Background: Transcutaneous measurement of carbon dioxide (Tc(co(2))) provides a non-invasive estimation of arterial carbon dioxide (Pa(co(2))). Nasal capnography (Pe'(co(2))) is used to assess ventilation during monitored anaesthesia care (MAC) with sedation since it can readily detect apnoea. We compared the agreement between Tc(co(2)) and Pe'(co(2)) with Pa(co(2)) and the ability to detect hypercarbia in patients under deep sedation.
Methods: Forty healthy female subjects receiving deep sedation for hysteroscopy were studied. A Tc(co(2)) (TOSCA 500, Radiometer, Inc., Westlake, OH, USA) electrode was applied to the earlobe and Pe'(co(2)) capnography was monitored using nasal side-stream sampling. All subjects received oxygen (3 litre min(-1)). Subjects were evaluated at intervals using a modified Ramsay sedation score until they reached a score >or=5. Arterial blood gas values were compared with Tc(co(2)) and Pe'(co(2)) values. Bland-Altman, linear regression, and receiver operator characteristics analysis were performed.
Results: The mean (sd) absolute difference between the Tc(co(2)), Pe'(co(2)), and the Pa(co(2)) were 0.43 (0.35) and 1.06 (0.8) kPa, respectively (P=0.002). Tc(co(2)) demonstrated a mean bias (2 sd) of 0.23 (0.07-0.4) kPa with Pa(co(2)) compared with -0.93 (-1.24 to -0.63) kPa for Pe'(co(2)). One minute before blood sampling, the sensitivity of the Tc(co(2)) monitor for detecting Pa(co(2)) >6.65 kPa was greater than for Pe'(co(2)) (66.7% vs 33.3%, P<0.01).
Conclusions: Tc(co(2)) demonstrated better agreement with Pa(co(2)) than Pe'(co(2)) for patients under MAC with deep sedation. Tc(co(2)) monitoring was more sensitive for detection of Pa(co(2)) >6.65 kPa than Pe'(co(2)).