Assessment for potential bias in multiplexed IL-10 and TNF-α from plex count

Multiplex arrays offer a high-throughput, cost-effective means for quantifying multiple analytes simultaneously, essential for large-scale biomarker research. However, the shared reactive environment in multiplex assays could lead to variations in results compared to single plex assays, potentially impacting outcomes. This study aimed to explore these differences by examining a "3-plex" cytokine panel vs. single plex assays for tumor necrosis factor-alpha (TNF-α) and interleukin-10 (IL-10).

Using the R&D Systems Luminex HS Cytokine Panel A, 72 serum samples were tested across several batches, including three "3-plex" and two "1-plex" batches each for IL-10 and TNF-α. We evaluated differences in cytokine values using paired t-tests, comparing within 1-plex, between 3-plex, and between 1-plex and 3-plex assays. Bland-Altman plots visually assessed absolute and percentage differences.

IL-10 values were similar between 1-plex and 3-plex assays, showing an average difference of 4.2 fmol/L (10.7 %), which was less than the within plex differences. Conversely, TNF-α showed a 16.7 % difference in the between plex comparison, compared to a 12 % difference in the within plex comparisons. Statistically significant differences emerged mainly for IL-10 across all comparisons. Bland-Altman analyses indicated pronounced variability at low analyte concentrations.

While the multiplex assays demonstrated variation at low analyte levels, especially for IL-10, such differences might not substantially affect comparability when mixed with single plex assays, particularly in datasets dominated by low concentrations.