Background: Patients hospitalized for ulcerative colitis (UC) are at high risk for colectomy. Despite growing interest in research using administrative data in inflammatory bowel disease, there is no available tool in such research to stratify disease severity or identify patients at high risk for colectomy.
Methods: Using the Nationwide Inpatient Sample (NIS) 2004, we identified patients hospitalized for UC flare through appropriate ICD-9-CM discharge diagnosis codes (556.x). Our primary outcome of interest was undergoing total colectomy (45.8). Multivariate logistic regression models were constructed to identify independent predictors of colectomy. From this, a cumulative risk score was developed. Hospitalizations were divided into 3 strata (low, intermediate, high) based on the odds of colectomy.
Results: There were a total of 15,142 hospitalizations with a discharge diagnosis of UC included in our study among whom 366 patients underwent total colectomy (2.4%). Anemia (odds ratio [OR] 2.13), requirement for blood transfusion (OR 2.22), malnutrition (OR 4.53), and total parenteral nutrition (OR 4.30) were independent predictors of colectomy as were transfer in from another hospital (OR 2.06) and admission to a teaching hospital (OR 1.73). The cumulative colectomy risk score ranged from 0-8, with significantly higher risk of colectomy in the high-risk stratum (13.7%) compared to the intermediate-risk (4.2%, P < 0.001) and low-risk (1.3%, P < 0.001) strata. The risk score performed well in its discriminative ability with an area under the curve of 0.70.
Conclusions: We propose a simple and novel risk score to stratify the severity of UC hospitalizations and predict colectomy in this population.