All the ICU patients were continuously studied during the first quarter of 5 consecutive years for infections according to a standard protocol. The investigators--the infection control officer and a well-trained infection control nurse--decided if the patient was infected by referring to medical and nursing record, temperature charts, laboratory and x-ray reports and, where necessary, by clinical examination. Definitions and criteria for infections comply with the CDC and the algorithms of the Senic Project. Only the first quarter of each year from 1980-1984 was analysed. The first quarter of 1980 was analysed retrospectively, the following years were examined prospectively. In 1984 a new ICU (ICU I) in addition to the old ICU (ICU II) was opened. The two ICUs differ in building construction but have similar patients, nursing staff and medical standards. The frequency of nosocomial infection was not affected by the different building constructions. The number of patients surveyed was 1009, 60% were males and 40% females. The average age was 45.5 years and the average period of stay about 4 days. 733 patients (72.6%) were intubated and artificially ventilated for 3 days. A fatal outcome resulted in 13.2% of all patients. 1129 nosocomial infections were registered in 311 patients, which means an infection rate of 32.8%. The most frequent nosocomial infections were those of the respiratory tract. Wound infections developed in 16.6%. The urinary tract was affected in 8.8%. Nosocomial septicaemias were observed in 8.7%. Catheter-associated infections were noticed in 6.7% of the patients. A fatal outcome resulted in 26% of the patients with nosocomial infections and in 6.9% of the non-infected patients, respectively.