Aims Catheter associated blood stream infection (CABSI) is a serious complication of central vessel catheterisation (CVC). During 2010 a series of policy changes were introduced including the use of multi-lumen umbilical venous catheters (UVC). Our aims were to assess whether these changes reduced the rate of CABSI or peripheral line use, and whether the CABSI rate varied by CVC type.

Methods CVCs inserted during 2009 and 2012 were identified using the electronic patient record (EPR). Line type, duration of catheterisation and patient demographics were collected from the EPR and cross referenced with the blood culture results. CABSI was defined using standard criteria.

Results In 2009 657 CVCs were placed in 311 babies providing 3992 catheter days (cd). In 2012, 682 lines were placed in 292 babies, providing 3792 cd. There were no significant differences in patient demographics. During 2009, the highest CABSI rate was seen during days with multiple CVCs in situ (39/1000 cd), in days of care with a single CVC the rate was higher in long lines (25/1000 cd) than in UVCs (17.5/1000 cd).

Between 2009 and 2012:

CABSI rate fell from 22.4/1000 cd to 6.1/1000 cd.

The proportion of babies with a CVC who developed CABSI fell from 23% to 7.5% (p < 0.0001).

There was a reduction in the proportion of CVC days provided by long lines from 51% to 45% (p < 0.01) and an increase in UVC days from 24% to 33% (p < 0.01).

The number of peripheral lines used in babies with CVCs fell by 21% (from 1557 to 1228).

The proportion of babies with a CVC who never required a peripheral line rose from 38/311 (12%) to 51/292 (17.5%) (P = 0.01).

Conclusion We have seen a significant reduction in CABSI since the introduction of our new policies. Some of this reduction can be attributed to the increased usage of multi-lumen UVCs however the magnitude of the effect suggests the whole care bundle has contributed. There has also been a reduction in the use of peripheral venous cannulation.