Distribution of Trichinella spiralis in the diaphragm of experimentally infected swine

The objectives of this study were to determine whether the distribution of Trichinella spiralis larvae within the diaphragm of infected swine was the same throughout five anatomically identifiable portions, and to assess the tenability of the Poisson model to describe the probability of detecting larvae in samples of digested tissue. To examine differences in levels of infection among the regions of the diaphragm, sixteen 2-month-old and fourteen 6-month-old swine were infected with approximately 300 T. spiralis larvae per animal. Since the infective dose was not adjusted for age (weight), higher infection rates resulted in the younger hogs. Because of this higher infection rate, the occurrence of positive samples judged to be free of larvae (false negatives), was lower for the younger hogs (0.9% vs 14.6%). Pooling over both age groups and across all five anatomical regions yielded the directional difference in false negative rates one would expect between 1 g and 5 g samples (viz., 12.3% for 1 g vs 2.6% for 5 g). In both age groups, levels of infection were significantly different among the five anatomical regions of the diaphragm (p=.06 at 2 mo; p=.0002 at 6 mo). The Crus muscle consistently yielded more larvae/g than the lumbocostal arch, the dorsal and ventral Pars costalis, and the Pars sternalis. Accordingly, the false negative rates were found to be lowest in the Crus muscle (5 of 59 or 8.5% in 1 g samples and 0 of 58 in the 5 g samples). We conclude that sampling from Crus muscle increases the ability to detect T. spiralis larvae in the diaphragm of infected pork carcasses. Moreover, we demonstrate that the Poisson model (random or "homogenous distribution" of larvae) grossly underestimates the observed false negative rates while predictions based on the negative binomial model provide close agreement.