The objectives of this study were to evaluate the efficacy of therapy of clinical mastitis following intramammary infusion of 2 durations of intramammary antibiotic therapy and to analyze clinical and microbiology data at enrollment or within 4 d of commencement of treatment as potential predictors of subsequent clinical and bacteriological cure. A total of 304 glands were randomly assigned to intramammary treatment with a combination of amoxicillin, clavulanic acid, and prednisolone either 3 times at 12-h intervals (3x12, n = 156) or 5 times at 12-h intervals (5x12 hourly, n = 148). At enrollment (d 0), assessments were made of the cow's demeanor, rectal temperature, teat-end hyperkeratosis, presence of heat or swelling of the gland, presence of clots in the milk, and California Mastitis Test (CMT) score. Cows were revisited on d 4, 14, and 21 and assessed for rectal temperature, the presence of clots in milk, heat or swelling of the gland, and CMT score. Milk samples were collected from enrolled glands for bacteriology (d 0, 4, 14, and 21) and for somatic cell count (d 14 and 21). Data were analyzed using generalized linear models. Treatment group was the key independent variable, but herd was included in the models. Other potentially confounding variables were included where associated at the bivariate level. A second series of analyses were undertaken to predict likelihood of clinical and bacteriological cure using the clinical signs cow age and days in milk, as well as bacteriological results at d 0 and 4. Although we found a higher clinical failure rate in the 3x12 than 5x12 hourly treatment group (28.2 +/- 5.0 vs. 13.4 +/- 3.6%, respectively), we noted no difference in bacteriological cure percentage (73.3 +/- 7.8 vs. 72.0 +/- 7.4%, respectively). The presence of heat or swelling at d 4, days in milk at initiation of treatment, age of the cow, and change in CMT score between d 0 and 4 were predictive of clinical and bacteriological cure. We concluded that increasing the duration of treatment resulted in significantly fewer clinical failures, but had no effect on cure proportion, somatic cell count, or new infection rate. Knowledge of cow age and days in milk and clinical signs were predictive of clinical and bacteriological cure proportion and may provide guidance for producers, under veterinary direction, as to when to extend or recommence therapy.
