Influence of the noise and artefact in automatically analysed long term electrocardiograms on different methods for time-domain measurement of heart rate variability

The authors evaluated the effects of the misrecognition artefact of automatic electrocardiogram (ECG) recognition on four methods for time-domain heart rate variability measurement which have been previously shown to provide clinically relevant prognostic data in survivors of acute myocardial infarction. Ambulatory long-term ECGs were recorded in 182 survivors of the acute phase of myocardial infarction, and these ECGs were analyzed in two ways: automatically and with visual checks and manual editing of the computerized recognition. Correlation coefficients were calculated for the two sets of results of each method. The precision with which patient stratification based on the manually validated data was reproduced with the automatically obtained data was also assessed for each method. The results shows that two methods were less influenced by errors in automatic recognition of long-term ECG than were the other two methods.<<ETX>>