0974 LVAD Pump Parameters and Pulmonary Vascular Resistance as Predictors of Obstructive and Central Sleep Apnea Severity

Abstract Introduction Sleep-disordered breathing (SDB), including both central and obstructive sleep apnea, is highly prevalent in advanced heart failure and may significantly influence hemodynamics in patients supported with continuous-flow left ventricular assist devices (CF-LVADs). Fluctuations in pump flow, speed, and pulsatility have been reported in association with untreated SDB, yet the impact and mechanism of this interaction remain unclear. As LVAD therapy expands, particularly as destination therapy, understanding cardiopulmonary predictors of SDB severity is crucial to optimize outcomes. Methods A retrospective analysis was performed on subjects supported by the same model of CF-LVAD who underwent a baseline in-laboratory polysomnography between January 2024 and February 2025 at Memorial Hermann Texas Medical Center. LVAD-specific parameters (such as pulsatility index), pulmonary vascular resistance (PVR) from right heart catheterization (RHC) when available, right ventricular systolic pressures, and the TRV-to-RVOT VTI ratio (an echocardiographic surrogate of PVR), obtained within 90 days of the polysomnography, were reviewed. All sleep studies were staged and scored using version V3 of the AASM scoring manual. Results 10 subjects were included for review; 70% were male, mean age 58.4 years (SD +/- 10.1), mean BMI 33.6 kg/m² (SD +/-3.2), mean obstructive AHI 11.3 events/h (included 1B hypopneas), mean obstructive RDI 25.8 events/h (included 1A hypopneas), and the mean oxygen saturation nadir was 84.2%. The average central apnea index (CAI) was 1.1 events/h (SD +/-2.5), and 40% of subjects had Cheyne–Stokes respiration (CSR). In a univariate regression analysis, the pulsatility index (p = 0.0003) and pump flow (p = 0.04) were associated with the overall CAI. No significance was observed between pump speed, pump flow, pulsatility index, AHI, RDI, or the presence of CSR. Analyses involving PVR were also non-significant, possibly due to the small sample size. Conclusion Higher pulsatility index and LVAD pump flows were significantly associated with higher CAIs, while the AHI, RDI, and the presence of CSR demonstrated no meaningful relationship with pump flow, pump speed, pulsatility index, or PVR in this cohort. The PVRs from RHC and echocardiography were limited; additional studies with larger datasets are needed to determine whether PVRs contribute to the burden of SDB in patients with LVADs. Support (if any)