| Laboratory methods: |
Assay procedures are similar but not identical. |
Testing laboratories should ensure full compliance with manufacturer’s instructions for use, especially if both manufacturers’ assays are used in one laboratory. |
| Quality assurance: |
Lot-to-lot variation and differences in laboratory staff proficiency may further reduce reproducibility of results. |
Continuous quality assurance should be practiced, including by ensuring laboratory staff proficiency, by regularly running well-characterized quality assurance specimens (recent, longterm and negative) and by monitoring the reactivity of kit-supplied specimens (controls and calibrators) over time. Participation in an external quality assurance programme like EQAPOL [27] is recommended. |
| Software: |
Although data capture and analysis software are similar, interpretive criteria for specific components differ. |
The data analysis software is specific to each assay and laboratories should use the software supplied by the manufacturer. |
| Conversion: |
Although it is possible to compute an approximate conversion factor, this does not perfectly predict equivalent ODn values. |
Rather than converting results, appropriately-derived MDRI and FRR estimates should be utilized for each assay. The same ODn thresholds may not be optimal. |
| Descriptive title: |
The names ‘HIV-1 Limiting Antigen Avidity EIA’ or ‘LAg assay’ do not distinguish between the two assays. |
Users should clearly identify the manufacturer of the kits used, as well as specimen type, in all publications and reports. |
| Assay performance: |
Despite differences in calibrator reactivity, and consequently in ODn values obtained on the same specimens, performance of the two assays for surveillance purposes was virtually indistinguishable. |
Both manufacturers’ assays are suitable for use, but they should not be mixed within studies, appropriate performance characteristic estimates must be used and care should be taken when comparing results. |
