Introduction

Crohn’s disease is a chronic, idiopathic, incurable disease of the gastrointestinal tract, and approximately 30 % of patients suffer from perineal Crohn’s disease (PCD) [1]. Skin tags, fissures, ulcers, abscesses, and fistulae produce symptoms in Crohn’s patients ranging from mild discomfort to fecal incontinence, chronic pain, and life-threatening sepsis. Multiple therapies, medical and surgical, have emerged to treat these conditions, attesting to the frequently refractory and recurrent nature of the disease. In treating PCD, the physician must balance aggressive control of septic foci, preservation of anorectal function, and a realistic appreciation of the chronicity of PCD.

PCD presents before or at the same time as intestinal symptoms in 36–81 % of patients with Crohn’s disease [2], whereas a small subset of patients has isolated perineal disease [3]. The incidence of PCD is higher in patients with more distal intestinal Crohn’s disease, and its presence has been noted to predict a more disabling disease course [4–6]. Perineal disease has also been noted to be more common in African-Americans and Hispanics as compared to whites [7] .

Despite multiple refinements in the treatment of Crohn’s disease over the years, many patients with PCD do poorly and require multiple procedures. A significant portion of these patients ultimately undergo fecal diversion, with reported rates of diversion ranging from 10 to 50 %. The need for diversion is higher in patients who have colonic disease and anal stenosis [4, 6].

The goal of the present study is to identify the population of patients ultimately requiring proctectomy/ fecal diversion and attempt to identify modifiable predictors for this unfortunate outcome. Our secondary goal was to describe the extent of utilization of medical care in all patients who presented to our institution with perineal Crohn’s disease requiring surgery.

Methods

Study population

Patients with Crohn’s disease who underwent operative treatment for perineal Crohn’s disease (PCD) between 2000 and 2009 at the Massachusetts General Hospital for Crohn’s disease-associated anal fistulas were identified and studied in a retrospective fashion. Patients with anorectal conditions other than fistula or abscess due to Crohn’s disease were excluded. Female patients with Crohn’s disease who presented with or developed rectovaginal fistulas were included. The patients’ operative reports, pathology, radiology and endoscopy reports, outpatient notes, and inpatient records were reviewed. The following data were collected: age, gender, duration of CD, age at the first intervention, history of prior bowel resection or anorectal operation, current or former infliximab therapy, inflammation status of bowel mucosa, and smoking status. In addition, we quantified the degree of utilization of medical care required in these patients by tallying the number of medical and surgical clinic appointments, visits for diagnostic workups, Emergency Room encounters, and inpatient admissions made by our cohort of patients to address their anal fistula symptoms. IRB approval was obtained to analyze these data. We then followed the course of these patients from their original presentation to January 1, 2012.

Classification of anal fistulas and their treatment

Patients’ fistulas were classified into simple and complex based on criteria determined by the American Gastroenterological Association Technical Review Panel [8]. Simple fistulas were superficial, located low in the rectal canal, and had a single external opening (i.e., intersphincteric/low transphincteric anal fistula). Complex fistulas were located high in the rectum, had more than one external opening, or involved an adjacent structure (i.e., multiple and/or high transphincteric fistula). Rectovaginal fistulas were classified as complex independent of their proximity to the dentate line.

All surgical procedures performed to address anal fistulas were recorded and categorized as either incision and drainage of abscesses alone, incision and drainage of abscess with fistulotomy, seton placement with or without incision, drainage of abscess, or finally surgery aimed at closing internal anal opening (e.g., endoanal advancement flap repair or anal plug insertion). All patients who ultimately underwent fecal diversion with either a loop or end ileostomy (with or without concomitant proctectomy) were identified.

Definition of fistula healing

Patients with no signs or symptoms of the disease 3 months after a surgical intervention were considered healed. Patients were considered unhealed if, at 3 months, they were found to have a recurrent or persistent fistula, unresolved pain, or persistent induration.

Statistical analysis

To analyze our primary outcome, we compared the patients who ultimately required fecal diversion to those who were able to be spared a stoma. Stoma + patients were compared to those who avoided diversion as far as demographics, surgical procedure, postoperative infliximab, and presence of proctitis (with as-appropriate chi square and t tests). Logistic regression was then utilized to predict the need for diversion, and the following a priori selected variables were included as covariables in this analysis: history of more than three operations, presence of rectovaginal fistula, perioperative infliximab therapy, smoking, presence of proctitis, and whether their surgery aimed at closing the internal anal opening.

Our secondary outcome—utilization of medical care—was reported as a cumulative number of surgeries, visits to the emergency room, the cumulative number of hospital admissions, and the cumulative number of specialist evaluations underwent by all 82 patients. In addition, means and ranges of the number of visits were reported per individual patient.

All statistical analyses were calculated by using SPSS v 16. p < 0.05 was considered significant in comparison tests.

Results

Our retrospective search covering surgical procedures performed in the time period between 2000 and 2009 initially identified 203 surgical procedures aimed at treating patients with perianal Crohn’s disease. This search did not identify the patients with very mild PCD who never had a procedure. Thirty-one patients were excluded when a careful review of their records revealed that the surgery was performed for a different diagnosis. We were left with 172 surgical procedures performed on 81 patients (some had more than one—mean 2.1 procedures/patient). The mean age of these patients was 31 years old (range 7–78) and 52 % were male. The mean duration of their Crohn’s disease was 15 years (range 1–56). Twenty-eight patients (34.6 %) were smokers at the time of their presentation. Thirty-five (43.2 %) had prior abdominal surgery.

First surgery

At the time of their first surgical intervention, 52 out of 81 patients (64.2%) patients complained of perianal drainage and 29 (35.8 %) had a concomitant anal abscess in association with their fistula. At first surgery, 22 % were found to have a simple (intersphincteric/low transphincteric) anal fistula, 61 % had complex (multiple and/or high transphincteric) fistulas, and 17 % were classified as other because their medical records were not specific enough to allow a reliable allocation. In women, 20.5 % had a rectovaginal fistula, and 49.4 % had an associated abscess. At this time, eight patients (9.9 %) had a diversion procedure. All had prior drainage procedures prior to the study period and were symptomatic despite draining setons. Thirty-one (38.3 %) patients underwent incision and drainage, 18 (22.2 %) underwent fistulotomy, 23 (28.4 %) had seton placement with or without concomitant superficial fistulotomy, and 1 patient had an anal plug insertion.

Following their first surgery, 42 patients (57.1 % of the 73 non-diverted patients) were considered healed at 3 months. These rates were highest in patients treated with fistulotomy (Table 1). There was no difference in healing between patients treated with and without postoperative infliximab therapy (52.4 vs. 57.1 %; p = 0.708).

Table 1 First surgical procedure performed at presentation and observed rates of healing
Full size table

All surgical procedures

Cumulatively, 81 patients underwent 172 surgical procedures. Forty-two patients healed after their first surgery. The remaining 39 patients who did not heal at first attempt underwent 91 further surgical procedures. Nineteen (23.5 %) of these patients needed three or more surgical procedures within this time period. Ultimately, only seven more patients (49 total) healed without fecal diversion, thus making our overall fistula healing rate 60 % without fecal diversion and 83.9 % with fecal diversion. Thirteen patients (16 %) retained permanent setons that allowed for reasonable symptom control without diversion. The rates of surgical success and complete fistula healing without diversion were highest in the group of patients who underwent immediate or delayed fistulotomy (i.e. after interim seton placement) (Table 2).

Table 2 All surgical procedures performed cumulatively and observed rates of healing
Full size table

Fecal diversion

Over the course of the study period, 19 patients (23.5 %) ultimately required fecal diversion. Patients treated with fecal diversion were similar to those who avoided diversion on initial presentation (Table 3). On multivariable logistic analysis, patients undergoing more than three operations (OR = 10.9, p = 0.006) and women with rectovaginal fistula (OR = 3.88, p < 0.01) were at a high risk for stoma. Modifiable factors such as infliximab (OR = 0.8, p = 0.779), smoking (OR = 0.54, p = 0.426), proctitis (OR = 1.57, p = 0.525), and surgery aimed at closing internal anal opening (OR = 6.4, p = 0.67) did not alter outcome.

Table 3 Predictors of fecal diversion on univariate analysis
Full size table

Overall utilization of medical care

Between surgery, preoperative visits, infliximab infusions, endoscopies, visits to primary care physicians, and evaluations by their respective gastroenterologists and surgeons, these 81 patients with perineal Crohn’s disease required numerous visits to their specialty providers. In addition to undergoing 172 surgical procedures (range 1–10; mean, 2.1), these patients also required a total of 2713 patient visits to see their physicians (mean of 33.5, range 1–118), 277 visits to the emergency room (mean 3.4, range 0–23), and 372 hospital admissions (mean 4.6, range 0–20). In total, these patients tallied 2200 hospitalized days, and 2713 outpatient visits (range 1–118; mean, 34.5) for PCD alone or PCD in combination with Crohn’s disease at other sites. This calculation did not include visits to other hospital emergency rooms, psychiatrists, and trips for imaging studies.

Conclusions

In this retrospective study of 81 patients treated for perianal Crohn’s disease, we found that the management of this disease comes at a significant social burden to the patients. We found that over a third of these patients had operations prior to the study period, and by the end of our study, the average patient underwent 2.1 operations and spent more than 30 days dedicated to the care of their disease on an inpatient or outpatient basis. Despite these heroic and costly efforts, by the conclusion of the study period, 19 of 81 patients (23.5 %) underwent fecal diversion. The diverted patients were more likely to be women with rectovaginal fistulas and patients requiring more than three operations. Other factors thought to influence healing in Crohn’s disease, such as infliximab treatment, smoking, and the presence of proctitis were not found to be predictive of the need for diversion.

Our data regarding rates of diversion are consistent with other recently published data on this complicated subject. Fry et al. reported that their patients with suppurative perianal Crohn’s required proctectomy in 12 % of cases [9]. Another series by Galandiuk et al.. found the rates of proctectomy to be as high as 50 %, but their 86 patients with perianal Crohn’s disease had higher rates of anal strictures [6]. Obviously, the rates of ultimate fecal diversion will vary based on patient characteristics, and our rate of 23.5 % is within the reported range [6, 9].

Similar to others, we have treated intersphincteric and low transphincteric fistulas with primary fistulotomy and higher complex fistulas with draining setons [10, 11] Infliximab was often used in these patients, as long as there was no evidence of residual undrained sepsis [12, 13]. Unfortunately, we found that patients with perineal Crohn’s did best if they healed after their first surgery. Patients who failed their first surgical attempt were mostly unsuccessful in healing their long-term disease. Thus, in our population, 42 patients healed after their first surgery. Despite numerous further procedures, only 7 more patients of the remaining 39 healed. In the interim, these patients required hundreds of specialty visits to see their doctors and had numerous expensive infusions of infliximab and uncounted imaging tests.

Our data regarding the overall impact of infliximab and subsequent surgery on the rates of healing of this disease once the first surgery fails is also unfortunately consistent with others, demonstrating poor rates of fistula closure. Our study found that the drug did not alter rates of ultimate diversion, though one could argue that it was only administered in 53 % of the patients who were ultimately diverted. Though this argument is valid in theory, we would suggest that our rates of overall fistula healing (60 %) is similar to a recent study of 29 patients with PCD treated with infliximab in all comers, where the authors report a 67 % success rate [14]. Similarly, another study of 22 patients with perianal fistulizing Crohn’s reported a 44 % fistula relapse rate in their patients who were treated with combination seton drainage and infliximab [12]. Finally, another larger retrospective study of 226 patients with PCD also reported a 60 % healing rate in these complicated patients [15]. Similar to our study, this study by Gaertner et al. also did not find infliximab infusions to impact outcomes in this difficult cohort of patients.

These data are sobering. In this era of rising medical costs, one cannot help but wonder whether or not we, as surgeons and caretakers, of this difficult condition owe it to our patients and to the society at large to develop a more rational algorithm towards this condition. It is important to educate patients regarding the sad truth that as the number of surgical interventions for perineal disease increase, so does the chance of them failing and needing diversion. Furthermore, what we desperately owe our patients is a concerted effort to seek a definitive cure of this debilitating condition through research and surgical innovation.