Subscribe to RSS
DOI: 10.1055/s-0040-1704062
RISK FACTORS FOR MORTALITY AMONG PATIENTS ADMITTED TO HOSPITAL FOR LOWER GASTROINTESTINAL BLEEDING: A PREDICTIVE MODEL
Publication History
Publication Date:
23 April 2020 (online)
Aims Lower GI bleeding (LGIB) is a common cause of hospitalization and death. A recent UK multicenter study has developed and validated a risk score for safe outpatient management. However, data on factors associated with mortality in higher risk patients are lacking. Aim of this study was to assess factors associated with mortality and to derive and validate a predictive model.
Methods Multicenter, Italian, prospective, observational study was conducted from 1st October 2018 to 28th October 2019 including adult patients with LGIB (ALIBI study). The study cohort was divided in derivation cohort (from 1st October 2018 to 25th June 2019) and validation cohort (from 26th June to 28th October 2019). Logistic regression was performed to identify factors independently related to death, by computing odds ratio (OR) with 95% confidence interval (CI). A predictive model was subsequently derived and validated.
Results Overall, 1198 cases of LGIB were collected in 14 hospitals. The derivation and validation cohort included 791 and 407 patients, respectively. In-hospital mortality occurred in 41 (3.4%) patients. Among the derivation cohort, increasing patient age (OR 1.10, CI 1.04-1.17), a higher Charlson Comorbidity Index (OR 1.24, CI 1.04-1.47), bleeding onset during hospitalization (OR 4.06, CI 1.62-10.2) and hemodynamic instability at presentation (OR 4.72, CI 1.87-11.89) were independent risk factors for in-hospital mortality. The model had good discrimination for higher risk of death as denoted by the C statistic (0.85, CI 0.78-0.9 in the derivation cohort; 0.74, CI 0.62-0.85 in the validation cohort) and was well calibrated according to the goodness-of-fit test (p=0.725 and p=0.117, respectively).
Conclusions Elderly patients with multiple comorbidities who develop LGIB during hospitalization and present with hemodynamic instability have a higher mortality risk. We developed and validated a novel clinical predictive model with good discriminative performance to identify patients with LGIB and higher mortality risk.