Background: In various fields of medicine the transcutaneous measurement (p(tc)CO₂) of carbon dioxide pressure (pCO₂) has been accepted as a reliable measuring method while for other disciplines the method has been doubted. Nevertheless, some minor therapeutic interventions, such as electroconvulsive therapy and breathing biofeedback used in psychiatry could benefit considerably from transcutaneous monitoring. The study presented here investigated the accuracy of transcutaneous measurement as compared to end-tidal (p(et)CO₂) and capillary (p(cap)CO₂) measurements in hyperventilation.
Methods: In this study 22 healthy volunteers underwent a hyperventilation procedure with assessment of p(tc)CO₂, p(et)CO₂, p(cap)CO₂, breathing and pulse rates. The three measurement methods of pCO₂ were compared pairwise by Bland-Altman diagrams before and at the end of hyperventilation. The time delay between p(et)CO₂ and p(tc)CO₂ was determined for each individual participant by a stepwise shifting the time course of p(et)CO₂ until an optimal congruence with p(tc)CO₂ was reached.
Results: The study group consisted of 10 men and 12 women with a mean age of 30.5 ± 9.4 years. The experimental procedure lasted an average of 27.7 min including 1.9 min for the baseline measurement and 11.6 min for the actual hyperventilation procedure. The course of pCO₂ from baseline to the end of the hyperventilation phase and back to normal baseline followed a U-curve in all individuals. The pCO₂ could be reduced by a mean of approximately 45 % for the 3 measurement methods from the individual baseline of 35 mmHg. The breathing and pulse frequencies increased by 26.8 % and 17.3 %, respectively. A total of 91 p(cap)CO₂ values (mean 4 values per person) and 1,218 pairs of p(et)CO₂ and p(tc)CO₂ values (mean 55 values per person) could be evaluated. The difference of the pCO₂ values for two measurement methods revealed similar standard deviations in the Bland-Altman diagrams of about 2.5 mmHg for the 3 pairwise comparisons. The absolute measurement differences between p(et)CO₂ and p(tc)CO₂, p(cap)CO₂ and p(tc)CO₂ as well as p(et)CO₂ and p(cap)CO₂ were averaged for each single participant and afterwards across all individuals resulting in values of 2.0, 2.5 and 2.3 mmHg, respectively. When the baseline and hyperventilation periods were examined separately in this analysis they did not show any relevant differences. The individually determined delays of p(tc)CO₂ in regard to p(et)CO₂ averaged 53 s. The time delay exceeded 1 min in only two study participants.
Conclusions: In the presented hyperventilation experiment the agreement between the p(tc)CO₂ values and the p(et)CO₂ and p(cap)CO₂ measurements were comparable to the agreement between the latter two. Altogether, the comparability of the measurement methods seemed satisfactory. The delay between the time courses of p(et)CO₂ and p(tc)CO₂ of <1 min for most participants was short and is acceptable for clinical practice. Because only healthy subjects were tested these results might not be the same in children and critically ill patients and could differ from the presented results.