Home oximetry studies for diagnosis of sleep apnea/hypopnea syndrome: limitation of memory storage capabilities.
BACKGROUND: Memory oximeters enable diagnostic studies for sleep apnea hypopnea syndrome (SAHS) to be performed in the home. However, memory capabilities may be limited. STUDY OBJECTIVES: To compare a pulse oximeter used at home with an 8-h memory, storing data every 12 s, and in the laboratory, with on-line recording every 2 s. DESIGN: Prospective cohort study. SETTING: Patients' homes and a sleep laboratory. PATIENTS: One hundred patients with suspected SAHS. MEASUREMENTS: Home oximetry and a laboratory full polysomnography. The number of >/= 4% dips in pulse oximetric saturation (SpO(2)) was calculated for each study. Daytime sleepiness was assessed by the Epworth Sleepiness Scale (ESS) score. RESULTS: The mean dips per hour were 5.3/h (range, 0 to 53/h) for home studies and 13.4/h (range, 0 to 106/h) for laboratory studies; the relationship between home and laboratory studies was as follows: home = (0.4 x laboratory) - 0.01 +/- 11.2; r(2) = 0.64. Mean difference was 8.4/h (- 2.5 to + 77.9/h), which correlated with the mean of the measurements. At a cutoff point of 10/h, 52 studies were both negative and 13 studies were both positive. Nineteen home studies were false-negatives. Sensitivity was 0.41, and specificity was 1.0. In these 19 studies, 7 patients had an ESS score > 10 and 4 patients had an ESS score > 14. To confirm that differences were due to different sampling rates, 16 additional patients had on-line data and stored data collected simultaneously in the laboratory. Mean dips per hour were 3.2/h (range, 0.1 to 18.3/h) for the stored data and 8.34/h (0.2 to 22.8/h) for on-line data; the relationship being stored was as follows: 0.5 on-line - 1.17 +/- 2.6; r(2) = 0.69. Mean difference was 5.2/h (0.04 to 15.4 h), which correlated with the mean of the measurements. CONCLUSION: Home studies using a memory storage pulse oximeter may underestimate the number of hypoxic dips, probably due to sampling rates. Clinically significant hypoxic SAHS may therefore be missed.