Bookmark and Share

Strain Controlled LVAD Unloading & Cardiac Remodeling in Heart Failure

Sonomicrometry Array Localization (SAL) Strain Maps

Dorsett hybrid sheep between 45 and 65 kg bred for laboratory use (Thomas Morris, Reisterstown, MD) were used for this study. All animals received treatment in compliance with the NIH Guide for the Care and Use of Laboratory Animals. The surgical procedure and post-operative care were carried out according to the approved protocol by the Institutional Animal Care and Use Committee of the University of Maryland at Baltimore.

The Surgery

In short, a left anteriolateral thoracotomy was performed. Polypropylene snares were placed around the first and second diagonal coronary arteries off of the left anterior descending artery. An ultrasonic flow probe was placed on the PA and the cable tunneled out to the skin. Two epicardial pacemaker wires were sewn to the left atrium. Sixteen sonomicrometry transducers (2mm in diameter) were sutured to the myocardium. Three groups of transducers were distributed between the LV base and apex, each having between 4-6 transducers placed circumferentially from the anterior wall to the posterior LV wall. One transducer was placed at the LV apex. The lead wires for the sonomicrometry transducers were brought through the skin via three skin buttons.

The snares were tightened, producing an infarct sized by discoloration of the epicardium. Myocardial infarction was confirmed by echocardiography and electrocardiogram changes. Sonomicrometric and hemodynamic measurements were made pre- and post-infarction and at two, six, and ten weeks. During the sonomicrometric data collection, the mVAD (if applicable) was stopped and the outflow graft was occluded via the implanted occlusion snare. The aortic pressure was maintained between 80-100 mmHg to control afterload.

Locations of SAL crystal placement on LV Free Wall

Locations of SAL crystal placement on
LV free wall.

Computational Information

Distances between all pairs of transducers (120 unique distances) were measured at a sampling rate of 200 samples/sec simultaneously with LV pressure, PA flow and arterial pressure. The instantaneous location of each transducer in a single 2D coordinate system was determined using a multidimensional scaling algorithm.

Once the 3D coordinate of each crystal was constructed, the strain field within the LV free wall was computed using finite element-based surface interpolation. The general approach to the surface fitting was based on the work by Sacks et al. [1] and utilizes finite element functions for surface interpretation. To fit the crystal array, a single nine-node C0-continuity quadratic Lagrangian element was used in the isoparametric coordinates r-s, with 1<r, s<1. The Almansi finite strain tensor was defined for each frame. Stretch in the area, representing the new change in the LV free wall area, was computed using Δ=λ1 λ2, where λ1 and λ2 are the principal stretches.


In the color strain map illustrated below, the negative strain (green to blue) effectively shows contractility of the myocardium at end systole. By comparing the sequence of the strain maps from pre-MI to 10 weeks after MI, the areal strain representation shows not only the regional function of the LV myocardium, but also the expansion of the dysfunctional region and the progression of the overall LV dysfunction. The systolic contraction was present in the majority of the region delimited by the crystals, except that at the apex pre MI. The region close to the LV apex became dysfunctional at 30 minutes after MI. The dysfunctional area became progressively enlarged and expanded beyond the initial infarct zone at two, six and 10 weeks after MI. The initial infarct zone also became more distensible. The size and shape of the dysfunction area indicated on the strain map (10 weeks) is consistent with the white scar area in the excised LV free wall.

Systolic strain map
Systolic strain map of the control animal at various time points throughout the study.
In the color strain map, the negative strain (green to blue) effectively shows
contractility of myocardium.  The positive strain (green to red) indicates the
dysfunctional myocardium and systolic bulging.

Remodeling strain maps of the sheep whose systolic strain was discussed above show regional remodeling by the deformation of the LV free wall between pre and post MI. Over the 10-week period there was progression of reduced contractility in the non-ischemic zone (NIZ) along with dyskinesis. The infarct zone experienced stretch (positive strain), while the remote zone exhibited distance shortening between crystals. The border zone myocardium (BZM) of the NIZ area experienced lengthening and hypocontractility.

Left Ventricular Assist Device

The Jarvik 2000 VAD (mVAD) intraventricular pump was placed within the apex of the ovine left ventricle and anastomosed to the descending aorta with a partial clamp occlusion. The impeller was driven by an external power source. Blood flow through the graft was monitored with a flow probe. Once the mVAD was placed and set to 9,000 rpms (3.5 l/min), the snares for inducing MI, the pacemaker, flow probe and crystals were placed in the same fashion as for the control group, with the exception that no apical crystal was placed. The same data collection and analysis were completed as for the control group above.

Insertion of the Jarvik 2000 VAD into the ovine model.
Insertion of the Jarvik 2000 VAD into the ovine model.  


The regional systolic areal strain maps of the LV free wall in one mVAD -supported sheep before MI and at 30 minutes, two weeks and five weeks after MI are shown in the figure below. Upon the initiation of MI, the dyskinetic area was similar to that found in the control animal (no mVAD). By two weeks, the early dyskinetic area had regressed. By five weeks, the dyskinetic area had regressed to approximately 20% of the original size. The hypocontractility of the NIZ was not seen in the mVAD animal.

Systolic areal strain maps of the LV free wall in one mVAD supported sheep
Systolic areal strain maps of the LV free wall in one mVAD supported sheep.

Remodeling strain maps of the mVAD sheep revealed that the infarct zone tended to be stretched, while the adjacent zone tended to shrink. The region in the LV free wall delimited by the crystals slightly increased its size at two and five weeks after MI. The striking finding in the mVAD sheep compared to the non-supported sheep was the dramatic reduction of the dyskinetic area associated with the initial infarct zone. Remembering that this study was done in a non-clinical scenario of immediate mVAD support with MI, this reduction of the infarct-associated dyskinesis represents the optimal effect of unloading. It is likely that this early benefit was based on the reduction of oxygen consumption that accompanied the reduced LV work [2]. We believe that complete unloading is more advantageous early after infarct (<48 hr) when maximum reduction of LV oxygen consumption is desired.

Our strain map analysis and representation of the regional function and LV remodeling post MI has a distinct advantage over the previously used segmental analysis. It enables us to identify the progression of LV remodeling and regional MI. The remodeling strain maps parallel the classic description of post MI scar expansion and remodeling but uniquely give us the associated regional strain.

Pressure Volume-Based LV Global Function with Conductance Catheter

We have explored the conductance catheter (CC) method as a complimentary method to access global LV function in conjunction with echocardiograms and LV volume. The CC method provides a continuous on-line measurement of LV volume by means of a multi-electrode catheter positioned in the LV. In combination with simultaneous measurement of LV pressure through a pressure sensor on the same catheter, the LV function can be assessed by means of pressure-volume relationship. The figure below shows examples of pressure-volume loops obtained by occluding inferior vena cava (IVC) before and after MI in one sheep.

Pressure-volume loop of one control sheep before and after MI.
Pressure-volume loop of one control sheep before and after MI.  

Regional Calcium Handling in the Post-MI LV

Myocytes are taken from the apex of the heart during crystal implantation. Single sheep ventricular myocytes are collected to meausre Ca2+ sparks, [Ca2+]i transients and cellular contractions. By combining regional strain data collected from the intact heart using piezoelectric crystals and localized single-cell isolation, we will be able to examine how MI leads to changes in regional strain and how this affects EC coupling in single cells.