[The reliability of computer-assisted long-term ECG analysis of pacemaker malfunction in patients with ventricular demand pacemakers].

The importance of the pacemaker follow-up clinic has markedly increased in the face of the currently available multiprogrammable pacemakers. In contrast to the common standard techniques in assessing pacemaker dysfunctions, Holter monitoring allows the detection of transient pace-maker dysfunctions during a long period of time. Especially computer-aided analysis provides a considerable progress, because different pacemaker dysfunctions can be detected during prolonged time periods, and--in contrast to visual analysis--a reliable survey of the real extent of transient pacemaker dysfunctions in the individual patient is assessed. The reliability of computer-aided analysis by a specially designed module was prospectively investigated in 100 consecutive patients with permanent ventricular inhibited demand pacemakers. It could be demonstrated that the positive predictive accuracy of this analysis was limited to 60.2% in detecting failures to sense and 63.2% in detecting inappropriate inhibitions, respectively. All detected failures to capture were false positive events. The positive predictive accuracy, therefore, was not calculated for this category of event. The overall positive predictive accuracy was 59.9%. In contrast, the sensitivity of computer-aided analysis was remarkably high. Possible causes of false positive and false negative findings are described. The reliability of pacemaker pulse detection was also investigated. Out of 100 analyzed Holter recordings five showed a temporary total loss of pacemaker pulses. Loss of single pacemaker pulses was found in 30 patients. False positive pacemaker pulses were seen in three patients. These results show that visual control and validation by an experienced physician are mandatory.(ABSTRACT TRUNCATED AT 250 WORDS)