The Effects on Human Health of Subtherapeutic Use of Antimicrobials in Animal Feeds (1980)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

APPENDIX G TRANSMIS S ION OF FOOD-BORNE D ISEASE S IMPLICATIONS OF THE SUBT~RAPEUTIC USE OF ANTIMICROBIALS , Jackson S. Risers The major food-borne bacterial diseases in the United States are caused by salmonellae, staphylococci, and Clostridium botulinum (Center for Disease Control, 1977b) (Table 1~. The number of out- breaks of food poisoning by salmonellae and staphylococci is with- out doubt greatly underestimated because many outbreaks are unre- ported and undiagnosed. Because of its acute onset and rapid course, staphylococcal food poisoning, in particular, is not often brought to the attention of any medical authority. The statistics of C. botulinum food poisoning, because it is a much more serious disease in terms of its life-threatening poten- tial, are much more accurate. However, since there is no known or suspected connection between _ botulinum food poisoning and the use of antibiotics at subtherapeutic levels in animal feeds, it will not be discussed further in this paper. Salmonellosis is ubiquitous. It occurs usually as entero- colitis, usually of only a few days duration, and it is frequently caused by food-borne salmonellae. This form of salmonellosis is acutely uncomfortable and may cause humans to be absent from work for one or several days. The disease frequently involves a visit to a physician and, occasionally, hospitalization for a short period. It is generally agreed that antibiotic treatment is con- traindicated since it does not shorten the course of the disease or lessen the severity of the symptoms (Hook and Johnson, 1972~. Moreover, such treatment is likely to prolong the fecal excretion of the organisms (Aserkoff and Bennett, 1969~. SALMONELLOSIS IN FOOD ANIMALS AND HUMANS Salmonellae infect both warm-blooded and cold-blooded ani- mals. In 1977, 5,243 Salmonella isolates from nonhuman sources were reported to the U.S. Department of Agriculture (USDA) and the Center for Disease Control (CDC). Of the total nonhuman iso- lates, 28% (1,474) were obtained from chickens and turkeys, 25% (1,313) were of porcine origin, and 14X (713) were of bovine origin (CDC, 1979~. Five of the 10 serotypes most frequently isolated Consultant, Agricultural Division of American Cyanamid Co., Princeton, N.J. 203

204 TABLE 1 Confirmed Outbreaks of Food-Borne Disease, by Etiology, 1976 Outbreaks t Cases Organism Number ~ Number %b Salmonella 28 21.2 1,169 32.7 Staphylococcus 26 19. 7 930 26.0 Clostridium botulinum 23 17.440 1.1 Clostridium perfringens 6 4.5509 14.2 Shigella 6 4.5273 7.6 Bacillus cereus 2 1. 563 1.8 Yersinia enterocolitica 1 0.8286 8.0 TOTAL 92 69.63,270 91.4 From Center for Disease Control, 1977b. Percent of all (bacterial, chemical, parasitic, and viral) food-borne disease outbreaks in the United States.

205 from animals (Table 2) were also among the 10 serotypes most f requently isolated from humans (Table 3 ~ . S. typhimurium, in- cluding var. Copenhagen, comprised 35. 5% of the Herman isolates reported and 19. 7% of the isolates from nonhuman sources. It is by far the most frequently isolated serotype. Although the greatest number of isolates from nonhuman sources was obtained from swine, the inc idence of Salmonella In healthy market-ready swine varies greatly trom place to place. In 1975, a survey was conducted at a slaughterhouse in each of three hog-producing areas (Gustafson et al., 1976~. The purpose of the survey was to determine the incidence of Salmonella in healthy swine at market and to see what percent of them might be harboring salmonellae with resistance to more than one antibiotic. Of 151 hogs sampled in Pennsylvania, 54 (35.77) had salmonellae and none had multiply resistant salmonellae. Of 251 hogs sampled in Iowa, 26 (10.3%) had salmonellae and one had multiply resistant salmonellae. Of 256 hogs sampled in Georgia, 215 had salmonellae (83.9%) and 9 (3.5%) had multiply resistant salmonellae. These data suggest that only a few hogs harbor multiply antibiotic- resistant salmonellae. THE EFFECT OF SUBTHERAPEUTIC TETRACYCLINES ON SALMONELLAE IN _ DOMESTIC ANIMALS In April 1973, the Food and Drug Administration (FDA) issued a statement requiring the manufacturers of tetracyclines and cer- tain other drugs to show, by April 20, 1974, that the use of these drugs in animal feeds did not increase the Salmonella reservoir in animals and poultry raised for meat (FDA, 1973~. low. The FDA Bureau of Veterinary Medicine (BVM) then issued a set of human health safety criteria for the drug manufacturers to meet in order to demonstrate that the Salmonella reservoir in meat ani- mals was not increased by the subtherapeutic levels of the drugs in feed. BVM and the manufacturers designed protocols for experiments to determine the effect of the drug on the quantity, prevalence, and duration of shedding and the resistance characteristics of the Salmonella in animals consuming subEherapeutic levels of antibiotics in their feed. By April 20, 1974, these experiments had been con- ducted and the manufacturers had reported the results to the BVM. The experiments that were published are described briefly be

206 TABLE 2 10 Someone Z Ma Serotypes Most Frequently Isolated from Nonhuman Sources, 1977 Serotype Number of isolates reported, by source b Feed Chickens Turkeys Swine Cattle TOTAL typhimurium ~10 67 17 194 310 1,033 anatum derby25 514 264 1 378 cholerae-suisd1 00 332 4 345 11 11107 69 24 314 agona3 2535 28 85 271 Heidelberg1 75113 3 1 236 saint-paul3 2182 7 1 219 san-die~o Panama 9 012 147 1 188 infantis 3 818 18 7 165 0 1104 1 0 112 . TOTAL 66 267 5921,063 4343,261 PERCENT OF TOTAL OF ALL SEROTYPE S ISOLATED 39 44 6881 6162 TOTAL OF ALL SEROTYPES ISOLATED 168 608 8661,313 7135,243 oFrom CDC, 1979. Also includes isolates from other animals and environmental samples. CIncludes var. Copenhagen. War. kunzendorf. . _

207 TABLE 3 10 Sa~oneZZa Serotypes Most Frequently Isolated from Human Sources, 1977 1 i SerotypeNumber Percent Rank in 1976 typhimurium9,690 35.3 1 newport2,187 8.0 4 Heidelberg1,741 6.3 2 enteritidis1,472 5.4 5 infantis1,304 4.7 6 agona1,229 4.5 3 saint-paul580 2.1 7 typhi549 2.0 8 montevideo470 1.7 12 oranienburg440 1.6 9 TOTAL19,662 71.6 TOTAL OF ALL SEROTYPES27,462 a From CDC, 1979. bIncludes var. Copenhagen. 100

208 EXPERIMENTS IN SWINE Gutzmann _ al. (1976) divided 30 5- to 6-week-old pigs, each weighing approximately 7 kg, into three groups of 10. One group was given chlortetracycline at 220 g/metric ton of feed. A second group was given AUREO S.P 250, at a level that supplied 110 g of chlortetracycline, 110 g of sulfamethazine, and 55 g of penicillin G per metric ton of feed. The third group was given the nonmedicated basal ration. On the sixth day of medication, all three groups were given in their feed 100 billion cfu (colony-forming units) per pig of a nalidixic- acid-resistant, tetracycline-sensitive strain of Salmonella typhi- murium of swine origin. A fourth group of 10 pigs from the same breeding was kept as an unmedicated, uninfected control. Fecal samples were taken at 1, 2, 6, 9, 15, 22, and 26 days after admin- istration to detect and enumerate salmonellae. As shown in Figure 1, chlortetracycline at 220 g/metric ton of feed substantially reduced the number of salmonellae shed. The effect of AUREO SeP 250 was somewhat less, but calculation of the total area under each curve, equivalent to the total number of sal- monellae shed by each group, showed a 22.5% reduction in the number of salmonellae shed by the group on chlortetracycline and a 9.6Z reduction in salmonellae shed by the group on AUREO S.P 250. The duration of shedding and number of animals shedding salmonellae was no different in the medicated than in the unmedicated groups. By plating the fecal samples on agar containing nalidixic acid and chlortetracycline, it was possible to learn whether the infecting salmonellae had acquired tetracycline resistance from the tetracy- cline-resistant E. cold which the pigs were known to have. Only a - few tetracycline-resistant Salmonella were found in six of 172 fecal samples, and these occurred in unmedicated as well as medi- cated pigs. Thus, it could be concluded that the drug had not selected for resistance. An experiment by another manufacturer (Evangelist) et al., 1975; Girard _ al., 1976) was conducted in a manner very similar to that described above except that oxytetracycline at 150 g/ton of feed or oxytetracycline at 150 g plus neomycin atj150 g/ton of feed was given to medicated groups. Figure 2 shows the quantity of Salmonella tYPhimurium shed. As in the previously described experiment the medicated groups shed substantially fewer salmonellae than did the unmedicated group.

209 7 6 in 8 a, 5 - c' lo- 4 _ o 3 ; _ 2 _ 1 < Unmedicated in\ I\ AUREO S' P 250 (to supply 110 9 \ \ chlortetracycline, 1109 sulfametha zine and 55 9 penicillin G/metric If feed) \ ~ hi, Chlortetracycline, ZO g/metric ton of feed \ _N in, in_  -_ 1 1 1 1 1 1 1 1 2 6 9 15 22 26 DAYS (Postinoculation with 1 X 1 o1 1 cfu S. typhimurium) FIGl~RF 1. Summary of Salmonella typhimurium isolations from swine. From Gutzmann_ al., 1976, with permis- sion from the authors and the American Journal of Veterinary Research.

210 J m - - 7r Unmedicated 6 4 \~ 1 lo Oxytetracycline _ 150 g/ton \ \ Neomycin + Oxytetracycline, ~50 9 each/ton - -  - I I I 7 10 14 21 28 1 1 , DAYS (Postinoculation with 1.4 X 1 o1 1 salmonellae) FIGURE 2. Number of salmonellae recovered from swine feces. From Evangelisti et al., 1975, and Girard et al., 1976.

211 There was no difference in duration of shedding between the medi- cated and unmedicated groups. Only two resistant isolates were de- tected in the medicated groups. None was seen in the unmedicated group. EXPERIMENTS IN CALVES . Layton et al. (1975) divided 24 4-week-old calves into three groups of eight calves each. One group was given 350 mg of chlortetracycline per head per day in feed, and the other group was given 350 mg of chlortetracycline plus 350 mg of sulfamethazine per head per day in feed. These levels were supplied by a weighed amount of a concentrate containing the desired amount of the anti- microbial. The third group was not medicated. After 5 days of medication, each calf 9as infected orally via stomach tube with approximately 6.5 x 10 cfu of a nalidixic-acid-resistant, anti- biotic-sensitive strain of Salmonella typhimurium of bovine ori gin. Fecal samples were obtained at 1, 2, 6, 9, 15, 22, and 26 days after infection and examined for the number of salmonellae. Figure 3 shows that the number of salmonellae from the medicated groups was lower than that from the unmedicated group at each sampling. The group on chlortetracycline shed 38.0Z fewer salmo- nellae. The group on the combination shed 28.6% fewer salmonel- lae. The duration and prevalence of shedding was slightly greater in the unmedicated than in the medicated group. Salmonellae that were resistant to chlortetracycline were recovered from five calves in the unmedicated group, three calves in the chlortetracycline- medicated group, and one calf in the group on the combination. Only a very small fraction of the salmonellae In any of these sam- ples was resistant, and in only two instances were resistant orga- nisms isolated more than once from the same calf. In no instance did the resistant organisms become established in a calf. In another experiment (Evangelist) et al., 1975; Girard et ale 1976), run in much the same way, three groups of 10 calves each, averaging approximately 85 kg per calf, were given feed containing oxytetracycline at 350 mg per head per day or oxytetracycline and neomycin in an amount calculated to give each calf 350 mg of each drug per head per day. One group was not medicated. The results of this experiment were very similar to the previously described experiment in calves. The number of salmonellae shed is shown in Figure 4. The duration and prevalence of shedding in the medicated groups were considerably less than in the unmedicated group. No tetracycline-resistant salmonellae were detected.

212 .~ / J l 105 104 in LU c' 1 03 1o2 lo1 _ ~\ ~ \` \Nonmedicated \ ~ \ Chlortetracycline + Sulfamethazine \<, ~ >~350 mg each/head/day ~ \ Chlortetracycline, 350 mg/head/da~', 1 1 1 1 1 1 1 1 2 6 9 15 22 26 DAYS (Postinoculation with 6.5 X 109 cfu S. typhimurium) FIGURE 3. Salmonella typhimurium isolated from feces of infected . calves (expressed as geometric means). From Layton et al., 1975, with permission from the authors and Zentral-  blatt fuer Veterinaermedizin Reibe B.

213 6 . ~ 2 ~ 2 LL J m 4 o /\lonmedicated Oxytetracycl ins, 350 mg/head/day r~~l ~~ I ---r~~~'~l~~ 2 4 7 10 14 21 28 Neomycin + \ Oxytetracycline, 350 mg each/head/day ~' \/ ~-~ DAYS (Postinoculation with 1.10 X 1 o1 ] salmonellae)  FIGURE 4. Number of salTnonellae recovered from calves feces. From Evangelisti et al., 1975, and Girard et al., 1976.

214 EXPERIMENTS IN CHICKENS In an experiment designed to meet the requirements of the BVM, Jarolmen et al. (1976) divided 112 1-day-old chicks into four groups of 28 each. Two groups were given feed containing 200 g chlortetracycline per ton of feed. The other two groups were given unmedicated feed. On the sixth day after medication had begun, each chick in one of the medicated and one of the un- medicated groups was infected orally with approximately 6.8 x 109 cfu of a nalidixic acid-resistant, antibiotic-sensitive strain of Salmonella tYPhimurium of chicken origin. Droppings from each bird in the infected groups were analyzed for Salmonella content at intervals for 57 days. Figure 5 shows the rate of disappear- ance and the number of Salmonella shed by medicated and unmedi- cated groups. Eight days after administration, the unmedicated group had shed 50 times as many salmonellae as did the medicated group. During the entire study the reduction in the number of salmonellae shed was highly signficant (P = 0.01~. Unlike the studies with swine and calves, many tetracycline- resistant salmonellae were isolated from both medicated and unmedicated groups. This was not unexpected since Walton (1966) had shown the in-vivo transfer of R factors in the guts of chick- ens that had been given 10 drops of an overnight broth culture of it-factor-containing E. cold by mouth immediately after hatching, while their intestinal tracts were still sterile, and the same amount of S. tYPhimurium bY mouth 6 hours later. Smith (1970) had also shown that resistance to tetracycline could be transferred from E. cold to S. typhimurium when newly hatched chicks were given large numbers of R-factor-conta~ning E. cold in their drinking water for 3 days and then given 10 viable _ typhimuri~m by mouth. However, he had also shown that transfer _ the occurred only when the I. cold used was a good colonizer of chickens' intestine and was able to transfer resistance readily in vitro. In the experiment by Jarolmen et al. (1976), the presence of chlortetracycline in the feed of the medicated chicks had insured that they would have large numbers of resistant E. cold in their intestines, and the unmedicated chicks had evidently become contam- inated because they were in the same room. The numbers of both sensitive and resistant S. typhimurium decreased at about the same rate, even though the medicated birds continued to receive chlor- tetracycline throughout the experiment. The performance, in teems of final weight and kilogram of weight gain per kilogram of feed,

215 6 5 In ._ Q o 4 ~5 - c' - 1 ~ c o o is a: Lit 2 1 o / - ~ _ A ' \\ Unmedicated by\ - , ~ \ \ Chlortetracycline \ \jOO g/ton \ \ \ 1 1 1 1 1 1 1 ~ 0 4 8 16 22 29 36 1 1 43 50 57 DAYS (Postinoculation with 6.8 X 109 cfu S. nyphimurium)  FIGURE 5. Salmonella typhimurium i so let ions from droppings of chickens inoculated wi th 6. 8 x 10 of u S. typhimurium Prom Jarolmen_ al., 1976, with permission from the authors and the Journal of Applied Bacteriology. .

216 was better in the medicated than in the unmedicated birds irre- spective of whether they had been inoculated (Table 4~. In an independent experiment (Evangelist) et al., 1975; Girard _ al., 1976) medication was started when the chickens were 8-days-old and thus had an established intestinal flora. One group of 10 birds that was given 200 g of oxytetracycline and another group of 10 birds given 200 g of oxytetracycline plus 200 g of neomycin per ton of feed shed far fewer S. typhi- murium and stopped shedding earlier than did the group of 10 unmedicated controls (see Figure 6~. Very few resistant S. typhimurium were detected. A second type of experiment more nearly approximated natural conditions. Jarolmen et al. (1976) divided 100 1-day-old chicks into two groups of 50 each and housed the groups in separate buildings. Two chicks in each group were given oral doses of approximately 3.6 x 106 cfu of a nalidixic-acid-resistant, anti- biotic-sensitive strain of S. typhimurium. One group was given feed containing 200 g of chlortetracycline per ton; the other group received unmedicated feed. Fecal samples were collected from each bird in each group once weekly for 8 weeks and analyzed for Salmo- nella. The medicated, inoculated birds stopped shedding salmonellae within the first 2 weeks, whereas the unmedicated inoculated birds shed for 6 weeks. Figure 7 shows the percent of contact birds, i.e., uninoculated birds that have been exposed to birds infected with salmonellae, that were positive for Salmonella in each group at weekly intervals. The Salmonella spread quickly throughout the unmedicated group and maintained a level of infection of 90% or more for 6 weeks, when it dropped to 507. In the medicated group, no more than 24% of the birds were infected and only 10% were positive at the end of the test. A tetracycline-resistant strain of S. ty- phimurium was isolated only once from a medicated bird, but it did not spread to other birds or recur in the same bird. These experiments provide evidence that the use of subthera- peutic levels of tetracycline in the feed of swine, calves, and chickens do not increase the number of animals shedding Salmonella or the total number of salmonellae shed nor does it prolong the duration of shedding or select for resistant strains when the in- fecting organism is sensitive to antibiotics. Williams et al. (1979) infected swine with S. typhimurium cells that were either sensitive or resistant to antibiotics. Some of the

217 TABLE 4 Final Weight and Ratio of Feed to Weight Gain of Chickens Inoculated with SaZmoneZZa typh~murium (6.8 x 109 colony forming units) Compared to Uninoculated Chickens When Receiving Medicated or Unmedicated Feeda Type of Feed Inoculated Final Weight, kg Feed: Gain Not Inoculated . . Final Weight' kg Feed: Gain Nonmedicated 1.95 2. 2084 1.90 2.2712 Chlortetracycline 2.01 2. 1758 2.04 2. 2440 a From Jarolmen et al., 1976.

6 ^ .` c = 0 4 ~ 3 m o 218 \ \ Nonmedicoted Iowan + 1 °~- 7~ g earn Oxvtetracyci lne, ~0 Ion 2 4 7 10 14 21 28 DAYS (Postlnoculallon with 2.g X 1011 salmonellae)  FICKLE 6. aver of salmo~llae recovered from chickens feces. From Eva~elisti ^~, 1975, and Cirard ~, 1976 e

219 1 00 80 Cal O 60 LO > 40 oh o Cal 1~ 20 C' LU Cal o \ Nonmedicated Ch lortetracycl ine, 200 g/ton -R 4 5 _ ~ 6 7 8 AGE OF BIRDS IN WEEKS FIGURE 7. Percent of contact birds positive for Salmonella typhimurium by di rect plating or af ter enrichment.

220 animals were given feed containing chlortetracycline at 110 mg/kg; others received unmedicated feed. The results obtained with the pigs infected with the sensitive strain were similar to those de- scribed previously. However, when the pigs were infected with an antibiotic-resistant strain and were then given feed containing chlortetracycline, they shed a greater quantity of salmonellae for a longer time than did the pigs on unmedicated feed. Salmonellae spread to uninfected pigs that were exposed to pigs with antibio- tic-resistant salmonellae. The spread was greater when the recip- ients were receiving medicated feed. Five of 14 pigs infected with the antibiotic-sensitive strain died during the experiment. None of the pigs infected with the antibiotic-resistant strain died. Garside _ al. (1960) studied the emergence of resistant S. typhimurium in both unmedicated chicks and those that had been medicated with chlortetracycline. Resistant strains that emerged in the medicated groups were used to infect other groups of birds that were either medicated or unmedicated. Table 5 summarizes the mortality in five experiments conducted by Garside and his collea- gues. The mortality in the birds infected with the resistant strains was not much less than it was in the groups infected with the sensitive strains, but the mortality in the groups infected with the resistant strains was reduced as much by medication with chlortetracycline as it was in the groups infected with sensitive strains. INFECTION OF PIGS ON THE FARM Williams _ al. (1969) have shown that animal protein meal, when used as a constituent of pig feed, may be a source of salmo- nellae for hogs. Of 311 samples delivered to one feed mill over a 10-month period, 68% contained Salmonella. Of 206 lots of meat meal, 86% contained Salmonella. Numerous other reports have incriminated the animal protein constituents of pig feed as sources of Salmonella. In 1977, 136 isolates of Salmonella from animal feeds were reported by various state health agencies to CDC (CDC, 1979~. Of these, 22 were S. derby, the third most frequently isolated serotype from meat animals (Table 6~. Only two isolates were obtained from vegetable protein. S. derby was not reported by the CDC as having been iso- lated from any food-borne outbreak of disease in humans in 1975, 1976, or 1977, although it was one of the three serotypes most fre- quently reported to the CDC as having been isolated from swine, cattle, turkeys, or chickens in each of those years (CDC, 1976b,

221 TABLE Mortality of Chicks Infected with Strains of SaZmoneZZa typh~muri~m that were Sensitive or Resistant to Chlortetracycline When Receiving _ed Treated with Chlortetracycline or Unmedicated Feeda Type of Feed Mortality, by sensitivity of ~aZmoneZZa Sensitive Resistant b nonmedicated 90/185 92/217 (48.6%) (42.4%) medicated 44/188 26/164 (23.4%) (15.9%) aFrom Garside et al., 1960. bNumber dead/total number .

222 TABLE 6 10 SahmoneZZa Serotypes Most Frequently Isolated from Food Animals, 1975 through 1977 a 1975 No. of Isolates Serotype senftenberg 1976 120 1977 Serotype No. of Isolates typhimurium 113 _ NOe of Serotype Isolates typhimurium 588 typhimurium86 derby33 cholerae-suis336 newport17 panama24 derby284 . . . derby6 cubana21 anatum211 meleagrid~s6 cholerae-suis10 saint-paul192 dublin5 newington10 heidelberg192 cholerae-suis3 dublin8 agona173 heidelberg3 anatum7 panama160 muenchen2 worthington7 infantis114 london2 heidelberg6 san-diego106 TOTAL250 239 2,356 PERCENTAGE: OF ALL SEROTYPES ISO LATED89 76 67 TOTAL OF ALL SEROTYPE S I SOLATET)28 0 316 3, 5 00 b From CDC, 1976b, 1977d, 1979. Includes var. copenhagen .

223 1977d, 1979~. The state agencies also reported that S. derby was one of the serotypes most frequently isolated from human dietary items in 1976 and 1977 (Table 7~. The figures reported by CDC have little significance since it probably receives reports on only a small number of the actual outbreaks of diseases. The fact that some few serotypes are reported year after year as the most frequently isolated serotypes suggests that they have some epidem- iological significance. This is particularly true of the isolates from humans for which eight of the 10 most frequently reported serotypes were the same in 1975, 1976, and 1977 (Table 8~. But because these serotypes are found so frequently in humans, it is highly speculative to assume that any particular serotype has come from animals. If phage typing were done and the phage type of the Salmonella from the animals were the same as the phage type of the Salmonella from the humans, the assumption would be more likely. In an epidemiological study of Salmonella infection in swine in Ontario, only 3.2% of 380 carcasses were positive for Salmonella (Groves et al.. 1970~. In a survey of five abattoirs, five car casses out of bO were positive in one abattoir and 10 out of 80 were positive at another. In the remaining three abattoirs, none of the carcasses were positive. However, 20.3% of the mesenteric lymph nodes from 462 hogs were positive. Of 190 environmental samples, 8.4Z were positive, and of 101 feed samples, 7.8% were positive. S. typhimurium, S. Heidelberg, S. muenster, and S. anatum were most frequently isolated from lymph nodes, but only S. typhimurium and _ Heidelberg were isolated from carcasses. Antimicrobial ance was detected in 11.7% of the 94 Salmonella isolates. ~, . ~. . ~_ _ resist Five of 14 isolates tested for R factors were positive. The other nine were resistant to only one antimicrobial agent. In one instance, S. Heidelberg was isolated from feed samples and swine none on ~ tare, trom swine trom the term slaughtered in an abattoir, from the abattoir environment, and from the carcasses and washings of edible pork products in the abattoir. These data provide circumstantial evidence that swine serve as a reservoir for salmonellosis in humans, but the humans working in the abattoir were not sampled to determine whether they might be a source of S. Heidelberg. Additionally, S. Heidelberg is one of the most frequently isolated serotypes: phage typing of isolates from animals and from cases of salmonellosis in humans would be necessary to prove that the infections were caused by salmonellae of animal origin.

224 TABLE 7 10 Sa1~none1~1'a Serotypes Most Frequently Isolated from Dietary Items for Human Consumption, 1975 through 1977 ~ 1975 1976 1977 No. of No. of No. of Serotype Isolates Serotype Isolates Serotype Isolates kentucky 74 panama 22 newington 19 l anatum 16 enteritidis 20 typhimurium 15 manhattan 16 derby 15 infantis 13 . typhimurium 13 typhimurium 14 dtlblin 8 . infantis13 heidelberg 11 weltevreden agona12 mokola 10 agona heidelberg12 orion 10 derby derby9 muenster 7 senftenberg 5 5 5 saint-paul7 anatum6 tennessee5 litchfield7 lanka6 bareilly4 TOTAL179 121 87 PERCENTAGE OF TOTAL ISOLATED FROM DIE TARY ITEMS77 70 66 TOTAL ISOLATED FROM DIETARY ITEMS232 173 131 a From CDC, 1976b, 1977d, 1979.

225 TABLE 8 10 SatmoneZ!a Serotypes Most Frequently Isolated from Humans, 1975 through 1977a 1975 _ 1976 of Serotype __ 1977 No. of No. of No. of Isolates Serotype Isolates Serotype Isolates typhimurium6,888 typhimurium7,347 typhimurium9,690 newport1,550 heidelberg1,962 newport2,187 enteritidis1,519 agona1,461 heidelberg1,741 heidelberg1,474 newport1,336 enteritidis1,472 agona1,333 enteritidis1,219 infantis1,304 infantis1,194 infantis1,014 agona1,229 saint-paul833 saint-paul545 saint-paul580 typhi551 typhi529 typhi549 oranienburg446 oranienburg465 montevideo470 javiana426 muenchen374 oranienburg440 TOTAL16,264 16,752 19,662 PERCENTAGE OF TOTAL SERO TYPES ISOLATED69 72 72 TOTAL SEROTYPES ISOLATED23,445 23,285 27,462 CDC, 1976b, 1977d, 1979. Includes var. copenhagen.

226 CONTAMINATION OF CARCASSES AND MEAT BY SalmoneZZa Contamination of hog carcasses by Salmonella was recognized before antibiotics were used in animal feed. An extensive study by Galton _ al. (1954) during 1951 and 1952 showed that 7% of 374 hogs on 28 Florida farms were positive for Salmonella. However, 18 of the 27 positive animals were from two farms. All pigs tested on 20 farms were negative. When hogs and carcasses were sampled for Sal- monella at five abattoirs, the investigators found that extensive contamination and infection of the hogs had occurred during trans- port to the abattoir from the farm. A total of 1,883 rectal swabs were taken at the same point in the processing line of the five abattoirs. Results among the abattoirs ranged from 11% to 80% posi- tive for Salmonella. There was a substantial difference in the efforts made by the abattoirs to prevent contamination and cross- contamunation of the carcasses. This was evident from the fact that only 1% of environmental samples taken from the cutting and sausage rooms of one plant were positive, whereas 66% of the samples taken from the same types of locations in another plant were positive. That pigs may be marketed virtually free of Salmonella infec- tion was shown by Edel et al. (1974~. They raised pigs in houses kept free of insects, birds, and rodents, and fed them pelletted feed. (Pelletting involves injecting steam into the feed to raise the temperature to 82 C. This temperature is maintained in the pellets long enough to kill most Salmonella.) Only 21 (1.6%) of 1,317 pigs slaughtered were Salmonella-positive, and of these, 14 (1.1%) were probably contaminated in the abattoir. A study of salmonellae in cattle, their feed, and their rela- tion to human infection was reported in England in 1965 (Ministry of Agriculture, Fisheries, and Food, 1965~. This investigation showed that S. dublin was the serotype most frequently infecting cattle - and calves but that there was also a high incidence of S. typhimur- ium. By the use of phage typing, the investigators showed that S. typhimurium spread from cattle, especially from calves, to humans. The infections in calves were exacerbated by shipping them long distances without proper food or rest, by subjecting them to chills, and by mixing them with calves from other sources. S. typhimurium was infrequently isolated from cattle feed. S. dublin, which was isolated from only two of 7,300 samples, was said to play only a small role in human infections. Finlayson (1977) discussed the relationship between human infec- tions and salmonellae in poultry products in Alberta, Canada. She reported that the great majority of human infections were "sporadic,"

227 involving only single cases or individual families, and that the vehicle of infection was seldom identified. There was circumstan- tial evidence that poultry products might be playing a role in human infections in that the serotypes most frequently isolated from humans were among those most frequently isolated from poultry. But again, this assumption would be strengthened if -the isolates were of the same phase type. Cleaning procedures varied consider- ably from plant to plant. Hence, Salmonella contamination of the plants varied from 46% to 89% of the environmental samples taken. Eighty-seven percent of the strains examined were sensitive to antimicrobial agents commonly used in treatment of humans and livestock. In birds, resistance occurred mainly in strains spec- ifically associated with turkeys (S. saint-paul and some types of S. typhimurium). Commoner Salmonella strains (S. California and _ infantis) were generally sensitive to antibiotics when isolated from turkeys. FOOD-BORNE OUTBREAKS OF SALMONELLOSIS In 1975 there were 25 reported outbreaks of salmonellosis in- volving an estimated 1,006 persons (CDC, 1976b), who comprised 4.3% of the 23,445 isolations reported that year. In 1976, there were 42 reported outbreaks affecting 1,915 persons. Of these, 571 had posi- tive cultures, which was 2.4% of the 23,285 reported isolates (CDC, 1977d). In 1977, there were 40 reported outbreaks involving 1,632 persons. Of these, 364 had positive cultures, or 1.3% of the 27,462 isolates. Of course, these figures are minimums since many outbreaks were not reported and, probably, many were not even diagnosed. From 1962 through 1977, 145 deaths were attributed to salmonello- sis--a case fatality ratio of 0.36% (CDC, 1979~. Table 9 lists the serotypes reported from 1975 to 1977 and the number of out- breaks from which they were isolated. S. typhimurium was isolated most frequently. Heidelberg, thompson, newport, and saint-paul were isolated in each of the 3 years. These serotypes, except for thompson, were among the seven most frequently isolated from humans in each of the 3 years. Fran 1975 through 1977 outbreaks of salmonellosis were traced to commercially prepared precooked roast beef. Several of these outbreaks involved more than one state. In 1975, four outbreaks in New Jersey caused by S. saint-paul were attributed to roast beef that had been cooked at a temperature too low to destroy Salmonella after being injected with a fluid containing spices, flavorings,

228 TABLE 9 Isolates from Food-Borne Outbreaks of Salmonellosis, 1975 through 1977a 1975 1976 1977 No. of No. of No. of Serotype Isolates Serotype Isolates Serotype Isolates saint-paul 6 typhimurium 15 typhimurium 8 typhimurium 2 typhi 5 infantis 4 enteritidis 2 heidelberg 5 agona 3 - typhi 2 enteritidis 3 heidelberg 2 thompson 2 saint-paul 2 san-diego 2 oranienburg 1 bovis-morbificans 2 typhi 2 tennessee 1 newport 2 bredeney 2 reading 1 typhimurium var. 1 london 1 heidelberg 1 copenhagan saint-paul 1 kottbus 1 poona 1 anatum 1 litchfield 1 muenchen 1 schwarzengrund 1 blockley 1 london 1 kottbus 1 dublin 1 schwarzengrund 1 bareilly 1 newport 1 blockley 1 thompson 1 singapore 1 javiana 1 weltevreden 1 bredeney 1 reading 1 TOTAL 25 san-diego 1 newport 1 minnesota 1 chester 1 thompson 1 oranienburg 1 infantis 1 muenchen 1 give 1 stanley 1 Group E 1 Group D 4 Group B 1 TOTAL 48 Other 2 TOTAL 44 a From CDC, 1976b, 1977d, 1979.

229 and, apparently, Salmonella. A second extensive outbreak, con- sisting of at least 200 identified cases in Pennsylvania, New Jersey, Connecticut, and Massachusetts, was due to S. bovis-mor- bificans (CDC, 1976a, 1977a). Epidemiological studies incrim- inated imported beef although Salmonella was not isolated from any of the beef at the site of purchase or processing. As a re- sult of this outbreak, the USDA issued a regulation on September 2, 1977 requiring that all commercially prepared roasts of beef be cooked to an internal temperature of 63 C. In a third outbreak, which occurred in June, July, and August, 1977 (CDC, 1977c), S. newport was isolated from both the interior and exterior of roasts of beef before they had been carved. These cases were reported in Pennsylvania, New York, and New Jersey, and were all attributed _ cooking the meat at too low a temperature. In the first outbreak, the contamination was due to injecting the meat with a contaminated fluid. Since no salmonellae were found in the meat consumed in the second outbreak, the source of contami- nation has not been proven. In the third, no source of contamina- tion was shown. In 1977, the food-borne outbreaks of salmonellosis involved meat, milk, ice cream, or cheese (CDC, 1979~. In July an outbreak in Michigan involving S. typhimurium was traced to home-made ice cream made with raw eggs (CDC, 1977e). Another outbreak involving S. typhimurium occurred in Kentucky where two children and their father were infected by raw milk used to make icing for a cake (CDC, 1977f). Salmonellae were isolated from the three patients, but not front the milk or icing. S. typhimurium with the same lysis pattern was isolated from a cow and a calf on the fans where the milk was produced. The organism was resistant to tetracycline, streptomycin, ampicillin, carbenicillin, and penicillin. There had been an oubreak of diarrhea! illness of several months duration in the dairy herd, but the report of the outbreak (CDC, 1977f) did not document whether it had been treated with antibiotics. Large amounts of ground beef, cold cuts, and frankfurters are consumed in the United States. Because the manufacture and sale of these products require more handling than do cuts of beef or pork, it is desirable to analyze CDC's food-borne disease surveillance data on these foods (CDC, 1975~. From 1966 through 1973, 2,464 food-borne outbreaks were reported to CDC. In 1,827 (74%) of the outbreaks the food vehicle was identified. Ground beef, cold cuts, or frankfurters were responsible for 91 (5%) of the 1,827; 65 (3.6%) were caused by ground beef; 6 (0.3%) by cold cuts; and 20 (1.1%) by frankfurters. In 21 of the 91 outbreaks the proximate cause was determined: one

230 was the result of a food-processing error before consumer pur- chase; the other 28 resulted from improper food handling after purchase. The etiologic agent was identified in nine of the 65 outbreaks attributed to ground beef. Three of these were due to salmonellae other than S. typhi, one was due to S. typhi, two to Clostridium perfringens, one to C. botulinum, and two to chemicals. Data concerning the place where ground beef was mishandled were available for 27 of the 65 outbreaks. The beef was mis- handled in food-service establishments in 18 outbreaks and in private homes in eight of the outbreaks. The other instance of mishandling occurred in a food-processing establishment. anatum. Of six outbreaks attributed to cold cuts, one was due to S. From 1972 through 1975, the USDA intermittently surveyed ground beef, luncheon meat, and frankfurters for salmonellae. Three contaminated beef patties were found. No salmonellae were found in luncheon meat or in frankfurters (CDC, 1975) (Table 10~. Data on food-borne outbreaks of salmonellosis in England and Wales for 1973 through 1975 are of interest for comparative pur- poses (Vernon, 1977~. Table 11 lists the 10 most frequently iso- lated serotypes from humans in England and Wales and in the United States. Six of the 10 most frequently isolated serotypes are the same in both lists as are three of the first four ranked. These four serotypes (S. typhimurium, S. newport, S. Heidelberg, and S. enteritidis) accounted for 60% and 55% of the total reported iso- lates in England and Wales and the United States, respectively. If one used the same logic that was used to link the isolates of Salmonella frog animals with those from humans, i.e., the fact that several serotypes are among those most often isolated from each other, then one might argue that salmonellosis in England and Wales is transferred from humans in the United States. It is doubtful that many epidemiologists would subscribe to that propo- sition. There were 1,685 reported outbreaks of salmonellosis in England and Wales from 1973 to 1975 of which 407 were general out- breaks (i.e., involving more than one family) and 1,278 were out- breaks within a single family. Hospitals accounted for 157 (39%) of the general outbreaks. There were 15,084 sporadic cases reported. The presumed causal agents were identified in 107 of the general and family outbreaks.

231 TABLE 10 Surveys of Salmonellae in Raw Beef Patties, Luncheon Meat, and Frankfurters, 1972 through 1975 a Product Sampl es Examined Sample s Positive Raw beef pa t ties Raw tr immings 6 90 1 Raw finished patties 735 3 Frankfurters Raw trimmings 842 56 Cooked finished frank- 690 0 furlers Sliced luncheon meat Raw trimmings 936 69 Cooked, sliced luncheon 456 0 meat Lax From CDC, 1975

232 TABLE 11 10 Most Frequently Isolated ~ ~from Humans in . En land and Wales and in the United States g England and Wales, 1973-1975 No. of Isolates Serotype United States 1977 , ~ pro tvr)e No. of Isolates typhimurium8,396 typhimurium9,690 agona3,685 newport2,187 enteritidis2,602 Heidelberg1,741 Heidelberg1,178 enteritidis1,472 anatum1,009 infantis1,304 Indiana931 agona1~229 newport908 saint-paul580 infantis698 typhi54 9 hadar512 montevideo470 bredeney494 oranienburg440 TOTAL20,4 13 19, 662 PERCENTAGE: OF ALL SEROTYPE S ISOLATED77 72 TOTAL SEROTYPES ISOIATED26,574 27,462 From Vernon, 1977, and CDC, 1979 I.

233 Meat was incriminated in 27 outbreaks of which 11 were attributed to ham and bacon. Poultry was responsibile for 49, milk and cream for 30, and sweet desserts for one. Several things seem clear from the above discussion on sal- monellosis. First, domestic animals are a significant reservoir of Salmonella. Second, this reservoir is expanded in the holding pens at slaughterhouses. Third, carcasses and meat are contami- nated with Salmonella during the killing and dressing process. . It is less clear how much of this Salmonella actually reaches the consumer. However, when it does reach the consumer, mishandling of meat and meat products can result in multiplication of the organisms to the point that the consumer becomes infected and an outbreak occurs. But examination of the information on which these conclusions are based produces no evidence that the use of subtherapeutic levels of tetracyclines or penicillin in animal feed played any role in the size of the reservoir, the spread of the salmonellae, or in any of the steps in transmission to the consumer. The factors that contribute to outbreaks of food-borne dis- ease were reviewed by Bryan (1978~. Inadequate cooling was the single most frequent factor: either failure to refrigerate or storing such large amounts that they cooled too slowly. Handling of cooked foods by infected persons and cross-contamination were other important factors. Antibiotic sensitivity is seldom men- tioned in the articles reviewed by Bryan, and in no instance was the subtherapeutic use of antibiotics in animal feed shown to be a factor. Occasionally, however, there were speculations that subtherapeutic use might increase the number of resistant Sal- monella reaching the human population. In the Kentucky outbreak discussed above (CDC, 1977f), resistance of the organism was probably due to treatment of the dairy herd for a prolonged epi- sode of diarrhea! illness, although this was not documented. ANTIBIOTIC RESISTANCE IN SALMONELLAE There have been a number of surveys of antibiotic-resistant _almonella in humans and animals in the United States. One such survey (Neu et al., 1975) compared the antibiotic resistance and it-factor transmissibility of 718 isolates from humans in the north- eastern United States with that of 688 isolates from animals in the eastern and Midwestern states. The isolates from animals, like those from humans, were obtained from sick individuals. The ani- mals, and perhaps the humans, had probably been treated with

234 antibiotics. The percentage of S. typhimurium, S. Heidelberg, and _ saint-paul frog animals with resistance to streptomycin, sulfisoxazole, and tetracycline was greater than that for the strains from humans (Table 12~. Resistance to kanamycin was high in the strains from the animals, but this may have been due to the fact that resistance to kanamycin and neomycin are often linked and neomycin is frequently used both therapeutically and prophylactically in animals. The percentage of S. saint-paul and _ enteritidis strains from animals with resistance to ampi- cillin was greater than that for the corresponding strains from humans. Ampicillin is not approved by FDA for use in animal feed but has been approved for therapeutic use in calves and swine since 1973. The therapeutic use of antibiotics has been shown to elicit antibiotic resistance somewhat more rapidly than do prophylactic levels. Luther et al. (1974) gave 6-day-old chicks oxytetracy- cline in drinking water and measured the development of resistance in the Escherichia cold of their intestinal flora. One group of birds received 50 ppm of oxytetracycline in drinking water. This amount is equivalent to 100 ppm in the feed, which is a subthera- peutic level (Table 13~. Resistance of 100Z of the isolates to tetracycline, dihydrostreptomycin, and sulfamethazine developed in 3 days. Another group of birds was given 500 ppm in drinking water. This is equivalent to 1,000 ppm in feed, a therapeutic level. In this group of chicks, resistance of all isolates to tetracycline and dihydrostreptomycin developed in 2 days. Some resistance to ampicillin emerged in the chicks on 50 ppm, pre- sumably because of linkage of resistance genes, but not in those on 500 ppm. On the basis of resistance patterns of salmonellae from humans and animals, Neu et al. (1975) suggested that animals might be contributing to the pool of resistant organisms in humans. Cherubin et al. (1972) thought that unlikely. They said, An urban reservoir of resistance transfer factor could exist equally as well as the postulated animal reservoir. It is difficult to reconcile the knowledge that the highest incidence of salmonellosis occurs in young children and infants in slum areas in New York City and the finding that the highest frequency of resistant Salmonella strains occurs in municipal hospitals serving these areas, with the hypothesis that Salmonella in gen- eral and resistant strains (or the resistance transfer factor) in particular, originate in farm animals. Unless one accepts the doubtful proposition that meat, poultry, eggs, milk and pro- cessed foods derived from them are preferentially consumed by the

235 TABLE 12 Resistance of Fourth 0neZ!a Serotypes from Humans and Animals to Individual Antimicrobial Agents a Percentage of Strains Resistant, by Serotype (No. of Strains Tested ) to phirrrurium heide Zb erg saint-p~uZ enteritidis (241) (484) (30) (72) (44) (77) (102) (27) Antibiotic human animal human animal human animal human animal . . . . Ampicillin 36. 9 31.030. 0 11.013.6 19. S 0. 9 7. 4 Cephalothin 2.1 1.63.3 00 1.3 0.9 0 Chloramphen~cal 2. 9 O. 80 0() O 0 0 Gent ami cin O OO OO O O O Kanamyci n 29.4 46. 030. 0 19.02. 2 36. 0 O. 9 7. 4 Streptomycin 45.6 64.016.6 32.011.3 45.0 2.9 0 Sulfisoxazole 30.9 73.06. 6 22.02. 2 27.0 2. 9 0 Tetracy~l~ ne 44.8 61.020.0 18.09.0 35.0 1.9 0 Trimethoprim 0 0 0 0 0 0 0 9 0 PERCENTAGE OF ALL STRAI NS IS OLATED WITH RESISTANCE TO AN ANTIBIOTIC a 57.6 80.0 30.0 41 15.9 61 S.8 7 Prom Neu et al., 1975, with permission from the authors and the Journal Infectious DO seases.

236 TABLE 13 Effect of Oxytetracycline (OTC) Water Medication on the Rate of Development of Escheriahia coZi Resistance In Vivo to Tetracyclines, Dihydrostreptomycin, Sulfamethazine, and Ampicillina No OTC OTC, 50 ppm OTC, 500 ppm Percentage Resistant Percentage Resistant Percentage Resistant Day Tc~ Dhs~ Size Ampe Tc Dhs Smz ~Tc Dhs Smz Amp -6 0 0 0 0 0 0 0 0 0 0 0 0 -2 3 6 6 6 0 0 0 0 0 0 0 0 o f - O O O O O O O O +1 0 0 3 0 55 50 55 23 66 66 0 +2 0 0 0 0 94 94 94 16 100 100 0 0 +3 19 19 16 9 100 100 100 3 100 100 0 0 +4 3 0 0 0 97 97 97 19 100 100 3 0 +5 0 0 0 0 100 100 100 13 100 100 0 0 +6 0 0 0 0 94 94 94 3 100 100 0 0 From Luther et al., 1974, with permission from the authors and Miles Labora tories, Ltd. Tc = tetracyclines t,Dhs = dihydrostreptomycin Smz = sulfamethaz ine Amp = ampicillin flashes indicate that no measurements were taken.

237 young and the poor, we would expect that the hypothesis of an animal origin would require that salmonellosis and resistant Salmonella isolates occur most frequently in adults and in the affluent. Another frequently mentioned survey of antibiotic resistance in animals is that of Pocurull _ al. (1971~. They tested 1,251 strains of Salmonella isolated from sick animals at the National Animal Disease Laboratory in Ames, Iowa. Of these, 935 (75%) were resistant to one or more of the 11 antimicrobials used in the sur- vey. Here again it is likely that this resistance resulted from therapeutic use rather than subtherapeutic use in feed since these isolates were obtained from sick animals. It is unusual for a veterinarian to take samples for antibiotic sensitivity testing before beginning treatment. Timoney (1978) provided further evidence that the pool of resistant Salmonella in humans may be different from that in calves. From 1973 to 1976 he tested 249 strains of S. typhimur- ium isolated from diseased animals in New York State for sensi- tivity to six commonly used antibiotics. Only 12% were sensitive to all six. Virtually all of the 135 strains isolated from calves during this time were resistant to kanamycin, neomycin, streptomy- cin, and tetracycline (Table 14~. An interesting observation was that resistance to ampicillin, which was first approved for therapeutic use in calves in 1973, increased steadily during the 4 years. Moreover, the resistance to chloramphenicol also increased during that 4-year-period although to a much smaller degree. Chlor- amphenicol is not approved for any use in livestock. The emergence of resistance to ampicillin in calves is reminiscent of a similar increase in resistance to this antibiotic in strains from humans in the northeastern United States during the late 1960's following widespread use of the drug after its introduction in 1964 (Neu et _ ., 1971~. The development of resistance to ampicillin in S typhimurium in calves and in humans would seem to be separate phe- nomena, and neither the use of antibiotics in feed nor therapeutic use in calves has influenced resistance in strains of S. typhimurium from humans in the northeastern United States. Further evidence of the separation of human and calf pools of S. typhimurium is reported by Cherubin _ al. (1980), who showed that resistance to ampicillin in S. typhimurium from two New York City hospitals has returned to 1965 levels (Table 15~. Resistance to tetracycline in these strains of S. typhimurium has also decreased, although not as markedly.

238 TABLE 14 Antibiotic-Resistant Strains of Sahnon~ZZa tuphinn~ - Isolated from Calves" - No. of Pergentage resistant Percentage with Year Isolates Amps colic Kma Nm e smf Tog transterAble R factors 1973 37 16 3 95 95 100 100 38 1974 40 33 3 87 83 97 95 49 1975 35 43 5 80 69 78 86 60 1976 23 74 9 96 87 100 100 87 a b d e From Timoney, 1978, with permission from the author and the University of Ch icago Pres s. Amp = ampicillin Coli = chloramphenico 1 Km = kanamycin Nm = neomycin Sm = s treptomycin Tc = tetracycline

239 TABLE 1 5 Antibiotic-Resistant SabnaneZZa typhi77rur:?~n Percentage of Resistant Isolates, by Study (Total of Isolates Tested ~ 1. ~ Antibi ~;~3~970 ~1973 '' 197019 - (128) (85) (241) (161) Ampicillin 23.4 12.9 36.9 4.3 Tetracycline 12.5 23.5 44. ~21.1 Streptomycin 27. 3 29.4 45. 6 59.6 Chloramphenicol O . O 0.0 2. 9 O. O From Neu_ al., 1971. bFrom Cherubin_ al., 1972. From Neu _ al., 1975. From Cherubin _ al., 1980. NOTE: Table reproduced with permission from the authors and the Journal of the American Medical Association 243:439-442, . February 1, 1980, O1980, American Medical Association.

240 There were no reports of Salmonella agona in humans before 1970, but by 1976, it had become the third most frequently iso- lated serotype (Table 16) (CDC, 1976b, 1977d, 1979~. Few isolates were reported from nonhuman sources until 1977 when 271 were re- ported. Of these, 173 were from pigs, cattle, chickens, and turkeys, 45 were from horses, and 10 were from miscellaneous animal sources. This suggests independent pools in humans and animals. S. agona was the second most frequently isolated serotype from humans in England and Wales from 1973 to 1975. It was also isolated frequently from poultry, sausage, and feedstuffs. SaZmoneZZcr typhi=La?ium IN ENGLAND An epidemic of antibiotic-resistant Salmonella typhimurium phage type 29 occurred in calves in England starting in 1964, peaking in 1965, and subsiding in 1966. This infection spread to humans. Of 2,544 cultures of S. typhimurium from humans sub- mitted in 1965 to the Enteric Reference Laboratory in the United Kingdom, 576 (237) were phage type 29 and 555 (22%) were resitant to antibiotics (Anderson, 1968b). In the testing of 1,772 cul- tures of S. typhimurium from bovines, 1,297 (73%) were phase type 29 and 1,294 were resistant to antibiotics. Thus, the phase type 29 strain played a much larger role in calves than in h~,rn~ns, but many human cases were involved. Since many of the infected human had been in close contact with infected calves, it is likely that the infection was transferred from the calves. most severe in calves intended for intensive (Anderson, 1968a). This was a highly profitable business and calves were bringing high prices. In this method of rearing the calves were gathered from various farms, often only hours after birth, shipped to a collecting point, and then transported to a farm where they were to be fed for several months before slaughter Some of the dealers were very careless in their management and hygiene. Calves were crowded into dirty vehicles and the condi- tions were ideal for the spread of disease. This epidemic was or confined rearing This was a highly profitable business and In thi s mothers of rearing the When the epidemic was at its height, large amounts of anti- biotics were used in an attempt to prevent or treat the disease. These attempts were largely futile because the organisms had become resistant to most of the commonly used drugs. This outbreak of disease can hardly be viewed as an indictment of the use of anti- b~otics in animal feed since it was really due to bad management. Specifically, calves were deprived of colostrum, without which no antibiotic will be effective in calves. Moreover, calves from various sources were mixed, which enabled the disease to spread

241 TABLE 16 SaZmone[Za agona Isolated from Human and Nonhuman Sources, 1969-1977 Number of Isolates Year of Study Human Nonhuman 19690 3 19704 5 197144 34 1972524 NAb 1973864 NA 19741,037 NA 19751,333 45 19761,461 40 1977 From CDC, 1976b, 1977d, 1979. b NA = Data not available 1,229 271 .

242 when the calves were moved to new farms. These animals were not kept adequately clean or warm nor were they fed or watered pro- perly. At the time of these outbreaks, antibiotics were not legally permitted in the feed of calves in England but were allowed in feed for pigs and chickens. Because of this epidemic the Swann Committee was appointed by the government of the United Kingdom. As a result of its recommendations (Swann et al., 1969), tetracy- clines, penicillin, and other antimicrobials were banned from the feed of pigs and chickens except on veterinary prescription. The ban on subtherapeutic use was not justified on the basis of evi- dence of incidents involving pigs or chickens since there had been no outbreaks of disease attributable to such use. The restrictions based on the Swann recommendations were ~m- plemented in March 1971. In 197S, there was an outbreak of anti- biotic-resistant S. typhimurium in calves in England that was similar to the 1965 outbreak (Threlfall et al., 1978~. Again, the disease was spread by crowding very young calves into dirty vehicles, transporting them long distances without adequate food, water, or warmth, and trying to cover these bad management prac- tices with heavy doses of antibiotics. Again, the motivating force was high prices for calves. It would seem that the Swann recommendations missed their mark. In summary, there is a large pool of Salmonella in domestic animals and a similarly large pool in humans. It is not clear that the second pool receives much contribution from the first. Food-borne outbreaks of salmonellosis are not uncommon, but what, if any, role antibiotic resistance plays in them seems impossible to determine from the evidence at hand. Finally, there seems to be no evidence that the use of tetracyclines and penicillin in animal feed play any adverse role in the development of salmo- nellosis in animals or humans or in food-borne outbreaks of sal- monellosis in humans. ANTIBIOTIC-RESISTANT Esoherzahia oofi IN HUMANS AND DOMESTIC ANIMALS In the Statement of Policy and Interpretation Regarding Animal Drugs and Medicated Feeds (FDA, 1973), the Commissioner of the FDA said that a theoretical hazard was posed by the possi- ble colonization of the intestines of humans by it-factor-bearing E. cold frog animals and by the possible linkage of an it-factor

243 plasmid with a plasmid for the production of enterotoxin in E. cold of animal origin, thus enhancing its pathogenicity. The first of these questions was addressed by Levy et al. (1976), who inoculated chickens in the cloaca with a strain of _ cold carrying a marked plasmid with resistance to tetracy- cline, streptomycin, and sulfonamides. These chickens were then placed in two cages with other, uninoculated chickens. The chickens in one of the cages were given feed containing oxytetracycline; those in the other were not. The resistant E. cold spread to other chickens in the cage with the feed con raining the antibiotic. It was not isolated from the chickens in the cage with unmedicated feed. Two other cages of uninocu- lated birds were in the same room. The birds in one of these cages were given feed containing oxytetracycline; the other plain feed. The E. cold with the marked plasmid was isolated frog birds in the cage on antibiotic feed but not from those on plain feed. On two occasions the labelled E. cold was recovered from fecal samples of people who worked in proximity to the chickens. An antibiotic-resistant E. cold that was isolated from one fecal sample taken from a young boy who cleaned the cages was resistant to streptomycin and sulfonamides but not to tetra- cycline. The second antibiotic-resistant E. cold was isolated from a young lady who took fecal samples from the chickens. This _ cold had the same resistance characteristics as that of the chickens. However, she also worked with the culture in the laboratory. Since the organism was isolated from her feces 7 days after she had handled the chickens, it is quite possible that she had acquired the organism in the laboratory rather than from the chickens. Because the organism was isolated only once from each person, it could not be said to have colonized their intestines. In a study of calves, Hirsh and Wiger (1977) used a labelled _ cold of bovine origin containing a labelled plasmid. This plasmid carried resistance to gentamicin, tetracycline, streptomy- cin, and sulfonamides. The calves were fed the organism and/or were given tetracycline twice a day orally. The tetracycline did not affect the prevalence of antibiotic resistance in E. cold from isolates from the feces of the calves: the prevalence was as great when the calves were not receiving tetracyclines as when they were. The marked organism with the marked plasmid was isolated once from each of three of the animal attendants. The plasmid did not trans- fer to the indigenous flora, did not persist, nor did it recur.

244 Two groups of investigators in England have studied the colonization of the human intestine by E. cold of animal origin. Shooter _ al. (1970) at St. Bartholomew's Hospital in London isolated antibiotic-resistant E. cold of the same O-serotype from animals and poultry at slaughter, from the meat as it came into the hospital kitchen, from the kitchen itself, especially from the water used to wash dishes, from the dishes, from the food as it was prepared for delivery to patients, and from the feces of the patients (Cooke et al., 1971; Shooter et al., 1970~. This study suffers from two serious flaws. First, Guinte (1963) has said, based on a study of more than 1,000 strains of E. cold from humans and animals, The results of our investigation show that with an extensive serological and biochemical study a relatively large number of serofermentation types can be found within one O-group, so that epidemiological conclusions based exclusively on the determination of the O-antigen can be quite deceptive. Shooter's study was based exclusively on the determination of the O-antigen. Guinbe also said, ...the results of the investigation reported above do not justify any conclusions on the problem of whether or not the strains of E. cold originating from animals are of any importance as pathogenic agents for man, or vice versa. Moreover, Shooter and his colleagues sampled the environment of the hospital kitchen, the articles used in the preparation of the food, and the food itself, but they failed to sample any of the people working in the kitchen to see what strains of E. cold they might possibly be contributing to the food. At the University of Bristol in England, a second group in- vestigated the possible colonization of human intestines by E. cold of animal origin. Initially, these investigators determined only the O-groups, but they later determined O-serotypes and resistance patterns of E. cold from animals and humans. If the two were the same, they also determined H and K antigens to establish conclu- sively the identity of the strains. In one study (Lipton et al., 1977), 15 chickens were purchased from a local retailer over a 3- month period and were distributed to five volunteers who prepared and ate them. After the chickens were thawed prior to cooking, swabs from them were cultured for E. colt. Prior to and after the chickens had been consumed, fecal samples from the volunteers were also cultured for E. colt. When strains from the chickens and the humans had the same O-serotype and antibiotic resistance pattern, the H and K antigens were also determined. No evidence of transfer occurred in 14 of the 15 chickens or in four of the five volunteers. Five strains of E. cold with similar serological and antibiotic resistance characteristics were isolated from one of the chickens and from one volunteer. Two of these strains were isolated only

245 on 1 day, the third after preparation of the chicken for eating. This could hardly be termed colonization. One strain was iso- lated the 10th and 11th days after preparation. There were 13 isolates of this strain from the human and only one from the chicken, suggesting that a strain from the human had contaminated the chicken. Two strains were isolated on the second, third, and fourth days after preparation in roughly equal numbers from each source. Therefore, if one defines colonization as the isolation of a particular strain on three consecutive days from a particular individual, these strains could be said to have colonized this in- dividual. These experiments suggest that the colonization of the human intestine by E. cold of animal origin is a rare event. R-FACTOR TRANSFER In Vivo The transfer of R factors in vivo is not easily accomplished in normal animals and humans. One experiment demonstrating this fact was conducted by Anderson et al. (1973~. They used E. cold isolated from the humans who were subjects of these experiments. The _ cold was isolated from the fecal flora of each subject, chromosomally labelled so it could be identified, and fed back to the same subject. Some strains also had an R factor introduced so that they might serve as donors in experiments on the in-vivo transfer of R factors. Anderson _ al. concluded that No it-factor transfer could be detected in the absence of antibiotic therapy, even though the R- factors concerned could be transferred in laboratory studies from the potential donors to many dif ferent enteric organisms as recip- ients, including the genetically marked potential recipients used in these experiments. When the subjects were treated with ampi- cillin or tetracycline, the organism containing the R factor would persist for a short time in the subjects' intestine allowing R- factor transfer to be demonstrated. This work was conducted with strains indigenous to the human gut flora as both donor and acceptor. An early attempt to demonstrate in-vivo transfer of antibiotic resistance from an E. cold of animal origin to an indigenous human strain was reported by Smith (1969~. The investigator characterized the principal resident E. cold in the alimentary tract of a volunteer and showed that it remained stable throughout the 2-year study. The organism did not contain an R factor but was capable of receiving an

246 R factor _ vitro. This volunteer ingested a series of 18 differ- ent it-factor-containing E. cold isolates--14 of animal origin and four of human origin--at different times during the 2-year period. The animal strains were from pigs, oxen, and fowl. They were shown to be capable of transferring their R factors to the resident strain _ vitro. In this study, Smith showed that none of the donor strains was capable of colonizing the alimentary tract permanently (Table 17~. When administered in a dose of 10 organisms, some of the strains of animal origin could be recovered from the feces for periods of 1 to 11 days, but when the dose was 106 organisms they could not be re- covered even 1 to 2 days after administration. When the strains were of human origin, colonization for as long as a month was achieved when the dose was 10 organisms. Smith also reported that when temporary colonization occurred, transfer of R factors to the resident _ cold took place at an extremely low rate. Only about one to 10 colonies of the' resident strain with an R factor were recovered after doses of 10 donor orga- nisms had been administered. These resident bacteria to which R factors had been transferred did not persist as a permanent part of the _ cold microflora. This suggests that the presence of the R factor reduced the survival potential of the strains in competition with their drug-sensitive forms. No it-factor transfer occurred with any donor strain when the dose of organisms was 10 cells. There is no doubt that carcasses are contaminated with coliform organisms at the abattoir. There are many references attesting to this. Those showing contamination with salmonellae at the abattoir are relevant for E. cold as well. A typical reference for the contamination of carcasses with antibiotic-resistant E. cold is that of Walton (1970~. The most important question is how much of the contamination persists in the retail store and in cooked food. Walton addressed this question in another study (Walton and Lewis, 1971) in which 25 samples from fresh minced beef and sausage and 25 from cooked boiled ham and roast pork obtained from 25 butcher shops were examined for contamination with antibiotic-resistant fecal colifonm organisms. Eleven sausages, three samples of minced beef, two of roast pork, and none of the boiled ham yielded antibiotic-resistant coli- fonm bacteria. All samples, fresh and cooked, were heavily gon- taminated with other types of bacteria. There were up to 10

Hi A) o Hi a, cd o lo' ¢ ED c} to En v o o A) - ~u ED · En lo: o of  x ED ED ED ·- to o c) us a) oO a) lo: o Or o us v ED ¢ so cO ED A] o p ct N o . _I O O Z L) JJ eq o ~ o · ~ o o z c: ~c c) o ~ v u) o o" z ~  0 o o ~ " bo o s~ o o ~ o o o o u~ L ~C~ 1 1- 1 1 o o o o o o z cn oo c ~ 1 1 1 1 o o o o o ~c ~o ~ 1 1 1 1 1 ~o o o ~o  o oo o o o ~oD 1 1 1 1 1 ~o ~o o o ~ s" v o o tn u) u' c~ - 0 ct ~ ~ ~s~ o ~v ooo o ~o ~'~ ~ ~^ :^ Lf, ~00 ~ ~ ~ ~ ~a, o o ~o ~v ~v ~v v ~v v om o ~0 o ~0 0 U~ ~ ~ ~o' ~ ~ C~ oo oo oo oo 0 oo C 0 X 0 P ~0 ~:e ~0 ~ :C 247 cn v tn ~D cr, ~ C) a, V U, tn ~n o ~ . ~=:

248 organisms per gram in the boiled ham and 105 organisms per gram in the roast pork. Approximately 40% of the fresh meat and from 4% to 12% of cooked meat samples carried drug-resistant coliform organisms. Of 100 samples, only two yielded E. cold with trans- ferable drug resistance. There was very little evidence of con- tamination of meats with drug-resistant fecal E. colt, but the meat was heavily contaminated with other organisms, suggesting that the contamination was not from the animals but from humans and the environment. In summary, it seems likely that the E. cold from animals does not play a significant role in colonizing the human intestine. If these organisms should do so, and if they contained R factors, it is unlikely that the R factor would be transferred to the in- digenous microflora. Carcasses are contaminated at the abattoir with fecal E. colt, some with transferable R factors, but it is likely that not much of the contamination reaches the retailer or the consumer. The organisms on the meat at the time of purchase by the consumer are more likely to be from humans or the environment. LINKAGE OF PLASMIDS FOR ENTEROTOXIN PRODUCTION AND ANTIBIOTIC RESISTANCE Falkow and Gyles (1973) have shown that a plasmid coding for enterotoxin production (Ens) could be transferred into an E. cold containing a plasmid for tetracycline resistance. They used 3- week-old weanling pigs in these experiments an] doses of donor and recipient organisms that were greater than 101 organisms. The recipient was given 1 or 4 hours after the donor. Neither organism persisted much longer than 48 hours in the intestines of the pigs in the experiments where transfer took place. The presence of a K88ac plasmid, which enhances colonization, seemed to prevent transfer of the Ent plasmid. The presence of 50 g of oxytetracycline per ton of feed seemed to have no effect on numbers of either donor or recipient recovered from the pigs. No attempt was made to show whether the Ent and resistance plasmids would transfer together. Walton (1977) used established donors of enterotoxin-plasmids to attempt the simultaneous transfer of Ent plasmids and R factors. The transfer plasmids was not de tected when antibiotic-resistant transconjugants were selected. of Ent plasmids from E. cold that contained R Gyles _ al. (1977) isolated a naturally occurring strain of E._ cold from a piglet with diarrhea. This strain, which was resistant to tetracycline, streptomycin, and sulfonamides and which was capable

249 of producing both heat-stable and heat-labile enterotoxin, could transfer all of these characteristics en bloc to other E. colt. However, such a strain would have no selective~advantage in pigs receiving antibiotics in their feed since a high proportion of the _ cold in those pigs would already be resistant to antibiotics. Falkow _ al. (1976) have shown that organisms of a strain of E. cold carrying an Ent plasmid will survive longer and in greater _ ~ numbers than would those of a homogenic strain without the Ent plasmid. The presence of a K99 plasmid, which confers the ability to colonize the intestinal tract of young calves, did not greatly increase the ability of the E. cold with the Ent plasmid to survive in the calves. Food-borne outbreaks of disease due to E. cold are not reported by the CDC in its annual summaries (CDC, 1977b). Such outbreaks are reported in England, but since they are difficult to identify most of them are probably not detected. Enterotoxigenic E. cold is generally recognized as the cause of traveler's diarrhea (Sack, 1978~. This disease is usually attributed to the eating of raw foods or the drinking of polluted water. It is usually not a very serious disease in otherwise healthy adults. Enterotoxigenic E. cold causes two more serious diseases, a choleralike syndrome and infantile diarrhea. Infantile diarrhea is a leading cause of death in children under 5 years of age. Sack et al. (1975) established that diarrhea! disease in 59 Apache children, all under 4 years of age, was associated with enterotoxigenic E. cold rather than the so-called enteropathogenic serotypes that are often associated with infantile diarrhea. The in- vestigators did not investigate the mode of spread of the disease since these children already had the disease when they were admitted to the hospital. Ryder _ al. (1976) investigated an outbreak of diarrhea! illness in a hospital special care nursery. Of 205 infants admitted to this nursery between December 1974 and September 1, 1975, 55 (27%) of them had diarrhea! illness. The diarrhea had a short incu- bation period, sometimes less than 48 hours, and a short duration (mean of 4 days), but it was associated with definite morbidity. Symptomatic, culture-positive infants had a mean hospital stay of 32 days compared to 15 days for asymptomatic culture-negative infants. This illness was associated with the presence of a strain of E. cold of serotype 078:K80:H12, which is resistant to chloramphenicol, ampi- cillin, carbenicillin, kanamycin, sulfonamides, cephalothin, and tetracycline but is sensitive to gentamicin and colistin. A factor specifying a pilus facilitating colonization was also shown to be

250 present. This strain produced a heat-stable enterotoxin. Of the possible risk factors examined in this outbreak, only oral consumption of formula was associated with illness, and it was only from this source that the epidemic strain was isolated. Unopened formula was sterile, but cultures taken from several locations in the nurseries and from the hands of one of the nurses were positive for the organism. A 5-day course of prophylactic colistin was ineffective in- preventing spread of the organism. There were no deaths in this outbreak, but several infants required Dehydration. Apparently, antibiotic resistance in enterotoxigenic E. colt, especially from adults, is not common (DuPont et al., 1978~. Table 18 shows the antibiotic susceptibility of 126 strains of enterotoxigenic E. cold isolated from children in Houston, Texas, students attending a Mexican University, and adults in Bangladesh. All of the strains were sensitive to nalidixic acid, oxolinic acid, and cinoxacin, which are related compounds that are not very impor- tant in the therapy of bacterial infections. Of the 126 strains that were isolated, 119 (92%) were sensitive to sulfamethoxazole- trimethopr~m. All of the strains from the Mexican students that produced stable toxin only were sensitive to the 10 antimicrobials tested. The greatest incidence of resistance occurred in strains that produced labile toxin in the Texas children. The authors sug- gested that naladixic acid or related compounds might be useful in preventing diarrhea due to enterotoxigenic E. colt. In summary, plasmids for antibiotic resistance and for the production of enterotoxin can coexist in an _ cold cell and can be transferred together experimentally. Cells containing both plasmids would have no survival advantage over cells containing the R factor alone in animals receiving antibiotics in their feed since a high percentage of the cells would already be resistant. Enteropathogenic E. cold causes three kinds of disease--a cholera-like disease, traveler's diarrhea, and infantile diarrhea. None of these is particularly common in the United States. There seems to be no connection between these diseases and the use of antibiotics in animal feed. STAPHYLOCOCCAL FOOD POISONING _ Staphylococcal enterotoxin is one of the two most frequently reported causes of outbreaks of food-borne disease. In 1974 and

o ~1 on Id on · - i on o :^ ID Q a, 0 U: Cal o ¢ to to 0 o o o ED CQ - _' U) a) _' Cal 1  ~1 V to U. D V £ TIC 3C ~0 Cal <: FEZ 1 ~1 ~1 . l 1 ~ E~ c £ ~1 o _: ~ ~ E Cal ¢ ~ ~1 ~1 £ £ In Al cn 0 ~n P4 o^ o o o o ~c o o - ' ~- - - ~D ~ O ~ ~c~ oo ^ ^ oo ~0 o ~ '- ~- - - - ~ mo m o o~ - cr' 0 o  o - a' o o ~ o ~o o oo - ~ pc o c, :~ u) ^^ ~ o oc u, - - u~ ~ o ^ o ~ ~ 0 - - ~ oD 8 o C) ~o o _ ~_ 5 C~ o o o o o o r~ ~0 0 0 ~ ~D~ ~ ____ _ + -~ o oo ~ ~ ,_ U~ ~ o o ~ o U. ___~  i,~,0 o C~ ~ o o o C o o ~ ~_4 __ _ n S~ o - a) ^^ ~o ~o ~ a' ~ 0 u ~s~ 0 ~ U. ~ ~0 0 o ~ ~ o ~ / ~o .,1 X ~o ~ ~ ~ c ~ + - ~ 0 0 - - ~ov ^ ~ ~ ~ ~ :^ ~ ~ ~ == e u-- 0= -v~e 00 ~ JJ V ~ :^ .,1 ~1 ~0c, ~ V £ ~ ~=- ~ ~ .~ eo ~ 0 0 ~0 CO 0 ~ ~ V ~ C - - v :^ v :- ~v :^ ~ ~0 ~ ~ e ~ ~ ~ V ~ ~ ~ ~ m~ ~ V 0 · ~O . ~ 0 · ~ . ~ "~ 0 v ~ 0 ~vm 0 ~ ~ n u a n n n 0 1 1 _1 1 1 ~ 1 1 06 1 O~ ~X - - ~ X - 0 ~ U, ~ ~ U, ~ ~ ~ C ~ ~0 P~ : ~ : ~eo ~ ~ ~ E ~ u~ o ~ ¢ ¢ ¢ ~m ~ ~.~.-o~ 251 V o V a, v ~ o X ~ V X o = - o X V ~ o n a n

252 1975 outbreaks of disease from Staphylococcus exceeded those from Salmonella, although Salmonella ranked first in 1976. Because staphylococcal enterotoxin is heat-stable, it is not destroyed by cooking. While meat products are frequently implicated in outbreaks of staphylococcal food-poisoning, investigation frequently reveals that the food has been contaminated by a food handler who is a carrier of an enterotoxin-producing strain of staphylococcus or who has a staphylococcus-infected wound (Bryan, 1978; CDC, 1977b). This contamination of meat by staphylococci of human origin was confirmed in a study by Sinell et al. (1975) in the Federal Republic of Germany. These investigators examined 3,065 cultures of staphylococci from slaughtered pigs, meat plant equipment, and other sources including human beings. Strains originating from swine produced significantly less (21%) enterotoxin than did strains from clinical specimens obtained in hospitals (approximately 40%~. Enterotoxin was not produced by any of 39 phage group II strains, a group that is considered to be specific to animals. The strains from animals were more frequently resistant to antibiotics than were the strains from the hospital. In 1973, before the German govern- ment banned the use of tetracycline in animal feed, there was a high incidence of resistance to tetracycline. In 1974, after the ban, a survey showed that the incidence of resistance to tetracycline in cultures from pigs had dropped but that the incidence of resistance to bacitracin had become very high. Strains of staphylococci iso- lated in a meat plant in 1974 continued to have a high incidence of resistance to both tetracycline and bacitracin. Domestic animals and pets can become carriers of strains of staphylococci of phage types that are frequently associated with human disease (Pagano et al., 1960~. During the school year 1956- 1957 an unusual number of senior students at the University of Pennsylvania School of Veterinary Medicine were afflicted with furuncles, deep cutaneous abscesses, cellulitis, and paronychia. The epidemic continued through the following 2 years. Staphylo- coccus aureus phage type 80/81 was repeatedly isolated from these . lesions. This strain was resistant to penicillin, streptomycin, the tetracyclines, and, occasionally, to erythromycin but was sensitive to chloramphenicol and novobiocin. A survey of students, faculty, and other employees of the school revealed that only the senior students who spent much of their time in the clinic and the faculty of the clinic had a high incidence of this strain in their external Hares. They frequently became carriers after contact with a student with an active lesion.

253 Sporadic culturing of swabs from lesions of animals in the clinic revealed the existence of type 80/81 strain on only three occa- sions, whereas random swabbing of the nostrils of the animals failed to show it. In 1959 a systematic swabbing of the nostrils of the animals showed the strain to be present in cows, horses, goats, and, most frequently, in dogs. There was no proof that the organism was transmitted from an animal to a human. The animal-to-person transfer of antibiotic-resistant S. aureus of human type 80/81 was suggested by Wallace et al. (1962~. Eight of 287 cows tested yielded staphylococci type 80/81 with resistance to streptomycin, tetracycline, and penicillin, and sensitivity to chloramphenicol, neomycin, novobiocin, oleandomycin, and carbomycin. The cows all had mastitis. Some cases were so severe that the ani- mals had to be destroyed. Three people in the family that operated the dairy studied by Wallace _ al. also yielded type 80/81 with the same resistance pattern when swabs were taken from their nares and from two boils on the father at different times and from one boil and the sore throat of an 18-month-old child. Here again the spread from animal to per- son is putative. The organism could have spread from the humans to the animals. Antibiotic resistance in staphylococci from mastitic cows is common because large amounts of antibiotics are used in intramammary infusions to treat the mastitis, which may clear up but frequently recurs. Another study on antibiotic-resistant S. aureus from dairy herds was reported by Devriese and H~mmez (1975~. In 1971 and from July 1972 to June 1973 they isolated 68 methicillin-resistant strains of _ aureus from milk samples from mastitic cows in 20 Belgian dairy herds. In the 1972-1973 survey 52 methicillin-resistant strains were isolated on nine farms, representing 50% of all S. aureus iso- lated on those farms. A determination of the biological and phage- types of the methicillin-resistant strains suggested that they were actually strains from humans and might have had a common human source, but that seems highly speculative. In summary, food poisoning due to Staphylococcus aureus entero- toxin is frequently reported in the United States, England, and Wales. It often involves meat products, but upon investigation it is fre- quently found that the food was contaminated by a human carrier or one with a staphylococcal lesion. There is no evidence that the use of antibiotics in animal feed is in any way involved.

254 Frequently, staphylococci of phase types usually considered to be of human origin have been found in pets and dairy cows. It is likely that these organisms were transferred from humans and were resistant when they were acquired by the cows, but they may have become resistant during treatment for mastitis by antibiotic intra- mammary infusion. It is highly unlikely that the resistance arose due to the use of antibiotics in the feed of the cows since that is a rather uncommon practice in the dairy industry in the United States. CONCLUSIONS Salmonellosis is the most important of the food-borne diseases that may be transmitted from animals to humans. There is a large pool of Salmonella in domestic animals, carcasses are contaminated with Salmonella at the slaughterhouse, and outbreaks of food-borne salmonellosis due to contaminated food of animal origin are a seri- ous threat to human health and an economic problem. It is not clear that the Salmonella on the food originated in the animals. Moreover, when antibiotic resistance has occurred, it does not seem to have been due to the use of antibiotics in animal feed. It seems likely that antibiotic-resistant E. cold in animals does not play a significant role in colonizing the human intestine. If these organisms do colonize temporarily, it is unlikely that they transfer their R factor to the human flora. Carcasses are contami- nated at the slaughterhouse with E. coli--some of which is resistant to antibiotics, but it is doubtful that more than a few of these organisms reach the consumer. The organisms on the meat are probably of human or environmental origin. Plasmids for antibiotic resistance and enterotoxin production can coexist in an _ cold cell and can be transferred together ex- perimentally. Cells containing both genes would have no survival advantage over cells containing the R factor alone in animals receiving antibiotics in feed since a high percentage of the cells would already be resistant. Food poisoning in the United States due to Staphylococcus enterotoxin is second only to salmonellosis as a food-borne disease. It often involves meat products, but upon investigation, it is fre- quently found that the food was contaminated by a human carrier or one with a Staphylococcus-infected lesion. Staphylococci of phase types usually considered to be of human origin have been found in pets and dairy cows. There is no evidence that the use of antibio- tics in animal feed is in any way involved.

255 REFERENCES Anderson, E. S. 1968a. Drug resistance in Salmonella typhimurium and its implications. Br. Med. J. 3:333-339. Anderson, E. S. 1968b. Transferable drug resistance. Sci. J. 4: 71-76. Anderson, J. D., W. A. Gillespie, and M. H. Richmond. 1973. Chemo- therapy and antibiotic-resistance transfer between enterobac- teria in the human gastro-intestinal tract. J. Med. Microbiol. 6:461-473. Aserkoff, B., and J. V. Bennett. 1969. Effect of antibiotic therapy in acute salmonellosis on the fecal excretion of salmonellae. N. Engl. J. Med. 281:636-640. Bryan, F. L. 1978. Factors that contribute to outbreaks of food- borne disease. J. Food Protection 41:816-827. Center for Disease Control. 1975. Current trends. Microbiologic standards for raw ground beef, cold cuts and frankfurters. Reported by the U. S. Department of Agriculture and the Center for Disease Control. Morbid. Mortal. Weekly Rep. 24:229-230. Center for Disease Control. 1976a. Salmonella bovis-morbificans in precooked roasts of beef. Reported by P. J. Checko, J. N. Lewis, R. Altman, K. Black, H. Rosenfeld, W. Parkin, the U.S. Department of Agriculture, and the Center for Disease Control. Morbid. Mortal. Weekly Rep. 25:333-334. Center for Disease Control. 197 6b. Salmonella Surveillance Annual Summary 1975. Center for Disease Control, Atlanta, Ga. 51 pp. Center for Disease Control. 1977a. Follow-up on salmonellae in precooked roasts of beef. Reported by J. N. Lewis, R. Altman, D. O. Lyman, W. E. Parkin, the U. S. Department of Agriculture, the Food and Drug Administration, and the Center for Disease Control. Morbid. Mortal. Weekly Rep. 26:394-399. Center for Disease Control. 1977b. Food Borne and Water Borne Disease Outbreaks. Annual Summary 1976. Center for Disease Control, Atlanta, Ga. 82 pp.

256 Center for Disease Control. 1977c. Multi-state outbreak of Sal- monella newport transmitted by precooked roasts of beef. Reported by P. J. Checko, J. N. Lewis, R. Altman, G. Halpin, R. Inglis, M. Pierce, K. Pilot, J. Prince, W. Rednor, M. Fleissner, D. O. Lyman, W. E. Parkin, the U. S. Department of Agriculture, and the Center for Disease Control. Morbid. Mor- tal. Weekly Rep. 26:277-278. Center for Disease Control. 1977d. Salmonella Surveillance Annual Summary 1976. Center for Disease Control, Atlanta, Ga. [23] pp. Center for Disease Control. 1977e. Salmonellosis associated with home-made ice cream--Michigan. Reported by R. Locey, P. Owens, G. Markakis, R. P. Daniels, D. F. Fuller, Jr., D. Muentener, N. S. Hayner, K. S. Read, and the Center for Disease Control. Morbid. Mortal. Weekly Rep. 26:94-99. Center for Disease Control. 197 7f. Salmonellosis--Kentucky. Re- ported by R. N. McLeod, W. L. Adams, L. M. Mullins, M. A. Shepherd, R. K. Bonner, B. F. Brown, N. J. Cambron, C. Hernandez, G. E. Killgore, J. W. Skaggs, [Kentucky] Epidemio- logical Notes and Reports 12~4~: 1, and the Center for Disease Control. Morbid. Mortal. Weekly Rep. 26:239. Center for Disease Control. 1979. Salmonella Surveillance Annual Summary 1977. Center for Disease Control, Atlanta, Ga. [21] PP Cherubin, C. E., M. Szmuness, and J. Winter. 1972. Antibiotic resistance of Salmonella. Northeastern United States--1970. N. Y. State J. Med. 72:369-372. Cherubin, C. E., J. F. Timoney, M. F. Sierra, P. Ma, J. Marr, and S. Shin. 1980. A sudden decline in ampicillin resistance in Salmonella typhimurium. J. Am. Med. Assoc. 243:439-442. Cooke, E. M., A. L. Breaden, R. A. Shooter, and S. M. O' Farrell. 1971. Antibiotic sensitivity of Escherichia cold isolated from animals, food, hospital patients, and normal people. Lance t 2:8-lO. Devriese, L. A., and J. Hommez. 1975. Epidemiology of methi- cillin-resistant Staphylococcus aureus in dairy herds. Res. Vet. Sci. 19:23-27.

257 DuPont, H. L., H. West, D. G. Evans, J. Olarte, and D. J. Evans, Jr. 1978. Antimicrobial susceptibility of enterotoxigenic Escherichia colt. J. Antimicrob. Chemother. 4:100-102. Edel, W., M. van Schothorst, P. A. M. Guin£e, and E. H. Kampelmacher. 1974. Preventive measures to obtain Salmonella-free slaughter pigs. Zentralbl. Bakteriol. Parasitenkd. Infektionskr. Hyg., I Abt. Orig. Reihe B 158:568-577. Evangelisti, D. G., A. R. English, A. E. Girard, J. E. Lynch, and I. A. Solomons. 1975. Influence of subtherapeutic levels of oxytetracycline on Salmonella typhimurium in swine, calves, and chickens. Antimicrob. Agents Chemother. 8:664-672. Falkow, S., and C. Gyles. 1973. Studies on In viva transfer of the Ent plasmid. Progress Report. FDA contract number 73-210 M #1. Food and Drug Administration, Rockville, Md. Falkow, S., L. P. Williams, Jr., S. L. Seaman, and L. D. Rollins. 1976. Increased survival in calves of Escherichia cold K-12 carrying an Ent plasmid. Infect. Immun. 13:1005-1007. Finlayson, M. 1977. Salmonella in Alberta poultry products and their significance in human infections. Pp. 156-180 in Proceedings of the International Symposium on Salmonella and Prospects for Control, June 8-11, 1977, University of Guelph, Guelph, Ontario, Canada. Food and Drug Administration. 1973. Antibiotic and sulfonamide drugs in the feed of animals. Fed. Reg. 38:9811-9814. Galton, M. M., M. V. Smith, H. B. McElrath, and A. B. Hardy. 1954. Salmonella in swine, cattle and the environment of abattoirs. J. Infect. Dis. 95:236-245. Garside, J. S., R. F. Gordon, and J. F. Tucker. 1960. The emer- gence of resistant strains of Salmonella typhimurium in the tissues and alimentary tracts of chickens following the feeding of an antibiotic. Res. Vet. Sci. 1:184-199. Girard, A. E., A. R. English, D. G. Evangelisti, J. E. Lynch, and I. A. Solomons. 1976. Influence of subtherapeutic levels of a combination of neomycin and oxytetracycline on Salmonella typhimurium in swine, calves, and chickens. Antimicrob. Agents Chemother. 10:89-95.

258 Groves, B. I., N. A. Fish, and D. A. Barnum. 1970. An ep~demio- logical study of Salmonella infection in swine in Ontario. Can. J. Public Health 61:396-401. Guin'ee, P. A. M. 1963. Preliminary investigations concerning the presence of E. cold in man and in various species of animals. Zentralbl. Bakteriol. Parasitenkd. Infektionskr. Hyg., I. Abt. Orig. Reihe A 188:201-218. Gustafson, R. H., J. D. Kobland, and P. H. Langner. 1976. Inci- dence and antibiotic resistance of Salmonella in market swine. P. M2 in Proceedings of the International Pig Veterinary Society, 4th International Congress, June 22-24, 1976, Ames, Ia. International Pig Veterinary Society, Ames, Ia. Gutzmann, F., H. Layton, K. Simkins, and H. Jarolmen. 1976. In- fluence of antibiotic-suppl~mented feed on occurrence and per- sistence of Salmonella typhimurium in experimentally infected swine. Am. J. Vet. Res. 37:649-655. Gyles, C. L., S. Palchaudhuri, and W. K. Mans. 1977. Naturally occurring plasmid carrying genes for enterotoxin production and drug resistance. Science 198:198-199. Hirsh, D. C., and N. Wiger. 1977. Effect of tetracycline upon transfer of an R plasmid from calves to human beings. Am. J. Vet. Res. 38:1137-1139. Hook, E. W., and W. D. Johnson. 1972. Nontyphoidal salmonellosis. Pp. 583-591 in P. D. Hoeprich, ed. Infectious Diseases. A Guide to the Understanding and Management of Infectious Pro- cesses. Harper and Rowe, Hagerstown, Md. Jarolmen, H., R. J. Sairk, and B. F. Langworth. 1976. Effect of chlortetracycline feeding on the Salmonella reservoir in chickens. J. Appl. Bacteriol. 40:153-161. Layton, H. W., B. F. Langworth, H. Jarolmen, and K. L. Simkins. 1975. Influence of chlortetracycline and chlortetracycline + sulfamethazine supplemented feed on the incidence, persistence and antibacterial susceptibility of Salmonella typhimurium in experimentally inoculated calves. Zentralbl. Veterinaermed. Reihe B 22:461-472. Levy, S. B., G. B. FitzGerald, and A. B. Macone. 1976. Spread of antibiotic-resistant plasmids from chicken to chicken and from chicken to man. Nature 260:40-42.

259 Linton, A. H., K. Howe, P. M. Bennett, M. H. Richmond, and E. J. Whiteside. 1977. The colonization of the human gut by anti- hiotic-resistant Escherichia cold from chickens. J. Appl. Bacterial. 43:465-469. Luther, H. G., W. G. Huber, D. Siegel, and H. G. Luther, Jr. 1974. Antibacterial feed additives: Residues and infectious drug resistance. Pp. 89-134 in W. W. Hawkins, ed. Drug-Nutrient Interrelationships. Nutrition and Pharmacology--An Inter- phase of Disciplines. Proceedings of the Miles Symposium '74, presented by The Nutrition Society of Canada, June 24, 1974, McMaster University, Hamilton, Ontario, Canada Ministry of Agriculture, Fisheries, and Food. 1965. Salmonellae in cattle and their feedingstuffs, and the relation to human infection. A report of the Joint Working Party of the Veteri- nary Laboratory Services of the Ministry of Agriculture, Fisheries, and Food, and the Public Health Laboratory Service. J. Hyg., Camb. 63:223-241. Neu, H. C., E. B. Winshell, J. Winter, and C. E. Cherubin. 1971. Antibiotic resistance of Salmonella in the northeastern United States, 1968-1969. N. Y. State J. Med. 71:1196-1200. Neu, H. C., C. E. Cherubin, E. D. Longo, B. Flouton, and J. Winter. 1975. Antimicrobial resistance and it-factor transfer among isolates of Salmonella in the northeastern United States: A comparison of human and animal isolates. J. Infect. Dis. 132: 617-622. Pagano, J. S., S. M. Farrer, S. A. Plotkin, P. S. Brachman, F. R. Fekety, and V. Pidcoe. 1960. Isolation from animals of human strains of staphylococci during an epidemic in a veterinary school. Science 131:927-928. Pocurull, D. W., S. A. Gaines, and H. D. Mercer. 1971. Survey of infectious multiple drug resistance among Salmonella isolated from animals in the United States. Appl. Microbiol. 21:358- 362. Ryder, R. W., I. K. Wachemuth, A. E. Buxton, D. G. Evans, H. L. DuPont, E. Mason, and F. F. Barrett. 1976. Infantile diarrhea produced by heat-stable enterotoxigenic Escherichia coli. N. Engl. J. Med. 295:849:853.

260 Sack, R. B. 1978. The epidemiology of diarrhea due to enterotoxi- genic Escherichia colt. J. Infect. Dis. 137:639-640. Sack, R. B., N. Hirschhorn, I. Brownlee, R. A. Cash, W. E. Woodward, and D. A. Sack. 1975. Enterotoxigenic Escherichia-coli-asso- ciated diarrhea! disease in Apache children. N. Engl. J. Med. 292:1041-1045. Shooter, R. A., S. A. Rousseau, E. M. Cooke, and A. L. Breaden. 1970. Animal sources of common serotypes of Escherichia cold in the food of hospital patients. Possible significance in urinary-tract infections. Lancet 2:226-228. Sinell, H. J., D. Kusch, and F. Untenmann. 1975. Enterotoxin-pro- ducing Staphylococci in meat processing plants: Origin, bio- chemical properties, resistance to antibiotics. Eur. J. Appl. Microbial. 1:239-245. Smith, H. W. 1969. Transfer of antibiotic resistance from animals and human strains of Escherichia cold to resident E. cold in the alimentary tract of man. Lancet 1:1174-1176. Smith, H. W. 1970. The transfer of antibiotic resistance between strains of enterobacteria in chickens, calves and pigs. J. Med. Microbial. 3:165-180. Swann, M. M., K. L. Blaxter, H. I. Field, J. W. Howie, I. A. M. Lucas, E. L. M. Millar, J. C. Murdoch, J. H. Parsons, and E. G. White. 1969. Report of the Joint Committee on the Use of Antibiotics in Animal Husbandry and Veterinary Medicine. Cmnd. 4190. Her Majesty's Stationery Office, London. 83 pp. Threlfall, E. J., L. R. Ward, and B. Rowe. 1978. Epidemic spread of a chloramphenicol-resistant strain of Salmonella typhimurium phage type 204 in bovine animals in Britain. Vet. Rec. 103: 438-440. Timoney, J. F. 1978. The epidemiology and genetics of antibiotic resistance of Salmonella typhimurium isolated from diseased animals in New York. J. Infect. Dis. 137:67-73. Vernon, E. 1977. Food poisoning and Salmonella infections in England and Wales, 1973-75. Public Health, London 91:225-235.

261 Wallace, G. D., W. B. Quisenberry, R. H. Tanimoto, and F. T. Lynd. 1962. Bacteriophage type 80/81 staphylococcal infec- tion an human beings associated with mastitis in dairy cattle. Am. J. Public Health 52:1309-1317. Walton J. R. 1966. In viva transfer of infectious drug resist , _ ance. Nature 211:312-313. Walton, J. R. 1970. Contamination of meat carcasses by antibiotic- resistant coliform bacteria. Lancet 2:561-563. Walton, J. R. 1977. The relationship between the transferability of enterotoxin plasmids and the simultaneous transfer of anti- biotic resistance plasmids. Zentralbl. Veterinasnmed. Reihe B 24:317-324. Walton, J. R., and L. E. Lewis. 1971. Contamination of fresh and cooked meats by antibiotic-resistant coliform bacteria. Lance t 2:255-257. Williams, L. P., Jr., J. B. Vaughn, A. Scott, and V. Blanton. 1969. A ten-month study of Salmonella contamination in animal protein meals. J. Am. Vet. Med. Assoc. 155:167-174. Williams, R. D., L. D. Rollins, D. W. Pocurull, M. Selwyn, and H. D. Mercer. 1978. Effect of feeding chlortetracycline on the reservoir of Salmonella typhimurium in experimentally infected swine. Antimicrob. Agents Chemother. 14:710-719. r