Risk Factors for Patient-Important Upper Gastrointestinal Bleeding

Rationale: Patient-important gastrointestinal bleeding is an endpoint developed by patients and family members; however, risk factors for this outcome are unknown. Objectives: We sought to identify risk factors for patient-important upper gastrointestinal bleeding among invasively ventilated adults. Methods: This preplanned regression analysis of an international trial database evaluated baseline and time-varying risk factors in the preceding 3 days for patient-important upper gastrointestinal bleeding, accounting for illness severity and the competing risk of death. Measurements and Main Results: Patient-important upper gastrointestinal bleeding occurred in the ICU among 131 of 4,821 (2.7%) patients. Baseline APACHE II score-hazard ratio (HR), 1.24 (95% confidence interval [CI] = 1.12, 1.37) per 5-point increase-and the following were associated with greater risk of patient-important bleeding: inotropes or vasopressors (HR, 2.05 [95% CI = 1.35, 3.12]), severe thrombocytopenia (platelet count, <50 × 10⁹/L) (HR, 2.21 [95% CI = 1.24, 3.94]) and platelet inhibitor drugs (HR, 1.69 [95% CI = 1.11, 2.56]). A lower bleeding risk was associated with pantoprazole (HR, 0.36 [95% CI = 0.25, 0.54]) and enteral nutrition (HR, 0.81 [95% CI = 0.68, 0.97]) for every increase of 500 ml/d. There was no interaction between enteral nutrition and pantoprazole (interaction P = 0.94). Allocation to pantoprazole was associated with a lower risk of patient-important upper gastrointestinal bleeding regardless of the volume of enteral nutrition (for 500 ml/d: HR, 0.36 [95% CI = 0.22, 0.58]; for no enteral nutrition: HR, 0.36 [95% CI = 0.18, 0.72]). The association of enteral nutrition and bleeding was similar with pantoprazole (HR, 0.82 [95% CI = 0.63, 1.07]) or without pantoprazole (HR, 0.81 [95% CI = 0.66, 1.00]). Conclusions: Several factors are associated with the risk of patient-important upper gastrointestinal bleeding during invasive ventilation.