Huimei Bi, Lei Zou, Jingyao Liu, Qingfeng Gao, Ci Wang, Liping Hou, Gang Wang, Jinxia Gao
FPA is a highly accurate and independent diagnostic marker for brucellosis and supports its application in large-scale screening programs. However, the subgroup ROC has truncation bias, and the lipid interference inference remains unvalidated.
OBJECTIVE: To evaluate the diagnostic efficacy of the fluorescence polarization assay (FPA) for detection of brucellosis.
METHODS: A total of 1,080 clinically confirmed brucellosis patients and 1,518 nonbrucellosis controls were enrolled. Participants were classified based on clinical manifestations and laboratory diagnoses. Serum samples were analyzed using the Rose Bengal plate agglutination test (RBPT), serum agglutination test (SAT), FPA, and enzyme-linked immunosorbent assay (ELISA), followed by comparative analysis. The overall diagnostic sensitivity of the two-stage screening algorithm across the full cohort was further calculated.
RESULTS: Multivariate logistic regression analysis identified FPA as an independent diagnostic predictor for brucellosis (Odds Ratio (OR) = 6.24, 95% Confidence Interval (CI): 5.18-7.51, p < 0.001). The area under the ROC curve (AUC) for FPA was 0.983 (95% CI: 0.978-0.987). At an optimal cutoff value of 81.5 millipolarization units (mP), FPA achieved a sensitivity of 93.80% and a specificity of 93.61%. Agreement analysis demonstrated high to perfect concordance between FPA and SAT (Kappa = 0.788), RBPT (Kappa = 0.875), and ELISA (Kappa = 0.752) (all p < 0.001). Using the 95th percentile of FPA values (P95 = 83 mP) as the mass screening warning threshold, samples were classified as negative (≤83 mP) or suspected positive (>83 mP). Among 976 suspected positive cases in the brucellosis group and 68 in the control group, secondary receiver operating characteristic (ROC) analysis yielded an AUC of 0.898 (95% CI: 0.861-0.936). The optimal cutoff value was 92.5 mP, with a sensitivity of 82.79% and a specificity of 94.12%. FPA values between 83 mP and 92.5 mP were defined as an equivocal zone; the total sensitivity of this two-step screening strategy was only 74.8%.
CONCLUSION: FPA is a highly accurate and independent diagnostic marker for brucellosis and supports its application in large-scale screening programs. However, the subgroup ROC has truncation bias, and the lipid interference inference remains unvalidated.