Emergence of Multidrug-Resistant Salmonella enterica Serotype Newport in Minnesota

Antimicrobial agents have been widely used in human and animal populations to control infectious diseases. In agriculture, antimicrobials are currently used for therapy, disease prevention, and growth promotion [1, 2]. However, in the past 2 decades, there has been a decrease in the number of antimicrobials that are effective for treating human infections, and antimicrobial resistance has become an important public health concern [1, 2].

The emergence of multidrug-resistant Salmonella enterica serotype Newport isolates has been well documented in humans [3, 4], but only sporadic cases and outbreaks in animals from a variety of geographic regions of the United States are available [5, 6]. The National Antimicrobial Resistance Monitoring System has systematically examined isolates of Salmonella from humans, and data has suggested that an increased percentage of infections are caused by multidrug-resistant S. Newport in humans [3, 4, 7]. These infections may have clinical implications and present a broader public health risk [7].

In this report, we evaluated antimicrobial resistance and PFGE subtypes of S. Newport isolates from humans and animals in Minnesota. A subset of the isolates was further characterized by screening for the presence of class 1 integrons.

Methods. Clinical laboratories in Minnesota are required to submit all human Salmonella isolates to the Minnesota Department of Health (Minneapolis) as part of the statewide, population-based, active laboratory surveillance established in 1996. Human isolates obtained from 1996 to 2003 were available for this study. S. Newport isolates cultured from animal tissues and feces at the Veterinary Diagnostic Laboratory of the University of Minnesota (Saint Paul) from clinical cases from 2000 to 2003 were available. Animal Salmonella isolates were serotyped at the National Veterinary Services Laboratories of Iowa (Ames) and were then forwarded to the Minnesota Department of Health.

The Minnesota Department of Health performed antimicrobial susceptibility testing by disk diffusion (Becton-Dickinson) for ampicillin (A), chloramphenicol (C), cephalothin (Cf), ceftriaxone (Cro), streptomycin (S), gentamicin (Gm), kanamycin (K), nalidixic acid, sulfisoxazole (Su), trimethoprim-sulfamethoxazole (Sxt), tetracycline (T), and ciprofloxacin. Isolates were considered to susceptible, intermediate, or resistant according to criteria of the Clinical and Laboratory Standards Institute [8]. The Etest for MIC was also used for isolates that had intermediate susceptibility to Cro. Isolates with a Cro MIC of 16–32 µg/mL were considered to have intermediate susceptibility to Cro (CroI), and isolates with a Cro MIC ⩾48 µg/mL were considered to be resistant to Cro (CroR). Multidrug-resistant was defined as resistance to ⩾5 antimicrobials. Isolates that were susceptible to all antimicrobials were described as pansusceptible.

PFGE and the analysis of profiles were performed by the Minnesota Department of Health using XbaI endonuclease (Promega) according to published methods [9]. Patterns were compared using BioNumerics software (Applied Maths) with the Dice coefficient and 1% matching criteria. A cluster was defined as >1 isolate that shared ⩾ 80% similarity. A subset of 149 representative human and animal isolates obtained during 2000–2003 were screened for the presence of a class 1 integron by PCR. The following primers were used: int1F 5′-GGC ATC CAA GCA GCA AGC, int1R 5′-AAG CAG ACT TGA CCT GAT, and sul1 5′-TGA AGG TTC GAC AGC AC [10]. The reactions were performed using 1.5 mmol/L MgCl₂, 200 µmol/L of dNTP, 0.2 µmol/L of each primer (int1F-int1R or int1F-sul1), 3 U Taq DNA polymerase (Qiagen), 1× buffer, and 150 ng of DNA for 35 cycles of 60 s at 94°C, 60 s at 56°C, and 60s at 72°C.

A χ² test for linear trend was performed to compare the proportion of animal and human isolates by year (EpiInfo software, version 3.3.2; CDC). A P value <.05 was considered to be statistically significant.

Results. The proportion of S. Newport isolates in Minnesota significantly increased during 1996–2003, from 1.5% of all S. enterica isolates from humans in 1996 to 8% in 2003 (χ², 62.1; P < .001). Similarly, a significant increase in isolates from animals, from 0% to 23%, was observed during the same time period (χ², 694.7; P < .001). Human isolates of S. Newport from 1996 to 2003 displayed a significant trend of increase in multidrug-resistant S. Newport (χ², 16.4; P < .001). Figure 1 illustrates annual percentages of antimicrobial resistance among S. Newport isolates. The percentages of CroR and CroI among human isolates increased from 1996 to 2003. Likewise, percentages of multidrug-resistant CroR and CroI isolates from bovine and other animals also increased. In humans, the most prevalent resistance phenotype (R-type) was ACCfCroSSuT (14% of isolates) from 2000 to 2003. The most prevalent R-type among bovine isolates was ACCfCroIGmKSSuT (36% of isolates). The highly resistant R-type ACCfCroGmKSSuSxtT was found in 2% of human isolates, 2% of bovine isolates, and 8% of S. Newport isolates from other animal species. None of the isolates was resistant to ciprofloxacin.

Figure 1

Antimicrobial resistance of Salmonella enterica Newport isolates from humans and animals. A, ampicillin; C, chloramphenicol; Cf, cephalothin; CroI, intermediate resistance to ceftriaxone; CroR, resistant to ceftriaxone; MDR, multidrug resistant (defined as resistant to ⩾5 antimicrobials); Pan, pansusceptible; S, streptomycin; Su, sulfisoxazole; T, tetracycline.

Eighty-one PFGE subtypes were identified in human isolates, 30 were identified in bovine isolates, and 16 were identified in isolates from other species. The most common PFGE subtypes in human isolates were NEW32 (6% of isolates), followed by NEW51, NEW37, and NEW122 (4% each). Among bovine isolates, the most prevalent subtypes were NEW58 (19% of isolates) and NEW32 and NEW66 (10% each). PFGE subtypes from other animal species were also diverse; the most common subtypes were NEW37 and NEW144 (13% each). A number of PFGE subtypes were common among human, bovine, and other animal species, including NEW32, NEW37, NEW51, NEW58, NEW66, NEW122, and NEW144. Some of the NEW32, NEW37, NEW58, and NEW66 isolates recovered from bovine, human, and other animal species had identical R-types (at least ACCfCroSSuT or ACCfCroISSuT). NEW51 and NEW122, which were isolated from human, bovine, and other animal species, were pansusceptible. NEW144 obtained from human and equine clinical cases in 2003 had identical R-types (ACCfCroIGmKSSuSxtT).

A subset of 96 human and 53 animal isolates were screened for class 1 integrons by PCR (table 1). Class 1 integrons were detected in 16% of the isolates. of these isolates, 67% were obtained from bovine, 29% from human, and 4% from other animal species. For the purpose of analysis, isolates were divided in 12 clusters of PFGE subtypes (A–L) based on 80% pattern similarity. Cluster analysis demonstrated that cluster A contained 82% of multidrug-resistant isolates and 92% of the class 1 integron isolates and that 84% of its isolates were resistant to at least ACCfCroSSuT or ACCfCroISSuT. In contrast, other clusters contained 93% of the pansusceptible isolates. The frequency of PFGE subtypes with class 1 integrons varied (29% of NEW58, 25% of NEW66, and 12% of NEW32). A total of 96% of the class 1 integron isolates were multidrug-resistant, and 79% of these isolates were resistant to at least ACCfCroSSuT or ACCfCroISSuT.

Table 1

Distribution of resistance phenotypes and class 1 integrons in human and animal Salmonella enterica Newport isolates based on PFGE clusters, Minnesota 2000–2003.

Discussion. We describe an increase in the number of isolates of S. Newport from both human and animal populations in Minnesota. A World Health Organization survey reported that S. Newport has recently emerged as one of the most prevalent human serotypes of Salmonella species worldwide [11]. In the United States, increased prevalence of S. Newport has also been observed among human patients [3, 4]. The recent emergence of S. Newport has important public health implications because of the concomitant emergence of multidrug-resistant isolates.

In this study, we showed that increases in the annual percentages of multidrug-resistant isolates were greater among isolates obtained from bovine and other animal species than among isolates from humans in Minnesota. However, this may be attributable to differences in disease surveillance among animals and humans. Refractory cases are more likely to be presented to veterinary diagnostic laboratories because of illness severity or farm outbreaks [12]. Isolates from human, bovine, and other animal species exhibited CroR or CroI phenotypes, as has been observed in other regions of the United States and as has been frequently associated with the presence of the plasmid-mediated bla_CMY gene [4–6]. Cro is an extended-spectrum cephalosporin that is commonly used for the treatment of pediatric infections [4]. Consequently, multidrug-resistant isolates with Cro resistance can have a serious public health impact, especially when treatment of ill infants and immunocompromised patients is performed empirically, without antimicrobial susceptibility testing [4–7].

Antimicrobial susceptibility and PFGE analysis of S. Newport isolates demonstrated phenotypic and genotypic heterogeneity of isolates. Twelve PFGE clusters with ⩾80% similarity were identified. of interest, cluster A contained most of the multidrug-resistant isolates, most of the class 1 integron isolates, and most of the isolates that had resistance to at least ACCfCroSSuT or ACCfCroISSuT, implying the clonal origin of the Minnesota isolates. Our findings agree with the findings of a previous report that analyzed a small number of S. Newport isolates from 25 states [5].

A combination of increased selective pressure from the use of antimicrobials in humans and animals and increased disease transmission can be factors in the emergence of antimicrobial resistance [1, 10]. Moreover, genetic elements, such as integrons, have also been reported to encode and transfer antimicrobial resistance to other microorganisms, contributing to the dissemination of antimicrobial resistance [5, 10]. The presence of large numbers of S. Newport isolates with ⩾80% genetic similarity, in addition to undistinguishable R-types and class 1 integrons, reinforces the clonal spread of antimicrobial resistance in animal and human populations in Minnesota, rather than the spread of horizontal antimicrobial resistance through integrons. This report documents a concurrent increase in the number of S. Newport isolates, the rate of multidrug-resistant S. Newport isolates, and the number of CroI or CroR S. Newport isolates obtained from animal and human populations in the same geographic region.

• Received December 30, 2005.
• Accepted April 4, 2006.

• © 2006 by the Infectious Diseases Society of America