A-320 Validation of a Multi-Analyte Immunoassay for Distinguishing Bacterial Versus Viral Infections in a Pediatric Cohort

Abstract

Abstract Background Clinical presentation of viral and bacterial infections or co-infections overlaps significantly. Pathogen identification is the gold standard for appropriate treatment. Recently, FDA approved a multivariate index test called MeMed-BV on the MedKey analyzer that distinguishes viral and bacterial infections based on a score derived from the differential expression of 3 host proteins. Here, we sought to validate the MeMed-BV immunoassay on the MedKey analyzer in our pediatric hospital, following guidelines from the Clinical and Laboratory Standards Institute. Methods The analytical performance of MeMed-BV was evaluated by precision (intra- and inter-assay), accuracy (sensitivity and specificity) and interference studies. A retrospective cohort study (n = 60) was conducted to assess the diagnostic accuracy using plasma samples from pediatric patients with acute febrile illness who visited the emergency department of the Texas Children's Hospital, Houston, TX, for whom a C-reactive protein test was requested. The following inclusion criteria were employed: 1) age ≤18 years during the clinical encounter; 2) peak fever >37.5°C with the onset of symptoms such as cough, sore throat, abdominal pain or nausea and vomiting; 3) duration of symptoms ≤7 days; and 4) a positive nucleic acid amplification test (for virus) or abnormal blood culture (for bacteria). Patients with 1) active malignancy, 2) congenital immunodeficiency, 3) active immunosuppressive or immunomodulatory therapy and 4) human immunodeficiency virus or hepatitis B/C virus infection were excluded. Clinical performance was evaluated by blinded adjudication from our emergency room physicians. Results MeMed-BV showed acceptable intra- and inter-assay precision with a range of <3 score units in both the high-score bacterial and the low-score viral samples. Accuracy studies revealed a sensitivity of 94% and specificity of 88% for identifying bacterial infections or co-infections. Our MeMed-BV results showed an excellent agreement (R = 0.998) with the manufacturer's laboratory data and compared well with ELISA studies. Gross hemolysis and icterus did not affect the assay, but gross lipemia showed a considerable bias in samples with a moderate likelihood of viral infection. Importantly, the MeMed-BV score performed better than routinely measured infection-related biomarkers like white blood cell counts, procalcitonin and C-reactive protein in classifying bacterial infections. Furthermore, it is important to note that the score but not individual levels of the biomarkers, should be used to discriminate between bacterial and viral infections. Conclusion MeMed-BV score on the MedKey analyzer demonstrated acceptable analytical performance and is reliable for distinguishing viral and bacterial infections or co-infections in pediatric patients. Future studies are warranted to examine the clinical utility, especially with respect to reducing the need for blood cultures and reducing the time to treatment for the patient.