Background: Adenosine is known to shorten atrial action potential duration and refractoriness, promoting atrial fibrillation (AF). It may also be used during non-contact mapping of atrial propagation to reduce the impact of far-field ventricular signals. However, the functional effect of adenosine on atrial propagation is poorly understood and may reveal mechanisms important in maintenance of persistent AF.
Methods: Seventeen patients undergoing elective catheter ablation for persistent AF using the AcQMap non-contact dipole density mapping system were studied. Recordings of AF propagation were made before and after administration of adenosine at one or more of three time points during the procedure: pre-pulmonary vein isolation (PVI), post PVI, or at the end of the case. The QRS-T signals were subtracted and AF propagation maps calculated. Data was exported and analysed using custom designed software to quantify localised rotational activation (LRA), localised irregular activation (LIA) and focal firings (FF) identified by the in-built AcQTrack tool. Areas of low frequency of activation patterns were removed to achieve adequate specificity using cut offs resulting in a relative 5%, 10%, 20% and 30% drop in the time each pattern was present over the recording. Patterns were quantified for the number of occurrences; percentage time patterns were present and the surface area over which patterns occurred. Global chamber dipole signals were also exported and phase mapping applied using the technique of sinusoidal recomposition and Hilbert Transform to calculate the global cycle length (AFCL) and identify phase singularities (PS).
Results: A total of 28 paired 5-second segments were analysed. Adenosine resulted in a shortening of global mean AFCL from 181 ± 19 to 163 ± 22 (mean difference 19 ms, 95% CI 10.4–26.6, p<0.0005) with a similar reduction in the lower 5% value of AFCL (representing the region with fastest activation as a surrogate for refractoriness) from 137 ± 17 to 126 ± 14 (11, 6.1–16.1, p<0.0005). Coefficient of variance (standard deviation/mean) of AFCL was unaffected.
Analysis of AF propagation patterns revealed an increase in the number of LIA occurrences at each cut off but with no clear difference in the time patterns were present or the surface area affected. At the 5% cut off adenosine resulted in an increase in LRA occurrences from 54 ± 31 to
88 ± 41 (34, 17–51, p<0.0005) with a corresponding 13% (6–21) increase in the proportion of the recording LRA was present (p = 0.001) but with no change in surface area affected. These findings were consistent across all cut offs. There was a similar increase in the number of PS identified with the use of adenosine from 33 ± 15 to 47 ± 15 (13, 7–21, p<0.0005) but with no significant change in PS duration (21, –25 to 66, p = 0.365). There were no differences in the frequency of FF.
Conclusion: The use of adenosine appeared to uniformly shorten global AFCL. Adenosine primarily promoted rotational activation patterns which occurred at a higher frequency and duration but covering the same atrial surface area. This was consistent with the higher number of phase singularities detected. This implies a possible mechanistic role of rotational activation patterns in AF maintenance. The lack of change in surface area in which these patterns are seen suggests that adenosine may reasonably be used during clinical AF mapping without adversely impacting the clinical utility of the mapping findings.