In a further report that included data from four of these six subjects, we showed that mean hypopnea and apnea duration (usually REM related) was prolonged by oxygen treatment. In the present study, however, mean Ve in REM sleep was similar on air and oxygen. Representative mean data obtained over six to ten minutes of each sleep stage give an overall picture of Ve rather than an analysis of individual disordered breathing events; minor differences between air and oxygen nights may have been obscured. Nevertheless, it is reassuring that oxygen desaturation can be greatly improved or abolished without an apparent adverse effect on overall ventilation. Here
Multiple linear regression has been shown to give more accurate VM coefficients (in equivalence to the spirometer) than the isovolume method, and it requires only data obtainable during quiet normal breathing and so is ideally suited to naive subjects. Instability of RIP calibration through “respiband” slippage was minimized by tightly securing the bands to the patient before the study. The awake breathing pattern is influenced by environmental factors, including the polysomnography apparatus and emotional stress. These effects were lessened in our study, however, by studying patients already acclimatized to the surroundings and by taking a period of wakefulness following sleep as “representative wakefulness” for analysis. Randomizing to air and oxygen on consecutive nights ensured that acclimatization was similar in both parts of the study.
No particular sleeping posture was imposed on the subjects as most were habituated to sleeping for long periods in one chosen position. No allowance was made, therefore, for possible variation in diaphragmatic contribution to breathing or altered tendency to hypoventilate in different postures. Changes in sleep state and, by implication, of respiratory muscle recruitment, may also have influenced these data, but the magnitude of this influence could not be assessed as calibration was possible only in awake subjects. Such problems are shared by all studies in respiratory noninvasive breathing pattern measurements during sleep.