(145) Identification of Gastrointestinal Perforations in Patients with Inflammatory Bowel Diseases: Challenges and Opportunities in Administrative Claims Data

Publication Number: 480

Background: Gastrointestinal (GI) perforation is a rare but serious complication of inflammatory bowel disease (IBD), which encompasses Crohn’s disease (CD) and ulcerative colitis (UC). Chronic inflammation and ulceration of the intestinal wall are among the etiologies can lead to perforation. Large longitudinal real-world data sources could be utilized to study the natural history of rare outcomes such as GI perforation.

Objectives: To identify incident GI perforation among patients with UC and CD and to characterize how the GI perforation cases differ between UC and CD.

Methods: The study population consisted of prevalent CD or UC patients (≥ 2 diagnoses for UC or CD ≥ 7 days apart) identified in the IBM® MarketScan® claims database (January 1, 2005-December 31, 2019), respectively. Patients aged ≥ 18 years with at least 12 months of continuous database enrollment before IBD index date were eligible. Patients diagnosed with both CD and UC were excluded to minimize misclassification. The algorithm for GI perforation was established and validated in a similar claims database in rheumatoid arthritis population (overall positive predictive value [PPV]: 94%, 95% confidence interval: 86%-98%). The algorithm stratifies GI perforation by location (upper GI: esophagus, stomach; lower GI: small intestine, large intestine, unspecified lower GI) with PPVs ranging from 89% to 100%. In this research, the number and percentage of GI perforation were stratified by location for UC and CD.

Results: The cohort included 66,692 patients with UC (mean age ± standard deviation [SD]: 50±16, 54% female) and 55,070 patients with CD (mean age ± SD: 46±16, 59% female) during the study years. Among these, 357 and 505 incident GI perforations occurred among patients with UC and CD, respectively. In the IBD population, upper GI perforations in esophagus and stomach were the less common (UC: 6.7%, 24/357; CD: 6.1%, 31/505) than previously validated in RA (35%, 244/704). The location for lower GI perforation differed between UC (17% small intestine, 53% large intestine, and 30% unspecified location) and CD (25% small intestine, 40% large intestine, and 35% unspecified location).

Conclusions: Unlike in RA, upper GI perforations were uncommon in IBD. Perforation most commonly occurs in the lower GI track for patients with IBD. Despite the high overall PPV, the current RA-based algorithm might not sufficiently ascertain the location and etiology for GI perforations in IBD. Future validation studies in IBD should focus on developing case definitions of GI perforation especially distinguish perforations that occur in or outside regions of bowel inflammation or strictures.