Introduction: Eosinophilic colitis (EoC) is a rare entity characterized by the presence of high eosinophilic infiltrate into the colonic wall in symptomatic patients, in the absence of known other causes of colonic eosinophilia. Recently proposed guidelines suggest 100 eosinophils per high-power-field in the ascending colon, 85 in the descending, and 65 in the sigmoid colon as diagnostic thresholds. Whereas ulcerative colitis (UC), is known to be associated with an increased risk of colorectal cancer (CRC), little is known of the association between EoC and CRC. We aimed to compare EoC patients to UC patients to assess their odds of developing CRC. Methods: Using the multi-institutional, health research network database TriNetX (Cambridge, MA), de-identified, aggregated clinical data were obtained on patients with a diagnosis of either EoC or UC. Patients with other identifiable causes of eosinophilia were excluded from the EoC cohort, including inflammatory bowel disease, food allergies, helminth infections, and other eosinophilic gastrointestinal disorders. A 1:1 propensity score matching method was used to stratify EoC and UC patients. Matched variables were age at diagnosis, sex, race, obesity, tobacco abuse, alcohol abuse, and family history of digestive tract malignancy. Odds ratios(OR) and confidence intervals (CI) were calculated for development of CRC and subsequent colonoscopies for each cohort. Results: A total of 1,310 and 195,477 EoC and UC patients were identified, respectively (Table). The higher proportion of EoC patients were female compared to UC patients (65.4% vs 54.4%), and were African American (11.9% vs 9.2%). More EoC patients reported tobacco usage (4.6% vs 2.5%) and were obese (11.8% vs 7.0%) compared to UC patients. A 1:1 matching ratio stratified 1,310 patients into each cohort. After adjusting for the aforementioned covariates, EoC patients had an OR of 1.304 (95% CI 0.828-2.054, p-value 0.2503) of developing CRC compared to UC patients (Figure). EoC patients demonstrated an OR of 0.754 (95% CI 0.583-0.976, p-value 0.0313) of undergoing subsequent surveillance colonoscopies after initial diagnosis of disease compared to UC patients. Conclusion: Patients with a diagnosis of EoC have a similar risk of developing colorectal cancer compared to UC patients. Despite this, EoC patients are less likely to undergo subsequent surveillance colonoscopy compared to UC patients. Hopefully, our data can be used for future prospective studies investigating the natural course of EoC.Figure 1.: Forest plot showing odds ratios and confidence intervals of development of colorectal cancer (CRC) and subsequent surveillance colonoscopies in EoC patients compared to UC patients Table 1. - Baseline patient characteristics of EoC (eosinophilic colitis) and UC (ulcerative colitis) patients before and after 1:1 propensity score matching Baseline Patient Characteristics Before Matching p-value After Matching p-value EoC(n=1,310) UC(n=195,477) EoC(n=1,310) UC=n1,310) Age at Diagnosis 54 +/- 17.2 51 +/- 17.6 < 0.0001 54 +/- 17.2 54 +/- 17.4 0.9451 Sex Female 857(65.4%) 106,402(54.4%) < 0.0001 857(65.4%) 865(66.1%) 0.7419 Male 453(34.6%) 89,075(45.6%) < 0.0001 453(34.6%) 445(33.9%) 0.7418 Race African American 157(11.9%) 17,934(9.2%) 0.0005 157(11.9%) 134(10.2%) 0.1527 Asian 32(2.4%) 3,583(1.8%) 0.1014 32(2.4%) 39(2.9%) 0.3997 Caucasian 980(74.8%) 145,147(74.3%) 0.6468 980(74.8%) 1,007(76.9%) 0.2178 Other/Unknown 141(10.9%) 28,813(14.7%) < 0.0001 141(10.9%) 130(10.0%) 0.6152 Other Covariates Alcohol abuse 18(1.4%) 3,241(1.7%) < 0.0001 18(1.4%) 151(11.5%) 0.8078 Family history of GI cancer 38(2.9%) 4,033(2.1%) < 0.0001 38(2.9%) 49(3.7%) 0.2818 Obesity 155(11.8%) 13,688(7.0%) 0.4222 155(11.8%) 12(0.9%) 0.2706 Tobacco use 60(4.6%) 4,807(2.5%). 0.0338 60(4.6%) 31(2.4%) 0.3931
Read full abstract