Prevalence, incidence, and treatment patterns of fistulizing Crohn disease: A US population-based cohort study
Plain language summary
Some people with Crohn disease (CD) get tunnel-like connections (called fistulas) between the gut and other organs that can be dangerous and might need to be fixed by surgery. There is little information about how many patients have fistulizing CD in the United States. Using a database of medical insurance records, we found that fistulizing CD was not common. Patients with fistulizing CD needed a variety of medical treatments, including biologic drugs.
Implications for managed care pharmacy
Here, we report the prevalence and incidence of fistulizing CD in the United States using an administrative claims database. We show that patients with fistulizing CD experience considerable surgical burden and require intensive use of medication. This research will help to inform health care practitioners and policymakers on the frequency and extent of fistulizing CD. In addition, the substantial medical burden of fistulizing CD reported here may also help to inform health care practitioners and policymakers on the importance of prevention, early recognition, and appropriate management of this severe complication.
Crohn disease (CD) is a chronic and relapsing inflammatory gastrointestinal disease, characterized by a variable disease course, with periods of recurrence and remission.1-3 Patients can develop severe complications that include the formation of fistulas between the bowel and adjacent organs, with the most common form being a perianal fistula.4,5 Development of fistulas is associated with considerable disease and treatment burden, which have a negative impact on patient quality of life and health care resource use.6,7 Up to half of all patients may develop a fistula within 20 years of initial diagnosis.8
Although there is limited evidence to guide the optimal treatment approach in patients with fistulizing CD, treatment strategies aim to induce and maintain endoscopic and clinical remission.1,4,9 Available treatments for CD include the use of antibiotics, immunosuppressants, and biologics. Patients have a high likelihood of surgery to treat complications arising from disease progression, with up to 62% of patients requiring surgery within 10 years of their diagnosis.4,7,9,10 It is estimated that up to one-third of patients with CD experience recurrent fistulas during their disease course despite extensive medication use and surgical treatments.6,11
Globally, population-based studies show that the prevalence and incidence rates of CD vary between newly industrialized and postindustrial countries. The cost burden of CD is rising for health care systems in many countries worldwide because of increased treatment costs, hospitalization rates, and inpatient visits.7,12 There are limited population-based studies characterizing patients with CD complications, particularly regarding the prevalence and incidence rates of fistulizing CD, and treatment patterns in the United States.2,6,12
We aimed to estimate the prevalence and incidence rates, and examine associated treatment patterns, for fistulizing CD based on claims data from a representative US population. Specifically, the primary objective was to estimate the crude, age, and sex-adjusted prevalence and incidence rates of fistulizing CD in the United States. Secondary objectives were to examine comorbidities in patients with continuous medical and pharmacy enrollment during the 12-month pre-index period and to examine associated treatment patterns (medications dispensed and medical procedures performed) during the follow-up period among incident cases.
Methods
STUDY DESIGN AND SETTING
This retrospective, observational cohort study was conducted using the Optum Clinformatics Data Mart (CDM) database—an administrative claims database of a commercially insured population in the United States. Data were used to estimate the prevalence and incidence rates of fistulizing CD, describe patients’ comorbidity profiles, and examine related treatment patterns among patients with incident fistulizing CD.
ASSESSMENT PERIODS AND STUDY POPULATIONS
The assessment period included a study period, baseline period, patient selection period, and follow-up period (Supplementary Figure 1, available in online article). The study period was between January 1, 2016, and December 31, 2020. The patient selection period was between January 1, 2017, and December 31, 2019 (end of data availability). This allowed for a 365-day baseline period (defined as the period prior to the cohort entry date, the earliest such period being between January 1, 2016, and December 31, 2016, and the latest being between January 1, 2019, and December 31, 2019) that was used to confirm eligibility, differentiate prevalent and incident cases of fistulizing CD, and allow for a possible 365-day follow-up period. The source population for estimating incidence and prevalence included all eligible patients in the database with or without CD.
Patients who had a diagnosis of CD (identified using International Classification of Diseases, 10th Revision [ICD-10] codes) (Supplementary Table 1) and at least 1 year of continuous insurance enrollment (with medical and pharmacy coverage) during the patient selection period (gaps of ≤30 days of enrollment were allowed), and had nonmissing age and sex data, were selected. The end of the first year of continuous insurance enrollment was defined as the cohort entry date, at which point patients had to be aged 18 years or older. The index date was the date of the first diagnosis of fistulizing CD on or after the cohort entry date.
Primary Objective Population. For estimating the prevalence and incidence rates of fistulizing CD, patients were required to have continuous insurance enrollment during the baseline period (365 days prior to [but not including] the cohort entry date) and the follow-up period (which began on the cohort entry date [inclusive] and was used to calculate the person-time at risk for estimating the incidence rate of fistulizing CD, including those with fistulizing CD at first presentation). Patients were censored on the date of the first occurrence of any of the following events: end of study period, discontinuation of insurance medical coverage (last date of enrollment), a diagnosis of fistulizing CD, or death. If patients did not have a diagnosis of fistula (with or without CD specified) or a procedure for seton placement, fistula plug, or fistulotomy prior to the cohort entry date, they were defined as incident cases; otherwise, they were considered existing cases. Prevalent cases comprised incident and existing cases. Diagnosis criteria were the presence of an ICD-10 CD-related fistula diagnosis code or a general fistula diagnosis code with a CD diagnosis code in any field of medical claims during the study period; or a fistula diagnosis (CD not specified) during the follow-up period and a CD diagnosis prior to the fistula diagnosis during the baseline or follow-up period. The full list of the ICD-10 Clinical Modification codes for CD and fistulas is provided in Supplementary Table 1.
Secondary Objectives Population. The secondary objectives population, used to describe the demographic characteristics, comorbidities, medications dispensed, and medical procedures performed among patients with fistulizing CD, comprised those patients in the primary objective population with fistulizing CD who had at least 365 days’ continuous enrollment before the index date. The baseline period was as defined above. The follow-up period for incident or existing cases began on the index date (inclusive), and patients were censored analogous to the criteria used previously on the date of the first occurrence of any of the following events: end of study period, discontinuation of insurance medical coverage (last date of enrollment), end of 12 months, or death. The proportion of patients with selected comorbidities was only evaluated during the baseline period, whereas the proportion of patients with medications dispensed and medical procedures performed was only evaluated during the follow-up period and reported quarterly and annually.
STATISTICAL ANALYSIS
All analyses were descriptive in nature; no formal statistical comparisons were performed.
For the primary objective, the crude prevalence of fistulizing CD was calculated by dividing the total number of patients with incident and existing fistulizing CD by the total number of eligible study participants during the defined follow-up period, as shown in the following formula:
This calculation was conducted for each year of the study period (2017-2019).
The crude incidence rate of fistulizing CD was calculated by dividing the total number of patients with incident fistulizing CD by the sum of the follow-up time at risk in person-years for all eligible study participants (formula shown below). Patients with existing fistulizing CD in the follow-up period were not considered at risk and so were excluded from the eligible study participants in the calculation.
In addition, the crude incidence rate was adjusted for age and sex using the 2018 US Census population. Furthermore, the crude prevalence and incidence rates were estimated and stratified by demographic characteristics (age, sex, calendar year, region of residence, and insurance type) and location of CD (small bowel/ileal, colon, ileocolonic, isolated upper, or gastrointestinal).
For the secondary objectives, the denominator population was the population of patients in the primary objective population with fistulizing CD who had at least 365 days’ continuous enrollment before the index date. The proportions of patients with comorbidities during the baseline period for incident and existing cases of fistulizing CD were calculated as the total number of patients with fistulizing CD with the diagnosis of the comorbidity of interest in each cohort divided by the total number of eligible patients with fistulizing CD in each cohort. Treatment patterns (medications dispensed and medical procedures) in incident or existing cases of fistulizing CD were calculated by the total number of patients (incident or existing) with fistulizing CD, with a dispensing of the specific medication or a claim for a surgery of interest during each period of evaluation, divided by the total number of patients (incident or existing) with fistulizing CD with the required continuous insurance enrollment during the evaluation period. Time from the index date to the first medical procedure in patients with incident fistulizing CD was estimated using the Kaplan–Meier method.
DATA SOURCE
The Optum CDM is a proprietary research database with data from more than 100 million geographically diverse patients and is representative of members with commercial insurance and Medicare Advantage plans across the United States. The database contains deidentified patient-level data from all health care sites (inpatient, hospital, outpatient, emergency department, and physician’s office) for all types of services provided, including specialty, preventive, and office-based care. The database also contains longitudinal medical and pharmacy administrative claims from commercial plans across the United States.
The database used in this study is compliant with the Health Insurance and Portability and Accountability Act of 1996. No identifiable protected health information was extracted or reported; therefore, no institutional review board approval or informed consent was required.
STUDY REPORTING AND APPROVAL
The results reported here are in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for cohort studies. This study used a database that was a secondary database consisting of anonymized and deidentified data; therefore, no institutional review board approval or informed consent was required.
Results
There were 20,193,078 eligible adult patients (with or without CD) in the Optum CDM database during the study period.
For the primary objective population, 5,082 patients had fistulizing CD diagnostic ICD-10 codes with at least 365 days of continuous insurance enrollment during 2017 to 2020; 2,399 of 5,082 patients (47.2%) had incident fistulizing CD and the remaining 2,683 of 5,082 patients (52.8%) had existing fistulizing CD.
For the secondary objectives population, 5,005 patients with fistulizing CD also had at least 365 days of continuous insurance enrollment prior to the index date of the first fistulizing CD diagnostic ICD-10 code. Of these, 2,137 of 5,005 patients (42.7%) had incident fistulizing CD and the remaining 2,868 of 5,005 patients (57.3%) had existing fistulizing CD (Supplementary Figure 2).
PREVALENCE AND INCIDENCE RATES
The overall crude prevalence and incidence rates of fistulizing CD between 2017 and 2019 were 25.2 (95% CI = 24.5-25.9) per 100,000 persons and 6.9 (95% CI = 6.6-7.1) per 100,000 person-years, respectively. The overall age- and sex-adjusted prevalence and incidence rates were 24.9 (95% CI = 24.2-25.6) per 100,000 persons and 7.0 (95% CI = 6.7-7.3) per 100,000 person-years, respectively (Table 1). The annual crude prevalence and incidence rates between 2017 and 2019 were generally stable, ranging from 12.8 (95% CI = 12.2-13.3) to 9.9 (95% CI = 9.4-10.4) per 100,000 persons and 4.8 (95% CI = 4.5-5.2) to 6.4 (95% CI = 6.0-6.8) per 100,000 person-years, respectively (Table 1). Overall, the prevalence and incidence rates increased with age, peaking at 30-39 years, after which they declined (Figure 1).
TABLE 1
Prevalence and Incidence Rates of Fistulizing CD
a Adjusted for age and sex using 2018 US Census data.
CD = Crohn disease.
Prevalence and Incidence Rates of Fistulizing Crohn Disease by Age Group
BASELINE CHARACTERISTICS AND COMORBIDITIES
Overall, patient demographic characteristics were similar across the incident and existing cases of fistulizing CD measured on the index date (Table 2). Notably, approximately half of all patients were female (incident cases, 51.6%; existing cases, 51.9%). Patients were most frequently residents from the south of the United States (incident, 39.5%; existing, 40.2%), had commercial insurance (incident, 67.2%; existing, 72.0%), and had both small and large intestinal involvement (incident, 24.2%; existing, 32.3%). Fistulas were most commonly reported in the anal (incident, 18.3%; existing, 30.1%) and intestinal (incident, 18.2%; existing, 12.4%) areas.
TABLE 2
Demographic Characteristics of Patients With Fistulizing CD at Index Date
a CD location was found using the last encounter during the 6-month baseline period (including index date) with evidence of a general CD diagnosis. If there was not an encounter during the baseline period or on the index date, the first encounter during a 6-month follow-up window was used instead.
b Fistula location was found using the last encounter during the 6-month baseline period (including index date) with evidence of fistula diagnosis. Diagnosis code description was used to determine location. If the diagnosis code did not specify location, the patient was placed in the unknown/unspecified group.
CD = Crohn disease.
The most common comorbidities reported in patients with fistulizing CD were cardiovascular disease (incident cases, 44.8%; existing cases, 40.2%), anxiety (incident, 21.0%; existing, 23.7%), depression (incident, 18.1%; existing, 21.0%), and obesity (incident, 18.4%; existing, 18.2%) (Supplementary Table 2).
TREATMENT PATTERNS FOR INCIDENT CASES
Medication use among patients with incident cases of fistulizing CD remained generally stable within the first-year post-index date, particularly after 3 months post-index date, except for a decreasing use of corticosteroids (3 months pre-index, 33.6%; 9-12 months post-index, 20.9%), 5-aminosalicylic acid (3 months pre-index, 15.6%; 9-12 months post-index, 9.9%), and antibiotics (3 months pre-index, 22.7%; 9-12 months post-index, 9.5%). The most commonly used biologic medication class was tumor necrosis factor α (TNF-α) inhibitors (3 months pre-index, 24.9%; 9-12 months post-index, 34.5%), whereas the least common medication class was antidiarrheals (3 months pre-index, 4.0%; 9-12 months post-index, 2.7%) (Figure 2).
Treatment Patterns Among Incident Cases of Fistulizing CD
Approximately one-third of patients (n = 736; 34.4%) with incident cases of fistulizing CD required surgery during the follow-up period. Most patients with fistulizing CD received surgical intervention within 3 months of the index date. The median time to the first surgery among patients with incident fistulizing CD was 179 days (interquartile range, 24-476; n = 2,137) (Figure 3).
Time From Index Date to First Surgery Among Incident Cases of Fistulizing Crohn Disease
Discussion
Population-based studies for characterizing patients, particularly regarding the prevalence and incidence rates of fistulizing CD, and treatment patterns in the United States, are limited. However, they are important in highlighting the current disease burden, resource use, and unmet need of this severe disease complication.2,6,7 By using a well-characterized database for patients with fistulizing CD representative of commercially insured patients across the United States, we found that fistulizing CD is uncommon in the United States. Overall, our data are in keeping with another recent population-based study that reported that patients with perianal fistulas had higher medication use than patients without perianal fistulas.7
Our data on the incidence rate in patients with fistulizing CD somewhat reflect those of patients in the overall CD population previously reported in the literature, with estimates varying according to the study design, study population, and geographical region. Studies estimating the prevalence and incidence rates in the United States are typically smaller, based on referral centers, and restricted to specific geographical regions; few are population-based, and even fewer are based on patients with fistulizing CD.12-15
Here, we report the overall adjusted prevalence and incidence rates in patients with fistulizing CD (2017-2019), which were 25.2 per 100,000 persons and 6.9 per 100,000 person-years, respectively. When adjusted by age and sex, the prevalence and incidence rates barely changed (24.9 per 100,000 persons and 7.0 per 100,000 person-years, respectively). A previous study used a logistic regression model to extrapolate local population prevalence and incidence data to the US population, based upon the US Census Bureau’s 2017 population estimate.11 From this, they estimated that there were 76,616 and 15,732 patients with prevalent and incident fistulizing CD in the United States in 2017, respectively, which was equivalent to prevalence and incidence rates of 23.4 per 100,000 persons and 4.8 per 100,000 person-years, respectively.11
The trend presented here of increasing prevalence of fistulizing CD with age, peaking at 30-39 years, is consistent with previous studies that represent the overall CD population.13,14 However, in contrast to a US population-based study by Kappelman et al that showed a general levelling off in patients with CD, we report a decline in prevalence in adults aged between 30 and 39 years.16 Others have described the median onset of CD at 30 years of age, with 2 peaks—the first between 20 and 30 years, and then a smaller peak around 50 years.1,17
Overall, the decline in incidence in patients with fistulizing CD presented in this study is in line with the general shift of stabilizing or falling incidence currently observed worldwide in patients with CD from industrialized countries.12 However, a recent study observed a decreasing cumulative incidence in patients with perianal or rectovaginal fistulas in the biologic era compared with the prebiologic era, noting that the reasons for the decline are currently unknown.15
An estimated 71%-84% of anal fistulas are complex and often require a medical intervention to manage this hard-to-treat population.11,18-20 Although not stratified as simple or complex fistula, our results indicated that the most common fistula location was anal fistula (incident cases, 18.3%; existing cases, 30.1%), which likely contributes to the overall disease burden of fistulizing CD in the United States. Similarly, our results align with those reported in the literature, with the most commonly reported type of fistula being anal fistula.8,11,21
Patients with CD, compared with those without, are known to have an increased lifetime risk for intestinal complications and extraintestinal manifestations.22 The most common comorbidities reported in this study for patients with incident or existing cases of fistulizing CD were cardiovascular disease (42.2%), anxiety (22.5%), depression (19.8%), and obesity (18.3%). The high rates of cardiovascular disease are relevant given that there is growing evidence supporting an association between CD and an increased risk of developing cardiovascular disease.23 Similarly, studies have reported increasing prevalence of obesity among patients with CD, with a potential impact on the effectiveness and response duration of biologic therapy, and rates of postoperative complications.24 In addition, the burden of CD increases in those with complications compared with those without. For example, patients with vs without complicated CD phenotypes (stricturing CD, penetrating CD, and stricturing and penetrating CD) reported higher rates of active CD-related luminal and extraintestinal manifestations (eg, aphthous ulcer, 15.4% vs 10.5%, P < 0.05; inflammatory bowel disease-related arthropathy, 25.8% vs 17.2%, P < 0.01) and more underwent surgeries (77.7% vs 12.2%, P < 0.001), despite being treated with more biologics and corticosteroids.25 Furthermore, direct and indirect health care costs per person per year (PPPY) are higher for patients with CD who require surgery (approximately $112,000/PPPY) or prolonged opioid analgesia (approximately $81,000/PPPY) than for those with moderate-to-severe CD (approximately $53,000/PPPY).26
Overall, it is well documented that patients with complicated CD experience considerable disease burden and treatment needs.5,8,21,25 Therefore, it is not surprising that, of the patients with fistulizing CD examined here (2017-2019), approximately half were prescribed biologic therapies within 1 year after diagnosis. Medication use among patients with incident cases of fistulizing CD remained generally stable over 12 months, particularly after 3 months post-index date, except for corticosteroids, antibiotics, and 5-aminosalicylic acid treatment, which decreased. The most widely prescribed biologic medication class over time was TNF-α inhibitors (9-12 months post-index date, 34.5%). In addition to having a substantial role in the treatment of CD, the high use of TNF-α inhibitors may be reflective of previous evidence supporting infliximab as an effective treatment for achieving and maintaining fistula closure in adults with fistulizing disease, although it is associated with frequent loss of response.5,9,24 The results of this study demonstrated that 34.4% of patients with fistulizing CD required surgery during the follow-up period, which was higher than that reported in a Norwegian population-based study in patients with CD (cumulative probability of 13.6% after 1 year of CD diagnosis).27 However, the latter study did not include perianal abscess drainage or fistulectomy in its definition of surgery, whereas the current study included these procedures, as well as colectomy, ileostomy, seton placement, and small bowel resection.
LIMITATIONS
The limitations of this study include those inherent in any retrospective analysis using administrative claims. For example, inaccuracies or misclassification of patients due to using administrative claims data, rather than medical records, to identify patients with fistulizing CD, clinical characteristics, specified treatments based on ICD-10 diagnosis codes, procedure codes, and national drug directory codes. There may be some misclassification bias as the presence of a diagnosis code on a medical claim may not always indicate disease, as the diagnosis code may be incorrectly assigned or included as rule-out criteria rather than actual disease. In addition, the high rate of unknown or unspecified code use may be due to the information not being recorded in the system; alternatively, the clinician may have been unsure of the CD type and thus coded it as unknown or unspecified. To mitigate misclassification in the study cohort, this study required diagnoses of both CD and fistula, if the fistula diagnosis was not CD specific, and used fistula-related procedures or surgeries to identify fistulas in addition to diagnosis codes. Furthermore, the study findings may not be generalized to all patients with CD in the United States, as the dataset was limited to a commercially insured patient population. The diagnosis codes on claims reflected information submitted by the physicians for reimbursement, rather than confirmed clinical conditions, and may be subject to data limitations and data entry error.
A strength of this study was its population-based data, covering patients with fistulizing CD across most regions in the United States. Therefore, the prevalence and incidence rate estimations may reflect a geographically inclusive US CD population, rather than a single region or a limited type of population.
Conclusions
This study reports age- and sexadjusted prevalence and incidence rates for fistulizing CD of 24.9 per 100,000 persons and 7.0 per 100,000 person-years, respectively. We showed that a considerable proportion of patients with fistulizing CD required intensive medical treatment and had a substantial surgical treatment burden. Further research is warranted to determine what treatments can most effectively treat or even prevent fistulizing CD to improve the overall management of patients with CD.
ACKNOWLEDGMENTS
All authors meet criteria for authorship as recommended by the International Committee of Medical Journal Editors (ICMJE) and made the decision to submit the manuscript for publication. The authors did not receive payment related to the development of the manuscript. In the preparation of this manuscript, Pallavi Patel, PhD, of Hyperion, OPEN Health Communications (London, UK) provided medical writing, editorial, and formatting support, which was contracted and funded by Boehringer Ingelheim.
Funding Statement
This study was funded by Boehringer Ingelheim. Boehringer Ingelheim was given the opportunity to review the manuscript for medical and scientific accuracy, as well as intellectual property considerations.
DATA AVAILABILITY STATEMENT
We conducted a retrospective study using claims data from the Optum research database, a database of commercially and noncommercially insured populations in the United States. All patient data were anonymized and deidentified; therefore, informed consent was not necessary. Restrictions apply to the availability of these data because of a contract between Optum and Boehringer Ingelheim, and data are thus unavailable to the public. For enquiries on the dataset analyzed in this study, please contact Optum (https://www.optum.com).
REFERENCES