William D. Black

Associations among antimicrobial drug treatments and antimicrobial resistance of fecal Escherichia coli of swine on 34 farrow-to-finish farms in Ontario, Canada

Logistic regression was used to model associations between antimicrobial treatment and resistance among fecal Escherichia coli of finisher pigs at the farm level. Four sets of potential risk factors representing different levels of refinement of antimicrobial use on farms were modelled on resistance to antimicrobials. Final models for each antimicrobial were constructed from treatment and management variables significant on initial screening, and corrections for overdispersion were made. In general, in-feed antimicrobial treatment of pigs was more consistently associated with an increased risk of resistance than individual-animal treatment. Antimicrobial treatment in starter rations was significant in final models of resistance to ampicillin, carbadox, nitrofurantoin, sulfisoxizole, and tetracycline. Treatment in grower-finisher rations was significantly associated with resistance to ampicillin, spectinomycin, sulfisoxizole, and tetracycline. There was little evidence that in-feed antimicrobials increased the risk of resistance to gentamicin, which is a drug used only for individual-pig treatment in this study population. These results suggest that antimicrobial medication of rations of post-weaning pigs selects for and maintains antimicrobial resistance among E. coli of finisher pigs. Although resistance was common on farms that did not medicate rations of post-weaning pigs, the results indicate that antimicrobial use does increase the risk of resistance to the antimicrobials studied.

INDIVIDUAL AND GROUP ANTIMICROBIAL USAGE RATES ON 34 FARROW-TO-FINISH SWINE FARMS IN ONTARIO, CANADA

Antimicrobial drug-use was assessed on 34 farrow-to-finish operations that marketed at least 500 hogs/yr. These operations either did not use any antimicrobials or used narrow-spectrum or broad-spectrum antimicrobials in rations of post-weaning pigs. Total antimicrobial use was measured for two months after obtaining inventories and records of all antimicrobials used. The collection of empty medication bottles and inventories of drugs on hand was convenient for producers and useful for estimating or validating recorded treatment rates, particularly for antimicrobials that were used only in one class of pig. Treatment records, however, underestimated by approximately 35% the amounts used for 27/29 farm-antimicrobial combinations. Rates of individual-pig treatment varied from 0-24.1 pigs treated/1000 pig-days, with a median of 5.29. Most individual animal treatments were given to piglets and sows at parturition and penicillin was the most commonly used antimicrobial. Gentamicin was administered to suckling piglets on 19 of the farms.

A dairy farm survey of antibiotic treatment practices, residue control methods and associations with inhibitors in milk

A mail survey was conducted of dairy producers who had received a positive bulk milk antibiotic residue test result in a two-year period (1987-88) of government monitoring (case farms) and farms that were negative for all tests conducted in the same period (control farms). Farmers were asked to complete questionnaires designed to determine dairy management practices, as well as, antibiotic handling and residue prevention methods. Using multiple logistic regression analysis, and adjusting for the size of the milking herd, the following factors were associated with increased risk of antibiotic residues in milk: the use of part-time assistance in milking, use of a milking parlor and increased estimated frequency of intramammary antibiotic treatments. Unconditionally, significantly more control farmers used separate equipment to milk treated cows rather than simply attempting to divert milk from the bulk tank. Controls were also more likely to vary the withholding time of milk for different drugs. Other significant differences between cases and controls with respect to residue prevention methods were observed, however, some of these may have been due to changes instituted on case farms after the antibiotic residue violations occurred. For example, significantly more case than control farmers reported using on-farm residue test kits and marking of treated animals as residue prevention methods and more case farmers believed that failure to keep good records of treatment was an important factor in residue occurrence. No significant differences were observed in the proportions of case and control farms that used medicated feed, in the number of people employed on the farm, or in the general knowledge of antibiotic residue prevention.

Prevalences of resistance to seven antimicrobials among fecal Escherichia coli of swine on thirty-four farrow-to-finish farms in Ontario, Canada.

Fecal specimens were composited and a hydrophobic-grid membrane-filter method was used to measure antimicrobial resistance to ampicillin 16 micrograms/ml, carbadox 30 micrograms/ml, gentamicin 4 mu/ml, nitrofurantoin 32 micrograms/ml, spectinomycin 16 micrograms/ml, sulfisoxazole 32 micrograms/ml and tetracycline 8 micrograms/ml among 8119 Escherichia coli isolates from 68 fecal samples collected on 34 farrow-to-finish swine farms marketing over 500 hogs/yr. The overall prevalences of resistance to antimicrobials among these isolates were: ampicillin 29%, carbadox 3.5%, gentamicin 0.6%, nitrofurantoin 27%, spectinomycin 28%, sulfasoxizole 38% and tetracycline 71%. Thirty to seventy-six per cent of the variations in prevalences were explained by between-farm differences.

Quinolone arthropathy in immature rabbits treated with the fluoroquinolone, PD 117596

To study the potential of the fluorquinolone, PD 117596 to cause arthropathy in experimental animals, immature rabbits were orally administered the drug for five days at 0, 100, 350, 500 and 750 mg/kg. Characterization of changes induced in major synovial joints was based on: macroscopic and histopathologic observations, transmission electron microscopic examinations and magnetic resonance imaging. Preferentially targeting the knee, PD 117596 produced vesicles and erosions in articular cartilage which resembled, morphologically, those described in other laboratory species. Lesion incidence was not clearly dose-related. In the perivesicular region, degenerate chondrocytes were intermixed with hypertrophic cartilage cells and chondrocyte clusters. Ultrastructurally, hypertrophic chondrocytes were the consequence of karyomegaly and RER proliferation. Matrix density was reduced due to collagen and proteoglycan loss. Joint structures were readily visualized by magnetic resonance imaging which identified thickened articular cartilage, surface irregularities consistent with ruptured vesicles and separation of opposing articular surfaces secondary to synovival effusions. The immature rabbit, although less sensitive than the juvenile dog to the arthropathic effects of quinolones, was nonetheless a good model to study this experimental osteoarticular disease.

Accumulation of dietary cadmium (Cd) in rabbit tissues and excretions: a comparison of lettuce amended with soluble Cd salt and lettuce with plant-incorporated Cd.

Quantifying the transfer of Cd from foods to mammalian target organs is key to estimating the health risk from this exposure; however, the bioaccumulation of Cd from foods is modified by many dietary components. Studies of dietary Cd absorption would be simpler if it were known that Cd added to foods as a soluble salt was as bioavailable as Cd incorporated during growth of the food species. Rabbits were fed, for 16 d, fresh lettuce containing cadmium incorporated into the lettuce during growth or added to the lettuce before feeding, or lettuce with no Cd but soluble Cd administered to the animals by gavage. There was a marked positive relationship between increased Cd dose and its accumulation in kidney; the slopes for the gavage and added treatments were not clearly different from the incorporated treatment; liver data were highly variable. In a 10-wk study of Cd-incorporated and -amended lettuce diets, for the incorporated and control diets there was less Cd accumulation in the kidneys, but not liver, per unit cumulative dose, than for the amended diet. Cd accumulation in the small intestine and Cd concentration in feces, both per unit daily dose, were smaller for the incorporated than for the control and amended diets; Cd concentrations in bile, urine, and serum, per unit daily dose, were higher in the control diet than values in the amended diet, which were higher than the incorporated diet. These differences could not be accounted for by variation in Fe or Zn contents of the diets. Thus, data suggest that Cd-amended diets overestimate bioaccumulation in kidney, an important target organ, by up to one-third, and that studies of short duration are not adequate to evaluate Cd bioavailability from food.

Cadmium bioavailability and bioaccessibility as determined by in vitro digestion, dialysis and intestinal epithelial monolayers, and compared to in vivo data

The objective of the study was to compare in vivo estimates of Cd bioavailability in two diet materials (lettuce and durum wheat grain) with bioaccessibility estimates from three in vitro methods. For both dietary materials, the Cd was either incorporated during growth or applied topically as a soluble salt just prior to experimentation. Simulated gastric/intestinal digestion using a physiologically based extraction technique (PBET) solubilized less than 56% (lettuce) or 13% (grain) of the Cd that was either incorporated into the plant tissues during growth, or added to the plant tissues before experimentation, as Cd(NO3)2 · H2O. Amended diets could not be distinguished from incorporated diets. More of the Cd solubilized from amended lettuce than from incorporated lettuce moved to the outside of MWCO 10 kD and 25 kD dialysis sacs; no difference between the amended and incorporated diets was observed for grain. The percentage of lettuce-Cd solubilized by the PBET and sorbed by Caco-2 cells was greater for incorporated than for amended lettuce; for Cd in grain, the reverse occurred. As expected, none of the in vitro estimates of bioaccessibility were the same percentage of Cd in the lettuce or grain as was measured as bioavailable in vivo. The in vitro assays all predicted that substantially less than 100% of the Cd in the foods would be bioavailable, as was identified in vivo, and simulating intestinal selectivity improved the comparison to in vivo. Some of the in vitro assays identified subtle differences between the diets (i.e., amended vs. incorporated) that were consistent with in vivo studies, and with speculated differences in Cd speciation; this suggests their potential usefulness for the study of modifiers to dietary Cd bioavailability.

Detecting oxytetracycline residues in chicken tissues using the Delvotest® P system

Oxytetracycline spiked chicken liver and kidney were tested for residues using Delvotest P and the results compared to standard plate assay using Bacillus cereus as test organism. Delvotest P detected all replicates at concentrations of ≥0.62 μg/g in liver and kidney. Between 0.41-0.2 μg/g, the responses were mixed (positive, doubtful, and negative); at 0.1 ≥ concentration, the responses were negative. The limit of detection of the plate assay was 0.2 and 0.31 μg/g for liver and kidney, respectively. Residue analysis on tissues from chickens dosed twice daily (for 4 d) with 25 mg/kg oxytetracycline was carried out using the Delvotest P and the plate assay. Delvotest P responses were all positive for 1-4 h after treatment in kidney and 1 and 2 h in liver and serum. Muscle never achieved concentrations that were consistently positive. The plate assay detected drug from 0.5-6 h in serum, from 0.5-8 h in liver and kidney, and at 4-6 h in muscle. Kidney tissue appeared to be the best for detecting oxytetracycline residues in chicken. The Delvotest P method was simpler to use and required less time; 3-4 h compared to about 18 h for the plate assay.

On the transformation technique in pharmacokinetic curve fitting

It is shown that when one nonlinear regression model is a reparametrization of a second model, the parameter estimates, and their standard errors, for one model can be obtained directly from those obtained from fitting the other model.

Detecting sulfamethazine residues in chicken tissues using the Delvotest® SP system

Sulfamethazine-spiked chicken liver and kidney were tested for residues using the Delvotest SP. The results were compared to a standard plate assay using Bacillus subtilis as the test organism. The Delvotest SP gave positive responses to all homogenized liver replicates spiked with sulfamethazine at or above 1.0 μg/g of liver. Mixed responses were obtained at 0.5 μg/g and negative responses at ≤0.25 μg/g. The plate assay had a minimum sulfamethazine detection limit of 1 μg/g and 0.5 μg/g of liver and kidney, respectively. Chickens were dosed with sulfamethazine (100 mg/kg of body weight) daily for 5 days, and tissues were tested for residues after treatment stopped. The Delvotest SP was positive for all serum and kidney samples from the end of treatment to 24 h, and for muscle and liver samples up to 8 h. The plate assay detected the drug up to 24 h in serum and kidney samples and up to 8 h in liver and muscle samples. Kidney and serum samples both appeared to be good tissues for testing sulfamethazine residues in chickens. Serum could be used for antemortem screening, whereas kidney samples would appear to be the best for postmortem screening of residues in chickens.