Monitoring of non-invasive ventilation (NIV) in a non-intensive care unit (non-ICU) setting requires a method of evaluating nocturnal PaCO2, such as transcutaneous CO2 monitoring (TcPCO2). However, changing the probe site after 4 h and recalibrating (as recommended) is time-consuming and impractical. Continuous (8-h) TcPCO2 monitoring at a lower electrode temperature (43 degrees C) in this setting has never been formally studied. Patients under intermittent NIV were studied (n = 28, aged 69 +/- 9 years, PaO2: 71 +/- 13 mmHg, PaCO2: 49 +/- 9 mmHg). After calibration and stabilization of TcPCO2 (Radiometer Tina TCM3 capnograph), arterial blood gases (ABG) were measured and compared with transcutaneous readings. In 10 patients who underwent continuous 8-h TcPCO2 recording, ABGs were also measured after 4 and 8 h. The correlation between TcPCO2 and PaCO2 was highly significant (r2 = 0.92, P<0.0001). Mean (TcPCO2 PaCO2) gradient (bias) was: -2.8 +/- 3.8 mmHg; limits of agreement were: (-10.4; +4.8 mmHg). TcPCO2-PaCO2 gradient was lowest (i.e. within-bias +/- 2 mmHg) between 40 and 54 mmHg, increasing below and above these values. Over 8 h, no significant drift of the TcPCO2 signal occurred (ANOVA). No discomfort or skin lesion was noted. In conclusion, with an electrode temperature of 43 degrees C, 8-h continuous monitoring of TcPCO2 was well tolerated, without any local side-effects or significant drift of TcPCO2 signal; when compared to previous reports, lowering the electrode temperature did not decrease performance for CO2 monitoring.