Effects of pressure variation and atrial fibrillation on CardioMEMS HF measured pulmonary artery diastolic pressure: comparison of device-averaged and visually inspected waveforms.

Objective: Heart failure (HF) management guided by implantable hemodynamic monitoring reduces hospitalization rates. Hemodynamic data from the CardioMEMS HF system includes device-averaged pulmonary artery pressures (PAP) and heart rate. Agreement of device-averaged values compared to the standard method of visual inspection of pressure waveforms at end-expiration is unknown. We evaluated the agreement between device-averaged and visually inspected end-expiratory PAP.

Approach: Twenty-one patients implanted with the CardioMEMS HF system were evaluated. Eight-hundred twenty-three PAP waveforms from the Merlin remote monitoring website were visually inspected and pulmonary artery systolic pressure (PASP) and pulmonary artery diastolic pressure (PADP) at end-expiration were recorded. Waveforms were evaluated for pressure variation (PV), defined as the difference between highest and lowest PASP measurement of  ⩾20 mmHg. Bland-Altman analysis quantified differences between device-averaged and visually inspected waveforms.

Main Results: All patients were NYHA functional class III, mean age was 67  ±  15 years and 15 (71%) had AF. Bland-Altman analysis of all waveforms revealed a mean-difference in PADP of  -1.4 mmHg, indicating that visually inspected values were higher than device-averaged values. For PV  ⩾20 mmHg, this value increased to  -2.8 mmHg. The mean-difference comparing waveforms from patients with or without AF was  -1.3 and  -1.6 mmHg, respectively. The 95% limits of agreement were  >50% wider for waveforms from patients with versus without AF (10.3 versus 6.7 mmHg).

Significance: There is good agreement between device-averaged and visually inspected waveforms when pressure variation is  <20 mmHg and for patients without atrial fibrillation.