Identifying obstructive sleep apnoea (OSA) patients according to endotype can help guide therapy and improve outcomes, delegates at ERS 2023 have heard.
Professor Danny Eckert
Current evidence indicates at least four key pathophysiological phenotypes, or endotypic traits, linked to OSA pathophysiology, which, if identified, could potentially direct treatment pathways, said Professor Danny Eckert, College of Medicine and Public Health at Flinders University, Australia.
While the leading trait of OSA pathogenesis is impaired pharyngeal anatomy, around 70% of patients have one or more non-anatomical characteristics that contribute to their disease, such as impaired pharyngeal dilator muscle function during sleep, unstable respiratory control and a low respiratory arousal threshold.
Categorisation of OSA via the three-point PALM scale (Passive critical closing pressure of the upper airway, Arousal threshold, Loop gain, and Muscle responsiveness; more information here) is an effective way of determining the key drivers of the condition, which can help to maximise the efficacy of existing therapies by ensuring they are used in the right patients, he said.
For example, patients whose sleep apnoea is driven by a low arousal threshold, i.e. those who wake up too easily during minor airway narrowing, tend not to do too well on continuous positive airway pressure (CPAP), both in terms of uptake and compliance, so may need a different approach of additional support. At the same time, those with pharyngeal dilator muscle issues might benefit from hypoglossal nerve stimulation, he noted.
A recently completed study co-led by Dr Eckert (link here) highlighted the potential benefit of a precision medicine approach to managing the condition.
In the study, OSA patients with an incomplete response to oral appliance therapy were phenotyped and given endotype-informed target combination therapy.
Initially, impaired anatomical end0types were treated with expiratory positive airway pressure (EPAP) valve and supine avoidance device therapy, and those with residual OSA then also received one or more non-anatomical interventions based on their endotype, such as O2 to reduce high loop gain and atomoxetine-oxybutynin to boost pharyngeal muscle activity.
If necessary, OAT was combined with EPAP and CPAP therapy; however, “except for a handful of cases, we managed to treat everyone without CPAP in this study,” Dr Eckert said.
Results showed that OSA was successfully controlled with combination therapy in all but one study participant and in 17 of 20 without CPAP, highlighting the potential of a personalised strategy to achieving the best treatment outcomes.
Dr Eckert’s team has now also developed a machine learning algorithm that can estimate loop gain, arousal muscle threshold, Pcrit (so PALM scale categorisations), as well as an algorithm that can predict response to treatment, which evidence suggests can increase the treatment success rate from about 50% to 80%, he said.
Simplified tools such as these machine learning algorithms are already in development and showing “considerable promise”, but in the meantime, clinicians can take into account certain factors to inform treatment selection, “for example, the apnea: hypopnea ratio and higher therapeutic CPAP levels are linked with increasing Pcrit, while three parameters from a standard diagnostic PSG can be used to estimate a low arousal threshold endotype,” he stressed.