(M1130-02-07) Embracing Singlicate Ligand Binding Biomarker Assays: A Case Study Comparing Duplicate vs Singlicate Analysis for the Determination of IL-6, IFN-gamma, and TNF-alpha in Human Sodium Heparin Plasma

Purpose: Most conventional LBAs pipette a sample into two adjacent wells, whereas LC/MS assays and other clinical assays are performed and validated using singlicates. The rationale for duplicates or even triplicates was based on historical assumptions regarding LBA imprecision which may no longer be valid for many LBAs today. We retrospectively investigated whether an electrochemiluminescent three plex ligand binding biomarker assay for IL-6, IFN-gamma, and TNF-alpha (from Meso Scale Discovery) in human sodium heparin plasma could have been executed in singlicate. This assay was validated in 2014 and has been used to support three clinical studies, which makes it an excellent case study to evaluate the feasibility of singlicate analysis.

Methods: Quality control sample data over time were evaluated through control charts to ensure method stability and representative data. Validation data were analyzed via standard linear mixed effects models¹ to estimate variability sources (run, aliquot, well) used to quantify method precision against well replication strategies across the QC range. Duplicate clinical sample data were summarized by mean and percent CV, then jointly modeled using a smoothing spline to accurately predict well CV over the range of observed mean concentrations, with specific interest near the LLOQ.

Results: Collectively for these analytes: (1) empirically ~ 1 out of 300 duplicate well CVs from clinical sample data failed the 20% acceptance criterion, in line with what would have been expected from the validation data, (2) predictions from the smoothing spline confirm method precision is as good or better than what was observed during validation, and (3) the expected method CV for singlicate compared to running in duplicate increases by ~1% (absolute) due to the reduced sampling but remains < 10% overall across the range. The assay continues to be fit for its intended purpose and reliable for decision-making.

Conclusion: This analysis supports the use of singlicates for upcoming studies using this assay. It is also our first step to change current practice and implement this approach routinely for ligand binding biomarker assays starting with assay validation. It is expected this approach will result in significant financial savings during sample analysis with reduced use of plates, critical reagents, labor, and time, while still ensuring high-quality, reproducible patient sample data.

References: McCulloch, C., Searle, S., Neuhaus, J., Generalized, Linear, and Mixed Models, (2008), Wiley-Interscience

Acknowledgements: No conflicts of interest


Validation Data


Clinical sample in-study well replicate CVs


Impact on Method CV Across Assay Range: Duplicates vs Singlicate