RESUMO
BACKGROUND: Immunoassays allow the specific detection and quantitation of biological molecules in complex samples at physiologically relevant concentrations. However, there are concerns over the comparability of such techniques when the same assay is performed by different operators or laboratories. An international intercomparison study was performed to assess the uncertainty involved in the estimation of a protein cytokine concentration using a fluorescent ELISA. METHODS: The intercomparison study method was based on a non-competitive sandwich immunoassay with an enhancement step to generate a fluorescent readout. The intercomparison was performed in two phases, with the uncertainty of the instrument determined separately from that of the assay. The 11 laboratories participating in the study represented national metrology institutes or nominated expert laboratories. RESULTS: Participants were asked to determine an undisclosed concentration of interferon using a supplied standard. The mean participant estimate and experimental standard deviation of the mean was 3.54+/-0.22 mg/L, with the spread of data ranging around +/-35% of the mean. The quantitation range of the ELISA and of participants' instruments displayed large variation that contributed to the overall uncertainty. CONCLUSIONS: Identified sources of uncertainty within the ELISA methodology included pipetting, data fitting, model selection and instrument/plate variation.