Abstract

The sensitivity and specificity of overnight recording of arterial oxygen saturation (SaO2) in routine clinical practice was evaluated in 41 subjects who were being investigated for possible sleep apnoea-hypopnoea syndrome. SaO2 was measured with an ear probe oximeter (Biox IIa) and chart recorder during an "acclimatisation" night immediately before a detailed polysomnographic study. The recordings were classified by two observers as positive, negative, or uninterpretable. Twelve of the 41 patients had the obstructive sleep apnoea syndrome when defined in terms of an apnoea-hypopnoea index greater than 15 events an hour on the second night. The sensitivity of nocturnal SaO2 on the acclimatisation night when the diagnostic criterion was an apnoea-hypopnoea index of greater than 5, greater than 15, and greater than 25/h was 60%, 75%, and 100% respectively. Corresponding values for specificity were 95%, 86%, and 80%. Oximetry alone therefore allowed recognition of a moderate or severe sleep apnoea syndrome. In routine practice an appreciable number of equivocal results is likely and repeat oximetry or more detailed polysomnography will then be required if clinical suspicion is high.