According to the traditional classification an obstructive apnea is defined by absence of respiratory flow and presence of respiratory efforts. However, when we examined these apneas according to the presence or absence of the magnified cardiac oscillation signal (Lemke et al, Pediatr. Res., 37:340A, 1995), it became clear that these obstructive apneas were actually mixed, with the intermittent presence of the cardiac oscillation signal on the respiratory flow. We have therefore hypothesized that purely obstructive apneas are non-existent and that what used to be called obstructive by the traditional method were indeed mixed apneas. To test this hypothesis we examine 1931 apneas ≥ 3 seconds in 15 infants (birthweight 1200±114g; study weight 1400±136g; gestational age 29±1 wk; postnatal age 29±4 d). We used a flow-through system to measure respiratory pattern. Of the 1931 apneas, 1481 (77%) were central and 450 (23%) were mixed by the cardiac oscillation method; 285 (15%) were ≥ 10 seconds and 55 (3%) were longer than 20 seconds. No obstructive apneas were observed. However, of the mixed apneas, 11% would have been classified as obstructive by the previous method, because there were sporadic respiratory efforts present in the absence of respiratory flow. These were clearly interspersed with periods of airway opening and presence of cardiac oscillations. These findings suggest that when apneas are measured with a more accurate index of airway obstruction, ie., absence of a magnified cardiac oscillation signal, obstructive apneas are non-existent. We conclude that in preterm infants apneas should be defined as central or mixed only and speculate that the primary mechanism responsible for all apneas is central.Supported by the Children's Hospital of Winnipeg Research Foundation.