Extracting concealed atrial premature contractions from ambulatory Holter data using a PC-based system

Atrial fibrillation may be cured by eliminating triggering atrial premature contractions (APCs.) Characterizing these APCs may help localize triggers, however this is difficult when the APC is obscured by the preceding T-wave. Therefore we sought to develop an inexpensive method to extract APCs from ambulatory Holter monitors in AF patients. We examined 24-hour Holter strips from patients with AF triggered by APCs. Triggering and nontriggering APCs were extracted using a PC-based T-wave subtraction algorithm. The extracted APCs were compared visually as well as by correlation software. The software accurately identified 29 episodes of atrial fibrillation and extracted 76 APCs in 12 patients. In all cases, visual inspection confirmed the presence of atrial fibrillation of APCs. Triggering APCs were similar to other triggering APCs in each patient while nontriggering APCs were less similar to other nontriggering APCs in each patient. Triggering APCs were different from nontriggering APCs. Automated computerized subtraction of T-waves using an inexpensive PC-based system may permit characterization of APCs during ambulatory monitoring.