Development and evaluation of a lateral flow-based portable optical system for determination of the pregnancy status of dairy cows

Our objectives were to develop and evaluate an integrated system consisting of a lateral flow immunoassay (LFIA) and an electronic portable imaging device for determination of pregnancy status of cows based on plasma concentrations of pregnancy-specific protein B (PSPB). Experiment 1 was conducted to test the performance of the LFIA for PSPB (PSPB-LFIA), whereas experiment 2 was conducted to evaluate the performance of the integrated system, including both the LFIA and imaging device. The PSPB-LFIA strips were made of nitrocellulose membrane with polystreptavidin, anti-mouse antibody, europium anti-PSPB conjugates, and biotin PSPB. After adding buffer and plasma in a 96-well plate, strips were dipped to initiate flow and were read in a fluorescence microscope to estimate PSPB concentrations based on the test-to-control line signal (T/C ratio). The T/C ratio of standards was linearly associated with PSPB (R-2 = 0.99 in both experiments) concentrations. To test the ability to identify pregnant cows of the PSPB-LFIA only or the integrated system, plasma samples were collected and transrectal ultrasonography (TUS) was conducted 29 to 35 d after AI in lactating Holstein cows (experiment 1: n = 83; experiment 2: n = 205). A cow was considered pregnant (Preg) if concentrations of PSPB in plasma obtained by ELISA were >= 2 ng/mL or if an embryo was visible by TUS. In experiment 1, the accuracy of the PSPB-LFIA compared with ELISA was 92.7% (91.2% sensitivity [Se]; 96.1% specificity [Sp]; 98.1% positive predictive value [PPV]; 83.3% negative predictive value [NPV]) and compared with TUS was 90.4% (100% Se; 78.9% Sp; 84.9% PPV; 100% NPV). The agreement between LFIA and ELISA (kappa = 0.84; 95% CI 0.71-0.96) or LFIA and TUS (kappa = 0.80; 95% CI 0.67-0.93) as methods to classify cows as Preg or nonpregnant (Non-Preg) was high. In experiment 2, the accuracy of the PSPB-LFIA compared with ELISA was 96.1% (93.8% Se; 100% Sp; 100% PPV; 90.5% NPV) and compared with TUS was 92.2% (99.0% Se; 84.7% Sp; 87.6% PPV; 98.8% NPV). The agreement between LFIA and ELISA (kappa = 0.92; 95% CI 0.86-0.97) or LFIA and TUS (kappa = 0.84; 95% CI 0.77-0.92) as methods to classify cows as Preg or Non-Preg was high. We conclude that a system integrating a fluorescence-based LFIA and an optical reader was effective for classifying cows as pregnant or not pregnant based on estimations of plasma concentrations of PSPB. This novel system serves as a platform for further development of on-farm pregnancy testing tools based on measurement of biomarkers of pregnancy in bodily fluids of cattle.