BACKGROUND: Central venous catheters (CVCs) are widely used for children with cancer and are a major risk factor for bloodstream infection. Early and specific diagnosis of CVC-associated bloodstream infection allows early targeted treatment, reducing the risk of CVC removal and avoiding the operative risks and trauma of reinsertion, but peripheral vein sampling, as used in adults, improves specificity but is not usually acceptable in children. OBJECTIVE: To improve the detection and treatment of CVC-associated bloodstream infection in children (aged 0-18 years) with cancer admitted with fever. METHODS: There were four main studies: (1) evaluation of the diagnostic accuracy of a quantitative molecular method for the detection of bacterial deoxyribonucleic acid (DNA), based solely on blood samples drawn through the CVC; (2) analysis of the prognostic risk of CVC removal and duration of intravenous (i.v.) antibiotic treatment days in relation to presenting clinical features, blood culture results and bacterial DNA test results; (3) systematic reviews of treatment options for CVC-associated infection and a questionnaire survey of current practice in paediatric oncology centres; (4) evaluation of the clinical effectiveness of different test-treatment strategies to reduce i.v. antibiotic treatment days and unnecessary CVC removals. RESULTS: (1) The bacterial DNA test detected two-thirds [95% confidence interval (CI) 44% to 83%] of children classified with probable CVC-associated infection - specificity was 88% (95% CI 84% to 92%). Although high bacterial DNA concentrations were associated with subsequent CVC removal and long duration of i.v. antibiotic treatment, the test did not improve the prediction of these outcomes over and above clinical signs of CVC-associated infection combined with blood culture results. (2) High DNA load was predictive of CVC removal and i.v. treatment duration, before blood culture results became available at 48 hours after sampling. (3) There was limited evidence that antibiotic lock treatment reduces the risk of recurrent CVC-associated infection or CVC removal (pooled relative risk 0.7, 95% CI 0.47 to 1.05), but prophylactic use of antimicrobial locks halved the risk of bloodstream infection (pooled incidence rate ratio 0.43, 95% CI 0.36 to 0.51). Contrary to this, the national survey of paediatric oncology centres found that locks are being used for treatment rather than prevention and that problems related to the formulation of lock solutions currently impede a shift to their prophylactic use in children. (4) Most i.v. treatment days would be saved by early stopping of treatment for children at low risk of infection. LIMITATIONS: The accuracy study was limited primarily by the lack of an adequate reference standard, and the main limitation of the series of systematic reviews was the poor quality of included studies and lack of randomised controlled trials of CVC removal or antimicrobial locks for treatment of infection. CONCLUSIONS: There is strong evidence to support the use of antimicrobial locks for prevention of CVC-associated infection; however, few of these studies involved children with cancer. The analysis does not support routine bacterial DNA testing on admission to detect CVC-associated infection, but repeated testing (as a marker of microbial load) should be evaluated in high-risk groups. Further research should determine the effectiveness of antibiotic locks for treating CVC-associated infection. TRIAL REGISTRATION: Current Controlled Trials ISRCTN68138140. FUNDING: This project was funded by the NIHR Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 15, No. 7. See the HTA programme website for further project information.