Computer‐Assisted Analysis of Holter Recordings

Early work in radiotelemetry and signal processing by Holter’ made it possible to detect changes in heart rate and cardiac rhythm in ambulatory subjects. Heart rhythm was identified by cycle-length-dependent superimposition of electrocardiographic complexes, with a 60-fold ratio of playback to recording speed to allow data analysis in compressed time. By 1964, Gilson et al. were able to report that simple arrhythmias, brief changes in electrocardiographic morphology, and a wide range of artifacts could be detected by these methods in apparently normal subjects.2 Useful association of transient arrhythmias with symptoms was observed by Corday et al. in patients studied with 10-hour continuous recordings,’ and limitations and advantages of early ambulatory recording methods were reviewed by Hinkle e f al. in 1967.4 Advancing technology soon made it practical to acquire longer samples of analog electrocardiographic data on a lightweight, portable, battery-powered recorder, using reel-to-reel tape as the storage medium. The advantage of extending continuous electrocardiographic monitoring to 24 hours to fully characterize highly sporadic and variable ambulatory rhythms was demonstrated by Lopes et al.’ While more prolonged recording periods may further increase the detection of sporadic arrhythmias,6 which may demonstrate striking spontaneous and activity-related ~ a r i a b i l i t y , ~ . ~ 24-hour continuous recordings have evolved as the most widely used compromise between technology and convenience in ambulatory populations.