Hispanic Ethnicity is Associated with Milder Disease Severity in Crohn’s Disease but not Ulcerative Colitis

accounting US Abstract Background: The incidence of inflammatory bowel disease (IBD) is rising among minority patients; however, few studies have examined racial/ethnic disparities in epidemiology, disease severity, and treatment course. Aim: To characterize differences in IBD disease severity and outcomes among IBD patients from minority communities. Methods: A cohort of adult patients with IBD was followed from January 2007- December 2012 in a large urban safety-net hospital. Fisher exact and Mann-Whitney rank sum tests were used to compare disease phenotype, severity, need for anti-TNF therapy, receipt of IBD-related surgery, and hospitalizations. Results: We identified 291 adult IBD patients; 148 with Crohn’s Disease (CD), 143 with ulcerative colitis (UC). Our cohort was racially diverse with 32% Caucasian, 37% Black, 28% Hispanic, and 2% Asian. 54% were male and the median age of the cohort was 44 years. Hispanic patients had a more benign disease course characterized by less IBD-related surgeries (mean 0.4 vs. 1.1, p=0.001), less hospitalizations (mean 2.1 vs. 2.9, p=0.04), and lower need for anti-TNF therapy (27% vs. 39%, p=0.05) when compared to non-Hispanic patients. On subgroup analysis, these differences were noted particularly among those with CD but not UC. Hispanic UC patients had a similar number of IBD-related surgeries (p=0.16) and hospitalizations (p=0.62), whereas Hispanic CD patients had less IBD-related surgeries (mean 0.8 vs. 1.8, p=0.01), less hospitalizations (mean 2.4 vs. 3.8, p=0.05), and lower need for anti-TNF (38% vs. 57%, p=0.09). lower disease severity in CD but not associated with disease severity in UC.


124
to the 2013 U.S. Census Bureau, are increasingly being identified with IBD [9]. Therefore, it is important to recognize epidemiologic differences in order to optimize treatment for these patients. Similar to what was seen in AAs, studies identifying phenotypic variations, treatment preferences, and disease outcomes in Hispanics have been conflicting [10][11][12][13]. One group found that Hispanics were diagnosed at a significantly older age than Caucasians and had a lower number of bowel resections per patient. Additionally, UC was found to be more common than CD in Hispanic IBD patients [14]. On the other hand, another group found that prolonged steroid exposure was more common among Hispanics with UC, and these patients had more UC-related surgeries and hospitalizations, although these differences were not significant in multivariate analysis [15]. Altogether, the existing data leave open several unanswered questions regarding IBD disease course and management among different ethnicities. The large urban safetynet health care system of Parkland Hospital serving 2.4 million in the Dallas County provided an ideal setting to investigate these questions given its widely racially diverse population. Our intent was to determine if ethnicity is a clinically significant variable influencing disease severity, need for IMM or anti-TNF therapy, surgery, and hospitalization in this ethnically diverse cohort of IBD patients.

Study Setting and Population
We performed a retrospective cohort study of adult IBD patients (> 18 years old) followed at Parkland Health and Hospital System, the safety-net health system for Dallas County, between January 2007 and December 2012. Given its integrated structure and function as a safety-net institution, patients admitted to Parkland Hospital often receive their continuity care through the Parkland Health System clinics. Parkland utilizes a single comprehensive electronic medical record (EMR), including laboratory and radiology results, for all inpatient and outpatient care. Patients were identified using ICD-9 codes for CD and UC as well as a prospectively maintained database of patients with IBD seen in the Gastroenterology clinic.
One author (C.C) reviewed all cases to confirm that patients truly had a confirmed diagnosis of IBD. Patients were excluded if 1) they had an indeterminate or alternate etiology of colitis (infectious, ischemic, diverticular-associated, etc.), 2) received steroids, IMM, or anti-TNF therapies for conditions other than IBD, 3) received anti-TNF therapy prior to establishing care at Parkland, 4) were lost to follow up within 6 months of IBD diagnosis, or 5) had insufficient data for assessment of our primary outcome. This study was approved by the Institutional Review Board at the University of Texas Southwestern Medical Center and Parkland Memorial Hospital.

Data Collection
Patient demographics, clinical history, laboratory data, and endoscopic data were obtained through review of EMR records.
One investigator (C.C.) extracted information using standardized forms, with another investigator (T.A.) available for questions.
Age, gender, race/ethnicity, and smoking history (ever vs. never), and family history of IBD were recorded. Dates of IBD diagnosis, any IBD-related surgeries, and IBD-related hospitalizations were documented. IBD related surgeries were defined as any bowel resection or intervention for peri-anal disease (i.e. incision and drainages of abscess, seton placement, fistulotomy, etc) that

Statistical Analysis
Our primary outcome of interest was ethnicity as a clinically significant variable in disease outcomes. In univariate analysis, Statistical significance was defined as p< 0.05 for univariate and multivariate analyses. All data analysis was performed using Stata 11 (Stata Corp, College Station, TX).

Patient Demographics
We initially identified 351 patients with a diagnosis of IBD

125
for another disease process (n=6), (ii) being lost to follow up within 6 months of IBD diagnosis (n=21), (iii) having had prior colonic surgeries (n=2), or, (iv) having insufficient data for assessment of our primary outcome (n=4). Of the final 291 patients remaining in the cohort,148 had CD and 143 had UC (Table 1). Overall, 54% were male, and the median age of patients was 44 years. Our cohort was racially diverse with 32% Caucasians, 37% Blacks, 28% Hispanics, and 2% Asians. Due to small sample size, Asians were not included in ethnic comparisons but were included in all "non-Hispanic" vs.

Racial/Ethnic Differences in Disease Presentation
We found significant racial/ethnic differences in age at IBD presentation (  (Table 4). Finally, we found no significant differences in rates of EIMs (p=0.50) between AAs, Hispanics, and Caucasians.

Racial/Ethnic Differences in Disease Severity And IBD-Related Outcomes
Considering IBD-related surgeries and hospitalizations, Hispanic patients had a more benign disease course ( Table 2).
Compared to non-Hispanics, they had significantly less IBD-related the CD population, but this did not reach statistical significance (Table 3).

Racial/Ethnic Differences on Multivariate Analysis
Hispanic ethnicity was associated with significantly fewer IBDrelated surgeries as mentioned above. However, in multivariate analysis upon adjusting for age at diagnosis, tobacco use, and need for anti-TNF therapy, Hispanic ethnicity was no longer significantly associated with total number of IBD-related surgeries (p=0.09) or hospitalizations (p=0.56) among CD patients. Adjusting for those same variables, Hispanic ethnicity was also not a significant factor in UC.

Discussion
Though current literature remains conflicting, ethnicity does seem to play a clinically significant role in disease activity.

128
inclusion of all IBD patients seen within one health care system may reduce selection bias and improve generalizability in our findings. Lastly, we utilize a step-up approach to biologic therapy in patients with moderate IBD. This then limits use of anti-TNF therapy to those who have failed IMMs thus categorizing this subgroup of IBD patients as those with severe disease. This stepup approach to biologic therapy also eliminates any variability in practitioners' perspective and bias in this setting. There are a number of limitations to this study that are important to mention.
The retrospective nature of our study design did not allow for assessment of treatment response among the ethnicities. However, this is not dissimilar to most studies to date, which have assessed use of anti-TNF and IMM therapy rather than actual treatment response [1,4-6, 8,10,13-15,20-23]. The clinical practice does not routinely utilize disease activity indices as part of clinical care, and therefore these were not available in our retrospective analysis.
Additionally, using biologic therapy primarily in those who have failed IMM therapy does not represent the most common practices among IBD experts. Moreover, the use of anti-TNF therapy as also a marker of disease severity may not be entirely accurate and confound other variables. The modest sample size of Asians as well as short recruitment period of five years may make it difficult to adequately compare the different ethnicities as well as assess disease severity and complications and may explain the lack of findings upon extrapolating to multivariate analyses. Another notable point is that our foreign-born Hispanic patient population is largely from Mexico which is different from the largely Cuban foreign-born Hispanic population studied by Damas and colleagues.
Altogether, this study contributes additional evidence that ethnic differences can influence the phenotypic expression and treatment needs of IBD. Future studies will be required to elucidate whether this is driven by underlying genetic differences or environmental influences that are disproportionately segregated among these groups.