Auto-CPAP: an effective and low-cost procedure in the management of OSAS?

Nasal continuous positive airway pressure (CPAP) is generally considered the first-line therapy in obstructive sleep apnoea (OSA) patients because it is effective at both correcting respiratory events and related arousals, and also improving morbidity [1]. Obstructive apnoeas and hypopnoeas can virtually always be eliminated by some level of CPAP <20 cmH2O [1, 2]. Determination of the prescription pressure (e.g. the lowest pressure that achieves a therapeutic result during an entire night), or "CPAP titration", has traditionally been conducted by a technician in the sleep laboratory during full polysomnography [2, 3]. Although this procedure was initially exclusively directed towards eliminating only frank apnoeas and hypopnoeas, it is now common for the sleep technician to titrate CPAP first to the abolition of apnoeas and hypopnoeas, then to the elimination of snoring, intermittent oxygen desaturations and microarousals on electroencephalogram (EEG) [3, 4]. Until recently, CPAP titration was always carried out in the sleep laboratory, but this is a costly and labour-intensive procedure. Many strategies have been attempted to reduce this expense, including prediction formulae [5, 6], split night studies [7], attended home titration, partner titration and home stepwise titration. An alternative is to automate the titration process and determine a single pressure suitable for subsequent at-home fixed pressure treatment. The validity of such a strategy in correcting sleep breathing disorders has been demonstrated when compared with manual titration [8, 9]. However, whether this permits a substantial reduction in or abolition of laboratory-based tests needed to titrate CPAP in OSA patients remains unknown. In this issue of the Journal, BERKANI et al. [10] have examined the feasibility and economic impact on OSA management of an automated titration based on snoring detection. Although an adequate detection of the respiratory abnormalities on which the algorithm of the titrating device is based is the first condition of success, we will not comment further on this aspect as our group was involved in the development of the device used in the present report. However, several methodological points should be noted. An auto-adjusted CPAP device can be used either for titration only in order to choose the single therapeutic pressure that will be used on a long-term basis, or for adjusting permanently to the patient's changing needs during the night. Ideally a single device should be able to perform both [3], but in the present report, only the titration function was tested [10]. The authors have set several conditions to prevent any excess of sensitivity of the CPAP machine. In order to avoid any inappropriate increase in pressure: 1) the pressure was set at a maximal value of 14 cmH2O, or subsequently at 16 if previous threshold was attained; and 2) only snoring detection was used. This may be adequate for titration, but may be inadequate for long-term autoadjusted CPAP treatment. As a result of changes occurring in the upper airway structure and function (i.e. reduction of upper airway (UA) oedema and restoration of UA reflexes), snoring detection might become less effective in driving an auto-adjusted CPAP on a long-term basis. Moreover, although the authors carefully selected permanent snorers, the two subjects in whom the CPAP pressure adjustment was suboptimal had UA anomalies e.g. laryngectomy and uvulopalatopharyngoplasty (UPPP), respectively. It should be emphasized that UPPP is usually a contra-indication to auto-CPAP, whatever the device and indeed CPAP with fixed pressure is often less well-tolerated after such a surgical procedure [11]. There is little published uniformity to the criteria used by a technician to decide on the "optimal" pressure based on an attended study. Most protocols stress the abolition of apnoea which is relatively simple. The reference to hypopnoea is much more complex owing to the variability of detection between laboratories but also within a laboratory in this specific context of manual titration. Some laboratories also titrate CPAP to try to abolish arousals which may result from intermittent upper airway obstruction not producing detectable hypopnoea or desaturation. An additional approach that has been proposed is to use oesophageal pressure monitoring to titrate the level of CPAP to the pressure which minimizes intrathoracic pressure swings indicative of high resistance, thus correcting any elevated upper airway obstruction. A similar but noninvasive method is to use the inspiratory flow contour (flow limitation) to effect the CPAP titration [12]. Most, if not all the signals mentioned above are poorly standardized in terms of techniques of measurement and reference values. This is particularly the case for snoring measurement during CPAP which would be needed to provide comparable results between laboratories. Variability in hypopnoea detection has already been mentioned and is probably critical during an "on-line" procedure such as manual titration. Arousal detection is an even more complex issue as both the criteria used [13] and the normal values [14] remain questionable. Oesophageal pressure values may depend upon the technique of measurement but are also critically affected by sleep stages and body position. Thus, there have been virtually no normal values published until now [15]. Finally, Sleep and Respiration Unit, University Hospital of Grenoble, Grenoble France.