Ebola vaccine development relies on antibody measurements to assess immunogenicity, especially in non-epidemic settings. While no definitive correlate of protection exists, anti-EBOV GP1,2 IgG antibody (Ab) levels are considered a key primary endpoint. The Filovirus Animal Nonclinical Group (FANG) ELISA assay has been widely used for measuring these Abs. However, low throughput, high cost, shortage in reagents, and high inter-laboratory variability limit its practicality. Thus, a switch from the FANG assay to a high-throughput alternative with a different quantification technique, the Luminex assay, was implemented during long-term follow-up in a large vaccine trial (PREVAC, NCT02876328). We aimed to establish a statistical conversion between the two assays and assess its impact on longitudinal anti-EBOV GP1,2 IgG analyses within the PREVAC trial.

PREVAC is a phase 2b, randomized, double-blind, placebo-controlled trial evaluating 3 different Ebola vaccine regimens in the general population in West Africa. Anti-EBOV GP1,2 IgG Abs were measured in two central laboratories from baseline (D0) to Month (M)12 (primary endpoint) with the FANG assay, expressed in EU/mL (ELISA Units). From M24 to M60, Abs were measured annually with the Luminex assay, expressed as net Median Fluorescence Intensity (MFI), in one central laboratory. To establish a conversion between the two assays, 2,467 samples which were previously tested with the FANG ELISA spanning the range of Ab levels during the first year of follow-up, were re-tested with the Luminex assay. The conversion process involved three key steps: (i) MFI to IU/mL conversion: Luminex Ab levels (net MFI) were calibrated with a WHO reference standard, which has known concentrations in IU/mL (International Units). A log-linear conversion formula from net MFI to IU/mL was derived. (ii) IU/mL to EU/mL conversion: Luminex Ab levels in IU/mL were converted to EU/mL with a previously established conversion factor (1 IU/mL = 27,135.90 EU/mL). (iii) Bridging between FANG and Luminex: Deming regression models were used to estimate conversion formulas between FANG (EU/mL) and Luminex (converted to EU/mL). Separate models were developed for each FANG laboratory to account for inter-laboratory variability. Ab levels < 100 EU/mL were excluded due to a non-linear association at lower levels, resulting in a censoring threshold. Results of the conversion were assessed using Bland-Altman plots. All analyses were done with SAS software (version 9.4).

A total of 12,361 samples were tested in duplicate with Luminex, including M24-M60 samples and 2,467 re-tested D0-M12 samples. After steps i) and ii), the Ab dynamics from D0 to M12 measured by Luminex closely mirrored those previously obtained with FANG. However, on average, D0-M12 Luminex Ab levels were 2.5 times higher than those obtained with FANG. Deming regression models to bridge the FANG to the Luminex range (step iii) resulted in the following formulas: For FANG Laboratory 1: Log10(FANGLuminex) = -1.123022 + 1.4756017 x Log10(FANGFang) For FANG Laboratory 2: Log10(FANGLuminex) = 0.1002727 + 1.0755272 x Log10(FANGFang)

Switching the Ab assay technique during a clinical study requires complex statistical adjustments. In the PREVAC trial, we established conversion formulas to analyze long-term Ab endpoints, addressing differences in assay characteristics and ensuring continuity in data interpretation.