Variability in Blood Pressure Assessment in Patients Supported with the HeartMate 3TM

Targeted blood pressure (BP) control is a goal of left ventricular assist device medical management, but the interpretation of values obtained from noninvasive instruments is challenging. In the MOMENTUM 3 Continued Access Protocol, paired BP values in HeartMate 3 (HM3) patients were compared from arterial (A)-line and Doppler opening pressure (DOP) (319 readings in 261 patients) and A-line and automated cuff (281 readings in 247 patients). Pearson (R) correlations between A-line mean arterial (MAP) and systolic blood pressures (SBP) were compared with DOP and cuff measures according to the presence (>1 pulse in 5 seconds) or absence of a palpable radial pulse. There were only moderate correlations between A-line and noninvasive measurements of SBP (DOP R = 0.58; cuff R = 0.47) and MAP (DOP R = 0.48; cuff R = 0.37). DOP accuracy for MAP estimation, defined as the % of readings within +/- 10 mmHg of A-line MAP, decreased from 80% to 33% for DOP <= 90 vs. >90 mmHg, and precision also diminished (mean absolute difference [MAD] increased from 6.3 +/- 5.6 to 16.1 +/- 11.4 mmHg). Across pulse pressures, cuff MAPs were within +/- 10 mmHg of A-line 62.9%-68.8% of measures and MADs were negligible. The presence of a palpable pulse reduced the accuracy and precision of the DOP-MAP estimation but did not impact cuff-MAP accuracy or precision. In summary, DOP may overestimate MAP in some patients on HM3 support. Simultaneous use of DOP and automated cuff and radial pulse may be needed to guide antihypertensive medication titration in outpatients on HM3 support.

Automated summary

This study found a moderate correlation between noninvasive measurements and arterial line measurements of blood pressure in left ventricular assist device patients, and Doppler opening pressure may overestimate arterial blood pressure in some patients. These findings suggest the need for the simultaneous use of multiple measurement methods in blood pressure control.