Response of LVAD pump parameters to an orthostatic challenge and its association with CPET outcomes

Introduction

Despite improved outcomes of left ventricular assist device (LVAD) therapy, individual prognosis remains variable. Cardiopulmonary exercise testing (CPET) outcomes hold a strong prognostic value for mortality in LVAD patients, but is an intensive and time-consuming test for both patient and healthcare provider. Developing an alternative method with similar prognostic relevance could increase the implementation of individualized, physiology-based risk stratification. In this study we therefore examined the relationship between CPET outcomes as a proxy for prognosis, and a simple cardiovascular provocation: an orthostatic challenge.

Methods

Patients with an LVAD were asked to perform an orthostatic challenge from supine to standing position. We assessed the pump parameters using the "HM3 Snoopy" device, which provided pump parameters at a 1Hz sampling rate. We studied the response of the MagLev impeller displacement and the pump flow. Based on pump flow response, patients were assigned to one of the three response types as found by Ruoff et al.

Results

65 patients were included (median age: 57 [IQR 44-65] years, 26% female, median months after implantation: 9 [IQR 6-14]). Specificity was 84% and sensitivity was 74% for device-based detection of transition from supine to standing using change in impeller displacement.

We observed three response types based on pump flow (Figure 1). The first response group consisted of 18 patients, for which minimum flow decreased ≥50% compared to baseline. The second response group consisted of 9 patients in which pulsatility (Flow_max - Flow_min) decreased ≥50%. The last response group comprised 38 patients where no flow response was observed. Between the different response types, there were no significant differences in age, sex, BMI, time after implantation and LVAD speed settings. PeakVO2 expressed as %predicted (%pred) peakVO2 differed significantly between the groups: response type 1 patients showed a better outcome (mean 58 (SD=14)%) compared to type 2 (mean 54 (SD=12)%), while type 3 showed the poorest performance (mean 48 (SD=15)%); p<0.05. Absolute peakVO2 showed a similar trend although not statistically significant (p=0.39): response type 1: peakVO2 median 16.1 IQR[14.0-19.6] mL/kg/min, type 2: median 14.4 IQR[11.7-18.1] mL/kg/min and type 3: median 13.8 IQR[10.8-17.6] mL/kg/min.

Conclusion

By evaluating LVAD pump parameter responses during an orthostatic challenge, we showed the possibility to detect change in position automatically using change in impeller displacement. Moreover, LVAD flow response was significantly associated with exercise capacity expressed by %pred peakVO2. These findings pave the way for further research focusing directly on the relation between orthostatic response and outcomes such as mortality and adverse events on LVAD. If validated, an orthostatic challenge could in the future serve as a simple but accurate risk stratification method during outpatient visits.For image description, please refer to the figure legend and surrounding text.