Pathological respiratory chemoreflex activation predicts improvement of neurocognitive function in response to continuous positive airway pressure therapy.

Introduction: There is a need for biomarkers predicting neurocognitive improvement following treatment of obstructive sleep apnea (OSA) with continuous positive airway pressure (CPAP). The role of sleep apnea endotypes as predictors are promising. Objective: To assess the relationship between a high loop gain biomarker, elevated low frequency narrow band (e-LFCNB), and improvements in neurocognitive function in the Apnea Positive Pressure Long-term Efficacy Study (APPLES). Methods: The e-LFCNB % metric was estimated on baseline polysomnography. Logistic regression analysis was performed to identify the potential association between e-LFCNB% of total sleep time and the observed improvement in neurocognitive function following the specified treatment. Results: A total of 362 subjects received CPAP and had e-LFCNB % measurements. For Sustained Working Memory Test-Overall Mid-Day (SWMT-OMD), e-LFCNB% > 2.35% correlates positively with the proportion of participants who showed an increase in test scores > 0.65 after 2 months CPAP treatment (OR: 2.617, 95% CI: 1.095-6.252, p: 0.030); e-LFCNB% > 9.45% correlates positively with improvement in test scores > 0.8 after 6 months CPAP treatment (OR: 2.553, 95% CI: 1.017-6.409, p: 0.046). For Buschke Selective Reminding Test sum recall (BSRT-SR), e-LFCNB% > 3.65% correlates positively with an increase in test scores > 12 after 2 months CPAP treatment (OR: 2.696, 95% CI: 1.041-6.982, p: 0.041). Results of the Pathfinder Number Test-Total Time (PFN-TOTL) were not significant. Conclusion: e-LFCNB% (probable high loop gain) may be a clinically useful predictor of cognitive improvement following CPAP.