Left ventricular function during support with an asynchronous pulsatile left ventricular assist device

Left ventricular assist devices (LVADs) are frequently used to maintain patients with severe heart failure until heart transplantation becomes possible. Some patients may experience recovery of LV function during such support. Therefore, it is essential to be able to monitor changes in LV function in this setting.

We studied LV function in 10 patients (median age 34 years, 9 male) who had LVADs implanted because of severe heart failure due to dilated cardiomyopathy a median of 4 months previously. Median pre-implant ejection fraction was 27% and all patients had been on maximal medical therapy, including intravenous inotropic support, prior to insertion of the LVAD.

During LVAD support there were cyclical variations in LV dimensions, fractional shortening (FS) and transmitral flow, related to changes in the phase relationship of the LV and the LVAD. The "best" FS occurred when LV systole coincided with device filling and the "worst" FS when LV systole coincided with device ejection. Median FS with the pump switched off was 18% (10% to 32%). Pump-off FS was significantly greater than the "worst" FS with the pump on (5%, p = 0.002), and similar to the "best" pump-on FS (19%, p = NS).

LV function could be studied echocardiographically during LV support and brief periods of interruption in support. Function varied according to the phase relationship of the LV and LVAD. The "best" FS measured during LVAD support was more closely related to the FS with the device switched off than the "worst" pump on FS. The "best" pump-on LV function is therefore most representative of intrinsic LV performance and can be used as a guide to recovery and the potential need for pump-off studies.