This retrospective nationwide study of 756 cases of S. aureus bacteremia in hematological patients, between 1977-1990, confirms that S. aureus bacteremia is disproportionately prevalent among, and a major cause of morbidity and mortality in, hematological patients. The case mortality rate attributable to S. aureus bacteremia (44%) was significantly higher than the rate (29%) found in patients without hematological disease. A decreasing mortality was, however, observed from the first to the last part of the study period (53 and 37% respectively). A logistic regression analysis showed that mortality was further dependent on age, the type of hematological disease and origin of infection/primary focus of infection. Analysis of clinical features confirmed that in comparison with other patients, S. aureus bacteremia in hematological patients more often was hospital-acquired, acquired without known portal of infection/primary focus of infection and much less frequently resulted in osteomyelitis or endocarditis. From phage typing and antibiotic susceptibility patterns no specific types of S. aureus were associated with hematological patients.