Signal quality evaluation in Ambulatory Impedance Cardiography

Abstract

A fundamental question regarding ambulatory impedance cardiography monitoring is evaluation of the rate of artefacts during impedance holter recordings. It is also important to know what is the average length of recordings, which are too noisy for automatic analysis. Another problem is the question of how representative are the mean values from periods of different length for evaluating trends. With the aim to evaluate these uncertainties 12 male patients with atrial fibrillation (aged 36-78 years) and 6 healthy males (24-68 years) as a control underwent examination (using ReoMonitor) lasting for 12 hours starting from about 18:00. The ratio of artefact free cycles to the total number of cycles was calculated for the same hour of recording (during daytime activity) when the averaging period varied from 5 to 60 minutes. Also the rates of artefact-free recordings were calculated for each 30-minute period during the daytime (18.00-23.00) and over the night period (23:00-04:00). The percentage of cycles recognised as normal according to the criteria of automatic determination of cardiac parameters from impedance cardiography signals varied from 20% during the daytime up to 90% during the night hours, when 5-minute period are analysed. In the worst case, the minimal individual value of the ratio (observed in 5-min periods) was 0.21 (when mean ± SD was 0.63±0.26 and range: 0.21-0.98). For the entire group of subjects (n=18) in two cases the minimal ratio was lower than 0.25 and in eight cases more than 0.75. It was demonstrated that the system might be used to collect signals in a laboratory and in the field to monitor changes in cardiovascular parameters. © Springer-Verlag 2007.