This single-center secondary data cohort study (NICU, RWTH Aachen) evaluated whether gestational age affects the accuracy of transcutaneous CO₂ (tcpCO₂) monitoring compared to capillary pCO₂ (pcCO₂) in 116 neonates during the first postnatal week, using 926 simultaneous paired measurements analyzed with linear mixed-effects models.
Extremely preterm infants (<28 weeks) had a median tcpCO₂ overestimation (ΔpCO₂) of 7.6 mmHg [IQR 2.4–13.9], versus 3.1 mmHg for very preterm (28–32 weeks), 0.6 mmHg for late preterm (32–37 weeks), and −0.5 mmHg for term infants. In multivariable analysis, gestational age, early-onset bacterial infection (EOBI), and capillary pCO₂ were independently associated with ΔpCO₂; birth weight was not after adjustment.
- Single-center study with a small number of extremely preterm infants (n=22), limiting generalizability. - Sensor temperature was set at 40.2°C (below the recommended 42°C), which is known to increase bias and was not adjustable in this dataset. - Blood sampling was clinically driven, not protocolized, potentially overrepresenting periods of instability and biasing agreement estimates.
When using tcpCO₂ monitoring in extremely preterm infants (<28 weeks), clinicians should expect a systematic overestimation of ~7–8 mmHg compared to capillary values — more frequent blood gas confirmation is warranted in this group. Early-onset bacterial infection further widens the gap and should prompt additional caution when interpreting tcpCO₂ readings.
