Crohn’s disease is a chronic IBD, characterized by transmural segmental inflammation that can occur in any portion of the GI tract from the oral cavity to the anus. It has long been recognized as a protean disorder that incorporates a range of disease phenotypes, including the development of fibrotic strictures, perforation, abscess formation, and fistulization.1 A percentage of patients with Crohn’s disease (CD) will develop a variety of perianal complications during the course of the disease, which range from mild problems requiring minimal intervention, such as anal skin tags, to severe fistulizing disease resulting in proctectomy. In addition to the morbidity associated with the perianal lesions themselves, the luminal disease follows a more severe disease course and patients with perianal Crohn’s disease (PCD) often require more intensive medical and surgical management overall.2,3
There has been a wide range in the reported frequency of perianal involvement in CD (10%–80%),4 and perianal symptoms can occur at any time relative to the diagnosis of CD itself.5,6 Factors responsible for this wide variation in incidence include the duration of follow-up, the source of the patients (community versus referral service), and discrepancies in the classification and inclusion of perianal lesions in different studies. Studies that include only anal fistulas and abscesses are more likely to report a lower rate of perianal involvement than those that include other perianal lesions such as skin tags, fissures, and hemorrhoids. Several classification systems have been proposed to cover the wide of variety of perianal lesions that can complicate CD. These include the Cardiff Classification7 and a system proposed by Buchmann et al8 in 1980. However, because of poor clinical utility, none have gained widespread acceptance prompting the American Gastroenterological Association (AGA) to publish an empiric classification in their recent Technical Review on Perianal Crohn’s Disease.9
Few previous studies have attempted to classify all symptomatic perianal lesions presenting in CD patients, and those that have done so have arisen from tertiary referral centers.10–12 Studies from tertiary referral center practices are likely to overestimate the frequency of perianal involvement, because they contain a higher proportion of more severe CD cases with a greater rate of perianal involvement in comparison with population-based studies. Although previous population-based studies have reported the overall rate of perianal involvement in CD, none have fully classified and reported all symptomatic perianal lesions. The present study aimed to determine the rate of all symptomatic perianal CD lesions in a population-based cohort classified according to the AGA guidelines and to investigate the temporal relationship of the onset of perianal disease to CD diagnosis.
MATERIALS AND METHODS
This study was performed as part of the Canterbury Inflammatory Bowel Disease Project, a large population-based study aimed at describing the epidemiology of IBD in the Canterbury region and to study the clinical, environmental, and genetic risk factors for IBD. The methods used in data collection in this project have been described previously13 and are summarized here.
The Canterbury region is located on the east coast of the South Island of New Zealand and had a population of 464,700 in 2005 when recruitment for the study was completed.14 Patients were initially recruited to the project by the use of a multifaceted approach. Over 3 years, surgeons and gastroenterologists referred all IBD patients from public and private specialist clinics. It was also possible for patients to self-refer for inclusion as a result of extensive advertising with general practices, specialist clinics, the Crohn’s and Colitis support group, and the public media. Capture-recapture techniques estimated that at least 91% of patients with IBD residing in the region were included in the cohort. IBD cases were confirmed with the use of standard diagnostic criteria.15
Demographic information and clinical data, together with IBD phenotype and IBD-related surgical procedures, were extracted retrospectively from patient medical records. All patients with symptomatic perianal lesions noted in the medical record at any time point relative to the diagnosis of IBD were included. This was achieved by meticulously reviewing the hospital and specialist medical records and, where necessary, those of general practitioners. Perianal lesions were classified according to the system proposed in the AGA technical review on perianal CD.9 This study defines PCD lesions as skin tags, hemorrhoids, fissures, ulcers, fistulas, abscesses, and/or strictures. Fistulas are defined as “simple” if they are superficial or low inter- or transphincteric with a single external opening and are not associated with a perianal abscess or stricture. In contrast, “complex” fistulas are high (high intersphincteric, transsphincteric, extrasphincteric, or suprasphincteric), may have multiple external openings, and may be associated with a perianal abscess, rectovaginal fistula, or stricture. Because of difficulties in accurately differentiating between anal fissures and ulcers from descriptions in the clinical record, these 2 conditions were recorded together.
All collected data was entered into a custom-built Microsoft Access (Microsoft, Redmond, WA) database. Kaplan-Meier curves were used to demonstrate the proportion of patients developing PCD over time. Statistical analysis was performed with the use of SPSS version 19 (SPSS, Chicago, IL). Ethical approval was obtained from the regional ethics committee, and all patients entering the study provided written informed consent.
A total of 1421 patients were recruited into the Canterbury IBD Project, of which 715 had CD. The median age was 40 years (range, 4–93 years) and 422 (59.0%) were female. The median follow-up from CD diagnosis was 6.3 years (range, 2 months to 65 years). One hundred ninety patients (26.6%) had symptomatic perianal lesions. The median age of the patients with perianal involvement was 37 years (range, 4–82 years) and 111 (58.4%) were female. Median follow-up from CD diagnosis was 9 years (range, 2 months to 45 years).
One hundred eighty-six patients had complete data on the date of first presentation with perianal disease. Figure 1 summarizes the temporal relationship of the presentation of perianal lesions to the diagnosis of CD. There was a broad range of onset of perianal symptoms, from 18 years before diagnosis to 33 years after diagnosis. In 32 patients (17.2% of the PCD group), the perianal lesion preceded the diagnosis of CD by more than 6 months. In 50 patients (26.9%), perianal disease presented from 6 months before to 6 months after the diagnosis of CD, whereas perianal disease was first observed more than 6 months after CD diagnosis in the remaining 104 (55.9%) patients. With the use of Kaplan-Meier methods, the cumulative probabilities of developing any PCD were 0.295 and 0.427 at 10 and 20 years after diagnosis (Fig. 2). When considering only perianal fistulas, the cumulative probabilities were 0.169 and 0.283 at 10 and 20 years after diagnosis (Fig. 3).
The frequencies of different perianal lesions, classified according to the AGA guidelines, are presented in Table 1. Perianal fistulas represented the most common lesion identified (50% of perianal CD patients), and more than half of these fistulas were classified as complex. Perianal abscesses were also common. Fissures, skin tags, strictures, and hemorrhoids made up the remainder of the lesions in decreasing order of frequency. The perianal lesion included an abscess and/or fistula in 131 patients or 18.3% of the total CD sample.
This study estimated the rate of perianal fistulae at 28.3%, and all perianal lesions at 42.7% at 20 years after diagnosis. Although the onset of perianal symptoms in CD was clustered around the time of diagnosis, the temporal relationship between perianal symptoms and CD diagnosis was highly variable. Perianal CD is now recognized as an important predictor of a severe disease course,16 such that PCD patients require significant medical and surgical interventions and consume considerable health care resources.17,18 Despite this, the reported rate and definition of PCD has varied widely in studies to date. The strength of this study was that it used stringent methodology to obtain population-based data capturing an estimated 91% of IBD cases within the region. To our knowledge, the classification provided is the only detailed classification of perianal CD to come from a population-based cohort to date.
A further strength of this study was the clear definition of PCD lesions with the use of the empiric classification proposed by the AGA in their recent Technical Review on Perianal Crohn’s Disease.9 This article recommends clinical examination to document the presence of any of perianal lesions. Perianal CD fistulas can be classified as either simple or complex on the basis of their anatomic and associated features. Simple fistulas are low (superficial, low intersphincteric, or low transsphincteric) without associated abscess or stricturing. Complex fistulas, on the other hand, are high (high intersphincteric, transsphincteric, extrasphincteric, or suprasphincteric) and may be associated with multiple external openings, an abscess, rectovaginal fistula, or stricture. This classification has significant clinical relevance, because simple fistulas are more straightforward to treat and have higher rates of healing.19
Despite these strengths, there are also several methodological aspects of this study that could decrease the accuracy of the estimated rate of PCD. First, this study assumed that all perianal lesions that occurred in patients with an eventual diagnosis of CD were related to CD. Many of the lesions associated with PCD also occur commonly in the general population in the absence of CD. Idiopathic anal fissure and cryptoglandular fistula-in-ano are 2 examples of such conditions. Retrospectively differentiating these lesions was not possible; hence, some patients with perianal lesions not necessarily related to CD may have been included. This dilemma is common to any retrospective study of CD and indeed remains an issue in prospective studies and clinical practice itself. In practice, any CD patients with perianal lesions, whether truly CD related or idiopathic, require special consideration. Different management approaches are required in CD patients; hence, the inclusion of all lesions is justified.
Second, the retrospective identification and classification of PCD lesions used in this study may be inaccurate in some cases. Classification was performed by using public and private hospital records, including inpatient and outpatient notes and operation and radiology reports. These clinical records were compiled by doctors of varying clinical experience, which may have led to inaccuracy in diagnosis. To address this, where doubt regarding classification existed, treating consultant gastroenterologists and surgeons were contacted for clarification. By use of this methodology, only 1 of 95 fistulas could not be classified according to the AGA guidelines.
No study has previously reported the rate of all symptomatic perianal lesions in a population-based cohort, although 3 studies from such cohorts report the frequency of perianal fistulas in the range 13.7% to 26%.20–22 Tang et al21 reported 21.7% of patients had perineal fistulas, with a median follow-up of 24 years; and, in the Stockholm County cohort, 13.7% were diagnosed with a perianal fistula.22. Unfortunately, the median follow-up in the later study was not documented, making comparison with the present results difficult. Schwartz et al20 used Kaplan-Meier methodology similar to that used here and found a cumulative incidence of perianal fistulas of 26% at 20 years in 169 patients in Olmsted County. The present, larger study confirms a very similar rate of perianal fistulae of 28.3% at 20 years and adds to this by estimating the overall rate of perianal involvement, including all symptomatic lesions, at 42.7% at 20 years after diagnosis.
This is also the first report from a population-based CD cohort to classify the range of different symptomatic perianal CD lesions. The most frequently cited data documenting the range of different perianal lesions comes from a referral center cohort published by Keighley and Allan.11 This series reported on 202 CD patients presenting to a CD follow-up clinic over 1 year who were examined to determine the presence of perianal lesions. One hundred ten of the patients had evidence of PCD. Keighley and Allan demonstrated a slightly higher rate of significant perianal lesions, including abscesses and fistulas, than that demonstrated in the present study. This would be expected from a referral center cohort in which severe CD, and therefore PCD, tends to be overrepresented. That the difference is not greater between the 2 studies may be explained by the longer duration of follow-up in the present study (median, 9 years) compared with Keighley and Allan’s data, which considered only patients presenting to a follow-up clinic over the course of 1 year. The present study also demonstrates that there is a wide variation in the timing of perianal disease presentation relative to the diagnosis of CD; the longer follow-up in this study thus provides a more representative estimation of the burden of PCD over time.
The timing of perianal symptoms relative to the diagnosis of CD has been explored in other cohorts with comparable results to those presented here. Williams et al5 reviewed 1098 patients with CD presenting to the Lahey Clinic in Massachusetts over a 10-year period. In the majority (64%) of the 242 patients with PCD, perianal manifestations developed after intestinal disease over a wide range of time from 2 weeks to over 10 years. Hellers et al6 investigated perianal fistulae in a population-based cohort from Stockholm and found that in 20% perianal fistulae preceded diagnosis by more than 6 months, in 38% perianal fistulae occurred between 6 months before and 3 months after the diagnosis, and, in the remaining 40%, occurred more than 3 months after diagnosis. The present study similarly demonstrated that a significant proportion of patients develop perianal symptoms before intestinal disease, with 17% presenting more than 6 months before a CD diagnosis and 32.2% before or at the time of diagnosis of CD. In addition, the present study, with longer follow-up, documents an even more dramatic range of first perianal presentation from 18 years before to 32 years after the diagnosis of CD.
This purely descriptive study was undertaken to accurately document the true rate of PCD and highlight the importance of a standardized definition of PCD, recognition of the relevance of the population under investigation, and the importance of adequate duration of follow-up. This approach has also already allowed meaningful investigation of the factors predisposing to PCD in this cohort.23 The challenge for future research remains to better define both these factors and those governing response to emerging medical and surgical therapies for PCD.
The authors thank the patients who were involved in this study, the Christchurch and Ashburton Crohn’s and Colitis Support Groups, and the Canterbury Ostomates Society. They also thank the specialists and staff of Christchurch and Ashburton Hospitals, Christchurch Gastroenterology Group, Christchurch Surgical Associates, Intus at the Oxford Clinic, and others who assisted with the recruitment of patients.
1. Ardizzone S, Porro GB. Perianal Crohn’s disease: overview. Dig Liver Dis. 2007;39:957–958
2. Tarrant KM, Barclay ML, Frampton CMA, Gearry RB. Perianal disease predicts changes in Crohn’s disease phenotype-results of a population-based study of inflammatory bowel disease phenotype. Am J Gastroenterol. 2008;103:3082–3093
3. Beaugerie L, Seksik P, Nion-Larmurier I, Gendre JP, Cosnes J. Predictors of Crohn’s disease. Gastroenterology. 2006;130:650–656
4. McLeod RS. Management of fistula-in-ano: 1990 Roussel Lecture. Can J Surg. 1991;34:581–585
5. Williams DR, Coller JA, Corman ML, Nugent FW, Veidenheimer MC. Anal complications in Crohn’s disease. Dis Colon Rectum. 1981;24:22–24
6. Hellers G, Bergstrand O, Ewerth S, et al. Occurrence and outcome after primary treatment of anal fistulae in Crohn’s disease. Gut. 1980;21:525–527
7. Hughes LE. Clinical classification of perianal Crohn’s disease. Dis Colon Rectum. 1992;35:928–932
8. Buchmann P, Alexander-Williams J, Alexander-Williams J. Classification of perianal Crohn’s disease. Clin Gastroenterol. 1980;9:323–330
9. Sandborn WJ, Fazio VW, Feagan BG, Hanauer SB. AGA technical review on perianal Crohn’s disease. Gastroenterology. 2003;125:1508–1530
10. Fielding JF. Perianal lesions in Crohn’s disease. J R Coll Surg Edinb. 1972;17:32–37
11. Keighley MR, Allan RN. Current status and influence of operation on perianal Crohn’s disease. Int J Colorectal Dis. 1986;1:104–107
12. Platell C, Mackay J, Collopy B, Fink R, Ryan P, Woods R. Anal pathology in patients with Crohn’s disease. Aust NZ J Surg. 1996;66:5–9
13. Gearry RB, Richardson A, Frampton CM, et al. High incidence of Crohn’s disease in Canterbury, New Zealand: results of an epidemiologic study. Inflamm Bowel Dis. 2006;12:936–943
15. Lennard-Jones JE. Classification of inflammatory bowel disease. Scand J Gastroenterol Suppl. 1989;170:2–6
16. Eglinton TW, Gearry RB. Clinical factors predicting disease course in Crohn’s disease. Expert Rev Clin Immunol. 2010;6:41–45
17. Eglinton T, Reilly M, Chang C, Barclay M, Frizelle F, Gearry R. Ileal disease is associated with surgery for perianal disease in a population-based Crohn’s disease cohort. Br J Surg. 2010;97:1103–1109
18. Lion M, Barclay M, Gearry R, Eglinton T. The cost of paediatric and perianal Crohn’s disease in Canterbury, New Zealand. NZ Med J. 2012;125:1349
19. Bell SJ, Williams AB, Wiesel P, Wilkinson K, Cohen RC, Kamm MA. The clinical course of fistulating Crohn’s disease. Aliment Pharmacol Ther. 2003;17:1145–1151
20. Schwartz DA, Loftus EV Jr, Tremaine WJ, et al. The natural history of fistulizing Crohn’s disease in Olmsted County, Minnesota. Gastroenterology. 2002;122:875–880
21. Tang LY, Rawsthorne P, Bernstein CN, Tang LY, Rawsthorne P, Bernstein CN. Are perineal and luminal fistulas associated in Crohn’s disease? A population-based study. Clin Gastroenterol Hepatol. 2006;4:1130–1134
22. Lapidus A, Lapidus A. Crohn’s disease in Stockholm County during 1990–2001: an epidemiological update. World J Gastroenterol. 2006;12:75–81
23. Eglinton TW, Roberts R, Pearson J, et al. Clinical and genetic risk factors for perianal Crohn’s disease in a population-based cohort [ published online ahead of print December 13, 2011]. Am J Gastroenterol. 2012;107:589–596