Risk stratification for arrhythmogenic events and sudden death in patients with organic heart disease, particularly those with coronary heart disease and a history of MI, continues to be one of the major tasks of clinical cardiologists, although advanced management strategies including thrombolysis, acute PTCA and surgical intervention dramatically reduced the percentage of sudden deaths following acute MIs, Noninvasive studies like resting and exercise ECG, echocardiography, signal averaging, 24-hour ECG, and radionuclide studies, as well as invasive techniques such as electrophysiologically programmed electrostimulation and coronary angiography, are being used routinely. Ambulatory Holter monitoring is an established noninvasive technique for risk stratification. There is evidence showing that its predictive potential for arrhythmogenic risks is enhanced, if more than one parameter is analyzed. Absence of ST segment changes and a normal HRV are the parameters signaling out low-risk patients. The use of additional parameters which escape electrocardiographic recording, like ventricular function and myocardial ischemia, improve the accuracy of predicting arrhythmogenic events. The most predictive combination of risk parameters is, however, still poorly understood. Future research should define normal ranges of parameters recordable by H-ECG, solve technical problems of recording data and analyzing them. In addition, the accuracy of measuring QT duration and documenting late potentials should be improved by more sophisticated methods. But it is unrealistic to expect that the QT interval will become amenable to automatic analysis in all patients. A fully automatic QT analysis without visually checking the measuring points at the tip and the end of the T wave for their consistency is hardly conceivable. The documentation of late potentials, in turn, is limited by artefacts caused by muscle contraction during physical activity. Clinical aspects, e.g., the predictability of arrhythmogenic events in patients with cardiomyopathies and valvular disease should be addressed. This will require studies combining the predictive potentials of rhythmologic and hemodynamic data.