Optimizing atrio‐ventricular delay in pacemakers using potentially implantable physiological biomarkers

Figures

• 

  FIGURE 1. Simplified atrioventricular (AV) delay optimization protocol [Colour figure can be viewed at wileyonlinelibrary.com] We measured continuous noninvasive beat‐by‐beat blood pressure (BP) and noninvasive laser Doppler perfusion during our AV delay optimization protocol. (a) A minimum of 6 transitions were carried out between a tested AV delay (e.g., 80 ms) and a reference AV delay (120 ms). This is repeated at a range of AV delays from 40 to 320 ms until intrinsic conduction occurred. These changes are reflected in both the BP and laser Doppler traces. (b) At each alternation, a relative change in BP and laser Doppler perfusion are calculated. The mean relative change is then calculated from the average of multiple transitions and the standard error of the mean is plotted for each. This is repeated for each tested AV delay

• 

  FIGURE 2. Examples of parabolas obtained following the atrioventricular (AV) delay optimization protocol for BP (left) and laser Doppler (right) measurements with examples of their corresponding signals (bottom panel). Figure 2a shows a clearly defined optimal AV delay and its corresponding good quality laser signal. Figure 2b shows no clear optimum with the corresponding poor quality (noisy) laser signal in spite of filtering. Note that when noise is large (2b), the orientations of the resulting parabolas are random, unlike the situation when noise is much smaller than the biological signal (2a) [Colour figure can be viewed at wileyonlinelibrary.com]

• 

  FIGURE 3. Association between optimal AV delays found using BP and laser Doppler. As expected, there is a significant correlation between the laser Doppler and BP derived optimal AV delays (Spearman's rho 0.82, p < .0001) [Colour figure can be viewed at wileyonlinelibrary.com]

• 

  FIGURE 4. Effect of number of replicates in the protocol on the agreement between laser Doppler and blood pressure optima. If we trim the optimization session data to just five replicates per AV delay transition, there is considerable disagreement between laser Doppler and blood pressure derived AV delay optima (black dots in left column with median shown as red dot). These results arise from 30‐fold bootstrapping analysis of the extended‐protocol patients. As we use progressively more of the data from each optimization session, the disagreement in AV delay optima between laser Doppler and blood pressure progressively falls [Colour figure can be viewed at wileyonlinelibrary.com]