Clinical Use of Loop Gain Measures to Determine Continuous Positive Airway Pressure Efficacy in Patients with Complex Sleep Apnea. A Pilot Study

Abstract

Rationale: Measures of unstable ventilatory control (loop gain) can be obtained directly from the periodic breathing duty ratio on polysomnography in patients with Cheyne-Stokes respiration/central sleep apnea and can predict the efficacy of continuous positive airway pressure (CPAP) therapy.

Objectives: In this pilot study, we aimed to determine if this measure could also be applied to patients with complex sleep apnea (predominant obstructive sleep apnea, with worsening or emergent central apneas on CPAP). We hypothesized that loop gain was higher in patients whose central events persisted 1 month later despite CPAP treatment versus those whose events resolved over time.

Methods: We calculated the duty ratio of the periodic central apneas remaining on the CPAP titration (or second half of the split night) while patients were on optimal CPAP with the airway open (obstructive apnea index < 1/h). Loop gain was calculated by the formula: LG = 2π/[(2πDR - sin(2πDR)]. Patients were followed on CPAP for 1 month. Post-treatment apnea-hypopnea index and compliance data were recorded from smart cards.

Measurements and main results: Thirty-two patients with complex sleep apnea were identified, and 17 patients had full data sets. Eight patients continued to have a total of more than five events per hour (11.8 ± 0.5/h) (nonresponders). The remaining nine patients had an apnea-hypopnea index less than 5/h (2.2 ± 0.4/h) (responders). Loop gain was higher in the nonresponders versus responders (2.0 ± 0.1 vs. 1.7 ± 0.2, P = 0.026). Loop gain and the residual apnea-hypopnea index 1 month after CPAP were associated (r = 0.48, P = 0.02). CPAP compliance was similar between groups.

Conclusions: In this pilot study, loop gain was higher for patients with complex sleep apnea in whom central apneas persisted after 1 month of CPAP therapy (nonresponders). Loop gain measurement may enable an a priori determination of those who need alternative modes of positive airway pressure.

Keywords: Cheyne-Stokes respiration; central sleep apnea; continuous positive airway pressure ventilation; mixed sleep apnea; obstructive sleep apnea.

Figure 1.

Flow chart of inclusion/exclusion of patients from the sleep study cohort. From the 3,247 patients’ baseline or split studies, 168 patients with primarily obstructive and some central events on the diagnostic evaluation (sleep study) were identified. Of these 32 patients (pts) with complex sleep apnea defined as persistent or emergent central events occurring during continuous positive airway pressure (CPAP) titration were identified. Of these, 18 patients were treated with CPAP and 17 patients had 4-week compliance downloads available for review. After 1 month of CPAP, nine patients had apnea–hypopnea index (AHI) < 5/h (responders) and eight patients had AHI > 5/h (nonresponders). ASV = adaptive servoventilation; BiPAP = bi-level positive airway pressure; CAI = central apnea index; CSA = central sleep apnea; OSA = obstructive sleep apnea; PAP = positive airway pressure; rx = treated with.

Figure 2.

Raw polysomnographic data from a single patient with complex sleep apnea, during a continuous positive airway pressure (CPAP) titration, showing persisting central apneas (but no obstructive apneas) while on optimal CPAP of 13 cm H₂O (left), with an inset (right) showing a magnification of the flow/effort channels from the same patient, exhibiting how duty ratio is measured (duty ratio is defined as the ventilation duration/cycle duration). Thirty-second epochs are denoted by the vertical blue lines.

Figure 3.

Regression analysis showing the relationship between residual apnea–hypopnea index (AHI, events/h) and loop gain after continuous positive airway pressure therapy.

Figure 4.

Mathematical relationship between loop gain and duty ratio (DR) illustrating mean data for responders and nonresponders. Note the nonresponders are 40% further from stable breathing (loop gain < 1.0) than responders.