All patients with an intravascular device in the intensive care units
at Prince of Wales and Prince Henry Hospitals between July and Septemb
er 1995 were followed for the development of line-related bacteraemia
per 1000 line days. Cases of sepsis related to an intravascular device
were identified using a case definition which incorporated clinical a
nd laboratory parameters. Data were collected prospectively for the da
tes of insertion and removal of devices for 188 lines inserted in 69 p
atients, The majority (90%) of lines had both date of insertion and re
moval documented allowing the calculation of the rate of primary bacte
raemia over 832 at-risk line days. Multiple concurrent lines were more
common (88.4%) than single lines, with one central and one or two per
ipheral lines being the most common (42.3%) combination. Five cases of
bacteraemia were diagnosed clinically and confirmed microbiologically
. The incidence density of primary bacteraemia was 6.0 per 1000 line c
lays (CI95% 5.7-6.3).