7938 Validation of Inflammatory Cytokine Multiplex Array With Variations in Target Panel

Abstract Disclosure: W.M. Sanders: None. C.A. Karvonen-Gutierrez: None. D.S. McConnell: None. Background: Multiplex arrays are an invaluable tool for conducting large-scale biomarker research as they provide a high-throughput, cost-effective method for quantifying a panel of endocrine markers. However, the shared reaction environment between antibodies of differing reactivities creates the potential for inducing a technical difference in the assay methods due to any variation in panel selection. Methods: We evaluated the R&D Systems Luminex HS Cytokine Panel A (Catalog # FCSTM09) for systematic differences when measuring a “panel” of IL-1β, TNF-α, and IL-10 vs. measuring “IL-10 only.” We selected 72 serum samples and assayed them in singleton 4 times: 2 separate “panel” batches, and 2 separate “IL-10 only” batches. Differences in log-transformed IL-10 results for sample replicates within and between the “panel” and “IL-10 only” batches were assessed for significance with paired t-tests (α=.01). We also assessed the differences graphically with Bland-Altman plots of differences and absolute percent differences.All batches included control materials from R&D Systems and internal lab serum pools (male and female). To retroactively assess the test batches, we graphically compared their quality control results to those of the subsequent, larger assay project we were piloting. Results: The “IL-10 only” assay yielded an average result of .879 pg/mL (SD .57) on batch 1 and .712 pg/mL (SD .58) on batch 2, with an average paired sample difference of .167 pg/mL, and an average absolute percent difference of 28.8%. The “panel” assay produced an average IL-10 result of .645 pg/mL (SD .50) on batch 1 and .720 pg/mL (SD .52) on batch 2, with an average paired sample difference of .075 pg/mL, and an average absolute percent difference of 15.5%. Comparing pooled IL-10 results from the “IL-10 only” batches with the “panel” batches, there was an average paired difference of .113 pg/mL and absolute percent difference of 16.2%. Paired t-testing of log-transformed results for within “panel,” within “IL-10 only,” and between pooled results from the two methods demonstrated that each had significant differences (p<.0001). The Bland-Altman plots further revealed that these differences were ubiquitously large for samples with low concentrations of IL-10.Test batch quality control results were consistent with those in the primary assay project. The serum pool QCs exhibited significant inter-assay variation – the male pool average result was .41 pg/mL (SD .10) vs. .359 in the test batches, and the female pool average was .57 pg/mL (SD .10) vs. .548 in the test batches. Conclusion: The multiplex assay had considerable inter-assay variation at low concentrations of IL-10, and low results were prevalent. Due to this, we determined that any methodological differences between the “panel” and “IL-10 only” methods were undetectable compared to ambient variation, and irrelevant given the sample characteristics. Presentation: 6/3/2024