[Value of various concepts in long-term electrocardiography]

With the introduction of long-term ECG by N. Holter more than 25 years ago a more sensitive detection of spontaneous arrhythmias and ST-changes became possible compared with the standard ECG at rest. Unfortunately, visual high-speed analysis of the tapes caused remarkable loss of stored informations due to the failure of rapid identification. Full disclosure of the ECG, therefore, is a valuable supplement to long-term ECG. As an isolated method, however, it cannot be recommended since the significantly reduced amplitude and paper speed of full disclosure likewise is responsible for a remarkable loss of informations. Only computer-aided analysis of the ECG provides more complete information and at the same time more quantitative data. Analysis of the accuracy of different computer-based systems showed, however, that a sensitivity as well as a positive predictive accuracy of 100% remains the exception to the rule. Moreover, the spectrum of cardiac arrhythmias and ST-changes identifiable by automatic analysis is small. For these reasons, interaction of the user is mandatory--in order to validate computer results as well as to detect ECG abnormalities unidentifiable by computer. The first is possible only to a limited degree with discontinuous ECG recording by event recorders, the last remains impossible. Thus, the promising concept of real-time cardiac monitors should be considered investigational or experimental at this time.