Diagnosis of sleep-disordered breathing in chronic heart failure

Abstract

Sleep-disordered breathing (SDB) is prevalent but underdiagnosed in chronic heart failure (CHF). The aim of this thesis was to investigate the utility of simple portable monitoring and clinical features to identify CHF patients with SDB. In addition, the influence of scoring criteria on diagnosis and classification of SDB in CHF was investigated. In the first study, clinical characteristics in CHF patients with SDB were compared to those without SDB. No specific symptom, anthropometric characteristic or measure of heart failure severity was reliably able to identify CHF patients with SDB. In the second study, the accuracy of heart rate variability (HRV) analysis and overnight pulse oximetry for diagnosis of SDB in CHF were investigated. The percent very low frequency increment (%VLFI) of HRV had low accuracy for diagnosis of SDB. In contrast, the >3% oxygen desaturation index measured by pulse oximetry had a high diagnostic accuracy, with sensitivity 0.97 and negative likelihood ratio 0.08 at a prespecified cutoff of >7.5 desaturations per hour. In the third study, the impact of hypopnoea scoring criteria on SDB diagnosis in CHF was evaluated. The prevalence of SDB changed significantly from 29% when hypopnoeas were scored with a corroborative ≥4% oxygen desaturation, to 46% when hypopnoeas were scored with a corroborative ≥3% oxygen desaturation or arousal from sleep (p<0.001). Respiratory event scoring criteria did not influence the classification of SDB as obstructive or central sleep apnoea. In summary, this thesis has shown that clinical features and HRV analysis can not be used to identify CHF patients with SDB. Overnight pulse oximetry has a high diagnostic accuracy and would be of greatest clinical use to rule out SDB in patients with CHF. The criteria used to score respiratory events during sleep studies can have a significant impact on the diagnosis and prevalence of SDB in CHF.