175 Evaluation of Electrical Impedance Spectroscopy to Determine Swine Herd Reproductive Status Using the Accuracy of Ultrasound and Hormonal Assays as Diagnostic Tools

Abstract

Abstract Management of fertility determines production efficiency. Important fertility measures include detection of estrus (DE), early pregnancy confirmation (EPC), and day of farrowing (DF) with few tools available to diagnose or predict. Electrical impedance spectroscopy (EIS) uses an array of frequencies to determine tissue resistance in relation to change. Our objective was to determine if EIS can predict reproductive events. The study was conducted for DE using weaned sows (n = 135) scanned on D 1-7; EPC used inseminated sows (n = 135) scanned on D 18-20; DF used pregnant sows on D 113 to farrowing. The EIS device used a four-electrode transducer with an internal processor and was connected to a mobile device for operation, reading, and data storage. The device obtained impedance (Ω) and phase data for analysis from 42 frequencies between 1,000 and 29,000 Hz. The device was disinfected and inserted into the vagina with scans requiring ~20 s. Data were uploaded to a server and stored on a website for labeling and analysis. Accuracy was based on DE, EPC on D 30, and DF. Data were analyzed in RStudio using GLM with logistic regression to generate log-likelihood estimates for r2 and associated P values. Predictive models for DE, EPC, and DF included EIS, day of measure, and parity. For DE, prediction was greatest 1 d before and on estrus (r2 = 0.98), moderate 3 to 4 d before estrus (r2 = 0.59), and not predictive ≥4 d before estrus. For EPC, the measures for EIS were not predictive due to limited numbers of non-pregnant sows (3%). For DF, prediction for 1 d before was (r2 = 0.98), for 2 d (r2 = 0.89), and for 3 d (r2 = 0.51). Our preliminary results suggest impedance of the sow vagina can be used to predict events, but more data for failures and days will be needed to improve prediction.