Stress ulcer prophylaxis in mechanically ventilated patients: integrating evidence and judgment using a decision analysis

Objective: Stress ulcer prophylaxis with a histamine-2 receptor antagonist can reduce the risk of gastrointestinal bleeding in mechanically ventilated patients but may also increase the risk of ventilator-associated pneumonia. We sought to clarify the tradeoffs involved in selecting a prophylactic strategy. Design: Decision analysis. Patients and participants: A decision tree was constructed for a hypothetical cohort of patients receiving mechanical ventilation for an expected duration of longer than 48 h, using probabilities estimated from the published literature. Interventions: Patients in the model could receive either prophylaxis with a histamine-2 receptor antagonist or no prophylaxis. Sensitivity analyses were preformed varying the estimated probabilities over their plausible ranges. Measurements and results: Both strategies were associated with approximately the same baseline expected mortality (16.6% for histamine-2 receptor antagonists and 16.9% for no prophylaxis, risk difference 0.3%). Varying the estimated probabilities resulted in only small changes in both the expected mortality and the absolute risk reduction associated with the preferred treatment. At the extremes of assumptions the absolute mortality reduction ranged from 0.1% to 3.3%. Conclusions: No single strategy of stress ulcer prophylaxis is preferred when mortality is used as the outcome. In the absence of a clinical trial demonstrating survival benefit the individual clinician's assumptions regarding the effect of prophylaxis on gastrointestinal bleeding and pneumonia and the attributable mortality of pneumonia vs. gastrointestinal bleeding will have a significant effect on the decision.