Comparison between in-house competitive and commercial ELISA for the detection of heat shock protein 70 in milk, blood, and saliva of dairy cows

Heat shock protein 70 (HSP70) has emerged as a potential physiological indicator of heat stress (HS), recognized for its substantial upregulation in response to increased temperatures and cellular stress. The accurate measurement of HSP70 is crucial for its application in research and on-farm monitoring. Commercial ELISA kits are widely available and exhibit high sensitivity; however, their limited working ranges require the dilution of high-concentration samples, which can potentially introduce variability. A recently developed in-house competitive ELISA provides a cost-effective alternative, previously validated for bovine HSP70 detection; however, its performance across various biological matrices has not been systematically evaluated against commercially available systems. A secondary objective was to determine whether HSP70 levels measured in less-invasive matrices such as milk and saliva, correspond to those obtained from blood. This study compared an in-house competitive ELISA to a commercial sandwich ELISA kit for detecting HSP70 in milk, blood, and saliva samples of dairy cows. Samples were analyzed with optimized methods, and assay performance was evaluated using descriptive statistics, Pearson correlations, and linear regression models. As expected for assays using different formats and calibration systems, the in-house ELISA indicated elevated absolute concentrations across all matrices; however, these differences were systematic and did not affect the strong agreement between methods. Blood samples had the lowest CV in both assays (2.76% in-house; 5.89% commercial), indicating greater repeatability than milk (5.69% and 10.24%) and saliva (7.5% and 9.95%). The consistently lower CV for the in-house method across all fluids indicates that it is more precise and can lead to higher statistical power than commercial assays for a given sample size. Strong positive correlations were observed between assays across all matrices (milk r = 0.91; blood r = 0.88; saliva r = 0.87). Regression analysis validated these relationships, with R² values of 0.777, 0.833, and 0.752 for milk, blood, and saliva, respectively. The findings indicate that the in-house ELISA closely aligns with the performance of the commercial kit. Despite the lower detectable minimum concentration of the commercial assay, the in-house assay provides a more economical alternative for evaluating physiologically high and moderate HSP70 levels; however, the commercial kit remains widely available for purchase. Among biological matrices, plasma exhibited the highest cross-assay correlation, whereas milk and saliva showed consistent relative trends, indicating their potential as noninvasive alternatives for HS monitoring in dairy cows.