Heather O Brien, Maryke Nielsen, Kenneth Meyler, Nicola O Sullivan, Robert Cunney, Richard J. Drew
results in a paediatric patient cohort and to review discordant results. Methods: A retrospective audit of all blood and CSF samples from a tertiary referral paediatric hospital sent to the Irish Meningitis and Sepsis Reference Laboratory over a four-year period. All PCR tests performed for H. influenzae which had a contemporaneous culture performed were included in the analysis. Results: For the blood PCR test, there were 10 positive samples out of 1,367 samples tested. The sensitivity was shown to be 60% (95% CI 14.6–94.73%) and the specificity was 99% (95% CI 98–99.7%). For the CSF assay, the sensitivity was 100% (95% CI 15.8–100%) with a specificity of 99% (95% CI 99.2–99.9%) and there were 5 positive samples by PCR out of the 1,224 samples tested. Ten patients had positive PCR results, with a negative corresponding culture (blood=7, CSF=3). Three of the 10 cases were deemed to be primary H. influenzae infections, while seven were deemed to be likely co-infection (respiratory syncytial virus=2, Influenza=2, measles=1, rotavirus=1, Staphylococcus aureus pneumonia=1). Conclusions: The incidence of invasive H. influenzae disease was low in this population. The sensitivity and specificity of the assay in CSF was excellent, but the sensitivity of the assay in blood was lower at 60%. Most patients with discordant PCR/culture results had viral co-infections. A more rational approach to requesting PCR is required in paediatric patients.