Previous research has suggested correlations between the end-tidal partial pressure of carbon dioxide (PETCO2) and the partial pressure of arterial carbon dioxide (PaCO2) in mechanically ventilated patients, but both the relationship between PETCO2 and PaCO2 and whether PETCO2 accurately reflects PaCO2 in neonates and infants are still controversial. This study evaluated remote sampling of PETCO2 via an epidural catheter within an endotracheal tube to determine the procedure's clinical safety and efficacy in the perioperative management of neonates. Abdominal surgery was performed under general anesthesia in 86 full-term newborns (age 1-30 days, weight 2.55-4.0 kg, American Society of Anesthesiologists class I or II). The infants were divided into 2 groups (n = 43 each), and carbon dioxide (CO2) gas samples were collected either from the conventional position (the proximal end) or a modified position (the distal end) of the epidural catheter. The PETCO2 measured with the new method was significantly higher than that measured with the traditional method, and the difference between PETCO2 and PaCO2 was also reduced. The accuracy of PETCO2 measured increased from 78.7% to 91.5% when the modified sampling method was used. The moderate correlation between PETCO2 and PaCO2 by traditional measurement was 0.596, which significantly increased to 0.960 in the modified sampling group. Thus, the PETCO2 value was closer to that of PaCO2. PETCO2 detected via modified carbon dioxide monitoring had a better accuracy and correlation with PaCO2 in neonates.