Determinants of the time-to-peak left ventricular dP/dt (Td) and QRS duration with different fusion strategies in cardiac resynchronization therapy

Abstract

Background: Cardiac resynchronization therapy (CRT) is helpful in selected patients; however, responder rates rarely exceed 70%. Optimization of CRT may therefore benefit a large number of patients. Time-to-peak dP/dt (Td) is a novel marker of myocardial synergy that reflects the degree of myocardial dyssynchrony with the potential to guide and optimize treatment with CRT. Optimal electrical activation is a prerequisite for CRT to be effective. Electrical activation can be altered by changing the electrical wave-front fusion resulting from pacing to optimize resynchronization. We designed this study to understand the acute effects of different electrical wave-front fusion strategies and LV pre-/postexcitation on Td and QRS duration (QRSd). A better understanding of measuring and optimizing resynchronization can help improve the benefits of CRT.

Methods: Td and QRSd were measured in 19 patients undergoing a CRT implantation. Two biventricular pacing groups were compared: pacing the left ventricle (LV) with fusion with intrinsic right ventricular activation (FUSION group) and pacing the LV and right ventricle (RV) at short atrioventricular delay (STANDARD group) to avoid fusion with intrinsic RV activation. A quadripolar LV lead enabled pacing from widely separated electrodes; distal (DIST), proximal (PROX) and both electrodes combined (multipoint pacing, MPP). The LV was stimulated relative in time to RV activation (either RV pace-onset or QRS-onset), with the LV stimulated prior to (PRE), simultaneous with (SIM) or after (POST) RV activation. In addition, we analyzed the interactions of the two groups (FUSION/STANDARD) with three different electrode configurations (DIST, PROX, MPP), each paced with three different degrees of LV pre-/postexcitation (PRE, SIM, POST) in a statistical model.

Results: We found that FUSION provided shorter Td and QRSd than STANDARD, MPP provided shorter Td and QRSd than DIST and PROX, and SIM provided both the shortest QRSd and Td compared to PRE and POST. The interaction analysis revealed that pacing MPP with fusion with intrinsic RV activation simultaneous with the onset of the QRS complex (MPP^*FUSION^*SIM) shortened QRSd and Td the most compared to all other modes and configurations. The difference in QRSd and Td from their respective references were significantly correlated (β = 1, R = 0.9, p < 0.01).

Conclusion: Pacing modes and electrode configurations designed to optimize electrical wave-front fusion (intrinsic RV activation, LV multipoint pacing and simultaneous RV and LV activation) shorten QRSd and Td the most. As demonstrated in this study, electrical and mechanical measures of resynchronization are highly correlated. Therefore, Td can potentially serve as a marker for CRT optimization.

Keywords: LV dP/dtmax; QRS duration; acute hemodynamic response; cardiac resynchronization therapy; fusion with native conduction; heart failure.

Figure 1

Mode of pacing and fusion with intrinsic RV conduction. A significant increase was found for both QRS duration and time-to-peak dP/dt with RV pacing compared to baseline LBBB. Multipoint pacing with intrinsic RV conduction significantly shortened QRS duration and time-to-peak dP/dt compared to the others (p < 0.01). RV, right ventricle; LBBB, left bundle-branch block; DIST, distal electrode; PROX, proximal electrode; MPP, multipoint pacing. * p < 0.01 compared to all others.

Figure 2

Effect on LV pre-/postexcitation on QRS duration and Td. Simultaneous pacing (SIM) with intrinsic RV conduction shortened both QRS duration and Time-to-peak dP/dt compared to pre-excitation and postexcitation. PRE, pre-excitation of the LV between 75 and 25 ms prior to QRS onset of RV pace onset; SIM, pacing the LV between 25 ms before and 25 ms after QRS onset or RV pace onset; POST, postexcitation of the LV 25–75 ms after QRS onset or RV pace onset. * p < 0.01 compared to all others.

Figure 3

Analysis of the interaction effect between mode of pacing, electrode configuration and LV pre-/postexcitation on the change in Td and QRS duration. The tables show the estimated marginal means and standard error of the linear mixed models analysis for each interaction for time-to-peak dP/dt (A), QRS duration (B), and dP/dt_max (C). The numbers are the estimated change from the reference selected as STANDARD*MPP*POST (marked in yellow). Colors that mark the estimates indicate a significant change compared to the reference (p < 0.01). Numbers are estimated marginal means ± SEM. DIST, distal electrode; PROX, proximal electrode; MPP, multipoint pacing; PRE—pre-excitation of the LV between 75 and 25 ms prior to QRS onset of RV pace onset; SIM— pacing the LV between 25 ms before and 25 ms after QRS onset or RV pace onset; POST, postexcitation of the LV 25–75 ms after QRS onset or RV pace onset, Td—time-to-peak dP/dt; dP/dt, first order derivative of pressure.

Figure 4

Agreement between mechanical and electrical measures during resynchronization with different modes of pacing and electrode configurations. (A) The relationship between the difference in Td and QRSd from the reference (STANDARD*MPP*POST, see Figure 3) as a result of resynchronization, and the Bland-Altman Plot (B) to demonstrate the agreement between the measurements. Linear relationships between the difference in Td (C) or the difference in QRSd (D) and the difference in dP/dt_max were not found. Td, time-to-peak dP/dt; dP/dt, first order derivative of pressure.

Figure 5

Diagramatic figure showing an ideal electrical activation diagram with the effect of LV pre-/postexcitation and electrodes on LVAT, RVAT and QRS duration. The diagram shows the theoretical change in LVAT and QRS duration relative to LV pre-/postexcitation based on pacing from the LV electrode, with ideally placed RV and LV electrodes. With near maximal LV pre-excitation, the LV is almost fully activated once RV is paced. The residual area, the area not yet activated from the LV as RV is paced, will be recruited from both the RV paced electrode and the LV paced electrode. Hence recruitment in the area will occur in a shorter time. This effect is more pronounced with less LV pre-excitation up to simultaneous pacing when the effect of combined activation of residual areas reaches its maximum, as the residual area recruited from both LV and RV electrodes is at its largest. LVAT and QRS duration shorten with less LV pre-excitation in this diagram down to simultaneous stimulation. Once the LV is postexcitated, the residual area shrinks and LVAT and QRS duration increase (as long as LVAT > RVAT). Shortening LVAT and QRS duration may also occur from multipoint pacing (MPP) with LV pre-excitation (blue arrow and stippled black line). It is clear from this diagram that the LV electrode position may largely affect LVAT and subsequently QRS duration and that intrinsic RV activation (that may shorten both RVAT and LVAT) will impact how this diagram reads. LV, left ventricle; RV, right ventricle; LVAT, left ventricular activation time; RVAT, right ventricular activation time.