The accuracy of analysis and susceptibility to faults of additional diagnostic functions in pacemakers were tested in 16 patients (eight men, eight women, average age 61.5 +/- 10.8 years, 14 VVI, 2 AAI) with diagnostic pacemakers (Quintech DPG 921). The average observation period was 19.7 months. With 109 follow-up intervals this meant an observation period of 25.9 patient-years. A stage-by-stage test structure over several periods of time was used in order to validate the data acquired through pacemaker Holter functions by means of control ECGs and mathematical test methods. The only fault to occur in the diagnostic functions was a loss of stored data. This was observed in 10% of follow-up intervals and was, on average, 28% (16-38%) of the total events. 44% (7/16) of the pacemakers were affected by this malfunction. Our results show that the diagnostic data are certainly correctly acquired. The clinical value of the diagnostic functions is clearly reduced by this potential loss of data.