The pacing site is thought to influence the functional response to CRT.5
However, the functional response to varying pacemaker sites in normal and infarcted ventricles has not been formerly demonstrated. We have applied ERNA to demonstrate the varying functional effects of different pacing modes.9
Here, we applied ERNA to determine the functional effects of different pacing sites.
Under controlled conditions, we evaluated systolic LV function, myocardial synchrony, and MAP, at baseline and with pacing from varying sites in normal ventricles and with a small distal infarction in an animal model. The rate was constant and the influence of varying A-V synchrony, an important influence on paced cardiac output,9
was eliminated. Phase imaging objectively displayed the sequence of ventricular contraction,8
confirmed the conduction sequence and served as an indicator of intra- and interventricular synchrony.9
The study confirms the functional importance of the focus of excitation especially in scarred ventricles. Regional scar has previously been shown to importantly influence the success of CRT.6
Not previously demonstrated, is the interaction of scar with different pacing sites, presenting a varying effect on LV function and synchrony. We previously demonstrated the influence of the ventricular tachycardia exit site on the pattern of electrical conduction and resultant systolic ventricular function and synchrony.11
We now demonstrate a link between the pacing site with its related electrical conduction pattern, and the variable resultant effect on systolic ventricular function and synchrony, in normal and scarred ventricles.
The Interaction of Activation and Contraction
How does the pacing site exert its influence on ventricular function? The findings support the influence of a varied activation sequence and conduction pattern derived from each pacing site which result in a specific related myocardial contraction sequence. This sequence relates to a specific contraction pattern, resulting in varying LV systolic function, synchrony, and hemodynamics. The magnitude and direction of the functional effect of pacing relates directly to the induced contraction pattern and related synchrony which is determined, in part, by the pattern of electrical activation. The electrical activation pattern is a product of the pacing site and the state of the intervening myocardium. Owing to the influence of scar on both conduction and contraction, these effects are most prominent in the presence of scar. These findings provide insight into the source of variable outcome in CRT, where scar is generally present and often widespread. Tying activation to contraction in a predictable manner is the fact that the pattern of electrical activation correlated closely with the mechanical contraction sequence on phase imaging (Figure ). This correlation is the subject of another manuscript.17
Phase analysis previously revealed the prognostic importance of synchrony in congestive cardiomyopathy.10
The functional benefit of CRT, varies with its ability to improve mechanical synchrony.19
In this model we demonstrated that the pacemaker site had a widely varying effect on wall motion abnormalities, synchrony, and hemodynamic changes, sometimes bringing improvement, sometimes worsening, to the function of the LV apex, with its limited scar. Decremental MAP during pacing in infarcted dogs, correlated only indirectly with the specific pacing site, but correlated more directly with the presence and severity of LV wall motion abnormalities, and induced dyssynergy, interventricular, and LV intraventricular dyssynchrony. This was most frequent in association with LVa pacing.
The evidence suggests that in this model, LV dysfunction is secondary to delayed conduction through an altered myocardium (Figure ) producing segmental LV wall motion abnormalities, frequent dyskinesis (Tables and ) and dyssynchrony. When activation originated distal to the infarction, from the LV apex, LVa pacing, both ventricles demonstrated retrograde depolarization and delayed conduction in the peri-infarct area, producing a dyskinetic, dyssynchronous contraction pattern, decremental function, and reduced MAP. Compared to the more physiologic conduction pattern of HS pacing, the pattern of retrograde ventricular activation related to this LVa pacing site would be expected to encounter the distal infarct area early on, requiring the most circuitous activation sequence to gain access to the bundle branches, with the greatest resultant conduction delay and ventricular dyssynchrony. The delay in activation with LVa compared to HS pacing was clear and the resultant wall motion abnormalities were evident (Figure , Tables and ). Delay in contraction and, presumably, conduction in the peri-infarct region, was also seen in relation to RVW pacing (Figures C and D). Possibly owing to their temporal resolution, such delay was sometimes seen better on phase than isochronous activation maps.
Relationship to Infarct Size
The lack of correlation between pathologic MI size and LV wall motion abnormalities or hemodynamics, appears to be due, in part, to the small numbers involved, and the varied distribution of infarct size. It is likely influenced as well, by the variable alteration in conduction through and around scar tissue. This leads to an altered contraction sequence and variable ventricular function which again suggests the importance of the pacemaker site to ventricular function and the relatively small effect of the limited scar present here.
CRT is delivered with endocardial, A-V synchronous, biventricular pacing. We studied the functional effects of a few representative isolated pacing sites which were chosen with consideration of their likely differences on ventricular function.
Epicardial pacing was applied to gain a selection of stable pacing sites as it is sometimes necessarily applied with CRT.22
Epicardial activation, conduction, and possibly hemodynamic effects may differ from the endocardial location. However, phase image documentation of earliest contraction at the pacing site and the demonstrated relationships of specific contraction patterns and resultant LV function with the designated pacing site, are most important here and not the specific form of pacing. The parallels with and implications for CRT are clear.
Although animal numbers are relatively small, the number of episodes studied is adequate to establish the relationships sought. This intensive protocol terminated with killing of the subject, so we sought to achieve our objective with fewest animals.
That large infarctions relate to more frequent and more severe wall motion abnormalities is well established. The lack of relationship here between infarct size and LV function is negatively affected by the single large infarction and other factors. Demonstrated here was the frequent occurrence of severe wall motion abnormalities in relation to small infarctions and their variation with the pacing site! These same principles must be demonstrated to apply to humans.
Although overlapping structures could lead to ERNA inaccuracies in both phase and LVEF calculation in dogs, structural separation permitted appropriate regions of interest, adequate reproducibility, and likely accuracy.
Phase angle is generated from a symmetrical cosine curve fit of the asymmetrical regional time versus radioactivity curve for each pixel of the ERNA. A multiharmonic fit of the ERNA time activity curve would increase the accuracy of the curve fit and would be required to translate phase angle into absolute timing of contraction onset. However, a multiharmonic fit would require greater data density than that which is available in a single image pixel and would require analysis of grouped pixels. Yet, multiharmonic fits demonstrated no advantage over the first harmonic method for the evaluation of serial phase or contraction sequence where the sequence of phase angle progression correlated well with electrocardiographic and electrophysiologic methods in this and prior studies.2
Differences in the sampling rate between the phase image and epicardial map presented an intrinsic obstacle to exact agreement of the two. However, the data well supported the comparison of their sequences. In the figures, appropriate intervals were empirically selected to demonstrate the relationship. Illustrations were designed to demonstrate the sequence of phase progression not obscure it.
In this preliminary study, arterial pressure was the only hemodynamic parameter measured. Thus, if hemodynamic changes were induced by reflex mediated mechanisms causing vasodilation or altering venous return, these would not be detected.