Cardiac remodeling in patients with centrifugal left ventricular assist devices assessed by serial echocardiography

The aim of the study was to characterize the remodeling process in a large cohort of patients supported with a centrifugal left ventricular assist device (cfLVAD) by standardized serial echocardiography.

the number of patients included is low and most of the data originate from patients supported with axial flow pumps. 4,5 The aim of our study was to evaluate the echocardiographic results of a large population of patients supported with the cfLVADs HeartWare HVAD™ or HeartMate 3™ who underwent standardized echocardiography at 6 and 12 months after implantation.

Study population
Between March 2018 and January 2020, 219 patients without congenital heart disease underwent implantation of a cfLVAD at our center. The patients postoperatively received a standardized echocardiographic assessment every 6 months in the outpatient clinic. The assessment also included an interview, physical examination, laboratory tests, ECG, and a 6-min walk test. Transthoracic echocardiography was performed using a standardized protocol (see below).
All data were captured and analyzed in a RedCap database. The study was reviewed and approved by the ethics committee at Charité University (EA2/229/19). The ethics committee waived the need for informed written consent for publication of the study data.

Surgical technique
In the majority of cases, LVAD implantation was performed via median sternotomy. In patients with previous cardiac surgery and a status post sternotomy, LVAD implantation was performed via left lateral thoracotomy. The implantation was carried out primarily using off-pump techniques; however, in case of hemodynamic instability, circulatory support was provided by ECLS. If the patients were already on mechanical circulatory support with ECLS or Impella® preoperatively, the support was continued during the LVAD implantation. If concomitant intracardiac procedures (e.g., valve surgery, left ventricular thrombectomy, or closure of patent foramen ovale) were necessary, the implantation was performed employing cardiopulmonary bypass.

Device and patient management
The device speed was clinically adjusted to optimize flow and organ perfusion prior to discharge and at every outpatient visit. During the hospital stay and in the later follow-up, echocardiography was performed serially to evaluate the position of the inter-ventricular septum as well as the left ventricular (LV) and right ventricular (RV) size and function. The goals of adjusting the device speed were to avoid suction events and to achieve not only a mid-position of the inter-ventricular septum but also, to the best extent possible, periodic valve opening.
Anticoagulation with warfarin plus antiplatelet therapy was administered as recommended by current guidelines. 6,7 All patients received heart failure medication as recommended by the heart failure guide-lines. Management of blood pressure included a target mean arterial pressure lower than 80 mmHg. Loop diuretics were given in case of fluid overload.

Echocardiography
Transthoracic echocardiographic results were obtained up to ten days preoperatively. For the 6-and the 12-month follow-up, echocardiography was performed by seven experienced physicians as per current guideline recommendations and following a standardized protocol (see Supplementary Table S1). 3

Statistical analysis
Continuous data are summarized as mean and standard deviation or, in case of skewed data, as median and interquartile range (IQR

Preoperative status and LVAD implantation
Most of the patients were in Interagency Registry for Mechanically

Follow-up
One hundred fifty-six patients presented for the 6-month follow-up (median 187 days [177-210]) and 144 patients at the 12-month followup (median 371 days [357-396]). During the above period, 62 patients died, seven underwent transplantation, and ten were lost to follow-up.
Most patients were in New York Heart Association Class I or II (78.5% at 6 months, 72.7% at 12 months) with a mean arterial pressure of 81.4 ± 11.5 mmHg and 82.1 ± 11.2 mmHg, respectively. All received intensive medical heart failure treatment. For details and device settings see Table 2.

Echocardiography
A total of 512 echocardiograms were reviewed for the study (216 preoperative, 156 at 6 months, and 140 at 12 months). In three patients no preoperative echocardiographic examination was available. All echocardiograms at 6 and 12 months followed the standardized pro- Table S1). There was no significant difference in echocardiographic results between patients supported with a Heart-Ware HVAD™ or a HeartMate 3™ device (Supplementary Tables   S2A,S2B,S2C  months. For details see Table 3.
• Right ventricular size and function LVAD implantation led to a significant decrease in tricuspid regurgitation peak gradient. There was no significant change in the proximal right ventricular outflow tract diameter (RVOTprox), the right ventricular end-diastolic area (RV-EDA), and the right ventricular end-systolic area (RV-ESA  Table 5.

DISCUSSION
To our knowledge, this is the largest study of echocardiographic results in patients with centrifugal, continuous-flow left ventricular assist devices.
Although guidelines provide a general framework for the echocar-   artificial pulsatility, and the optimal speed differs between manufacturers and device types. Furthermore, echocardiography in patients with cfLVADs is even more challenging than in patients with afLVADs. The device itself is located at the apex and is larger than the inflow cannula of axial flow pumps. This often hinders apical 4-chamber, 2-chamber and 3-chamber views without foreshortening. Also, the device causes characteristic Doppler artifacts that may preclude sufficient Doppler interrogation.
We implemented a standard protocol for LVAD surveillance echo that includes a limited number of measures that are assessable in most patients with cfLVAD (see Supplementary   Table S1).  ical circulatory support led to an increase in TAPSE. 13 It has also been shown that cardiac surgery can cause a decrease in longitudinal right ventricular function which may be compensated by increased transversal motion. 14,15 In our study TAPSE significantly decreased after cfLVAD implantation but right ventricular fractional area change remained stable, indicating a significant change in the right ventricular contractile pattern.

TA B L E 5 Heart valves
Preoperative and postoperative tricuspid regurgitation (TR) is associated with poor survival and may be a risk factor for the development of early and late right heart failure. [16][17][18] Preoperative TR was very common in this study (see Table 4 and Figure 2). Only 13 patients underwent concomitant tricuspid valve repair; one patient underwent tricuspid valve replacement. However, mechanical circulatory support led to an impressive improvement in tricuspid regurgitation. This may be explained by the decrease in right ventricular afterload caused by pulmonary decompression. Interestingly, many patients who had ms-sTR during the follow-up had no relevant TR preoperatively (see Figure 2). One cause could be a leftward shift of the septum during support, leading to dilation of the tricuspid annulus, as is often seen in patients with severe left ventricular unloading.

CONCLUSION
LVAD therapy can lead to reverse left ventricular remodeling and improvement of mitral and tricuspid regurgitation. However, right ventricular function does not improve and prevalence of AV regurgitation progressively increases during mechanical support. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.