Document resource
Introduction Some patients with a normal or borderline obstructive sleep apnoea (OSA) home sleep study have significant OSA when assessed by polysomnography (PSG). We hypothesized that this could be due to poor detection of hypopnoeas by domiciliary equipment and sought to determine if a high respiratory disturbance index (RDI) on the home study could predict this and prevent unnecessary PSG referrals.Methods All patients screened for OSA by home sleep study (WatchPAT, Itamar Medical, NL) and subsequent PSG were retrospectively reviewed over 12 months. Patients (n=90) had a WatchPAT (WP) apnoea/hypopnoea index (AHI) <10 and no significant central sleep apnoea. Relationships between AHI, RDI and hypopnoea index (HI) were assessed by Spearman Rank and agreement was assessed by Bland-Altman plot. Differences in confounding factors (BMI,%REM sleep,%supine sleep) were assessed using a Wilcoxon Signed Rank test. CPAP prescription after PSG and long-term compliance were assessed as secondary outcomes.Results Patients were 43M:47F, mean age 45 (range 22–77), median AHI-PSG 13.6 (IQR 5.4–25.8). There was a preponderance of hypopnoea-dominant OSA, with a median%HI-PSG (of total AHI) of 91.6 (IQR 78.1–98.1). There was no relationship or agreement between AHI-WP vs AHI-PSG, HI-WP vs AHI-PSG, or RDI-WP vs AHI-PSG. However, the difference in AHI between the two studies correlated significantly with HI-PSG (r2=0.643, p<0.0001) with strong agreement overall ( figure 1). Patients spent more time supine during their home study 53.5% (IQR 34.8–70.7) vs PSG 33.9% (IQR 13.3–53.2) (p<0.0001) and in REM sleep 23.3% (IQR 17.3–27.5) vs 12.5% (IQR 7.8–15.3) (p<0.0001) but there was no difference in BMI between studies. CPAP was requested in 54% of patients after PSG, but successful long-term compliance was only achieved in 45% of these.Abstract P34 Figure 1A Bland-Altman plot comparing the difference in AHI between the home WatchPAT study and PSG [AHIPSG - AHlWP] versus the HI by PSG [HIPSG]. The mean difference (red dotted line) was 0.217 with most data clustered closely around this, suggesting good agreement and inferring that the difference in AHI between the two studies in most cases could be explained by a lower detection rate of hypopnoeas by the WatchPAT device. There was one clear outlier, but no explanation was found for this.Conclusions The difference in AHI between home test and PSG seems to relate to an underestimation of hypopnoeas by WatchPAT rather than differences in known confounders of OSA. It is not possible to obviate the need for PSG in some patients based on RDI-WP as this did not reflect a HI. Hypopnoea-dominant OSA patients may be less likely to adhere to treatment long-term but requires exploration in a future comparative study.