The aim of this study was to describe changes in regional intramucosal PCO(2) (Pr(CO(2)) measured with capnometric recirculation gas tonometry [CRGT]) in patients with acute respiratory failure, who proceed from mechanical ventilation to weaning. In addition, we compared the predictive power for the weaning outcome of CRGT measurements obtained during mechanical ventilation to the frequency/ tidal volume (f/VT) ratio. A total of 24 patients (31 weaning trials) were included in the study, but four of the 24 patients (17%) were excluded because of extubation failure. Of the remaining 27 weaning trials in 20 patients, 12 (44%) were unsuccessful. Changes observed in patients with weaning failure (increase in Pr(CO(2)) from 60.4 +/- 15.0 mm Hg in mechanical ventilation to 67.4 +/- 21.0 mm Hg, in weaning) were significantly different (p = 0.046) from those observed in patients with weaning success (fall in Pr(CO(2)) from 61.5 +/- 15.0 mm Hg in mechanical ventilation to 56.3 +/- 16.7 mm Hg in weaning). However, absolute values of Pr(CO(2)) were not significantly different between patients with weaning success and failure, neither during mechanical ventilation (success, 61.5 +/- 15.0 versus failure, 60.4 +/- 15.0 mm Hg, p = 0.848) nor during weaning (success, 56.3 +/- 16.7 versus failure, 67.4 +/- 21.0 mm Hg, p = 0.135). The best single predictor for weaning outcome was the f/VT ratio measured early during weaning (area under the curve: 0.844 +/- 0.081; adjusted odds ratio for threshold value </= 105: 42.0, 95% CI 3.8 to 469.1, p = 0.002). CRGT could confirm a significant increase in Pr(CO(2)) during weaning in patients who finally failed the weaning trial. However, differences between patients with weaning success and failure were small and CRGT did not replace or improve the predictive power of the f/VT ratio for weaning outcome.