Study objectives: To evaluate the initial measurement of arterial oxygen saturation (SaO2) as a predictor of outcome in acute childhood asthma compared with other factors of past and present asthma history.
Design: Prospective observational double-blind study.
Setting: The emergency department of an urban pediatric hospital with a 1988 annual census of 50,000 children.
Type of participants: Two hundred eighty children with recurrent wheezing that was diagnosed by a physician as asthma, who presented to the ED with wheezing.
Intervention: SaO2 was measured on arrival in the ED, and a detailed history of the present attack and past asthma was recorded. Children were treated according to then-current practice guidelines. Parents were contacted by telephone to determine the outcome of the attack; a "poor outcome" was defined as admission to hospital or re-presenting with ongoing symptoms to receive medical care if sent home from the ED. A "worst outcome" was defined as receiving IV aminophylline and steroids after failing to respond to repeated bronchodilation and oral steroids.
Measurements and main results: The proportion of children at each percent SaO2 who had a poor outcome increased with decreasing SaO2 (r = .97). Likelihood ratios for a poor outcome were 35 (confidence interval [CI], 11 to 150) for an SaO2 of 91% or less compared with 96% or more and 4.2 (CI, 2.2 to 8.8) for an SaO2 of 92% to 95% compared with 96% or more. An SaO2 of 91% or less predicted with a sensitivity of 100% and a specificity of 84% those children with a worst outcome who required IV therapy. Other factors of current or past asthma history failed to predict outcome.
Conclusion: We have shown that in acute childhood asthma, the initial level of SaO2 reflects severity as it predicts the likelihood of poor outcome. This predictive quality of SaO2 is independent of current or past clinical factors.