Objective To examine the reporting of hospital-acquired bloodstream infection (HABSI) and central line-associated BSI (CLABSI) rates in neonatal intensive care units (NICUs).

Design Systematic review of evidence published after 2000 reporting HABSI cumulative incidence, crude HABSI and/or CLABSI rate and total patient-days and/or central line-days for single NICU.

Setting Inpatient.

Patients Neonates admitted to NICU.

Main Outcome Measures To consider the reporting of and relationship between cumulative incidence of BSI and HABSI and/or CLABSI rates.

Results 18 studies fulfilled inclusion criteria. There was a wide variability in reporting of HABSI indicators and risk-adjustment strategies with reported crude HABSI and/or CLABSI rates showing an approximately sevenfold variation between centres. Information about NICU size and level of care was not always available. Many studies provided insufficient information about case mix, such as surgical care provision and prematurity. The proportion of total patient-days that were central venous catheters (CVC)-days ranged from 11.7% to 85.4%. Only six studies reported HABSI and CLABSI incidence. Comparing HABSI and CLABSI ranking, we found a relationship between rates.

Conclusions We found significant variability in HABSI rate reporting. Although there appears to be an association between CLABSI and HABSI rates, non-CVC-related BSIs are likely to be highly relevant in some NICUs. If confirmed, and given CLABSI rates are more challenging to collect, it may be more appropriate to use HABSI rates for monitoring NICU healthcare-associated infection (HAI) in some settings. A European network of NICUs using a standardised methodology is required to determine the feasibility and reliability of different risk-adjusted measured of HAI rates.