Phenotyping Using Polysomnography
Authors List
Thomas M Tolbert1, Anna E Mullins1, Daphne Valencia1, Ankit Parekh1, Andrew W Varga1, Indu Ayappa1, and David M Rapoport1.
1Division of Pulmonary, Critical Care, and Sleep Medicine, Icahn School of Medicine at Mount Sinai, New York, NY.
Rationale
Obstructive sleep apnea (OSA) may be driven by the combination of predisposition to upper airway collapse and unstable ventilatory control (high loop gain). Small physiologic studies have shown that subjects with high loop gain on ambient air demonstrate a decrease in loop gain with the administration of supplemental oxygen. Phenotyping using polysomnography (PUP) is an algorithmic method of determining loop gain by automated flow signal analysis. We tested whether PUP-estimated loop gain would decrease with supplemental oxygen administration in subjects with OSA who underwent continuous positive airway pressure (CPAP) drops.
Methods
Subjects were drawn from a previously performed study of patients with moderate-to-severe OSA who underwent polysomnography (PSG) with CPAP drops to selectively disrupt slow-wave sleep (SWS). CPAP was titrated to eliminate respiratory events and flow limitation and allow stable sleep onset. On the observation of SWS, CPAP was decreased to a lower pressure with the intent of causing obstructive respiratory events to disrupt SWS. This procedure was repeated on two separate nights in random order, one on ambient air (CPAPair) and one with the addition of supplemental oxygen during CPAP drops (CPAPoxygen). For the current study, all PSG studies were manually reviewed. Events were marked as beginning with the decrease in CPAP during a drop and ending when breath size returned to baseline—either due to subject arousal or upper airway muscle activity or to CPAP increase back to baseline. PUP analysis was applied to the CPAP flow signal and restricted only to these events. Events occurring outside CPAP drops or not specifically attributable to a drop in CPAP were excluded from PUP analysis. Periods during which CPAP was less than 2 cmH2O were excluded given an unreliable or absent CPAP flow signal at pressures near zero. Parameters were compared across studies (CPAPair vs CPAPoxygen) using paired Student t-testing for parametric distributions and paired Wilcoxon rank sum tests for non-parametric distributions.
Results
Twelve subjects with severe OSA (age 42.9±11.0 years, 10 males:2 females, BMI 37.6±6.6 kg/m2, AHI3A 66.2±18.3 events/hour) had sufficient CPAP drops for PUP analysis. The burden of hypoxemia, measured by NREM sleep time with oxygen saturation <90% (NREM T90) was greater during CPAPair than CPAPoxygen (1.74±2.9 minutes vs 0.27±0.54 minutes, p = 0.018). Loop gain was greater during CPAPair (0.57±0.19) than CPAPoxygen (0.49±0.11), but the difference was not statistically significant (p = 0.149). The two subjects with the greatest change in NREM T90 (a decrease of more than 5 minutes in both) had the greatest falls in loop gain between CPAPair and CPAPoxygen (loop gain decreases of 0.33 and 0.40).
Conclusions
In a small group of subjects with severe OSA who underwent CPAP drops during SWS, PUP did not detect a significant change in loop gain with the addition of supplemental oxygen. A significant change in PUP-estimated loop gain may occur only in subjects with significant baseline hypoxemia that is corrected by supplemental oxygen.
Thomas M Tolbert1, Anna E Mullins1, Daphne Valencia1, Ankit Parekh1, Andrew W Varga1, Indu Ayappa1, and David M Rapoport1.
1Division of Pulmonary, Critical Care, and Sleep Medicine, Icahn School of Medicine at Mount Sinai, New York, NY.
Rationale
Obstructive sleep apnea (OSA) may be driven by the combination of predisposition to upper airway collapse and unstable ventilatory control (high loop gain). Small physiologic studies have shown that subjects with high loop gain on ambient air demonstrate a decrease in loop gain with the administration of supplemental oxygen. Phenotyping using polysomnography (PUP) is an algorithmic method of determining loop gain by automated flow signal analysis. We tested whether PUP-estimated loop gain would decrease with supplemental oxygen administration in subjects with OSA who underwent continuous positive airway pressure (CPAP) drops.
Methods
Subjects were drawn from a previously performed study of patients with moderate-to-severe OSA who underwent polysomnography (PSG) with CPAP drops to selectively disrupt slow-wave sleep (SWS). CPAP was titrated to eliminate respiratory events and flow limitation and allow stable sleep onset. On the observation of SWS, CPAP was decreased to a lower pressure with the intent of causing obstructive respiratory events to disrupt SWS. This procedure was repeated on two separate nights in random order, one on ambient air (CPAPair) and one with the addition of supplemental oxygen during CPAP drops (CPAPoxygen). For the current study, all PSG studies were manually reviewed. Events were marked as beginning with the decrease in CPAP during a drop and ending when breath size returned to baseline—either due to subject arousal or upper airway muscle activity or to CPAP increase back to baseline. PUP analysis was applied to the CPAP flow signal and restricted only to these events. Events occurring outside CPAP drops or not specifically attributable to a drop in CPAP were excluded from PUP analysis. Periods during which CPAP was less than 2 cmH2O were excluded given an unreliable or absent CPAP flow signal at pressures near zero. Parameters were compared across studies (CPAPair vs CPAPoxygen) using paired Student t-testing for parametric distributions and paired Wilcoxon rank sum tests for non-parametric distributions.
Results
Twelve subjects with severe OSA (age 42.9±11.0 years, 10 males:2 females, BMI 37.6±6.6 kg/m2, AHI3A 66.2±18.3 events/hour) had sufficient CPAP drops for PUP analysis. The burden of hypoxemia, measured by NREM sleep time with oxygen saturation <90% (NREM T90) was greater during CPAPair than CPAPoxygen (1.74±2.9 minutes vs 0.27±0.54 minutes, p = 0.018). Loop gain was greater during CPAPair (0.57±0.19) than CPAPoxygen (0.49±0.11), but the difference was not statistically significant (p = 0.149). The two subjects with the greatest change in NREM T90 (a decrease of more than 5 minutes in both) had the greatest falls in loop gain between CPAPair and CPAPoxygen (loop gain decreases of 0.33 and 0.40).
Conclusions
In a small group of subjects with severe OSA who underwent CPAP drops during SWS, PUP did not detect a significant change in loop gain with the addition of supplemental oxygen. A significant change in PUP-estimated loop gain may occur only in subjects with significant baseline hypoxemia that is corrected by supplemental oxygen.
