RESUMO
BACKGROUND AND OBJECTIVES: Transcutaneous PCO2 and PO2 measurement systems offer non-invasive blood gas trend monitoring. The aim of this prospective study was to assess bias and precision of a transcutaneous PCO2 and PO2 measurement system incorporating a novel pO2 sensor (Sentec OxiVenT™) in neonates ≥34 weeks of gestational age (GA) admitted to intensive care. METHODS: Transcutaneous PCO2 and PO2 were compared to arterial and capillary blood gas measurements. Bias and precision were calculated by fitting linear mixed models to account for repeated measurements, and influence of clinical covariates on bias and precision was assessed. RESULTS: We obtained 611 paired transcutaneous and blood gas measurements in 110 patients (median GA 38.3 [interquartile range 36.1-39.7] weeks; age 9 [4-15] days; weight 3,000 [2,500-3,500] g). Transcutaneous PCO2 showed significant bias to arterial PCO2 (+0.61; 95% confidence interval 0.46, 0.76 kPa), but not to capillary PCO2 (-0.23; -0.46, 0.002 kPa). Bias of transcutaneous PO2 was significant to arterial PO2 (-2.50; -2.94, -2.06 kPa), while no significant bias compared to capillary PO2 was observed (+0.17; -0.30, 0.64 kPa). Precision intervals were ±1.8/2.0 kPa for arterial versus capillary PCO2 and ±4.9/3.3 kPa for arterial versus capillary PO2 comparisons, respectively. Further, sensor operating temperature (43°C vs. 42°C), soft tissue oedema, vasoactive drugs, weight, and GA significantly altered bias (p < 0.05). CONCLUSIONS: The tested transcutaneous blood gas measurement system showed no significant bias compared to capillary PCO2 and PO2, acceptable bias to arterial PCO2, and limited agreement with arterial PO2. Precision intervals were wide for all comparisons.
