The diagnosis of serious bacterial infections in the children’s emergency department

Irwin, Adam
(2016) The diagnosis of serious bacterial infections in the children’s emergency department. PhD thesis, University of Liverpool.

[thumbnail of IrwinAda_Feb2016 _2005599.pdf] Text
IrwinAda_Feb2016 _2005599.pdf - Unspecified
Available under License Creative Commons Attribution.

Download (5MB)


Background Acute febrile illness is a common presentation to the children’s Emergency Department (ED). Difficulty discriminating between Serious Bacterial Infections (SBI) and self-limiting infections results in delayed treatment of SBI, and over-treatment of self-limiting infections. Aims/Objectives To define the aetiology of bacteraemia in the children’s ED, to evaluate a universal molecular diagnostic for the diagnosis of bacteraemia, and to derive and validate risk prediction models for SBI in this setting. Methods A prospective diagnostic accuracy study of clinical and biomarker variables in febrile children presenting to the ED which incorporated a case-control study evaluating 16S rRNA followed by sequencing for the diagnosis of bacteraemia. An 11 year retrospective time series analysis described the aetiology of bacteraemia presenting to the children’s ED. The study had full ethical approval. Results Time series analysis of bacteraemia presenting to the ED between 2001 and 2011 (n=575) estimated an annual 10.6% reduction in vaccine-preventable infections, and an annual 6.7% increase in Gram-negative infections. The rate of healthcare-associated bacteraemia increased from 0.18 to 0.50 per 1000 ED attendances, and the proportion of isolates susceptible to empirical antibiotics declined from 96.3% to 82.6%. Episodes of Gram-negative bacteraemia received antibiotics 1h later than episodes of vaccine-preventable bacteraemia. 1101 children were recruited to the diagnostic accuracy study. 146 children were included in an evaluation of 16S rRNA PCR in whole blood (SepsiTest) followed by sequencing for the diagnosis of bacteraemia. 120 ‘high-risk’ children were selected alongside 26 ‘low-risk’ children. SepsiTest identified 9/16 (56%) cases of bacteraemia. Combination with blood culture yielded a sensitivity of 75%, and specificity of 66%. SepsiTest identified 17/120 bloodstream infections with Viridans Group Streptococci in the high-risk group, and none in the low-risk group (p=0.06). A risk prediction model combining clinical variables with the biomarkers CRP, Procalcitonin and Resistin discriminated well between Pneumonia, ‘other SBIs’ and no SBI (AUC 0.84 and 0.77 respectively). External validation of published models was performed and improvements in classification achieved by the addition of Procalcitonin and Resistin. The addition of biomarkers had particular value in ruling-out ‘other SBIs’ (NRI for non-events 5.3%). Conclusion Serious Bacterial Infections in the children’s ED are increasingly healthcare-associated, and remain difficult to recognise. Broad-range molecular tests which are culture independent may have a role as adjuncts to conventional microbiology but require ongoing evaluation. Meanwhile, risk prediction models improve discrimination between SBI and self-limiting infections and should be tested in robust impact studies.

Item Type: Thesis (PhD)
Additional Information: Date: 2016-02 (completed)
Divisions: Faculty of Health and Life Sciences
Depositing User: Symplectic Admin
Date Deposited: 22 Aug 2016 06:51
Last Modified: 17 Dec 2022 02:14
DOI: 10.17638/02005599