An objective tool to quantify atrial fibrillation substrate in freely moving rats

Introduction: The utility of rodents in AF research is growing exponentially. However, due to the small cardiac dimensions, rodent atrial recordings are often contaminated by ventricular signals limiting the capacity to analyze the properties of the arrhythmic waveforms. Recently, we introduced an implantable miniature quadripolar-electrode adapted for comprehensive atrial studies in ambulatory rats. Our system enables repeated testing of atrial refractoriness (AERP) and AF susceptibility. Here, based on this technical achievement we aimed to develop an unbiased computerized tool to "purify" the atrial signals and objectively analyze their level of irregularity.

Materials and methods: Nine rats were sequentially tested 1, 4 and 8 weeks post implantation of a miniature quadripolar-electrode on the right atrium and three subcutaneous peripheral ECG leads. AF triggering protocols contained twenty bursts (20s, 100Hz, double threshold). Pre-burst ventricular contamination was sampled and automatically subtracted from the pre- and post-burst atrial signals based on QRS detection in the ECG. Thereafter, the "pure" atrial signal was analyzed in 1s windows using the Lempel-Ziv complexity algorithm with a threshold of 20% maximal amplitude. Complexity-ratio was calculated for each post-burst interval by normalizing it to the pre-burst complexity value.

Results: Complexity-ratio values were markedly higher for positive vs. negative waveforms (Fig1A, p<0.001). ROC curve analysis indicated an optimal complexity-ratio cutoff of 1.236 which detected arrhythmic events with high sensitivity and specificity (Fig1B). Automated analysis of all post-burst recordings indicated a gradual increase in signal complexity over time (Fig1C, p<0.05), in agreement with the AERP shortening and AF substrate increase that we recently reported.

Conclusion: This novel methodology, which is inherently sensitive to the level of irregularity, accurately detects atrial arrhythmic waveforms. In combination with the superb capabilities of our implantable device this automated algorithm could markedly facilitate AF studies in small animals leading to the development of novel therapies.