Prevalence of Salmonella in dairy herds in Alberta

Abstract

Fifty dairy herds in Alberta were tested for the presence of Salmonella. Four (8%) dairy herds had at least 1 cow shedding Salmonella. Different isolates were identified by serotyping, phage typing, and antibiotic resistance patterns. Pulsed-field gel electrophoresis patterns were determined for unique isolates.

Salmonellosis is a major cause of human foodborne illness in North America. Salmonella infections are of increasing concern due to the emergence and increased prevalence of multiantibiotic resistant strains, such as Salmonella enterica subsp. enterica serovar Typhimurium DT104 (1,2,3). Data on the prevalence of Salmonella shedding in dairy cattle in Alberta are not available. This study was initiated to determine the baseline prevalence of Salmonella shedding in dairy herds in Alberta.

Two hundred randomly chosen dairy herd improvement listed dairy farms were solicited to participate in this study. Fecal samples were collected throughout Alberta during the summer of 1999 from 50 of the responding dairy farms. The 2248 samples collected were combined into 750 pools representing 2 or 3 cows from the same herd. Fecal pools were processed for Salmonella within 36 h of collection. Pooled samples were mixed thoroughly and 5 g from each fecal pool was processed in each of 2 recovery pathways, as described previously (4). This double pathway isolation method has been shown to have a high analytical sensitivity (about 3 Salmonella bacteria per 5 g, unpublished data).

Salmonella were detected in 5 (0.67% ± 0.57%) of the 750 pools tested. Thus between 5 (1 cow/pool, 0.22% ± 0.19%) and 15 (3 cows/pool, 0.67% ± 0.34%) individual cows of the 2248 tested were shedding at the time of sampling. Although the single sample protocol used in this study undoubtedly underestimates the true prevalence of intermittent Salmonella shedding, the method mirrors that used in other surveys. For instance, a similar survey of dairy cattle in the United States showed that 5.4% of cows tested were shedding Salmonella (5). The 5 fecal pools that were positive for Salmonella originated from 4 (8% ± 7.52%) of the herds sampled. Wells et al (5) found that 27.5% of the dairy herds tested in the United States had at least 1 cow shedding Salmonella. Kabagambe et al (6), however, reported that herds of ≥ 100 cows were 5.8 times more likely to be shedding Salmonella than were herds of < 100 cows. Wells et al (5) also found that 56.5% of herds with 400 milking cows had at least 1 cow shedding Salmonella, while only 4.8% of herds of < 100 milking cows were shedding. At the time samples were collected for this study, dairy herds in Alberta averaged 94.4 milking cows/herd (7). The herds sampled in this study varied from 44 to 182 head, with an average of 78.8 cows. Thirty-six of the tested herds had < 100 cows and 2 (5.6%) of these herds had a positive fecal pool. Fourteen herds had ≥ 100 cows and 2 (14.3%) of these herds had at least 1 positive pool.

Three isolates from each positive pool were submitted to the Health Canada Office International des Epizooties (OIE) Reference Laboratory for Salmonellosis (Guelph, Ontario) for serotyping and phage typing. The Salmonella serotypes isolated in this study were Heidelberg DT29, Typhimurium var Copenhagen (not phage typable), Typhimurium var Copenhagen DT104, 6,8:-:-, and 6,8:-:enx (Table 1). No samples were taken from cows known to have salmonellosis. The serotypes isolated in this study, however, were similar to the serotypes isolated in the United States from cows with salmonellosis (Typhimurium and Typhimurium var. Copenhagen) (5). By contrast, the 5 Salmonella serotypes most commonly isolated from clinically healthy dairy cows in the United States were Montevideo, Cerro, Kentucky, Anatum, and Meleagridis (5). These serotypes were not isolated during the course of this study.

Table 1.

The isolates that were serotyped were also screened for antibiotic susceptibility (Sensititre Custom Designed MIC Panel CMV6CNCD; Trek Diagnostic Systems, Westlake, Ohio, USA), according to the manufacturer's instructions. Sensitive, intermediate, and resistant breakpoints were defined according to published guidelines (8,9). No Salmonella isolate recovered in this study was sensitive to all antibiotics in the test panel. In comparison, 36% of Salmonella isolates from cattle compiled during the 1998 United States National Antibiotic Susceptibility Monitoring Program for Veterinary Isolates (10) were sensitive to all antibiotics tested; however, the report does not distinguish between beef and dairy cattle. Two (28.6%) Salmonella isolates recovered in this study were resistant only to tetracycline, 2 (28.6%) isolates were resistant to 2 antibiotics, and 2 isolates (28.6%) were resistant to 4 antibiotics (Table 1). Only 1 (14.2%) isolate, Salmonella Typhimurium var Copenhagen DT104 had the R-type penta-resistant (ACSSuT) phenotype (Table 1). By contrast, Salmonella isolates from clinically healthy and clinically ill cattle in the United States were reported to be resistant, respectively, to ampicillin, 1.1% and 32.2%; chloramphenicol, 1.6% and 5.5%; kanamycin, 1.7% and 29.0%; streptomycin, 4.1% and 33.9%; sulfamethoxazole, 2.9% and 30.6%; tetracycline, 8.0% and 36.6%; and ticarcillin, 3.6% and 31.7% (10).

In this study, we considered isolates to be different if they were isolated from different pools, had different serotypes, or had different antibiograms. Seven different Salmonella isolates were recovered during the course of this study. Pulsed-field gel electrophoresis (PFGE) was carried out on these isolates, as described previously (4). Two Heidelberg DT29 isolates recovered from the same pool were indistinguishable from each other by PFGE (100% relatedness), although there were differences in their antibiotic resistance profiles (Table 1). Salmonella Typhimurium var Copenhagen (not phage typable) and Typhimurium var Copenhagen DT104 were isolated from different farms and were not related to each other by PFGE (< 64% relatedness). Salmonella 6,8:-:-, and 6,8:-:enx were recovered from 2 pools from 1 herd (Table 1) and were indistinguishable from each other by PFGE (97% to 100% relatedness). The antibiotic resistance profiles of these isolates were also similar (Table 1).

In summary, this study has shown that S. enterica is present in dairy herds in Alberta at levels that are comparable with those in herds of similar size in the United States. The low prevalence of Salmonella found in this study is reassuring, but the proportion of multiresistant isolates among these few isolates is of concern and warrants further investigation.