654. Evaluation of the febridx host response point-of-care test to differentiate viral from bacterial etiology in adults hospitalized with acute respiratory illness during influenza season

Abstract

Background: antibiotics are overused in patients hospitalized with acute respiratory illness (ARI). Diagnostic uncertainty regarding microbial etiology contributes to this practice and so a host response test that can distinguish between viral and bacterial infection has the potential to reduce unnecessary antibiotic use. The FebriDx is a low cost, rapid, host response POCT that uses fingerpick blood samples to distinguish between viral and bacterial infection but has not been evaluated in hospitalized adults with ARI.

Methods: we took fingerpick blood samples from adult patients with ARI, hospitalized during influenza season, and tested them on the FebriDx. Respiratory samples were tested for viruses on the FilmArray Respiratory Panel (FARP). The FebriDx was evaluated for ease of use, failure rate and accuracy of the results (Viral, Bacterial, Negative).

Results: 149 patients were approached and 10 patients declined fingerpick testing. A valid result was obtained from 124/139 (89%) overall. Common user comments included test failure due to difficulty of getting blood to fill the capillary tube and difficulty in interpreting the results lines due to the variability of color change. 111/124 (89%) were tested for viruses by FARP. 69/111 (62%) had viruses detected. Of 69 patients with viruses detected, 41 (59%) had influenza, 12 (17%) rhino/enterovirus and 16 (23%) other viruses. 44/69 (64%) had a viral FebriDx result. For influenza-positive patients 34/41 (83%) had a viral FebriDx result, 1/12(8%) of rhinovirus-positive patients had a viral FebriDx result and 9/16 (56%) of patients with other viruses detected had a viral FebriDx result. These are interim results. Full results for 200 patients will be available at presentation.

Conclusion: the use of the FebriDx POC was associated with a failure rate of ~10% and problems with the interpretation of result lines. FebriDx was not sufficiently accurate in differentiating viral and bacterial infection when using detection of virus by PCR as the definition of viral infection; however, FebriDx had a high PPV for all viral detection and for influenza detection in this cohort and could have utility in hospital emergency departments.

More information

Identifiers

ORCID for Nathan Brendish: orcid.org/0000-0002-9589-4937

ORCID for Tristan William Clark: orcid.org/0000-0001-6026-5295

Catalogue record

Export record

Altmetrics