Development and implementation of a customised rapid syndromic diagnostic test for severe pneumonia

Evidence before this study The problems of microbial culture-negative pneumonia are well described in the literature. We conducted Medline searches from 1996 to December 2020 using the MeSH terms ‘polymerase chain reaction’, ‘pneumonia’ and ‘critical care’. Most reported studies concerned viral PCR only or single organism PCR and the great majority of critical care-based studies were retrospective studies on stored samples. We identified one study which retrospectively examined the impact on antimicrobial prescribing of a non-quantitative assay designed for blood cultures applied to respiratory samples. This analysis showed a reduction in antibiotic use. A small randomised trial examining blood-based PCR in patients with abdominal and pulmonary sepsis demonstrated faster time to result and no difference in time to therapy change or rate of therapy change. Notably several studies recorded that the narrow spectrum of organisms covered by molecular testing may lead to inappropriate withholding of antibiotics if the tests were used to guide therapy. Clinical trials largely excluded the most severe cases of pneumonia, including those requiring mechanical ventilation. Recent studies have demonstrated the difficulty in using novel tests to change entrenched antimicrobial prescribing patterns amongst critical care physicians. Our group has previously investigated TaqMan array card (TAC) technology for syndromic diagnosis, but found that limited pathogen coverage led to limited impact on clinical decision making.

Added value of this study This observational study demonstrated that a customised, broad-spectrum syndromic test for severe pneumonia could lead to substantial change in treatment. This is in the context of an Intensive Care Unit (ICU) with an established antimicrobial stewardship programme and daily input from microbiology and molecular specialists. The assay was proven to be accurate, with almost all organisms detected confirmed by culture and/or metagenomic sequencing, as well as excellent performance against a panel of organisms used for national reference standards.

Implications of all available evidence The use of multiplex and multi-organism molecular approaches offers rapid, sensitive testing which detects organisms in excess of those identified by conventional culture. Thus, a suitably broad range of targets, ideally including antibiotic resistance genes, is required to safely and effectively optimise antimicrobial therapy, and rapid diagnostics. Tests with a narrower range, or which cannot be customised to local microbial flora, may have limited impact on antimicrobial prescribing. Further work is required on determining cut-offs for distinguishing colonisation from infection, alongside integration with host-markers of inflammation and clinical impact.