Article Text

Download PDFPDF

Managing a sleep disorder service

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


1L. Davies, 1B. Sekar, 1J. Bhat, 2K. E. Lewis. 1Hywel Dda NHS Trust, Llanelli, UK, 2Swansea School of Medicine, Swansea, UK

Introduction Several formulae have attempted to predict the presence and severity of the obstructive sleep apnoea-hypopnea syndrome (OSAHS) with variable success. Many have been used to differentiate OSAHS from the background population rather than symptomatic clinic attenders, and few have been retested as our population becomes older and more obese.

Methods A retrospective review of a prospectively recorded database for a sleep-disordered breathing clinic at a district general hospital. We selected a random sample of 534 subjects with complete data regarding sleep studies referred between 2004 and 2008 to a respiratory service with symptoms of daytime tiredness and usually snoring or apnoeas. We defined OSAHS as daytime sleepiness and a 4% dip rate >15 events/h or apnoea-hypopnea index >20 events/h on limited channel overnight sleep study (Embletta or Visilab). We compared variables traditionally associated with OSAHS in those with OSAHS versus those without eventual OSAHS (Mann-Whitney U test/χ2 test). We then entered potentially discriminating factors as independent variables using the log10 transformation of the respiratory disturbance index (RDI) (for the whole sample) as the dependent variable, in a linear regression model to look for predictors of the severity of sleep disturbed breathing.

Results Our multiple linear regression model, entering these factors suggests that only body mass index (BMI) (p = 0.045) and collar size (p = 0.047) were significant predictors. (Age (p = 0.308), Epworth (p = 0.268) and systolic BP (p = 0.226)). The overall regression effect was highly significant (F(5,239) = 5.33, p<0.001), but R2 = 0.10. Mean values are shown in table 1.

Abstract P157 Table 1

Conclusion We have identified similar differences in baseline characteristics that help identify those with OSAHS from, for example, sleepy snorers, with BMI and collar size appearing the most important variables in predicting the severity of sleep disturbed breathing in those referred to a sleep clinic. However, together these variables still only account for 10% of the variance in (log)RDI, suggesting that other factors such as upper airway shape are important and sleep studies are ultimately still needed.

View Abstract