Survival analysis of two Mycoplasma hyopneumoniae eradication methods

Mycoplasma hyopneumoniae is an important respiratory pathogen causing significant losses in the swine industry. Eradication of this bacterium from herds results in increased pig performance, productivity, and animal welfare. The objective of this study was to compare the time-to-detection of M. hyopneumoniae in breed-to-wean farms after the application of one of two methods for M. hyopneumoniae eradication. The two methods compared in this study were: 1) Herd closure and medication, and 2) Whole herd medication without extended closure. Fifty-six breed-to-wean farms located in the US Midwest constituted the cohort for this investigation. Herd closure and medication was applied in 45 farms, while whole herd medication was applied in 11 farms. Two mutually exclusive events were recorded for each farm, either detection of M. hyopneumoniae, which was considered the event of interest, or end of follow-up, which was the right-censored event. Farms were monitored until recording the event of interest, or until the end of follow-up, whichever occurred first. Detection of M. hyopneumoniae was assessed by identification of antibodies against the bacterium in sentinel pigs using a commercially available ELISA assay within 6 months post-eradication completion. Moreover, clinical presentation of disease was recorded if observed post-eradication completion. The censored event occurred at the end of the study in November 2016 (administrative censoring). Time-to-detection of M. hyopneumoniae was analyzed with a Cox proportional hazards model. The proportional hazards assumption was assessed using graphical methods. A sensitivity analysis to evaluate the assumption of outcome-independent censoring was also performed. The cumulative incidence of M. hyopneumoniae detection at the end of follow-up was 18.6 % (95% CI: 6.5%, 46.8%) for herd closure and medication, and 36.4% (95% CI: 15.5%, 70.3%) for whole herd medication. An interaction term between the type of eradication method and follow-up time was included in the model to account for the non-proportional hazards. An overall effect of eradication method was present (P = 0.0442). The hazard ratio associated to the time-invariant effect of eradication method was 29.2 (95% CI: 0.95, 894; P = 0.053). The hazard ratio associated with the interaction term was 0.88 (95% CI: 0.65, 1.2; P = 0.405). Under these conditions, eradication using herd closure and medication reduced the likelihood of detecting cases of M. hyopneumoniae in breed-to-wean farms compared to whole herd medication. Detection of M. hyopneumoniae was concentrated during the first 64 months of follow-up in herd closure and medication, and in the first 8 months in whole herd medication.