Human Multidrug-Resistant Salmonella Newport Infections, Wisconsin, 2003–2005

We conducted a retrospective study of Salmonella Newport infections among Wisconsin residents during 2003–2005. Multidrug resistance prevalence was substantially greater in Wisconsin than elsewhere in the United States. Persons with multidrug-resistant infections were more likely than persons with susceptible infections to report exposure to cattle, farms, and unpasteurized milk.


D uring the past decade, multidrug-resistant (MDR)
Salmonella enterica serotype Newport strains in the United States have increased substantially (1). The prevalence of the most common MDR S. Newport phenotype, Newport-MDRAmpC, increased from 1% of human S. Newport isolates tested in 1998 to 21% of isolates tested in 2003 (2). Newport-MDRAmpC is resistant to at least chloramphenicol, streptomycin, sulfamethoxazole/sulfisoxazole, tetracycline, amoxicillin-clavulanic acid, ampicillin, cefoxitin, ceftiofur, and cephalothin. This phenotype exhibits decreased susceptibility to ceftriaxone (2), a critically important antimicrobial agent for treating invasive salmonellosis in children (3).
Studies suggest that dairy cattle are a major US reservoir for MDR S. Newport (4)(5)(6). However, data documenting the prevalence of MDR S. Newport among infected human case-patients in dairy-intensive states are limited. To assess the prevalence of resistance among S. Newport isolates in Wisconsin, which in 2002 had the greatest density of milk cows in the United States (7), we evaluated antimicrobial susceptibility data from S. Newport infections among Wisconsin case-patients during 2003-2005. We also compared information on potential exposures for case-patients with Newport-MDRAmpC and susceptible infections.

The Study
Surveillance data were electronically compiled for laboratory-confi rmed S. Newport infections among Wisconsin residents with illness onsets from January 1, 2003, through December 31, 2005. Providers and local health departments reported hospitalization status; travel history; and exposure to raw milk, cattle, horses, reptiles, and dead animals. The study population included case-patients whose isolates were tested for antimicrobial drug susceptibility at the Wisconsin State Laboratory of Hygiene. Identifi cation and susceptibility testing were conducted on isolates from stool, urine, and blood samples.
Serotype identifi cation was performed according to the Kauffmann-White scheme (8). Slide and tube agglutination were used for identifi cation of O (somatic) and H (fl agellar) antigens, respectively. All isolates were tested for susceptibility to ampicillin, amoxicillin-clavulanic acid, cefoxitin, ceftriaxone, cephalothin, chloramphenicol, ciprofl oxacin, gentamicin, kanamycin, nalidixic acid, streptomycin, sulfi soxazole, tetracycline, and trimethoprim-sulfamethoxazole, by using the Kirby-Bauer disk diffusion method. Results were interpreted according to Clinical and Laboratory Standards Institute (CLSI) guidelines (9). Antimicrobial agents were categorized into CLSI antimicrobial subclasses, and each isolate was assigned to >1 categories according to its antimicrobial resistance phenotype and the number of subclasses to which it was resistant (National Antimicrobial Resistance Monitoring System for Enteric Bacteria [NARMS], pers. comm.; Table 1). Pansusceptible isolates were defi ned as isolates that had no detected antimicrobial drug resistance. Because isolates were not tested for ceftiofur resistance, our defi nition of Newport-MDRAmpC did not include resistance to this drug.
The prevalence of each type of resistance among S. Newport isolates from Wisconsin case-patients was compared with that reported elsewhere in the United States, by using 2003 and preliminary 2004 NARMS data. Data were analyzed by using Epi Info 2002, version 3.3.2 (Centers for Disease Control and Prevention, Atlanta, GA, USA); to assess associations between antimicrobial resistance and reported exposures, odds ratios and Mantel-Haenszel and Fisher exact 2-tailed p values were calculated where appropriate.
Serotyping and antimicrobial drug susceptibility testing were conducted on S. Newport isolates from 268 casepatients. Median age was 34 years (range <1-96 years); of 267 case-patients for whom sex was reported, 57% were female. Resistance patterns are provided in Table 1. Among the 5 (2%) quinolone-resistant isolates (2 resistant to nalidixic acid and ciprofl oxacin, 2 resistant to nalidixic  Table 1; Figure).
Of 194 case-patients for whom hospitalization status was reported, 46 (24%) had been hospitalized. Of case-patients with Newport-MDRAmpC and pansusceptible infections, similar proportions were hospitalized (26% and 24%, respectively) and had known hospitalization status (72% and 73%, respectively). Two case-patients died: an 84-year-old woman and a 37-year-old man for whom salmonellosis was not considered the probable cause of death. The 2 associated isolates were pansusceptible.
Persons infected with Newport-MDRAmpC were signifi cantly more likely than persons infected with pansusceptible S. Newport to be male and to have had contact with cattle, to have drunk unpasteurized milk, and to live on or have visited a farm or petting zoo ( Table 2). Reported exposure to reptiles was signifi cantly associated with pansusceptible infection ( Table 2). No association was found between hospitalization and resistance (odds ratio [OR] 1.09, p = 0.81).

Conclusions
We describe a substantially greater prevalence of MDRAmpC resistance among Wisconsin case-patients with S. Newport infections that occurred during 2003-2005, compared with data reported elsewhere in the United States (NARMS, personal communication, 2007). This fi nding is of particular concern because Newport-MDRAmpC exhibits decreased susceptibility to ceftriaxone, a third-generation cephalosporin that is the treatment of choice for invasive salmonellosis in children (3). Additionally, because the bla CMY-2 gene that confers ceftriaxone resistance in Newport-MDRAmpC is located on a plasmid that was readily transferred between Escherichia coli in laboratory assays (10), propagation of Newport-MDRAmpC could increase the spread of CMY-2 plasmids to other bacteria.
Few published data are available on the prevalence of Newport-MDRAmpC in other dairy-intensive states. Minnesota, which in 2002 had the eighth-greatest density of milk cows in the United States (7), reported a signifi cant increase in MDR S. Newport among human isolates during 1996-2003, including an increase in the percentage of isolates with decreased susceptibility to ceftriaxone (13). However, NARMS reported a similar increase in Newport-MDRAmpC prevalence nationally during 1998-2003 (2). Analyses of unpublished data provided by the Minnesota Department of Health indicated that 22% of 147 human isolates tested had antimicrobial drug resistance profi les consistent with the Newport-MDRAmpC phenotype during 2003-2005; this prevalence is much lower than that among Wisconsin case-patients who were ill during the same period. Although differences in enteric disease surveillance could partially explain this discrepancy, Newport-MDRAmpC's emergence in dairy cattle is likely to be associated with several factors.
Our fi ndings underscore the need for intensive Newport-MDRAmpC surveillance in major dairy states. Efforts to promote the conservative and appropriate use of ceftiofur and other antimicrobial drugs in dairy cattle are indicated; they should be complemented by strategies to discourage transmission of MDR Salmonella among cattle, such as separating ill from parturient animals and disinfecting environmental niches (14). Furthermore, providers should be discouraged from prescribing antimicrobial drugs to patients with low-risk Salmonella infections (15), and public health messages should emphasize the importance of pasteurizing milk and cooking meat appropriately.