Several threshold values for predicting weaning outcome from mechanical ventilation have been proposed. These values, however, have been obtained in nonhomogeneous patient populations. The aim of the present study was to determine the threshold values in chronic obstructive pulmonary disease (COPD) patients and compare them to those reported for nonhomogeneous patient populations. The initial weaning trial included 81 COPD patients. Fifty-three of them underwent a successful weaning trial, whereas 28 failed it. The latter were enrolled into the present investigation, and were restudied during a subsequent successful trial. The weaning indices used were those reported in the literature. The threshold values obtained were within 10% of those reported for a nonhomogeneous patients population only for tidal volume and effective compliance. The classification error was <20% for maximal inspiratory pressure (MIP), occluded inspiratory pressure swing (deltaPI)/MIP, rapid and shallow breathing (respiratory frequency/tidal volume), and compliance, rate, oxygenation, pressure index (CROP), whereas the area under the receiver operating characteristic curves was >0.9 only for deltaPI/MIP and CROP. In conclusion, the threshold values obtained in chronic obstructive pulmonary disease patients who failed the first weaning attempt differed from those previously reported. Although a gold standard weaning index is not available for chronic obstructive pulmonary disease patients, the occluded inspiratory pressure swing/ maximal inspiratory pressure and compliance, rate, oxygenation, pressure index may be candidates for such a role.