Mo1967 Development of a Prediction Model to Assess the Risk of Chronic Gastrointestinal Ischemia in Referred Patients

Background and aim: Chronic gastrointestinal ischemia (CGI) is a challenging disease entity. Clinical symptoms may differ amongst patients. The aim of this study was to establish predictors for the diagnosis of CGI based on self-reported variables and to combine these in a prediction model. Patients and Methods: We analyzed data of a prospective cohort study. Between November 2006 and March 2013 self-reported symptoms were collected by a structured questionnaire of 431 consecutive patients referred to an academic hospital for evaluation of possible CGI. All patients received the standard work-up of CGI, consisting of radiological imaging of the gastrointestinal arteries, and functional testing for detection of mucosal ischemia by means of visible light spectroscopy (VLS) or tonometry. The results were discussed in a multidisciplinary expert panel leading to a consensus diagnosis, which was monitored during follow-up. Predictors for the diagnosis of CGI were obtained by comparing the self-reported symptoms in the questionnaire to the diagnosis of CGI. Multivariable logistic regression analysis was used to combine the strongest predictors in a prediction model. A simple score was developed based on the prediction model to distinguish low, intermediate and high risk patients for CGI. Results: Postprandial pain, exercise-induced pain or weight loss was present in 93% of patients. The majority of patients (n=288, 67%) was diagnosed with CGI and had persistent clinical response after treatment. Self-reported risk profiles showing strong association with CGI were female gender (OR 2.5, 95% CI 1.63.9), age > 60 years (OR 1.4 for 10 year increase, 95% CI 1.0-2.0), concomitant diabetes mellitus (OR 2.0, 95% CI 0.99-4.0), smoking (OR 1.5, 95% CI 0.98-2.3), and use of alcohol (OR1.3, 95% CI (0.96-1.7). Consequently, a c-statistic of 0.68 for the combination of predictors was obtained. Based on a 6-point scoring system patients were categorized as lower (predictive risk 51 79%) for CGI, indicating whether further diagnostic work-up is required (see Table 1). Conclusions: We present a scoring system for the presence of CGI on clinical features and risk profiles alone for patients suspected of CGI. This tool may be useful for clinicians to assess the risk of CGI and to decide whether further diagnostic work-up by means of radiological imaging of the gastrointestinal arteries and functional testing is indicated and worthwhile.