Development and validation of a dual sensing scheme to improve accuracy of bradycardia and pause detection in an insertable cardiac monitor.

BACKGROUND

Undersensing of premature ventricular beats and low-amplitude R waves are primary causes for inappropriate bradycardia and pause detections in insertable cardiac monitors (ICMs).

OBJECTIVE

The purpose of this study was to develop and validate an enhanced algorithm to reduce inappropriately detected bradycardia and pause episodes.

METHODS

Independent data sets to develop and validate the enhanced algorithm were derived from a database of ICM-detected bradycardia and pause episodes in de-identified patients monitored for unexplained syncope. The original algorithm uses an auto-adjusting sensitivity threshold for R-wave sensing to detect tachycardia and avoid T-wave oversensing. In the enhanced algorithm, a second sensing threshold is used with a long blanking and fixed lower sensitivity threshold, looking for evidence of undersensed signals. Data reported includes percent change in appropriate and inappropriate bradycardia and pause detections as well as changes in episode detection sensitivity and positive predictive value with the enhanced algorithm.

RESULTS

The validation data set, from 663 consecutive patients, consisted of 4904 (161 patients) bradycardia and 2582 (133 patients) pause episodes, of which 2976 (61%) and 996 (39%) were appropriately detected bradycardia and pause episodes. The enhanced algorithm reduced inappropriate bradycardia and pause episodes by 95% and 47%, respectively, with 1.7% and 0.6% reduction in appropriate episodes, respectively. The average episode positive predictive value improved by 62% (P < .001) for bradycardia detection and by 26% (P < .001) for pause detection, with an average relative sensitivity of 95% (P < .001) and 99% (P = .5), respectively.

CONCLUSION

The enhanced dual sense algorithm for bradycardia and pause detection in ICMs substantially reduced inappropriate episode detection with a minimal reduction in true episode detection.