70 cases of acute dihydralazine-associated hepatitis with centrolobular or confluent necroses, registered in the files of the Berlin-Friedrichshain Institute of Pathology, between 1981 and 1985, were classified into 3 types of diagnostic probability for differential diagnosis versus virus hepatitis. Classification was conducted according to recommendations given by a working group of pathologists, specialised in liver pathology. 42 cases out of this material had come from Prenzlauer-Berg Hospital, Department of Infectious Diseases, and were re-examined under clinical aspects. 6 of them were discarded from evaluation. Type I proved to be of high diagnostic reliability, as was seen from 61% of all cases. Only 3 cases had to be discarded from that group and were associated with other drugs, such as halothane, methyldopa, and propranolol. The following clinical parameters proved to be of particular value for definite assessment of drug-induced hepatitis: time of exposure (for analysis of co-medication), time of recovery, and re-exposure test. Only circumstantial evidence so far can be provided for all histological types to causative relationship between drug ingestion and hepatitis. Compliance with mandatory notification should be ensured in all cases, since suspicious cases are explicitly included. Higher sex-related disposition of women to drug-induced hepatitis was confirmed in our material, with the female-to-male ratio being 3:1.