To study the possible influence of sleeping position on arterial oxygen saturation, measured by pulse oximetry (SpO2), 7-h overnight recordings of breathing movements and ECG were performed in 43 infants (median age 2.4 months, range 0.2-11 months) at increased risk of sudden infant death syndrome (SIDS). Infants were randomly allocated to start sleeping either in their usual sleeping position or in the opposite position. After 3.5 h, all infants were gently turned over. Thus, each infant served as their own control. Recordings were analysed for sleep time, baseline SpO2 (only during regular breathing), and the number and duration of desaturations (a decrease in SpO2 to < or = 80%). In the prone position, a significantly higher proportion of time was spent asleep (median 79% versus 70%; p < 0.05). Median baseline SpO2 was 98.8% (91.7-100%) in the prone and 99.0% (92.0-100%) in the supine position (ns). A total of 191 desaturations were found in 29 recordings; 96 in the prone and 95 in the supine position (ns). One infant subsequently died of SIDS while sleeping in the prone position. He had a relatively high number of desaturations (n = 12) which all occurred in the prone position. These results confirm earlier studies which could not find a significant influence of sleeping position on baseline oxygenation. The occurrence of desaturations in the prone position only in the infant who subsequently died requires further investigation.