The rapid shallow breathing index (RSBI) is commonly used clinically for predicting the outcome of weaning from mechanical ventilation. We compared the RSBI and its predictive accuracies measured under 5 ventilatory strategies before weaning trials. Ninety-eight patients were included and divided into successful (n=71) and failed (n=27) groups based on their weaning outcomes. The RSBI was randomly measured when patients spontaneously breathed 21% O2 with no ventilator support (the control strategy) or were connected to ventilator breathing with 21% or 40% O2 and 0 or 5 cm H2O of continuous positive airway pressure (CPAP). We found that the RSBI values did not exhibit significant differences among the 4 ventilator strategies, but all were higher than that of the control; this remained valid in the non-chronic obstructive pulmonary disease (COPD) subgroup, but not in the COPD subgroup. Values of the area under the receiver operating characteristic curve of the RSBI for the 5 strategies were 0.51-0.62 with no significant difference between any 2 strategies. The incidences of adverse reactions (respiratory rateor =35 breaths/min or oxygen saturationor =89% foror =1 min) were relatively high for the 21% O2-0 and 5 cm H2O CPAP groups (20 patients each) and low for the 40% O(2)-5 cmH2O CPAP group (2 patients). We concluded that RSBI values increased with the use of a ventilator, but not with additional applications of 40% 02 and/or 5 cm H2O CPAP. Their accuracies for predicting weaning outcome were unaltered by any of these interventions, but the incidence of adverse reactions increased with the use of the ventilator and decreased with additional 40% O2 supplementation.