Prognostic Characteristics of Hospitalized Patients Without Overt GI Bleeding With Subacute Drop in Hemoglobin

Abstract

Introduction: Patients admitted to the hospital with a principal diagnosis that does not include GI bleeding who subsequently develop overt GI bleeding with corresponding drop in Hb are known to have higher morbidity and mortality. Little is known about these patients who have a subacute drop in Hb without evidence of GI bleeding. Although the causes of a subacute drop in Hb is likely multifactorial, GI consultation and subsequent endoscopy is often requested and performed. Our aim is to identify prognostic factors that affect the likelihood of finding a GI source in hospitalized patients with a principal diagnosis that does not include GI bleeding who subsequently develop a subacute drop in Hb. Methods: The electronic health records from our community hospital were used to identify all patients admitted from October 2016 to May 2018 who underwent any of the CPT codes associated with endoscopy. From these patients, we excluded those with an admitting diagnosis for GI bleed, GI hemorrhage, melena, hematochezia, and hematemesis using their ICD-10 Codes. This left a group of patients who had undergone an endoscopy during a hospitalization for a non GI bleed admitting diagnosis. Data mining was undertaken to include patients who had a subacute drop in their Hb defined as drop in Hb ≥ 3 over their hospitalization. Patients with history of significant GI bleeding defined as hospitalization and blood transfusion were also excluded. From this study group, characteristics were obtained and compared to endoscopic findings. The intent is to perform a multivariate analysis on those characteristics that are statically significant or presented most frequently. Results: Characteristics will include age, admitting diagnosis, use of anticoagulation, surgeries, number of phlebotomies, +FOBT, history of chronic anemia, liver disease and kidney disease. Using the endoscopy report, we will determine if a GI etiology was discovered. Each of the above characteristics will be compared to the outcomes in a univariate manner. Characteristics that are statistically significant or occur with most frequency will be evaluated in a multivariate analysis to identify characteristics that may be prognostic. Conclusion: The focus of this investigation is to identify characteristics that can be used to determine the likelihood of endoscopy revealing the cause of subacute drop in Hb in patients hospitalized with a principal diagnosis other than GI bleeding. These characteristics can be used in future prospective studies.