Perianal Crohn's disease findings other than fistulas in a population-based cohort§

Authors


  • Presented in part at the Annual Meeting of the American Gastroenterological Association, New Orleans, LA, May 1–5, 2010 (Gastroenterology 2010;138(5 suppl 1):S-199).

  • Supported by the Mayo Foundation for Medical Education and Research, and made possible by the Rochester Epidemiology Project (Grant Number R01 AG034676 from the National Institute on Aging).

  • §

    Peyrin-Biroulet: Study concept and design, acquisition of data, analysis and interpretation of data, drafting of the article. Loftus: Study concept and design, analysis, and interpretation of data, drafting of the article, critical revision of the article, obtaining funding, study supervision. Tremaine: Critical revision of the article. Harmsen: Analysis and interpretation of data, critical revision of the article. Zinsmeister: Analysis and interpretation of data, critical revision of the article. Sandborn: Study concept and design, critical revision of the article, obtaining funding.

Abstract

Background:

The cumulative incidence of and risk factors for perianal Crohn's disease (CD) for findings other than fistulas are unknown.

Methods:

The medical records of 310 incident cases of CD from Olmsted County, Minnesota, diagnosed between 1970 and 2004, were reviewed for evidence of perianal disease findings other than fistulas. Cumulative incidence was estimated using the Kaplan–Meier method, and associations between baseline factors and time to first event were assessed using proportional hazards regression. Four types of lesions were studied: anorectal strictures, deep anal canal ulcers, anal fissures, and perianal skin tags.

Results:

The 10-year cumulative probability from time of diagnosis was 5.8% (95% confidence interval [CI], 2.6%–8.8%) for anorectal strictures, 6.6% (3.6%–9.6%) for deep anal canal ulcers, 10.5% (6.8%–14.1%) for anal fissures, and 18.7% (13.9%–23.3%) for perianal skin tags. The cumulative probability for any perianal lesion other than fistulas was 21.3% (16.5%–25.8%) at 5 years and 29.2% (23.5%–34.5%) at 10 years. Baseline factors associated with time to first perianal lesion other than fistulas were age (hazard ratio [HR] per 10 years, 0.9; 95% CI, 0.8–0.98; P = 0.026), female gender (HR, 1.7; 95% CI, 1.1–2.7; P = 0.013), and presence of extraintestinal manifestations (HR, 1.7; 95% CI, 1.03–2.8; P = 0.038).

Conclusions:

Perianal lesions other than fistulas occurred frequently during the clinical course of CD. Female gender and extraintestinal manifestations were associated with increased risks for perianal lesions other than fistulas, while older age at diagnosis was associated with a slightly decreased risk. (Inflamm Bowel Dis 2011)

Perianal Crohn's disease (CD) encompasses both fistulizing lesions (fistulas, abscesses, and rectovaginal fistulas) and findings other than fistulas (anal fissures, deep anal canal ulcers, anorectal strictures, perianal skin tags, hemorrhoids, and cancer). Primary lesions are caused by intestinal inflammation and include superficial fissures, deep cavitating anal canal ulcers, and lymphedema.1, 2 Secondary lesions are fistulas and abscesses. Strictures are associated with primary or secondary perianal lesions or are a long-term consequence of the inflammatory process.1, 2 Hemorrhoids are incidental lesions that are not directly related to CD.

Most attention to perianal CD has focused on the diagnosis, epidemiology, and treatment of perianal abscess and fistula. By contrast, few data exist on perianal lesions other than fistulas. Available data suggest that perianal lesions other than fistulas may occur frequently during the course of CD.2 However, only referral center studies have reported the epidemiology of perianal CD findings other than fistulas in adults. Anorectal strictures and anal canal ulcerations were reported in 7%–9%3–5 and 10%–50%4–11 of patients with CD patients, respectively. A recent pediatric inception cohort study found that among 276 children with CD, 13 (4.7%) had only perianal skin tags and anal fissures within 30 days of diagnosis.12 Overall, the cumulative incidence of and risk factors for lesions other than fistulas are unknown.

Perianal CD lesions other than fistulas can be disabling and often requires medical or surgical treatment.2 A better knowledge of their natural history may help physicians managing these lesions. The aim of this study was therefore to estimate the cumulative incidence of and assess factors potentially associated with perianal CD findings other than fistulas in a well-defined population-based cohort.

PATIENTS AND METHODS

Patient Selection

The population-based inception cohort of subjects with CD in Olmsted County, Minnesota, has been previously described13, 14 and was updated through July 2009 for this study.

Study Setting

Olmsted County is situated in southeastern Minnesota and had ≈124,000 people at the 2000 U.S. Census. In 2000, 89% of the population was non-Hispanic white. Although 25% of county residents are employed in healthcare services (versus 8% nationwide), and the level of education is higher (30% have completed college versus 21% nationwide), the residents of Olmsted County are otherwise socioeconomically similar to the U.S. white population.15

Rochester Epidemiology Project

The Rochester Epidemiology Project (REP) is a unique medical records linkage system developed in the 1960s and funded in part by the National Institutes of Health. It exploits the fact that virtually all of the healthcare for the residents of Olmsted County is provided by two organizations: Mayo Medical Center (Mayo Clinic and its hospitals, Rochester Methodist and Saint Marys); and Olmsted Medical Center (a multispecialty clinic and its hospital, Olmsted Community Hospital). Diagnoses generated from all outpatient visits, emergency room visits, hospitalizations, nursing home visits, surgical procedures, diagnostic studies, autopsy examinations, and death certificates for all county residents seen at these two institutions as well as a small number of additional healthcare providers serving the local population have been recorded in a central diagnostic index since 1908.15, 16 In any 3-year period, over 90% of county residents are examined at one of the two major healthcare systems.15 Thus, it is possible to identify all diagnosed cases of a given disease, and all diagnostic examinations, including radiological studies, performed in these subjects. The resources of the REP have been used to identify Olmsted County residents diagnosed with CD from 1940 to 2004.13, 14

Case Ascertainment

With approval from the Institutional Review Boards of Mayo Clinic and Olmsted Medical Center, the records of 310 patients who had not denied permission for research access to their medical records and who were first diagnosed with CD between January 1, 1970 and December 31, 2004 were reviewed. Demographic information including date of birth, date of symptom onset, date of diagnosis, disease type, and extent were previously determined.13, 14 Beginning with the date of diagnosis, all clinical, endoscopic, and proctologic records were reviewed to identify the presence of perianal lesions other than fistulas. Four types of lesions were studied: anorectal strictures, deep anal canal ulcers, anal fissures, and perianal skin tags. Patients were followed through their medical records from the date of symptom onset to the date of last follow-up in their medical record or up to July 2009.

Statistical Analysis

The cumulative incidence (1 minus survival free) of each type of perianal lesion was estimated using the Kaplan–Meier method, and 95% confidence intervals (CIs) estimated using the log transform method (with modified lower limit). The associations between baseline demographic and clinical factors and time from diagnosis to initial occurrence (including those present at diagnosis) of specific findings (anorectal strictures, deep anal canal ulcers, anal fissures, perianal skin tags, and separately, the earliest of these) were assessed using Cox proportional hazards regression models. The risks for specific findings were summarized as hazard ratios (HR; 95% CI) calculated from the estimated coefficients (and their standard errors) in these models.

RESULTS

Baseline Characteristics (Table 1)

Half of the 310 patients were male. Corresponding to the Montreal Classification,17 only 10.6% of patients were diagnosed below the age of 16 years (A1), 57.1% of patients had an age between 17 and 40 years (A2), and approximately one-third of patients had an age at diagnosis above 40 years (A3). Approximately one-third of patients each had ileitis, colitis, or ileocolitis at the time of diagnosis. Approximately 8 out of 10 patients had nonstricturing, nonpenetrating disease (B1), almost 5% of patients had stricturing disease (B2), and 15% had penetrating disease (B3). Perianal disease was noted in 16.7% of patients prior to or within 90 days of CD diagnosis. Half of patients were nonsmokers, while one-third of patients were current smokers. A familial history of IBD was reported by 15% of patients. Extraintestinal manifestations at the time of diagnosis were noted in 15% of patients.

Table 1. Baseline Characteristics of the 310 Patients at the Time of Crohn's Disease Diagnosis
CharacteristicN (%)
Gender 
 Male156 (50.3)
 Female154 (49.7)
Age 
 Less than 16 years (A1)33 (10.6)
 Age between 17 and 40 years (A2)177 (57.1)
 Age more than 40 years (A3)100 (32.3)
Disease location 
 Small bowel (L1)96 (31.2)
 Colitis (L2)102 (33.1)
 Ileocolonic (L3)103 (33.4)
 Gastroduodenal (L4)7 (2.3)
 Not documented2
Disease behavior 
 Nonstricturing/nonpenetrating (B1)248 (81.3)
 Stricturing (B2)14 (4.6)
 Penetrating (B3)43 (14.1)
 Not documented5
Perianal disease prior to or within 90 days of diagnosis 
 Yes51 (16.7)
 No254 (83.3)
Smoking history 
 Current smokers98 (32.9)
 Former smokers45 (15.1)
 Nonsmokers155 (52.0)
 Not documented12
Family history of IBD 
 Yes43 (14.4)
 No255 (85.6)
 Not documented12
Extraintestinal manifestation 
 Yes46 (15.1)
 No259 (84.9)
 Not documented5

Cumulative Incidence of Perianal Lesions Other Than Fistulas

Among the 310 patients, 22 subjects had an anorectal stricture at the time of or after the diagnosis of CD. Only two patients developed anorectal strictures within the first year of diagnosis. In one patient the stricture was not passable by the examining finger or endoscope. The cumulative probability of developing an anorectal stricture from the time of diagnosis was 1.0% (95% CI, 0.2%–2.2%), 5.8% (2.6%–8.8%), and 17.5% (8.2%–27.8%) at 5, 10, and 30 years, respectively (Fig. 1a).

Figure 1.

Cumulative probability of developing individual perianal findings other than fistulas from the time of diagnosis of CD: (a) anorectal strictures; (b) anal canal ulcers; (c) anal fissures; and (d) perianal skin tags.

Beginning with the date of diagnosis, a total of 21 patients developed at least one deep anal canal ulcer. Eleven individuals had an anal canal ulcer within the first year of diagnosis. In 15 patients (71.4%), the anal canal ulcer was symptomatic. The cumulative probability of having or developing an anal canal ulcer at or from time of diagnosis was 5.6% (95% CI, 3.0%–8.3%), 6.6% (3.6%–9.6%), and 8.4% (4.5%–15.6%) at 5, 10, and 30 years, respectively (Fig. 1b).

Thirty-four patients had or developed anal fissures at the time of (n = 5) or after (n = 29) the diagnosis of CD. Sixteen individuals had anal fissures within the first year of diagnosis. In 29 out of 34 patients (85.3%), anal fissures were symptomatic. The cumulative probability of having anal fissures at or following the date of diagnosis was 7.3% (95% CI, 4.3%–10.3%), 10.5% (6.8%–14.1%), and 16.0% (9.8%–24.9%) at 5, 10, and 30 years, respectively (Fig. 1c).

A total of 63 patients had or developed perianal skin tags at the time of or after CD diagnosis. Twenty-eight individuals had perianal skin tags within the first year of diagnosis. In 32 of these patients (50.8%), the perianal skin tags were symptomatic. The cumulative probability of having perianal skin tags at or following the date of diagnosis was 14.7% (95% CI, 10.6%–18.7%), 18.7% (13.9%–23.3%), and 32.2% (22.1%–42.3%) at 5, 10, and 30 years, respectively (Fig. 1d).

For the 93 patients who developed at least one of anorectal stricture, deep anal canal ulcer, perianal fissure, and/or perianal skin tags, the cumulative probability of having or developing any perianal lesion other than fistulas (earliest of the four) was 21.3% (95% CI, 16.5%–25.8%), 29.2% (23.5%–34.5%), and 44.8% (34.0%–57.4%) at 5, 10, and 30 years, respectively (Fig. 2).

Figure 2.

Cumulative probability of developing any type of perianal lesions other than fistulas from time of diagnosis.

Risk Factors for Perianal Lesions Other Than Fistulas

We assessed potential associations between time to perianal lesion other than fistulas and six baseline characteristics: age, gender, disease location, smoking status, family history of inflammatory bowel disease, and the presence of extraintestinal manifestations. In univariate proportional hazards regression analyses, the only baseline factor associated with time to first anal canal ulcer was the presence of extraintestinal manifestations (HR, 2.9; 95% CI, 1.2–7.2; P = 0.02). None of the factors studied was associated with time to first anal fissure. The baseline factors univariately associated with time to first anal canal ulcer or anal fissure were age (HR per 10 years, 0.8; 95% CI, 0.7–0.96; P = 0.018) and the presence of extraintestinal manifestations (HR, 2.5; 95% CI, 1.4–4.5; P = 0.003). The baseline factors univariately associated with time to first perianal skin tags were female gender (HR, 2.9; 95% CI, 1.7–5.1; P < 0.001), former cigarette smoker status relative to nonsmokers HR, 0.2; 95% CI, 0.05–0.87; P = 0.03), and the presence of extraintestinal manifestations (HR, 2.9; 95% CI, 1.7–5; P < 0.001). The baseline factors univariately associated with time to earliest of any perianal lesion other than fistulas were age (HR per 10 years, 0.8; 95% CI, 0.7–0.97; P = 0.017), female gender (HR, 1.9; 95% CI, 1.2–2.8; P = 0.004), and the presence of extraintestinal manifestations (HR, 2; 95% CI, 1.2–3.3; P = 0.005) (Table 2).

Table 2. Univariate Associations Between Baseline Factors and Time to Earliest of Any Perianal Lesion Other Than Fistulas in 310 Patients from Olmsted County, Minnesota
CharacteristicHazard Ratio95% CIP-value
Gender   
 Male1.0Ref. 
 Female1.91.2-2.80.004
Age1.00.99-1.90.43
Disease location   
 Small bowel (L1)0.70.4-1.30.27
 Colitis (L2)1.0Ref. 
 Ileocolonic (L3)1.00.6-1.60.1
 Gastroduodenal (L4)0.90.2-40.96
Disease behavior at the 90-day baseline   
 Nonstricturing/nonpenetrating (B1)1.0Ref. 
 Stricturing (B2)001
 Penetrating (B3)0.70.3-1.30.2
Smoking history   
 Current smokers1.30.8-20.26
 Former smokers0.50.2-1.10.08
 Nonsmokers1.0Ref. 
Family history of IBD   
 Yes1.00.5-1.70.9
 No1.0Ref. 
Extraintestinal manifestation   
 Yes21.2-3.30.005
 No1.0Ref. 

In a multiple variable proportional hazards regression model, the baseline factors independently associated with time to earliest of any perianal lesion other than fistulas were age (HR per 10 years, 0.9; 95% CI, 0.8–0.98; P = 0.026), female gender (HR, 1.7; 95% CI, 1.1–2.7; P = 0.013), and the presence of extraintestinal manifestations (HR, 1.7; 95% CI, 1.03–2.8; P = 0.038) (Table 3).

Table 3. Multiple Variable Model Retaining Significant Baseline Factors for Time to Earliest of Any Perianal Lesion Other Than Fistulas in 310 Patients from Olmsted County, Minnesota
CharacteristicHazard Ratio95% CIP-value
Gender   
 Male1.0Ref. 
 Female1.71.1-2.70.013
Age per 10 years0.90.8-0.980.026
Extraintestinal manifestation   
 Yes1.71.03-2.80.038
 No1.0Ref. 

DISCUSSION

This is the first study investigating the cumulative incidence of and risk factors for perianal CD other than fistulas in a population-based cohort. One pediatric population-based study assessed the frequency of perianal lesions at CD diagnosis.12 All other available studies on the epidemiology of nonfistulizing perianal lesions were referral center-based.2 In addition, most studies focused on prevalence and clinical course of these lesions or only reported their crude incidence rate.2 The cumulative incidence of perianal CD other than fistulas remained unknown.

In this population-based cohort the cumulative probability of developing anorectal strictures from time of diagnosis was 5.8% (95% CI, 2.6%–8.8%) at 10 years. This figure is consistent with findings among 754 patients with CD seen between 1960 and 1965 at the Mayo Clinic, where ≈7% patients had anorectal stenosis.5 Anorectal stricture occurred in 12 of 160 patients (7.5%) with CD involving the colon and seen at the Mount Sinai Hospital, New York.3 Among 202 patients with CD who had been examined in Birmingham, United Kingdom, during the year 1984 to assess the frequency of perianal disease, 19 (9%) had anal stricture.4

In this population-based cohort the cumulative probability of having a deep anal canal ulcer or an anal fissure from time of diagnosis was 17.1% (95% CI, 12.5–21.6) at 10 years. Among 754 patients with CD seen between 1960 and 1965 at Mayo Clinic, Rochester, ≈10% patients had anal fissures.5 Among 233 new patients with CD referred to Saint Mark's Hospital, London, UK, between 1977 and 1983, 61 (26%) had anal fissures.7 Among 202 patients with CD who had been examined during the year 1984 to assess the frequency of perianal disease, 12 (6%) had anal ulcers and 38 (19%) had anal fissures.4 Among 153 patients attending a CD follow-up clinic at the Department of surgery in Birmingham, UK, 53 (35%) had developed deep anal canal ulceration.8 The charts of 1098 patients seen at the Lahey Clinic (Burlington, MA) between 1957 and 1978 were reviewed; anal fissures were noted in 29% of patients.9 By reviewing the case notes of 151 consecutive patients with CD presenting to surgeons, Hobbiss and Schofield10 identified 23 (15%) patients who had anal canal ulceration. In a retrospective study, among 225 patients with luminal CD anal canal ulcerations were present in 50% of patients.6 Among 101 patients consecutively referred to a French referral center for active CD between 1991 and 1994 and prospectively evaluated by experienced proctologists, 42.5% had anal canal ulceration.11

While anal canal ulcerations are classically described as painless, pain has been reported in up to 70% of patients in referral center-based series.7, 18 We found that 71.4% and 85.3% of patients with CD had symptomatic anal canal ulcers and anal fissures, respectively. This is line with the results of a retrospective study from a U.S. referral center showing that anal fissures were symptomatic in 84% of cases18 and a French prospective study in which anal canal ulcers were associated with unremitting pain in 56% of cases.11

In this population-based cohort, the cumulative probability of having perianal skin tags was 18.7% (95% CI, 13.9–23.3) at 10 years. Among 153 patients attending a CD follow-up clinic at the Department of surgery in Birmingham, UK, 37 (24.2%) had perianal skin stags associated with anal fissures or fistulas.8 Among 202 patients with CD who had been examined during the year 1984 to assess the frequency of perianal disease, 75 (37%) had perianal skin tags,4 and 86% of these lesions were asymptomatic.4 Recently, perianal skin tags were photographed in 170 consecutive patients with IBD and perianal skin tags in the course of 1 year.19 Perianal skin tags were found in 75.4% of patients with CD.19 In contrast to anorectal strictures and anal canal ulcerations, perianal skin tags are less often symptomatic and usually do not require specific management. In this population-based cohort, the proportion of symptomatic perianal skin tags was relatively high, occurring in half of the 63 patients who developed this type of lesions after CD diagnosis. It is therefore possible that physicians pay less attention to these lesions and do not systematically mention their presence on proctology and endoscopy reports. Therefore, even though we confirmed that the most common manifestation was perianal skin tags, the frequency of perianal skin tags might have been underestimated in our population-based study.

To our knowledge, the risk factors for having perianal CD other than fistulas have never been formally assessed. Due to the relatively small number of events, we were only able to assess risk factors for developing any type of lesions other than fistulas (anorectal strictures, anal canal ulcers, anal fissures, and/or perianal skin tags). Among baseline characteristics, we identified female gender as having the strongest association with lesions other than fistulas in multivariate analysis. Females were more likely to develop any lesion other than fistulas. Two risk factors were positively associated with the development of lesions other than fistulas, namely, age and presence of extraintestinal manifestations. However, due to low statistical significance, this remains to be confirmed in independent population-based cohorts.

In conclusion, perianal lesions other than fistulas occurred frequently during the clinical course of CD, with a cumulative risk of almost 30% at 10 years. Female gender, extraintestinal manifestations, and age were negatively or positively related to the presence of perianal lesions other than fistulas.