Original research

September and October, 1998

Effect of tylosin on an experimental Salmonella infection in pigs

Thomas R. Shryock, PhD; Robert A. Elliott, PhD; Travis H. Bennett, MS; Rodney P. Basson, PhD; Richard E. Bowen, DVM, PhD

TRS, RAE, THB, RPB, REB: Animal Science Research, Elanco Animal Health, 2001 W. Main Street, GL14, Greenfield, Indiana 46140; E-mail: trs@lilly.com

Also available as an Adobe Acrobat PDF (apx. 190k)


Objective: To evaluate the effect of tylosin on an experimental Salmonella typhimurium infection.

Method: Tylosin was added to the ration of pigs infected with Salmonella (INFMED pigs) at 100 g per ton and fed for a period of 56 days post-inoculation and compared to an infected, nonmedicated control group (INF). Fecal samples were taken at least weekly for 56 days from both INF and INFMED pigs. Representative Salmonella isolates from each timepoint and pig were tested for susceptibility to 12 different antimicrobial compounds. At the termination of the study, samples of liver, spleen, ileocecal lymph node, and colon contents were examined for the presence of Salmonella.

Results: There was no difference in the overall mean log10 counts of Salmonella between the INFMED and INF groups (P=.75), nor in the prevalence of positive fecal samples between the two groups (P=.66). Medication significantly reduced the mean duration of Salmonella shedding in INFMED pigs compared to INF pigs (P=.05). There was no difference between treatment groups in the number of Salmonella isolations from tissues taken at necropsy nor on the antimicrobial susceptibility of Salmonella isolates from the fecal or tissue samples.

Implications: Feeding tylosin to pigs experimentally inoculated with S. typhimurium had no effect on the quantity of Salmonella in the feces or on prevalence of Salmonella in the population of pigs tested. Treatment with tylosin did not change susceptibility to 12 antibiotics. However, treatment with tylosin did reduce the duration of Salmonella shedding in the feces. These findings suggest that tylosin did not reduce the competitive exclusion capacity of the normal intestinal bacteria.

Keywords: swine, Salmonella, tylosin

Received: November 19, 1997
Accepted: May 22, 1998

Antibiotics and antimicrobial agents fed to pigs have been consistently associated with an improvement in feed efficiency and growth rate.1 In order to receive United States regulatory approval for feed additives used as performance enhancers, the manufacturer of the product must investigate whether feeding the product will have an adverse microbiological effect on the animals or on humans. Feeding an antibiotic might:

  • increase the number of Salmonella in the environment by increasing the quantity in the feces,
  • elevate the prevalence of animals harboring the pathogen,
  • extend the duration of shedding, or
  • cause the acquisition of new resistance determinants.2-4

The rationale for using Salmonella as an indicator of gut effects is based on the possibility that performance-enhancing antibiotics could suppress susceptible intestinal Gram-positive microflora, thereby disrupting their competitive exclusion protection, and allow a relative increase in Salmonella concentrations.

Macrolide antibiotics are administered to livestock for both performance enhancement and therapy.5 Tylosin is a macrolide antibiotic that has been approved for use in swine feeds in the United States at 10-40 g per ton to increase rate of weight gain and improve feed efficiency, or at 100 g per ton to control swine enteric diseases, such as porcine proliferative enteropathy.6 Tylosin is active primarily against Gram-positive bacteria, but is also active against Mycoplasma, spirochetes, chlamydiae, and some noncoliform Gram-negative bacteria.7

The purpose of this study was to determine the effect of feeding the highest approved concentration of tylosin--representing the greatest selection pressure for resistance emergence--on the quantity, prevalence, and duration of Salmonella shedding, as well as to determine the antimicrobial susceptibility of Salmonella after an experimental infection in pigs.

Materials and methods

Study design

Thirty-five 7-week-old Chester White-Hampshire-Duroc pigs were obtained from a local breeder, weighed, and identified by ear tag. The pigs, whose mean weight on the first day of the trial was 11 kg (25 lb) (Figure 1), were stratified on the basis of weight and sex and assigned to one of the following treatments:

  • INFMED pigs (n = 10) inoculated with Salmonella by oral gavage on 2 consecutive days (days 0 and 1 of the trial) with 10 mL of the inoculum suspended in a buffered salt solution (hemaglutination or HA buffer), and subsequently fed tylosin-treated ration;
  • INF pigs (n = 10) inoculated with Salmonella in the same manner as the INFMED pigs, and subsequently fed nonmedicated control ration; and
  • CONTROL pigs (n = 15) that were not infected with Salmonella and received a nonmedicated control ration.

Each inoculated animal received a total of 5.2 x 1010 colony-forming units (cfu) in the first dose and 5.6 x 1010 cfu in the second dose, as determined by triplicate plate counts of the inoculum suspensions. Feed and water were withheld from the pigs from approximately 24 hours before the first inoculation until approximately 2 hours afterward. Control group pigs were then provided with ad libitum access to feed for the duration of the study. Feed and water were made available for 1 hour to the INF and INFMED pigs and then were withheld from approximately 20 hours before the second inoculation until approximately 1 hour after the second dose was administered.

Pigs were housed by treatment in 35 separate, contiguous rooms in three isolation buildings, with one pig per room. Within each building, pigs were placed so that a CONTROL-group pig was housed in the room between each INFMED and INF pig.


Building sides were each handled as a unit by a single technician, who wore new disposable coveralls, plastic boots, and disposable gloves each time he entered a room. Traffic of other visitors and personnel was restricted and controlled. The animals were cared for in the same order, from one end of a building side to the other, throughout the study.

Salmonella strain

The Salmonella inoculation strain was a swine-origin S. typhimurium (298-1NA) that displayed a typical reaction on a triple sugar iron (TSI) agar slant and was agglutinated by group-B somatic typing serum. It was susceptible to amikacin, ampicillin, carbenicillin, cephalothin, chloramphenicol, gentamycin, kanamycin, triple sulfa, tetracycline, and trimethoprim/sulfamethoxazole. The strain was categorized as intermediate in susceptibility to streptomycin and resistant to nalidixic acid by disk diffusion antimicrobial susceptibility tests as described below. The tylosin minimum inhibitory concentration (MIC) was 1024 µg per mL.

Performance measures

Individual body weights were taken the day before pigs were placed in the isolation rooms and on the day the trial ended (Figure 1).

Rectal temperatures were taken daily from 4 days prior to the first inoculation to 7 days after the second inoculation. General health conditions were noted throughout the experiment.

Feed disappearance was determined by weighing back feed daily from 5 days prior to the first inoculation to 6 days after the second inoculation, and then weekly for the remainder of the study.


Prior to inoculation, all pigs were fed a complete corn-soybean ration containing 16% protein. After inoculation, the infected INF and CONTROL groups continued to be fed this ration. The INFMED group received the same base ration medicated with 100 g tylosin premix per ton of feed. Each batch of feed was tested for and determined to be free of the presence of Salmonella and/or contaminating concentrations of antibiotics and antimicrobials.

Fecal sampling periods

Fecal samples were obtained from each pig at 14 and 7 days prior to and on the days of inoculation. Samples were cultured to confirm the absence of indigenous Salmonella.

Fecal samples were taken for Salmonella quantitation or detection at 3, 5, 8, 10, 12, 15, 22, 29, 36, 43, 50, and 57 days of the study.

At the end of the 57-day study period, all surviving animals in each of the three groups were euthanized, and necropsy was performed by a veterinarian. Approximately 1-g samples of liver, spleen, ileocecal lymph node, and colon content were aseptically collected for Salmonella isolation by an enrichment procedure. Necropsy samples were also taken from one pig that died during the trial.

Bacteriological culture

Three-gram samples of pretreatment fecal and feed samples were cultured for extraneous Salmonella contamination, using a brilliant green broth (BGB) enrichment, and subsequently plated onto a BG agar, as previously described.8

Fecal samples (3 g) and necropsy samples (1 g) were cultured by the same enrichment procedure, followed by plating onto MacConkey agar supplemented with 50 µg per mL nalidixic acid, as previously described.8

A maximum of five Salmonella-suspect colonies per sample were taken from the quantitative count agar plates, or from positive enrichment plates when quantitative plates were negative or not used, and inoculated onto TSI slants. The slants were incubated overnight at 37 degrees C and observed for TSI reactions typical of Salmonella. Isolates giving typical reactions were serologically typed with specific diagnostic serum to confirm the identity as group B Salmonella.

Confirmed S. typhimurium isolates were tested for antimicrobial susceptibility by the disk diffusion method.9 Susceptibility was determined to the same group of antibiotics as those identified during the inoculation strain characterization. Owing to the high MIC of tylosin, and given that enteric Gram-negative bacteria are impermeable to macrolides, no further testing was warranted.

Statistical analysis

Because pigs were housed individually, the pig was the experimental unit in all statistical analysis. Duration of shedding was not submitted to repeated-measures analysis.

SAS computer software (SAS, Cary, North Carolina) was used to compare study groups and Salmonella counts using repeated-measures ANOVA. Analysis to compare duration of Salmonella shedding was based on the last-day counts (log10 > 0.7), because all surviving animals were shedding at least qualitatively detectable amounts of Salmonella on the last sampling day. The effect of treatment on Salmonella prevalence was analyzed using a log-linear model (Grizzle-Starmer-Koch procedures).


Feed assays

Feed tests demonstrated an acceptable homogeneity and stability of tylosin in the feed, the absence of other contaminating antibiotics, and all feed samples were negative for Salmonella contamination.

Clinical data

Pig temperature

Initial pre-inoculation temperatures fluctuated around 38.3-38.9 degrees C (101-102 degrees F). Mean body temperature decreased to 38 degrees C (~100.5 degrees F) on Day 0 (first day of challenge). Body temperatures greater than 39.4 degrees C (103 degrees F) were observed in only eight of the INFMED- and INF-group pigs on days 1-5 of the study, at which time temperatures rapidly returned to pre-inoculation levels (Figure 2).

Pig weight

Pig weight gain did not differ significantly among study groups. There were no remarkable clinical findings other than the death of one INFMED pig, which was attributed to acute pneumonia, peritonitis, and enteritis.

Feed disappearance

Feed disappearance of all pigs was reduced immediately after inoculation (Figure 3). Mean feed disappearance in the INFMED and INF pigs returned to approximately three-fourths of the pre-inoculation level by day 4 of the study. By day 8 of the study, feed disappearance for the infected pigs no longer differed significantly from that of the non-infected controls. Thus, the INFMED pigs consumed the same amount of feed as the INF pigs, with feed disappearance in both groups returning to that of the CONTROL group by day 8 of the study.

Salmonella isolations

No Salmonella were detected in any of the fecal samples taken prior to inoculation, including samples taken on the first day of inoculation (day 0). A nalidixic-acid-resistant Salmonella was isolated from a single CONTROL-group pig on day 57. All other CONTROL-group samples were negative for the inoculated Salmonella strain.

Salmonella quantitation

Quantitative Salmonella log10 counts were obtained from the majority of the animals for the first four samplings (Figure 4). Some pigs then exhibited intermittently quantifiable fecal samples.

Initial average log10 counts were 5.20 for INFMED and 5.0 for INF pigs. Subsequently, log10 counts decreased over time, but remained over 1.0 in both infected groups throughout the observation period. The overall mean log10 counts (2.12 for INFMED group pigs and 2.14 for INF-group pigs) did not differ significantly (P =.75).

Duration of Salmonella

The duration of Salmonella shedding in INFMED-group pigs was significantly shorter than that of INF-group pigs (P =.05) (Table 1).

Prevalence of Salmonella

Salmonella prevalence did not differ significantly between INFMED- and INF-group pigs (P =.66) (Table 1).

Salmonella isolation from tissues

All samples from the pig that died on test were positive. Numerically similar results were obtained from pigs in the INFMED and INF groups (Figure 5).

Antibiotic susceptibility of isolated Salmonella

A total of 110 fecal samples from INFMED pigs yielded 107 isolates for susceptibility testing. INF pigs provided 110 isolates from 115 samples. The CONTROL group had only one isolate from 180 samples. Isolates from five necropsy samples from the pig that died during the study and from 33 enrichment-positive samples taken at necropsy at the end of the study were also tested. Results were not obtained for two isolates (pig No. 366 on day 22 and pig No. 361, colon content at necropsy) nor for two antibiotics of one isolate of pig No. 377 on day 22. Except for the expected resistance to nalidixic acid, none of the isolates developed resistance to any of the antimicrobial agents.


Persistent colonization was successfully established in all infected pigs after they were administered a high dose of Salmonella. Detecting Salmonella in only one of 15 CONTROL pigs on the last day of sampling confirmed the sensitivity of the test system to detect contamination resulting from failed biosecurity measures or from a "silent" carrier pig. Thus, it is unlikely that cross-contamination or carrier status among the experimental groups confounded our results.

The proportion of culture-positive INF pigs during this study is consistent with results reported for other induced-model infections.10,11 Other researchers have investigated Salmonella colonization and persistence within tissues, mortality, and pig temperatures using induced S. typhimurium model infections in swine.10-12 Wood, et al.,11,12 reported minimal recovery of Salmonella from the liver and spleen, with 45% positive ileocolic lymph nodes and 71% recovery from the cecum, 3% mortality,11 and body temperatures that spiked immediately after inoculation and returned to baseline by day 4.11,12 The consistency of our results to these frequencies of recovery, mortality, and temperature patterns substantiates the validity of the experimental model.

The results of the present study on tylosin are similar to those reported for other feed additive antibiotics evaluated in Salmonella infection models. Evangelisti, et al.,10 found only one instance of an oxytetracycline-resistant isolate when tested against eight antibiotics during their evaluation of the effect of 150 g per ton oxytetracycline on S. typhimurium in swine. Furthermore, Evangelisti, et al.,10 observed no difference in the quantity of Salmonella shedding, prevalence, or duration of infection between treated and nontreated pigs. Chlortetracycline at 55 mg per kg did not increase or prolong shedding of a tetracycline-susceptible S. typhimurium in a model infection,8 whereas virginiamycin fed at 55 mg per kg tended to numerically increase and prolong shedding, although the differences were not significant. In another study that investigated the effect of feeding a combination of aureomycin (100 g per ton), sulfamethazine (100 g per ton), and penicillin (50 g per ton) on pigs orally inoculated with S. choleraesuis, medicated pigs had reduced mortality and clinical illness associated with increased weight gain and feed efficiency compared to the nonmedicated group.13 Moreover, fewer Salmonella were isolated from medicated pigs than from nonmedicated pigs; thus, prevalence and duration of shedding were reduced in the presence of the antibiotic.


  • The use of tylosin at the highest approved dose (100 g per ton) did not cause increased Salmonella excretion and reduced the duration of shedding.
  • The use of tylosin in the presence of Salmonella did not appear to disrupt the competitive exclusion capacity of the normal gut microflora.
  • The use of tylosin did not select for Salmonella with new resistance phenotypes.


The expert assistance of Patti Lawhorn, Alan Zimmermann, and J. Mitchell Staples in the preparation of this manuscript is gratefully acknowledged.


1. Zimmerman DR. Role of subtherapeutic levels of antimicrobials in pig production. J Anim Sci.1986;62(3):6-17.

2. Solomons IA. Antibiotics in animal feeds: Human and animal safety issues. J Anim Sci.1978;46(5):1360-1368.

3. Van Houweling CD, Gainer JH. Public health concerns relative to the use of subtherapeutic levels of antibiotics in animal feeds. J Anim Sci.1978;46(5):1413-1424.

4. Corpet DE. Microbiological hazards for humans of antimicrobial growth promoter use in animal production. Revue Med Vet.1996;147(12):851-862.

5. Kirst HA. Macrolide antibiotics in food-animal health. Exp Opin Invest Drug.1997;6(2):103-117.

6. McOrist S, Morgan J, Veenhuizen MF, Lawrence K, Kroger HW. Oral administration of tylosin phosphate for treatment and prevention of proliferative enteropathy in pigs. Amer J Vet Res.1997;58(2):136-139.

7. Prescott JF, Baggott JD. Lincosamides, macrolides, and pleuromutilins. In: Antimicrobial Therapy in Veterinary Medicine. Second ed. Iowa State University Press: Ames, Iowa.1993.179-204.

8. Jones FT, Langlois BE, Cromwell GL, Hays VW. Effect of feeding chlortetracycline or virginiamycin on shedding of Salmonellae from experimentally infected swine. J Anim Sci.1983;57(2):279-285.

9. Anon. Methods for antimicrobial susceptibility tests for bacteria that grow aerobically. Wayne, Pennsylvania: NCCLS Document M2-T3. 1983;19087.

10. Evangelisti DG, English AR, Girard AE, Lynch JE, Solomons IA. Influence of subtherapeutic levels of oxytetracycline on Salmonella typhimurium in swine, calves and chickens. Antimicrob Agents Chemother.1975;8:664-672.

11. Wood RL, Rose R. Populations of Salmonella typhimurium in internal organs of experimentally infected carrier swine. Am J Vet Res.1992;53(5):653-658.

12. Wood RL, Pospishil A, Rose R. Distribution of persistent Salmonella typhimurium infection in internal organs of swine. Am J Vet Res.1989;50(7):1015-1021.

13. Schwartz KJ, Lucas TE. Induced Salmonella choleraesuis infection in pigs: The effect of Aureo SP-250 versus no drug on clinical disease, duration of shedding, and tissue localization. Agri-Practice.1994;15(2):7-14.