Heart rate variability: how to exclude extrasystoles from the analysis?

The evaluation of heart rate variability (HRV) is critical in patients with frequent supraventricular or ventricular arrhythmias. Currently available ambulatory ECG monitoring systems usually exclude extrasystoles (ES) from HRV analysis. However, exclusion of ES alone leads to overestimation and false negative results. Therefore, the authors analysed recordings from patients two weeks after acute myocardial infarction with frequent arrhythmias using different exclusion criteria for RR-intervals before and after ES with regard to HRV parameters in time domain 24 hours ambulatory ECG recordings were interactively and prospectively analysed and an annotated RR-interval file loaded on diskette by the analysis system (Zymed 1610). The subsequent analysis of HRV was performed on Apple McIntosh IIfx by software written in C. Intervals around ES could be flexibly in- or excluded in various ways. The additional exclusion of the subsequent interval after ES results in a highly significant reduction of 34% for root mean square of successive differences, 28% for proportion of adjacent interval with >50 ms difference, 21% for coefficient of variation, and 20% for average of 5 min standard deviations. The reduction for other parameters of HRV was below 10%. The error was related to the frequency of arrhythmias and was higher for ventricular in comparison to supraventricular arrhythmias. Additional exclusion of any further intervals did not reveal significant changes. It is concluded that HRV parameters, especially those parameters based on successive differences, can be considerably overestimated by inadequate exclusion of frequent arrhythmias. However, the exclusion of extrasystoles and their subsequent intervals is a useful method to provide reliable results of HRV analysis in the time domain