Article Text

This article has a correction. Please see:

Download PDFPDF

Clinical management of patients with COPD
P263 Reducing Investigation Time For the Diagnosis of the Obstructive Sleep Apnoea Syndrome (OSAS) by Combining Oximetry with a Screening Algorithm
  1. JA Benjamin
  1. Royal Glamorgan Hospital, Llantrisant, Wales, UK


Introduction Oximetry and limited channel sleep studies (LCSS) for the diagnosis of OSAS have sensitivities of 87% and 82–94% respectively and specificities of 65% and 82–100% respectively¹. Our centre finds oximetry useful for confirming OSAS in high probability patients but not at ruling it out. LCSS seems useful for both. The average time taken for oximetry is 15 minutes compared with 75 minutes for LCSS (times were pooled from 5 Welsh sleep centres). We proposed an investigative approach that combined a screening algorithm that would allocate high probability patients to oximetry and low probability patients to LCCS. High probability subjects with negative oximetry also proceeded to LCSS. This 2-step approach has been recently tested with a subsequent sensitivity of 97% and specificity of 87%2. The aim of our study was to assess if our interpretation of this approach reduced investigation time for the diagnosis of OSA.

Methods We recruited consecutive sleep referrals over a 5 month period. A screening algorithim was employed. The following risk factors were used to score referrals(1) Snoring, Apnoea (2) Daytime somnolence (3) Hypertension, Obesity, Large neck size 0.2/3 positive factors made the referral high probability; 2/3 negative-low probability. A referral with inadequate information went straight to LCSS. A high probability patient with subsequent negative oximetry proceeded to LCSS.

We compared the total and average time taken with our algorithm against a strategy (control arm) of LCSS for every patient. The study was powered for 50 patients.

Results N=50. Median age 53 years, 34 were male. See figure 1 for outcomes.

Abstract P263 Figure 1

Outcomes of 50 consecutive sleep referrals

For the control arm, assuming every patient had LCSS, the total investigation time for the 50 patients was 3750 minutes with an average time of 75 minutes per patient.

For the algorithm arm, the total time was 2715 minutes with an average time of 54.3 minutes per patient.

Average time saving – 20.7 minutes per patient.

Conclusion We propose that allocating high probability subjects to oximetry and using LCSS for low probability subjects and for oximetry negative high probability subjects results in significant time and resource savings.


  1. Thorax 2005; 60 (ii):37.

  2. Thorax 2011; 66.

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.

Linked Articles

  • Corrections
    BMJ Publishing Group Ltd and British Thoracic Society