To determine whether gastric intramural pH (pHi), an indirect measure of gastric mucosal ischemia, can be used to predict the success of weaning from mechanical ventilation. Gastric mucosal ischemia (and, therefore, acidosis) may develop in patients during unsuccessful attempts to wean them from mechanical ventilation because blood flow from nonvital areas (for example, splanchnic bed) is diverted to meet the increased demands of respiratory muscles. Cohort study. Intensive care unit. Twenty-nine patients receiving assisted mechanical ventilation for respiratory failure who were thought by their physicians to be weanable from mechanical ventilation. Simultaneous samples of arterial blood and gastric juice were obtained from patients during assisted mechanical ventilation, as well as during weaning trials. The predictor variable, pHi, was calculated using the following equation: 6.1 + log HCO3/(gastric PCO2 0.0307). The outcome was success or failure of weaning, decided by physicians blinded to the study. Patients who could not be weaned from mechanical ventilation had a substantially reduced gastric pHi (7.36 during mechanical ventilation compared with 7.09 during weaning [difference, 0.27; 95% CI, 0.12 to 0.42; P < 0.01]). Patients who were successfully weaned from mechanical ventilation showed no change in pHi (7.45 during mechanical ventilation compared with 7.46 during weaning [difference, 0.01; CI, 0.01 to 0.03; P = 0.29]). The sensitivity and specificity of pHi in predicting weaning success or failure were both 100% (CI, 81 to 100 and 72 to 100, respectively). Gastrointestinal acidosis may be an early sign of weaning failure. Measurement of pHi, which is simple and rapid, may be of practical value in predicting the likelihood of weaning success or failure during weaning trials.