The population selected consisted of healthy, nonobese, middle-aged subjects (mean age, 45.3±9.1 years) who were mostly male (81.2%). Both men and women were accepted without bias and done so in a consecutive manner, since the response to treatment in this study was expected to be similar. Objective sleep and respiratory analysis was done by polysomnography on three separate occasions for each subject.
The Pes data presented are, in this study, one of two objective measures in the polysomnogram. Three subjects at baseline polysomnography and five subjects at completion refused the Pes monitor. Two of these subjects refused Pes both pretreatment and posttreatment; one had OSAS, and the other was not classified. Chervin and Aldrich recognized that utilization of Pes monitoring was associated with small but significant (p=0.05) decrements in some of the sleep variables, one of which was sleep efficiency. These differences, when compared with matched subjects without Pes monitoring, were considered of doubtful clinical significance by these authors.
The SEI was the second objective sleep variable in our study that showed a statistically significant change with a pretreatment mean of 0.74±0.14 and a posttreatment mean of 0.84±0.09 (p=0.002). Hoffstein et al, in a study of snoring and sleep architecture, reported sleep efficiency to be correlated significantly with a snoring index. Although snoring did not affect sleep architecture in their study, it was suggested that heavy snoring may adversely affect daytime function. In our study, snoring was heavy and habitual, accompanied by subjective sleepiness. Subjective evaluation of daytime sleepiness in our study by ESS was found to improve after treatment (p^O.00001), as did snoring (p<0.0001), with an associated improvement in SEI scores (p=0.002), which is consistent with the findings that Hoffstein et al reported. Sleep parameters that included SEI were reported by Conway et al in a 6-week and 1-year follow-up of responders after uvulopalatopharyngoplasty. At follow-up testing, SEI improved significantly with a p<0.05. Concomitant normalization of sleep architecture and respiratory parameters also improved with a p=0.01. In the responders of the Conway et al group, a consistent improvement in the polysomnographic parameter of SEI was seen using a three polygraphic evaluation sequence that was similar to the sequence report in our study.