Bacteraemia caused by hospital-type micro-organisms during hospital stay.

A sharp transition between community-type and hospital-type pathogens at the second or third day of hospital stay is often assumed. This study aimed to test whether such a threshold phenomenon exists for bloodstream infections and to examine the relationship between the proportion of infections caused by hospital-type pathogens and length of stay in the hospital. Blood stream infections were studied in a referral and a university hospital in west Denmark, and a university hospital in central Israel during three study periods (1994-1996, 1992-1995, 1989-1995 in the three hospitals respectively). No threshold effect at 2-3 days stay in the hospital could be demonstrated. However the percentage of Pseudomonas aeruginosa bloodstream infections increased constantly in the three hospitals from 1%, 1% and 7% during the first 2 days to 7%, 4%, and 14% during the third week of hospital stay (P<0.01 for all three comparisons-chi(2)for linear trends). For Candida sp. the increase was from 0%, 2%, 1% during the first 2 days to 3%, 4%, and 9% during the third week, P<0.05. Methicillin-resistant Staphylococcus aureus in Israel increased from 26% of the total number of S. aureus during the first 2 days to 69% during the third week, P<0.0001. For penicillin-resistant S. aureus in Denmark, the percentages were 84% and 100%, P<0.05.The percentage of infections caused by hospital-type pathogens increased almost linearly during the first 3 weeks of hospital stay, with no threshold effect. This trend should be taken into account when prescribing empirical therapy for nosocomial infections.