Characteristics and predictors for gastrointestinal hemorrhage among adult patients with dengue virus infection: Emphasizing the impact of existing comorbid disease(s)

BACKGROUND Gastrointestinal (GI) bleeding is a leading cause of death in dengue. This study aims to identify predictors for GI bleeding in adult dengue patients, emphasizing the impact of existing comorbid disease(s). METHODS Of 1300 adults with dengue virus infection, 175 (mean age, 56.5±13.7 years) patients with GI bleeding and 1,125 (mean age, 49.2±15.6 years) without GI bleeding (controls) were retrospectively analyzed. RESULTS Among 175 patients with GI bleeding, dengue hemorrhagic fever was found in 119 (68%) patients; the median duration from onset dengue illness to GI bleeding was 5 days. Gastric ulcer, erythematous gastritis, duodenal ulcer, erosive gastritis, and hemorrhagic gastritis were found in 52.3%, 33.3%, 28.6%, 28.6%, and 14.3% of 42 patients with GI bleeding who had undergone endoscopic examination, respectively. Overall, nine of the 175 patients with GI bleeding died, giving an in-hospital mortality rate of 5.1%. Multivariate analysis showed age ≥60 years (cases vs. controls: 48% vs. 28.3%) (odds ratio [OR]: 1.663, 95% confidence interval [CI]: 1.128-2.453), end stage renal disease with additional comorbidities (cases vs. controls: 1.7% vs. 0.2%) (OR: 9.405, 95% CI: 1.4-63.198), previous stroke with additional comorbidities (cases vs. controls: 7.4% vs. 0.6%) (OR: 9.772, 95% CI: 3.302-28.918), gum bleeding (cases vs. controls: 27.4% vs. 11.5%) (OR: 1.732, 95% CI: 1.1-2.727), petechiae (cases vs. controls: 56.6% vs. 29.1%) (OR: 2.109, 95% CI: 1.411-3.153), and platelet count <50×109 cells/L (cases vs. controls: 53.1% vs. 25.8%) (OR: 3.419, 95% CI: 2.103-5.558) were independent predictors of GI bleeding in patients with dengue virus infection. CONCLUSIONS Our study is the first to disclose that end stage renal disease and previous stroke, with additional comorbidities, were strongly significant associated with the risk of GI bleeding in patients with dengue virus infection. Identification of these risk factors can be incorporated into the patient assessment and management protocol of dengue virus infection to reduce its mortality.