A case control study of factors and infections associated with clinically apparent respiratory disease in UK Thoroughbred racehorses

A matched case control study was used to determine infections and other factors associated with clinically apparent respiratory disease in young racehorses in training in the UK. A total of 170 cases, defined as horses with sudden onset coughing, nasal discharge or pyrexia, were identified and matched to 632 non-affected controls by trainer and time period. Factors examined included age, sex, time since entry into the training yard, time since last race and different infections including tracheal and nasopharyngeal (NP) bacteria and viruses. Multivariable conditional logistic regression (CLR) modelling was used to evaluate the risk of being a case for variables after adjustment for other factors. Three analyses were conducted using clinical cases as outcomes, which were compared with: (i) controls without evidence of subclinical inflammatory airway disease (IAD) (ii) controls with evidence of subclinical IAD and (iii) all controls irrespective of IAD status. A fourth analysis was conducted comparing the two groups of controls, i.e. those with and without IAD. Younger horses and those that had entered training more recently were at increased risk of suffering episodes of clinically apparent respiratory disease. Among the infections, increasing numbers of Pasteurella/Actinobacillus spp. in tracheal washes were associated with increasing risk of clinical disease. Tracheal infection with Streptococcus zooepidemicus was associated with both clinical respiratory disease and subclinical IAD when compared with controls with no evidence of IAD. This explained the lack of association between clinical cases and S. zooepidemicus when all controls were used. Tracheal isolation of Mycoplasma felis was also associated with clinical disease after controlling for other factors. An inverse association was identified between risk of clinically apparent disease and isolation from tracheal washes of the transient, non-pathogenic bacteria Staphylococcus and Acinetobacter spp. There was no significant association identified between clinical disease and infection with equine herpesviruses-1 and -4, rhinoviruses-1 and -2 or adenovirus. Equine influenza was significantly associated with clinical respiratory disease but it was a very rare infection in this well-vaccinated population, only occurring in three cases. (C) 2003 Elsevier Science B.V. All rights reserved.