Bilateral volume reduction surgery (VRS) improves lung function for selected patients with emphysema. However, predictors of outcome are not well defined. We reviewed the preoperative characteristics of the first 47 consecutive patients who underwent bilateral VRS at the Massachusetts General Hospital in order to define potential predictors of unacceptable outcome. Preoperative data included spirometry, plethysmography, diffusion of carbon monoxide (Dco), maximum inspiratory pressure (MIP), maximum expiratory pressure, resting arterial blood gases (ABG), cardiopulmonary exercise testing with ABG and lactate sampling, and radionuclide ventriculography. Prepulmonary and postpulmonary rehabilitation 6-min walk tets (6MWT), and preoperative chest CT scans were also obtained. Twenty-two subjects were male and 17 of the subjects were on the lung transplant list. Patient characteristics included age of 60.5 +/- 7.5 years, FEV1 of 0.67 +/- 0.20 L, total lung capacity of 7.56 +/- 1.7 L, Dco of 7.40 +/- 4.1 mL/min/mm Hg, and PaCO2 of 41.6 +/- 6.4 mm Hg (mean +/- SD). The FEV1, vital capacity, MIP, resting room air PaCO2, prepulmonary and postpulmonary rehabilitation 6MWT, and PaCO2 at maximum oxygen consumption correlated with length of hospitalization (p < 0.05). Based on analysis of 41 of 47 patients for whom there were complete data, the inability to walk more than 200 m on the 6MWT before or after preoperative pulmonary rehabilitation, and resting PaCO2 > or = 45 mm Hg were the best predictors of an unacceptable outcome. If either of these characteristics was present, six of 16 vs zero of 25 died (Fisher's Exact Test, p = 0.0025, one-tailed) and 11 of 16 vs four of 25 had hospital courses > 21 days (p < 0.002). Both the 6MWT < 200 m and resting PaCO2 > or = 45 mm Hg alone correlated with death (p = 0.004 and p = 0.012, respectively) and the resting PaCO2 > or = 45 mm Hg correlated with hospital days > 21 (p = 0.0002). In conclusion, the data suggest that the inability to walk at least 200 m in 6 min before or after pulmonary rehabilitation and a resting room air PaCO2 > or = 45 mm Hg are excellent preoperative predictors of unacceptable postoperative outcomes.