RESUMEN
BACKGROUND: Children with sleep apnea have increased morbidity if the sleep apnea is untreated. Polysomnography (PSG) is used to detect sleep apnea, but in children, there are technical difficulties that make the diagnosis more difficult. Cerebral near infrared spectroscopy (NIRS) has the ability to detect tissue hypoxia by measuring regional oxygen saturation (rSO2 ). We hypothesized that when used as an adjunct to PSG testing, cerebral NIRS would better detect the effects of sleep apnea in children than arterial pulse oximetry (SpO2 ). Specifically, we aimed to show that the apnea/hypopnea index (AHI) calculated with rSO2 from the NIRS would be greater than that calculated with SpO2 . METHODS: Forty-seven patients under 18 years of age who underwent PSG testing for sleep apnea were evaluated. Cerebral NIRS was utilized in addition to PSG. The AHI was calculated using SpO2 and compared to the AHI calculated using the rSO2 . A pediatric pulmonologist who was unaware of the NIRS data evaluated each patient for sleep apnea. Data are median (interquartile range). RESULTS: The median AHI(rSO2 ) was 2.4(1.2,5.1), significantly greater (P < 0.0001) than the AHI(SpO2 ) of 0.7(0.4,1.2). Four patients were diagnosed with sleep apnea; however, only one had an AHI(SpO2 ) ≥ 5, a typical threshold for the diagnosis of sleep apnea. All 4 subjects had an AHI(rSO2 ) ≥ 5 but 10 patients without sleep apnea had a value ≥5. The sensitivity and specificity for using the AHI(rSO2 ) to diagnose sleep apnea was 100% and 76.7%, respectively. CONCLUSION: Consistent with the ability of NIRS to detect tissue hypoxia, we found that the AHI calculated with rSO2 was greater than that using SpO2 . We conclude that NIRS has potential as a valuable adjunct to PSG in evaluating patients for sleep apnea and warrants further investigation for this purpose.