We performed an epidemiological investigation owing to the appearance between February and March 1988, of subsequent MRSA infections in our ICU; we also started a checking scheme. At the same time we performed a retrospective-perspective study of 701 patients (310 newborns 391 not newborns) admitted into ICU from 1-1-1987 to 30-6-1988. This study allowed us to settle the presence of MRSA in ICU and the manner in which they become infected. From our study the 4.9% of our patients contracted hospital infections and the 34.2% of these were MRSA infections. The 39.7% of our colonized patients were MRSA. It is really significant the frequency of MRSA colonization (p less than 0.0005) among newborn patients compared with other age patients; in the meantime there is not statistical significance in the frequency of MRSA hospital infections among the two groups. The results of our study show that infections were probably transmitted by the hands of the members of staff and that the reservoir of MRSA are tracheostomized hosts or long term ventilated patients.