Bacterial flora and risk of infection of the ovine teat duct and mammary gland throughout lactation V.S. Mavrogianni a , P.J. Cripps b , G.C. Fthenakis a, * a Veterinary Faculty, University of Thessaly, 43100 Karditsa, Greece b Faculty of Veterinary Science, University of Liverpool, Neston, South Wirral CH64 7TE, United Kingdom Received 11 January 2006; received in revised form 19 October 2006; accepted 28 November 2006 Abstract We collected samples of teat duct material and mammary secretion from ewes in three farms (flock A, polyparous n = 7; flock B, polyparous n = 6, primiparous n = 4; flock C, polyparous n = 4): 14 samples immediately after lambing (before sucking of lambs), 244 samples during the suckling period and 156 samples during the milking period. Conventional bacteriological techniques were used. The results were modeled using survival analysis, initially by the Kaplan–Meier method and then by the Cox Proportional Hazards method. Then, we calculated the minimum true risk of an ‘‘at- risk’’ teat or mammary gland being infected and analyzed these data with STATA using the GLLAMM program for Generalised Linear Latent and Mixed Models. During the suckling period, bacteria were isolated from 52 (21%) duct material and 19 (8%) secretion samples; respective results for the milking period were 20 (13%) and 9 (6%). There was an increased risk of duct rather than secretion samples being infected (P < 0.001). There was a significant difference among flocks in isolating bacteria from duct (P < 0.01) or secretion (P < 0.001) samples during suckling period, but not during hand-milking period (P > 0.4 and 0.1, respectively). There were no differences between isolation of bacteria from duct (P > 0.5) or secretion (P > 0.7) samples among primiparous and polyparous animals. Most bacterial isolates were staphylococci. Persistent isolation of the same bacterial species from duct material samples obtained from a particular ewe was recorded with five Staphylococcus spp. and two Mannheimia haemolytica isolates. The results indicate that infections of the teat duct can take place easily; however, not all infections result to infection of the mammary gland. The results support experimental evidence that defence mechanisms of the healthy teat are able to limit the infection. Maintenance of healthy teats contributes to effective defence mechanisms, and www.elsevier.com/locate/prevetmed Preventive Veterinary Medicine 79 (2007) 163–173 * Corresponding author. Tel.: +30 2441 066070; fax: +30 2441 066077. E-mail address: gcf@vet.uth.gr (G.C. Fthenakis). 0167-5877/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.prevetmed.2006.11.014