Pooled-sample testing for detection of Mycoplasma hyopneumoniae during late experimental infection as a diagnostic tool for a herd eradication program

Abstract

Early and accurate detection of Mycoplasma hyopneumoniae infection in live pigs is a critical component to measure the success of disease eradication strategies. However, the imperfect sensitivity of in vivo diagnostic tools, change in sensitivity over the course of infection, and expected low prevalence level at the end of an eradication program create a challenging diagnostic scenario. Here, the individual and pool sensitivities for detection of M. hyopneumoniae during the chronic phase of infection was determined using deep tracheal catheter samples, the in vivo sample type with the highest reported diagnostic sensitivity. Fifty samples from known infected pigs collected at 113 days post-M. hyopneumoniae intra-tracheal inoculation, were diluted in known negative samples to form pools of 1:3 and 1:5. Samples were tested for M. hyopneumoniae by a species-specific PCR. Ninety-eight percent (49/50) of individual samples, 84 % (42/50) of pools of 1:3, and 82 % (41/50) of 1:5 were detected positive for M. hyopneumoniae. To apply the sensitivity estimates for detection of M. hyopneumoniae in a low prevalence scenario, sample sizes with associated sample collection costs were calculated for individual and pooled testing using algorithms within the program EpiTools One-Stage Freedom Analyses. Assumptions included a ≥95 % population sensitivity, infinite population size, prevalence levels of ≥0.5 %, ≥1 %, ≥2 %, ≥3 %, ≥4 %, or ≥5 %, 100 % specificity, along with the mean and lower confidence limit of the individual or pool sensitivity for each pool size, when appropriate. For instance, following completion of a herd eradication program, if a low risk approach is targeted, sample size estimates for ≥2 % prevalence using the lower limit of the diagnostic or pool sensitivity 95 %CI may be followed. If samples were to be tested individually, 167 individuals would be sampled at a cost of 6,012 USD. If pooled by 3, 213 would be sampled (testing cost 3,266 USD), and for pools of 5, 220 individuals would be sampled (testing cost 2,464 USD). Population sensitivity was also calculated for a range of testing scenarios. Our study indicated that pooling samples by 3 or 5 was a cost-effective method for M. hyopneumoniae detection in low prevalence scenarios. Cost-effective detection was evidenced despite the increased sample collection costs associated with large sample sizes in order to offset decreased testing sensitivity attributable to pooling. The post-eradication sample collection scheme, combined with pooling, suggested lower cost options than individual sampling for testing to be applied at the end of an eradication program, without significantly compromising the likelihood of detection.