Abstract
Antibody responses to Chlamydia trachomatis (CT) antigens may be useful tools for surveillance of trachoma by estimating cumulative prevalence of infection within a population. Data were compared from three different platforms-multiplex bead array (MBA), enzyme-linked immunosorbent assay (ELISA), and lateral flow assay (LFA)-measuring antibody responses against the CT antigen protein plasmid gene product 3 (Pgp3). Sensitivity was defined as the proportion of specimens testing antibody positive from a set of dried blood spots from Tanzanian 1-9-year olds who were positive for CT nucleic acid of all nucleic acid amplification test (NAAT)-positive individuals (N = 103). The sensitivity of the LFA could not be determined because of the use of dried blood spots for this test; this specimen type has yet to be adapted to LFA. Specificity was defined as the proportion of sera from U.S. and Bolivian 1-9-year olds that had previously tested negative by the Chlamydia microimmunofluorescence (MIF) assay testing negative to Pgp3-specific antibodies (N = 154). The sensitivity for MBA and ELISA was the same-93.2 (95% confidence interval [CI]: 88.3-98.1). Specificity ranged across platforms from 96.1 (95% CI: 91.8-98.2) to 99.4% (95% CI: 98.2-100). ELISA performance was similar regardless of whether the plates were precoated or freshly coated with antigen. Sensitivity and specificity of control panels were similar if the cutoff was determined using receiver operator curves or a finite mixture model, but the cutoffs themselves differed by approximately 0.5 OD using the different methodologies. These platforms show good sensitivity and specificity and show good agreement between tests at a population level, but indicate variability for ELISA outcomes depending on the cutoff determination methodology.