Saline tonometry has been replaced by automated air tonometry (TONOCAP). As with saline tonometry there are some pitfalls to consider. We investigated the influence of different filling handicaps to the tonometry catheter sampling balloon on measurement of regional PCO2 (PrCO2). In an in vitro set-up, PrCO2 was measured using the TONOCAP at intervals of 10 minutes from a 8F tonometry catheter sampling balloon placed in a container with constant PCO2 (PcCO2). Catheter alarms displayed by the TONOCAP device were noted. The following experiments were performed: A) control measurement without modifications; B) adding deadspace of 7 ml into the sampling line; C) placing the sampling balloon in a 3 ml syringe; D) cross-clamping of the sampling balloon. Each experiment was performed four times using two catheters on two TONOCAP devices. Differences between PcCO2 and PrCO2 were calculated as Pc-rCO2. Results are presented as mean +/- SD and were compared using ANOVA with Scheffe's correction. Pc-rCO2 in the control set-up (A) amounted to 0.16 +/- 0.22%. Increasing catheter deadspace (B) resulted in a higher Pc-rCO2 of 3.2 +/- 0.57% (P < 0.0001). Restriction of balloon expansion (C) caused an Pc-rCO2 gap of 1.17 +/- 0.36% (P < 0.0001). With filling restriction (D) the Pc-rCO2 gap increased to 3.87 +/- 0.29% (P < 0.0001). The catheter alarms provided were not able to indicate all catheter problems and initial alarms disappeared although the problems continued. We conclude that impaired catheter deadspace to balloon volume relation can negatively influence accuracy of PrCO2 measurement by the TONOCAP.