Assessment of mechanical dyssynchrony in cardiac resynchronization therapy

Cardiac resynchronization therapy (CRT) significantly reduces morbidity and mortality in patients with symptomatic severe heart failure and evidence of interventricular conduction delay by ECG. Unfortunately, one third of patients do not respond to CRT and selection criteria may need to be improved. Assessment of mechanical dyssynchrony by echocardiography has been suggested to add value in selection of CRT candidates. However, current methods for dyssynchrony analysis may not accurately reflect an activation delay amenable to CRT and controversy remains for the role of mechanical dyssynchrony.

This thesis was based on the assumption that benefit from treatment with CRT requires a significant activation delay of the left ventricle (LV). It was hypothesized that echocardiographic methods for evaluation of mechanical dyssynchrony, reflecting this fundamental pathophysiologic change, could predict response to CRT. In addition, it was hypothesized that this approach would provide improved diagnostic value with regards to dyssynchrony analysis compared to conventional time-to-peak measurements.

Three studies were performed and consisted of: 1) A prospective study in 67 consecutive patients who fulfilled standard criteria for CRT and had left bundle branch block (LBBB) by ECG. Patients underwent 2D-strain echocardiography (speckle tracking analysis) one day prior to CRT, at day one and six months after implantation and the mechanics behind CRT-response was studied. Strain patterns thought to reflect a complete LBBB were characterized and the predictive ability of this approach was tested with regards to echocardiographic response at six months (> 15% reduction in LVESV) and compared to current time-to-peak indices. 2) A retrospective study in 131 consecutive patients from two centers. Patients all had a Tissue Doppler Imaging (TDI)-dyssynchrony study prior to implantation. Baseline mechanical dyssynchrony was determined by cross-correlation analysis (XCA), a more quantitative method for comparison of contraction patterns, and the association with long-term outcome (survival free from LVAD or heart transplantation after four years) was determined and compared to current time-to-peak indices. In addition, subgroup analysis of the relation to QRS-duration was performed. 3) A prospective study of 33 consecutive CRT-recipients. Patients were VV-optimized at day one after implantation. At six months, TDI and 2D-strain analysis were performed at six different interventricular pacing intervals in steps of 20 ms to investigate the performance of different indices of mechanical dyssynchrony (time-to-peak indices and XCA, respectively) and the relation between mechanical dyssynchrony and hemodynamic performance by LVOT VTI.

1) An LBBB-related strain pattern was highly predictive of LV re-modeling response to CRT at six months and significantly added to other known predictors of outcome (etiology and QRS > 150 ms). A reversal in strain-ratio between the early and late-activated myocardial regions was observed at day 1 only among responders suggestive of an important role in promotion of remodeling. 2) Mechanical dyssynchrony at baseline measured by XCA was independently associated with improved long-term outcome in CRT-recipients. Patients with lack of mechanical dyssynchrony and QRS between 120-150 ms showed particularly poor outcome. 3) VV-optimization after six months of CRT was beneficial in both responders and non-responders. Improvements in mechanical synchrony by any method translated into improved hemodynamics. XCA showed the best feasibility, reproducibility and correlation to hemodynamic performance. In all three studies, new methods for dyssynchrony assessment performed better compared to conventional indices.

This thesis confirms the importance of mechanical dyssynchrony for outcome to CRT and demonstrates the value of dyssynchrony assessment for prediction of response to CRT as well as in optimization of device programming. New methods, which better reflect a significant LV activation delay, provides an improved tool for dyssynchrony analysis compared to conventional techniques.