Login Search
Printed from: http://www.napcrg.org/Conferences/PastMeetingArchives/PastAnnualMeetingAbstracts/2013AnnualMeeting
2013 Annual Meeting

2013 NAPCRG Annual Meeting

Search by:

PD24 Systematic Review of the Diagnostic Accuracy of Capillary Refill Time for Serious Illness in Children

Peter Gill; Carl Heneghan; Caroline Jones; James Taylor; Matthew Thompson; Ann Van Den Bruel; Susannah G Fleming, DPhil

11/11/13 10:15 AM - 11:15 AM Governor General II

Context: Capillary refill time (CRT) is a simple, easily performed test which is recommended as part of the initial assessment of unwell children by a number of guidelines. Objective: To review the diagnostic accuracy of CRT for identifying serious illness in unwell children. Design: Systematic review. Search strategy: Medline, Embase and CINAHL were searched for English language papers using a strategy developed in collaboration with an information specialist. All papers were screened by 2 authors, and assessed for quality using criteria based on the QUADAS-2 checklist. Inclusion and exclusion criteria: We included studies reporting measurement of CRT and at least one outcome measure on 20 or more children under 18 years of age. We excluded studies of predominantly preterm neonates, or children with significant pre-existing cardiorespiratory disease. Instrument: Capillary refill time test. Outcome measures: dehydration, severity of infection, or presence of specific bacterial infections such as meningitis, pneumonia, bacteraemia, or urinary tract infection. Results: Of 1930 references identified, 10 papers were included: diagnostic accuracy was reported on 5 studies with 499 children for dehydration, and 5 studies with 18,544 children for outcomes related to infection. Meta-analysis using the hierarchical summary ROC model was carried out to assess diagnostic accuracy for dehydration. With the exception of one study reporting prediction of dehydration in children with malaria, CRT had high specificity for both dehydration and infection-related outcomes. Analysis of likelihood ratios and prevalence shows that CRT>2s considerably increases the probability of serious infection or dehydration in susceptible populations, but that a normal CRT does not reduce the probability of serious illness. Conclusions: CRT shows promise as a “red flag” to identify a subset of children with serious and life-threatening illnesses.