Evaluation of Herbal Remedies as Alternatives to Antibiotic Therapy in Dairy Cattle

1.) Four commercial dairy farms and one research herd were visited in late 2010-early 2011 for collection of the initial set of milk samples at dry off using aseptic techniques (National Mastitis Council, 2004). During the visit, farm owners and employees in the milk parlor were trained how to collect milk samples aseptically. Briefly, the teats and udder were cleaned of dirt and debris, teats were premilked to remove foremilk and stimulate milk letdown, teats were rinsed with a disinfectant (30 seconds) and dried with individual towels, teat ends were scrubbed (15 seconds) with 70% alcohol until the cotton ball was clean after contacting the teat, a milk sample was taken while being careful not to touch the teat end to anything in order to maintain sterility, and samples were stored immediately on ice or in refrigeration. These samples were used to assess bacterial content of the milk. An additional set of samples from each teat were taken to be shipped to the Dairy Herd Information Association (DHIA) laboratory to assess somatic cell count. 1a.) Each dairy was provided with a milk collection kit including sterile milk collection vials, 70% alcohol, cotton balls, freezer packs, racks for holding tubes, paper towels and permanent markers. Farmers were encouraged to pre-label vials before milk collection for ease of labeling and sample handling. 2a.) Four different treatments were assessed on the four commercial herds enrolled in the study: no dry off treatment (NOT), Phyto-Mast (PHY), Cinnatube (CIN), Phyto-Mast and Cinnatube (PC). The Center for Environmental Farming Systems (CEFS) herd in Goldsboro, NC had five treatment groups, with one group for conventional treatment with an antibiotic and an internal teat sealant (penicillin-dihydrostreptomycin plus bismuth subnitrate; CON). At CEFS, cows were pasture-based and split into two management groups, one of which received organic health treatments and the other received conventional treatments. All cows received a conventional concentrate supplement but were primarily fed on pasture. Every herd was divided into treatment groups balanced by breed, calving date, and lactation number. Four hundred and fifty cows were eligible for the experiment, which would give over 80% power to detect differences between all treatments for cure rate, new infection rate, and somatic cell score difference according to preliminary data collected in 2009-2010 at CEFS. 2b.) Duplicate milk samples were taken from all four quarters of each cow at CEFS at dry off immediately preceding administration of treatment and 3 to 5 days after calving (at freshening) using the aseptic techniques recommended by the National Mastitis Council (National Mastitis Council, 2004). A single set of milk samples were taken from cows at the commercial herds at dry off and again 3 to 5 days after calving. Samples were processed at the Milk Quality and Mastitis Laboratory at the College of Veterinary Medicine at North Carolina State University. If one quarter sample was contaminated and a duplicate was available, the duplicate was also cultured. The protocol for milk culture was, briefly: vortexed milk samples were plated (0.01 ml) on trypticase soy agar + 5% sheep blood plates. Any unusual appearance or smell of the milk sample was recorded. Plates were incubated at 36°C for 48 hours and observed for growth at 24 hours and 48 hours. Number of colonies and color of colonies were recorded. Individual colonies were selected for identification and Gram stained for initial screening. Streptococci were distinguished from enterococci using the CAMP test, esculin hydrolysis and growth on bile esculin agar with azide (Hardy Diagnostics, Santa Maria, CA). Coagulase-negative staphylococci were distinguished from Staphylococcus aureus by mannitol fermentation and coagulase testing. Gram negative rods were identified using morphology on MacConkey agar (Hardy Diagnostics, Santa Maria, CA) and oxidase testing. If required, the API20Strep and the API 20E identification systems (bioMérieux, Durham, NC), were used for further differentiation. 2c.) The research herd at CEFS was the only herd on DHIA test for the duration of this study, making it impossible to obtain milk production records for individual cows from the commercial herds. However, somatic cell counts (SCC) were obtained for the majority of quarters of every cow sampled. Statistical analysis was performed at the quarter level using SAS (version 9.2, Cary, NC). Prior to model selection, diagnostic tests were run to test for normality and outliers. Following this examination, all SCC variables were transformed to linear somatic cell score (SCS) to more closely approximate the Normal distribution; this transformation was performed using the formula to obtain the base 2 logarithmic transformation as recommended by Shook (1982). The generalized linear mixed model procedure (PROC GLIMMIX) was used, incorporating farm and cow as random effects. Model selection involved univariate analysis on all measured variables to qualify for insertion into the model (P < 0.20), then addition of main-effects interactions and finally backward stepwise selection, eliminating non-significant variables (P > 0.05) while keeping treatment forced into the model.