A clinical lab experience with an automated HIV Antigen/Antibody (Ag/Ab) combined assay

OBJECTIVES: To describe the diagnostic performance of a new fourth-generation HIV Ag/Ab chemiluminescent immunoassay, available on the new LIAISON® XL analyser, in a clinical setting.
METHODS: Through February 2012-October 2013, 12,438 samples of serum, received at our laboratory for screening for HIV infection were routinely tested with LIAISON® XL Murex HIV Ab/Ag assay (HIV-XL), which employs HIV-1, HIV-1 group O, and HIV-2 antigens and anti-p24 monoclonal antibodies in two coupled reagent cartridges, providing information of the overall Ab/Ag reactivity and detail of the specific reactivity for anti-HIV/HIV p24 antigen.
Each serum with positive result or with negative result displaying a value close to the cut-off were sent to the regional AIDS-Reference Laboratory (RefLab) to perform confirmatory assays (PCR, Immunoblot).
A previous verification of the HIV-XL demonstrated 100% sensitivity with a challenge panel of hundred positive sera provided by the RefLab.
Performed external quality control was from United-Kingdom National External Quality Assessment Service (NEQAS).
RESULTS: Out of the clinical samples, 12,312 non-reactive samples (including 6 negative results displaying a value close to the cut-off further confirmed true HIV negative), 64 Ab HIV reactive samples (all confirmed HIV-1 positive by immunoblot), including 4 samples reactive also for Ag HIV (confirmed positive by Ag assay/PCR), 42 Ab HIV reactive samples tested negative by immunoblot, and 20 Ag HIV reactive samples tested negative by the kit used for the Ag p24 detection in our HIV Reference Lab, have been found. All the 43 NEQAS specimens tested, 16 reactive and 27 non-reactive, were correctly classified.
These results, considered all together, provide a calculated positive predictive value of 57.5% with an estimated specificity of 99.5% (with 95% confidence interval of 99.36-99.62%), and a calculated negative predictive value of 100% with an estimated sensitivity of 100.0% (with 95% confidence interval of 95.49-100%).
CONCLUSIONS: In our experience HIV-XL showed excellent performance associated to all the advantages of a fully automated/random access instrument.