Predictors of sleep hypoxemia in children with cystic fibrosis

Objectives: To identify the determinants of nocturnal hypoxemia in children with CF using clinical parameters, polysomnography (PSG), and lung function. Hypothesis: Sleep hypoxemia in children with CF is predicted by both apnea hypopnea index (AHI) and percent predicted forced expiratory volume in one second (pFEV1). Design: Retrospective case series. Methods: Children aged 5-18 years were included based on (i) a diagnosis of CF; and (ii) availability of concurrent PSG and pFEV1 data. The impact of (i) demographic and clinical parameters; and (ii) PSG and pFEV1, on the total sleep time spent with arterial oxygen saturation below 90% (TSpO ₂ < 90) was measured using regression analysis. P-value <0.05 was considered significant. Results: The mean age was 11.6 years (95% confidence interval: 9.5, 13.1). Twenty of 35 (57%) were boys and the mean body mass index percentile was 42.1 (31.5, 52.6). The most common ethnicity was white (66%). OSA was diagnosed in 50%. Neither demographic predictors nor clinical variables predicted the severity of hypoxemia (R ² = 0.23, P = 0.09). While pFEV1 and PSG variables accounted for significant proportion of the overall variance in TSpO ₂ < 90 (R ² = 0.53, P < 0.001), pFEV1 was identified as the single best predictor of sleep hypoxemia. A pFEV1 cut-off of 53% indicated a sensitivity of 0.80 and a specificity of 0.87 in predicting sleep hypoxemia. Conclusions: pFEV1 is the best predictor of sleep hypoxemia in children with CF and referred for PSG. No demographic or clinical predictors of hypoxemia were identified in this population.