Energy expenditure measurements in ventilated preterm infants are difficult because indirect calorimetry underestimates energy expenditure during gas leaks around uncuffed endotracheal tubes routinely used in preterm infants or during nasal continuous positive airway pressure (CPAP). We, therefore, developed a breath collector that simultaneously sampled expired air expelled at the ventilator outlet and escaping via the tube leak from the infant's mouth and nose. Our breath collector was combined with a proprietary calorimeter (Deltatrac II). In vitro validation was done by methanol burning (Vo 2 , 13.8 mL/min; Vco 2 , 9.2 mL/min) during intermittent positive pressure ventilation (IPPV) with two commonly used ventilators (Sechrist IV-100B and Infant Star). Measurement error was determined at different ventilator flows, peak inspiratory pressures of 12-24 cm H 2 O, and during a complete tube leak. The mean measurement error with both ventilators was low (Vo 2 Ϯ 3%, Vco 2 Ϯ 2%) even during a complete tube leak and did not increase with peak inspiratory pressure. The system response time was 2 min. In vivo measurements at the bedside were performed in 25 preterm infants (body weight, 537-1402 g). Energy expenditure during IPPV was 40 Ϯ 9 kcal/kg per day and 46 Ϯ 15 kcal/kg per day during nasal CPAP. The tube leak in the preterm infants studied during IPPV was 0 to 47%, and during nasal CPAP 84 to 97%. In conclusion, indirect calorimetry performed with our breath collector was accurate during IPPV and nasal CPAP and was unaffected by tube leaks. Nutritional intake has to meet energy requirements. Although the EE of spontaneously breathing growing preterm infants is well documented (1, 2), there is a great need for EE measurements in critically ill and ventilated preterm infants (3). These infants can have large variations in EE due to the underlying disease, surgery, therapeutic interventions, or medication. The reason for the paucity of data is the methodologic difficulties for indirect calorimetry in ventilated preterm infants. The routine use of uncuffed endotracheal tubes leads to unpredictable leakage of expired air through mouth and nose (tube leak) in up to 50% of ventilated neonates (4), and the commonly used nasal CPAP represents a maximal tube leak. In previous measurements of EE, many ventilated preterm infants with tube leaks had to be excluded from analysis (5, 6), and no measurements during nasal CPAP have been reported. We, therefore, developed an indirect calorimetry system with the following capabilities: 1) accurate measurements in intubated preterm infants regardless of tube leak, 2) accurate measurements during nasal CPAP, 3) applicable in combination with pressure-controlled constant-flow ventilators, 4) applicable during increased Fio 2 , 5) accurate measurements of low absolute values of Vo 2 and Vco 2 , and 6) easy to use at the bedside without undue disturbance of the patient.The purpose of the present study was to validate this calorimetry system in vitro and to evaluate its performance in ...