used for descriptive purposes and were categorized to test for associations with the outcomes of interest. The associations between demographic and clinical variables and the outcomes of interest, ostomy reversal, and colectomy were tested using Chi-square test with the significance at a 0.05 level. All analyses were conducted using Stata 13 [StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP].

Results

Population

Thirty patients with colonic and/or perianal CD who had fecal diversion for refractory or symptomatic disease were included in the study (Table 1). Of these, 12 or 40% underwent ostomy reversal during the study. Overall, 50% were female and nearly one-quarter were African-American. The majority of patients (n = 25) were diagnosed before age 40, and about half were current or former smokers. The duration of disease was similar between groups. There were no significant differences between the group of patients who retained their stoma versus those who ultimately underwent stoma reversal in gender, race, age at diagnosis, smoking status, or disease location or phenotype. Median duration of disease was about 8 years in both groups.

Table 1 Demographic and clinical characteristics
Full size table

Indication for temporary fecal diversion was refractory colitis in 33%, perianal disease in 37%, and both in 30% of patients. Fecal diversion was accomplished with a loop ileostomy in 63% of patients and with end stoma formation in 37% of patients.

Pre-diversion Characteristics

The indication for temporary diversion was statistically different between groups. A greater proportion of patients in the diverted group had both colonic and perianal disease compared to those who achieved intestinal continuity (44 vs 8%, p = 0.03). Among patients with perianal disease, those who had complex fistulae were more likely to remain diverted throughout the study (83%, p = 0.02). Patients who remained diverted also had a higher percentage of prior perianal interventions, namely fistulotomy, drainage of abscess, and seton placement (Table 2). Only 1 of 6 (17%) patients with anal stricture achieved stoma reversal.

Table 2 Comparison of patients with stoma and those who had stoma reversal
Full size table

A majority of patients (87%) had been treated with anti-TNF therapy prior to fecal diversion, of which a large proportion received >1 anti-TNF agent. 50% of patients with complex perianal disease had received two or more biologics before ileostomy, and two-thirds were on combination therapy at the time of ileostomy formation, demonstrating the refractory nature of the disease.

Post-diversion Characteristics

Over one-third (37%) of patients had successful fecal stream restoration within 2 years of temporary diversion (Fig. 1). 70% of patients with colonic only and 25% of patients with perianal with or without colonic disease underwent stoma reversal. The majority of patients had loop ileostomy or colostomy. Patients treated with adalimumab preoperatively and patients on chronic narcotics were more likely to remain diverted (p < 0.01 and p < 0.05, respectively).

Fig. 1
figure 1

Rates of colectomy and ostomy reversal over time

Full size image

Characteristics of Patients with Colectomy

Six patients (20%) underwent colectomy during the course of the study period. Of these, all patients had exposure to anti-TNF therapy and 67% had received combination therapy immediately prior to surgical diversion. After fecal diversion, 5 of these patients had further optimization and alteration in therapy, and 67% (n = 4) received combination therapy (Table 3). One-half of the colectomy patients had anal strictures. Kaplan–Meier curve illustrates cumulative colectomy and stoma reversal rate in patients after diversion (Fig. 2). The number of patients had colectomy increased over time, whereas the stoma reversal rate was starting to plateau by 2 years. Patients who achieved restoration of intestinal continuity did so within the first year after diversion.

Table 3 Characteristics of colectomy patients
Full size table
Fig. 2
figure 2

Natural history of patients after fecal diversion: 2-year follow-up

Full size image

Discussion

In our study, successful restoration of fecal stream after temporary fecal diversion in patients who had optimization of medical therapy was 37%. Among these patients, one ultimately required colectomy for pelvic abscess, and three had recurrent symptoms. Therefore, symptom-free success rate occurred in 26%. A majority of patients who achieved restoration of intestinal continuity did so within the first year after diversion. Our overall colectomy rate was 20% despite optimization of therapy using combination of immunosuppressant and anti-TNF agents (in two-thirds of patients), more than one anti-TNF therapy in some, and natalizumab or non-FDA-approved ustekinumab in others. Patients at greatest risk of remaining diverted were those with severe complex perianal disease, surgical perianal disease, prior adalimumab exposure, and pre-diversion chronic narcotic use.

Historically, specialists caring for patients with Crohn’s disease used temporary fecal diversion as a “bridge to [more definitive] surgery.” The goal of a temporary stoma has been to “buy time” to improve the debilitating symptoms of colonic or perianal disease while improving nutritional markers and initiating medical therapies to control the underlying inflammation. In the 1970s, early clinical response to fecal diversion was high, yet low rates of stoma reversal of 7–5% plagued this approach. However, the introduction of biologic therapies, namely anti-TNF agents, revitalized the idea of true temporary diversion. Several published reports from 2000 to 2009 reported high early clinical response rates of 70–80%, but continued low stoma restoration rates and/or need for rediversion [9, 12, 13, 24,25,26].

In our patients treated at a tertiary referral center, successful symptom-free fecal stream restoration was achieved in 26%. Over 50% of our patient population had already been treated with >1 anti-TNF agent prior to fecal diversion. Despite changes in therapy, including adding an immunosuppressant, changing to an alternative anti-TNF agent, or trialing a different biologic agent, about 17% of all patients required colectomy for medically refractory disease post-diversion. One additional patient required rediversion after colectomy. Other published reports have estimated the rate of fecal stream restoration to be 16–31% in the post-biologic era [12, 19, 27]. A systematic review of existing literature through part of 2015 found that 16.6% of fecal stream diversion led to successful restoration [18]. Although these rates appear bleak at first glance, Sauk et al. [27] observed that the percentage of patients needing fecal diversion has decreased in the biologic era, suggesting that anti-TNF agents may be playing a role in better controlling disease.

Risk factors for failure of fecal stream restoration in our study included surgery for indication of perianal or colonic and perianal disease, complex perianal disease, need for surgical intervention to control perianal disease, and prior adalimumab and chronic narcotic use. Other published reports have also found an association between complex perianal disease and higher failure rates [9]. Presence of colonic disease and anal canal stricture may incur a 30-fold increased likelihood of permanent stoma [24]. Other risk factors have included increased number of surgical procedures and rectal involvement [28]. Pre-diversion immunomodulator use has also been associated with lower likelihood of fecal stream restoration in one study [27]. We did not find an association between prior immunomodulator use and risk of diversion failure. We believe failed stoma restoration in those treated with chronic narcotics may be an indicator of disease severity. Although we do not know what percentage of patients received narcotics to treat perianal disease, patients are often treated with these agents due to the debilitating and painful nature of fistulizing complications and abscesses. Lastly, all patients who underwent colectomy had prior exposure to more than two anti-TNF agents, which indicates that treatment was optimized in these patients before definitive surgery was recommended.

Our study had several strengths. First, we collected data over a span of 2 years post-fecal diversion allowing us to make more robust conclusions regarding the effectiveness of fecal diversion over time. In fact, we were able to establish that most patients achieved the optimal outcome of fecal stream restoration within 1 year of stoma creation and that this effect starts to plateau by the second year. Second, we also collected detailed data regarding medical therapies instituted pre- and postoperatively. Only one other study to our knowledge has discussed the time course of prior anti-TNF and immunomodulator use and its impact on temporarily diverted patients [27]. Third, the patients in our cohort were managed at a tertiary referral center where they receive specialized care by IBD gastroenterologists and colorectal surgeons. At our center, we maintain an IBD repository of our IBD patients, which helps ensure that our data are inclusive of all of our Crohn’s patients who underwent fecal diversion.

The limitations of our study included the retrospective nature of data collection, our small sample size, and the two patients who were lost to follow-up. Although our longer-term follow-up was adequate to have a better sense of disease progression and diversion success and failure rates, two additional patients ultimately underwent colectomy outside of the study time course (by 4 years).

In conclusion, the long-term outcomes of temporary fecal diversion have not changed much in the biologic era. Those at risk for a permanent stoma continue to be individuals with complex perianal disease. Other factors such as prior perianal surgical interventions, treatment with more than one biologic agent, and chronic narcotic use as markers of disease severity may also play a role. Based on our retrospective review, we believe that temporary diversion is in fact “temporary” as it provides time for symptom control, improvement of nutritional parameters, and may lead to acceptance of a permanent stoma. It may also buy time for about one-quarter of patients to attain fecal stream restoration after optimization of medical therapy is achieved. In the future, new therapies may thwart the need for fecal diversion in this patient population, as the need for fecal diversion itself in some patients is a marker for medically refractory disease.