The laryngeal mask airway (LMA) has become a popular tool for airway management in selected adult and pediatric patients undergoing routine surgical procedures. The relationship between end-tidal and arterial carbon dioxide during controlled ventilation via the LMA in infants under 10 kg has not been reported. After induction of general anesthesia, the LMA was placed in 12 healthy infants and mechanical ventilation initiated. After maintaining steady-state level of end-tidal carbon dioxide (minimum 5 min), an arterial blood sample was obtained and end-tidal carbon dioxide level noted. The laryngeal mask was then removed, the trachea intubated, and mechanical ventilation resumed with initial ventilatory variables. After reaching a steady-state level of end-tidal carbon dioxide, a second arterial sample was obtained and end-tidal carbon dioxide level noted. The mean end-tidal carbon dioxide and arterial partial pressure of carbon dioxide obtained during ventilation were 42.2 +/- 7.9 and 47.1 +/- 11.0 (LMA) and 37.4 +/- 4.6 and 42.6 +/- 6.7 (endotracheal tube), respectively. Analysis of differences between partial pressure of carbon dioxide and end-tidal carbon dioxide using the Bland and Altman method revealed bias+/-precision of 4.9 +/- 3.9 and 5.3 +/- 3.2 with ventilation via the laryngeal mask and endotracheal tube. Our data indicate that, while ventilating infants under 10 kg with LMA, end-tidal carbon dioxide is an accurate indicator of arterial partial pressure of carbon dioxide.