LIMITATIONS
While we attempted to obtain imaging on all patients, 7% lacked CTs. Patients implanted urgently may not have preoperative imaging, while patients who expire or are transplanted during the index hospitalization may not have postoperative CTs. We also did not serially assess IFC position at pre-specified intervals. Follow-up scans were only obtained if patients had pump-related complications potentially related tode novo or worsening malposition. Shih13 found no major change in Heartmate 3 pump-spine angle on x-ray during 12-month follow-up.
We also did not assess LV unloading using invasive hemodynamics. Our center does not routinely obtain these studies outside of the initial hospitalization. We used LVAD FI as a surrogate for cardiac output, but did not have a similar measure for LV filling pressures. We also did not have serial echoes at pre-specified time points, and instead used two broad intervals: early (1-3 months post-LVAD) and late (6-12 months).
Finally, while the population studied (N = 114) is one of the largest malposition series presented, these are single-center, single-device data covering a ten-year timespan. We chose to only include the HVAD in this series because we had extensive experience with and a well-established surgical protocol for that device. During the study period, only six Heartmate 3 patients met criteria, and we were still in the learning stages of using the LTHS technique for it. However, both pumps have similar profiles and similar-size integral IFCs so there is no reason to think our results would not apply to that device as well.