Optimising automation of a manual enzyme-linked immunosorbent assay

OBJECTIVE: Enzyme-linked immunosorbent assays (ELISAs) are widely used to quantify immunoglobulin levels induced by infection or vaccination. Compared to conventional manual assays, automated ELISA systems offer more accurate and reproducible results, faster turnaround times and cost effectiveness due to the use of multianalyte reagents. DESIGN: The VaccZyme™ Human Anti-Haemophilus influenzae type B (Hib) kit (MK016) from The Binding Site Company was optimised to be used on an automated BioRad PhD(TM) system in the Immunology Laboratory (National Health Laboratory Service) in Tygerberg, South Africa. METHODS: An automated ELISA system that uses individual well incubation was compared to a manual method that uses whole-plate incubation. RESULTS: Results were calculated from calibration curves constructed with each assay. Marked differences in calibration curves were observed for the two methods. The automated method produced lower-than-recommended optical density values and resulted in invalid calibration curves and diagnostic results. A comparison of the individual steps of the two methods showed a difference of 10 minutes per incubation cycle. All incubation steps of the automated method were subsequently increased from 30 minutes to 40 minutes. Several comparative assays were performed according to the amended protocol and all calibration curves obtained were valid. Calibrators and controls were also included as samples in different positions and orders on the plate and all results were valid. CONCLUSION: Proper validation is vital before converting manual ELISA assays to automated or semi-automated methods.