Cardiorespiratory screening for sleep-disordered breathing

To the Editor:

Sleep-disordered breathing (SDB) is associated with an increasing mortality 1, 2. The prevalence of SDB, in particular central sleep apnoea and Cheyne–Stokes respiration, is remarkably high in heart failure patients 3. Therefore, screening for SDB in heart failure patients is an emerging clinical problem.

Waiting times for in-hospital polysomnography (PSG), which still represents the gold standard for SDB diagnosis, are increasing. Consequently, cardiorespiratory polygraphy (PG) devices for the diagnosis of SDB have been introduced. Two studies of PG devices have previously been published in the European Respiratory Journal. Dingli et al. 4 compared data of PSG with those obtained with a portable PG device (Embletta: Medcare, Reykjavik, Iceland) to detect obstructive sleep apnoea. Using simultaneous measurements they found a close agreement in total apnoea/hypopnoea index (AHI) scores (29.2±3.7·h^-1 in PSG versus 27.2±3.4·h^-1 in PG). Quintana-Gallego et al. 5 assessed the benefit of ambulatory PG (Apnoescreen II; Erich Jaeger, Wuerzburg, Germany) to establish the diagnosis of SDB in heart failure patients. They reported a close correlation in the results of PSG versus PG measurements with a high sensitivity and specificity. In both studies, recordings were reviewed by sleep specialists.

Analysing software for PG recordings may further contribute to an increase in the number of screened patients. The use may be suitable especially in high volume cardiology centres with large numbers of heart failure patients. We used the Embletta device as introduced by Dingli et al. 4 to screen SDB in 104 consecutive patients. Somnologica for Embletta™ (Version 3.3; Medcare) was used for a first ever analysis (automatic analysis). All recordings were then reviewed by two sleep specialists (manual analysis), blinded to the automatic analysis results. Default settings and standard definitions were used for the detection of apnoea (complete cessation of airflow for ≥10 s) and hypopnoea (≥50% reduction in respiratory airflow accompanied by a decrease of ≥4% in arterial oxygen saturation lasting for ≥10 s).

For the most widely used parameter in SDB, the AHI, the correlation between automatic and manual analysis is shown (fig. 1⇓); as well as sensitivity, specificity, and positive and negative predictive values (table 1⇓) for several AHI cut-off values. In addition a Bland–Altman plot is presented (fig. 2⇓), as recommended by Flemons and Littner 6.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

Correlation of automatic versus manually obtained apnoea/hypopnoea index (AHI) in 104 consecutive patients screened for the presence of sleep-disordered breathing. –––: AHI≥5·h^-1; – – –: AHI≥15·h^-1; - - -: AHI≥30·h^-1; y = 0.9685x + 4.8129; r²  =  0.6664.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2—

Bland–Altman plot of mean differences and limits of automatic versus manually obtained (AHI) values in 104 consecutive patients screened for sleep-disordered breathing. The co-efficient of variation (2 sd) is 18.54 events·h^-1, with a trend toward increased error at larger AHI levels. ····: mean difference; – – –: ±2sd; y = -0.1876x -1.4248; r²  =  0.0812.

In addition to the results of Dingli et al. 4 and Quintana-Gallego et al. 5, the current data show that automatic analysis of PG recordings is effective for ruling out sleep apnoea, with a 96.4% specificity, when using the clinically relevant AHI cut-off of ≥15·h^-1. However, the sensitivity of 72.9% with automatic (not reviewed) analysis lead to a relevant underestimation of SDB in ∼25% of patients. Somnologica for Embletta™ software tended to underscore AHI, when compared with manual review by two independent sleep specialists.

In agreement with Quintana-Gallego et al. 5 we believe that cardiorespiratory polygraphy is an adequate tool for cardiology screening, especially heart failure patients, for the presence, type and severity of sleep-disordered breathing, but recordings have to be carefully reviewed by specially trained personnel.