Emergence of Domestically Acquired Ceftriaxone-Resistant Salmonella Infections Associated With AmpC β-Lactamase FREE

Each year, an estimated 1.4 million persons in the United States are infected with Salmonella.¹ Many of these infections occur in children; 7% of the 2840 culture-confirmed Salmonella infections reported to the Centers for Disease Control and Prevention in 1998 through Foodborne Diseases Active Surveillance Network (FoodNet) occurred in persons aged 18 years or younger.² Although most Salmonella infections result in mild-to-moderate gastroenteritis, which resolves spontaneously, systemic infections, including bacteremia and meningitis, also occur.³ Invasive infections commonly occur in children, particularly in infants.⁴ Ten percent of culture-confirmed infections from blood or the central nervous system reported to the Centers for Disease Control and Prevention were in infants younger than 1 year. Furthermore, an estimated 595 persons die each year as a consequence of Salmonella infections.¹

Antimicrobial agents are commonly used empirically for the treatment of patients with moderate-to-severe diarrhea and may be lifesaving for persons with systemic Salmonella infections. Fluoroquinolones (eg, ciprofloxacin) are often used for treating adults with Salmonella infections but are not approved for use in children in the United States. Children with Salmonella infections, particularly invasive infections, are commonly treated with expanded-spectrum cephalosporins (eg, ceftriaxone) because of their favorable pharmacokinetic properties and low prevalence of resistance.^5- 8 A limited number of cases of ceftriaxone-resistant Salmonella have been reported from South America, North Africa, Asia, Europe (France, Spain, Turkey, Greece, Hungary, United Kingdom, and Russia), and the United States.^9- 16 A national survey of 10% of culture-confirmed Salmonella infections in 1995 found 3 of 4003 isolates resistant to ceftriaxone; each of these infections was acquired outside the United States.¹⁷ A recent article reported the first case of a domestically acquired ceftriaxone-resistant Salmonella infection in the United States in a child who apparently acquired the infection from cattle.¹⁶ We reviewed national surveillance data to determine the magnitude and molecular mechanisms of this emerging problem.

After serotyping, public health laboratories in the 17 participating state and community health departments in the National Antimicrobial Resistance Monitoring System, the national surveillance system for antimicrobial resistance in enteric bacteria, forwarded every 10th Salmonella isolate received at their laboratory between 1996 and 1998 to the Centers for Disease Control and Prevention for antimicrobial-susceptibility testing. The population served by the 17 sites in 1998 was 97 million persons, 37% of the US population. Minimum inhibitory concentrations (MICs) were determined for amikacin, ampicillin, amoxicillin-clavulanic acid, apramycin, ceftiofur (an extended-spectrum cephalosporin used in veterinary medicine), ceftriaxone, cephalothin, chloramphenicol, ciprofloxacin, gentamicin, kanamycin, nalidixic acid, streptomycin, sulfamethoxazole, tetracycline, and trimethoprim-sulfamethoxazole by using the Sensititre broth microdilution system (Trek Diagnostics, Westlake, Ohio). When established, National Committee for Clinical Laboratory Standards were used; resistance to ceftiofur and apramycin was defined as MICs of 8 µg/mL or higher and 32 µg/mL or higher, respectively. Isolates collected between 1996 and 1998 with decreased susceptibility to ceftriaxone (MIC ≥16 µg/mL) were confirmed as Salmonella and tested for resistance to cefoxitin, cefotaxime, and ceftazidime by Episolometer test (E-test) according to the manufacturer's recommendations (AB-Biodisk, Solona, Sweden), for the full range of MICs for ceftriaxone by broth microdilution according to the National Committee for Clinical Laboratory Standards, and by further molecular characterization.¹⁸ Ceftriaxone-resistant Salmonella isolates had MICs of 64 µg/mL or higher and intermediate isolates had MICs of 16 to 32 µg/mL by broth microdilution. Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were used as controls.

With the assistance of state and local health departments, we attempted to interview patients with ceftriaxone-resistant Salmonella infections. Using a standard questionnaire, we asked patients about their illness, including symptoms, clinical outcomes, and hospitalization. Patients were also asked about activities before illness onset, including use of antimicrobial agents for another illness, farm visitation, and international travel in the 5 days before illness onset.

Salmonella serotype Typhimurium isolates collected between 1996 and 1998 were subtyped by bacteriophage typing¹⁹ and pulsed-field gel electrophoresis.²⁰ The molecular mechanism of ceftriaxone resistance among Salmonella isolates with MICs of 16 µg/mL or higher were characterized by β-lactamase extraction and isoelectric focusing, amplification of the ampC gene from Citrobacter freundii by polymerase chain reaction, plasmid profile analysis, and conjugal mating experiments with E coli C600N. Crude preparations of β-lactamases produced by these isolates were obtained by submitting cells grown in 100 mL of Trypticase soy broth (Difco, Detroit, Mich) to 5 cycles of freezing in 0.2-mol sodium acetate (Sigma, St Louis, Mo) and thawing followed by centrifugation.

Isoelectric focusing was performed at room temperature on a Bio-Rad mini isoelectric focusing 111 (Bio-Rad, Richmond, Calif). Enzymes were visualized by activity staining after the gel was overlaid with a 500-µg/mL solution of nitrocephin (BBL, Cockeysville, Md).²¹ The isoelectric points of enzymes from the Salmonella isolates were estimated by comparison with TEM-1, TEM-10, SHV-5, and P99 β-lactamases.

Plasmid DNA was isolated by previously described methods.²² The polymerase chain reaction primers used to determine the presence of the ampC gene from Citrobacter freundii have been previously described.^10- 11 The double-disk diffusion assay, using ceftriaxone, ceftazidime, cefotaxime, cefoxitin, and amoxicillin-clavulanic acid as indicator antimicrobial agents, was used to screen for the presence of extended-spectrum β-lactamases according to previously published methods.²³ All Salmonella isolates (donor strain; ampicillin-resistant, nalidixic acid-susceptible), excluding 1 isolate that was nalidixic acid–resistant, were mated to E coli C600N²⁴ (recipient strain, ampicillin-susceptible, nalidixic acid–resistant) using standard methods.¹¹ Counter selection was performed on Luria-Bertoni Broth agar containing 50 µg/mL of ampicillin and 30 µg/mL of nalidixic acid (Difco).

Fifteen (0.4%) of 4093 Salmonella isolates submitted from 1996 to 1998 had ceftriaxone MICs of 16 µg/mL or higher; 13 were resistant (MIC ≥64 µg/mL) and 2 were intermediate (MIC, 16-32 µg/mL). The prevalence of ceftriaxone-resistant Salmonella was 0.1% (1 of 1326) in 1996, 0.4% (5 of 1301) in 1997, and 0.5% (7 of 1466) in 1998. Twelve (80%) of the 15 Salmonella isolates with ceftriaxone MICs of 16 µg/mL or higher were Salmonella serotype Typhimurium, the other isolates were serotypes Newport, Thompson, and Cubana. Seven of the 12 Salmonella Typhimurium isolates did not react with any of the typing phages (untypable), 2 isolates reacted with phages but did not conform to any defined pattern (reacts but does not conform), 2 isolates were definitive type 104 (DT104), and 1 was definitive type 21 (Table 1).

None of these Salmonella Typhimurium isolates had a pulsed-field gel electrophoresis pattern that was indistinguishable from that of another isolate in the group. Even the 2 DT104 isolates (Figure 1, lanes 10 and 14) had highly similar but not identical pulsed-field gel electrophoresis patterns.

Each of the 13 ceftriaxone-resistant isolates was also resistant to ampicillin, cephalothin, and ceftiofur; 12 (92%) were resistant to amoxicillin-clavulanic acid. Some ceftriaxone-resistant isolates were additionally resistant to aminoglycosides (92% to streptomycin, 62% to kanamycin, and 23% to gentamicin), sulfonamides (77%), tetracycline (77%), and chloramphenicol (69%). All ceftriaxone-resistant isolates were susceptible to amikacin, apramycin, and ciprofloxacin. One isolate was resistant to nalidixic acid and had decreased susceptibility to ciprofloxacin (MIC, 0.12 µg/mL). All 13 ceftriaxone-resistant isolates collected between 1996 and 1998 were intermediate or resistant to ceftazidime and cefotaxime.

All 13 ceftriaxone-resistant isolates were from stool cultures. Patients lived in 8 states (California, Colorado, Kansas, Massachusetts, Maryland, Minnesota, New York, Oregon). The age range of the patients was younger than 1 year to 52 years (age missing for 1 patient); 10 (77%) were 18 years or younger and 5 (38%) were aged 1 year or younger.

Ten of 11 interviewed patients with a ceftriaxone-resistant Salmonella infection reported having gastrointestinal tract symptoms. Patients with gastrointestinal tract symptoms reported having diarrhea (82%), abdominal cramps (82%), fever (73%), blood in stool (55%), and vomiting (36%). Duration of diarrhea ranged from 1 to 14 days, with a median of 5 days. All interviewed patients sought medical care. Six of these patients (55%) reported that the Salmonella infection caused interference with daily activities. The number of days of illness ranged from 1 to 14 days (median, 4 days). Five patients (45%; age range, <1-52 years) received antimicrobial agents for their Salmonella infection (3 received ciprofloxacin, none received cephalosporins), and 3 patients (27%) received intravenous fluids. None of the patients took an antimicrobial agent after illness onset but before collection of the specimen that yielded ceftriaxone-resistant Salmonella. Three patients (27%) were hospitalized; the length of hospitalization ranged from 1 to 5 days. No patients had an invasive infection. One patient was evaluated with colonoscopy. None of the patients died.

Ten (91%) of the 11 patients interviewed with ceftriaxone-resistant Salmonella infection reported being in good health before their Salmonella infection; 1 patient reported an underlying immunocompromising condition. Nine (82%) of the 11 patients interviewed did not take antimicrobial agents for another illness in the 5 days before Salmonella illness onset. Of the 2 interviewed patients who took an antimicrobial agent, 1 took a cephalosporin (cephalexin). Ten (91%) of the 11 patients did not travel outside the United States before illness onset. Two patients (18%) visited a farm during the 5 days before illness onset; farm animals included horses, cows, chickens, and pigs. One patient drank unpasteurized milk in the 5 days before his illness onset.

Thirteen of 15 Salmonella isolates with ceftriaxone MICs of 16 µg/mL or higher from 1996 to 1998 were cefoxitin-resistant and were phenotypically classified as having an AmpC β-lactamase based on a negative double-disk diffusion test result. The double-disk diffusion test detects whether a β-lactamase is affected by the β-lactamase inhibitor clavulanic acid. These 13 isolates, which included 2 intermediate isolates (MIC, 32 µg/mL), also expressed a β-lactamase with an isoelectric point of more than 9.0, further suggesting that the ceftriaxone-resistance was mediated by an AmpC β-lactamase. Primers specific for the ampC gene of Citrobacter freundii amplified an appropriate 631–base pair product in all 13 of these isolates. These same primers were used to amplify and characterize a CMY-2 β-lactamase recently found in a ceftriaxone-resistant Salmonella Typhimurium from Nebraska.¹⁶ Six of these 13 isolates also expressed a TEM-1-like β-lactamase with an isoelectric point of 5.4. Confirming this observation, all 6 isolates were positive for an expected 952–base pair polymerase chain reaction product using primers specific for the TEM family of β-lactamases. Five of the 13 Salmonella isolates were able to transfer ceftriaxone resistance to E coli C600N. All C600N transconjugants expressed the AmpC β-lactamase (isoelectric point >9.0), which suggested that this β-lactamase was mediating the ceftriaxone resistance and was encoded on a transferable plasmid. The transferable plasmids ranged from 160 to 170 kilobase. The remaining 2 ceftriaxone-resistant Salmonella isolates (AMO2544 and AMO4707) expressed a putative extended-spectrum β-lactamase as assessed by a positive double-disk diffusion test result. Isolate AMO2544 expressed 2 β-lactamases, 1 with an isoelectric point of 5.4 and the other with an isoelectric point of approximately 8.0. Isolate AMO4707 also expressed 2 β-lactamases with isoelectric points of 5.4 and approximately 7.6.

Ceftriaxone resistance has emerged among Salmonella isolates in the United States. In 1998, 0.5% of Salmonella isolates tested were resistant to ceftriaxone. Since there are an estimated 1.4 million Salmonella infections annually in the United States, this suggests that several thousand Salmonella infections each year are caused by a ceftriaxone-resistant strain. Although antimicrobial agents are not indicated for the treatment of uncomplicated salmonellosis, ceftriaxone is commonly used empirically to treat children with severe Salmonella infections, including Salmonella bacteremia and meningitis. Continued emergence and dissemination of ceftriaxone resistance, therefore, is likely to have clinical consequences.^25- 26

Ceftriaxone-resistant Salmonella infections are now being acquired domestically in the United States. This is in contrast to a large national survey in 1995, which reported no domestically acquired ceftriaxone-resistant infections.¹⁷ Furthermore, no cephalosporin resistance was present in the previous national study conducted in 1990.²⁷ In our study, 10 (91%) of 11 interviewed persons with ceftriaxone-resistant infections reported no recent international travel.

The sources of the ceftriaxone-resistant infections described here are not known, but as with other Salmonella infections they are likely to be contaminated food, particularly foods of animal origin. Ceftriaxone resistance is evident in Salmonella isolates, particularly Typhimurium, from food-producing animals in the United States; in 1998, 2 (2%) of 98 Salmonella Typhimurium isolates collected from chickens and cattle in slaughter plants were resistant to ceftriaxone.²⁸ Recent reports of ceftriaxone-resistant Salmonella from cattle in Nebraska and a subsequent human infection and ceftriaxone-resistant Salmonella in pigs, cattle, and humans in Iowa are additional evidence of a food-producing animal reservoir of ceftriaxone-resistant Salmonella.^16,29 Interestingly, in our investigation, 2 (18%) of 11 interviewed patients visited a farm in the 5 days before their illness onset and 1 of those patients consumed unpasteurized milk. Furthermore, 2 ceftriaxone-resistant Salmonella isolates were DT104, a strain of Typhimurium that has caused illness in food-producing animals and humans in the United States.³⁰

The emergence of ceftriaxone resistance is occurring among several strains of Salmonella rather than through the emergence of a single clone. This suggests horizontal dissemination of a resistance determinant. Our demonstration of conjugal transfer of the ceftriaxone-resistance determinant by some of the ceftriaxone-resistant isolates also supports this hypothesis.

In contrast to ceftriaxone-resistant isolates outside the United States, which have involved TEM or SHV-derived extended-spectrum β-lactamases, most of the ceftriaxone-resistant determinants in our study were due to an AmpC-type resistance gene in the BIL-1, LAT1, or CMY2 family of cephamycinases.^31- 32 This finding of a similar molecular mechanism of resistance among different strains of Salmonella supports horizontal dissemination. Interestingly, the recent reports of ceftriaxone-resistant Salmonella in Nebraska¹⁶ and Iowa²⁹ also had an AmpC mechanism of resistance (CMY-2). AmpC plasmid-mediated β-lactamases have been reported in clinical strains of Klebsiella pneumoniae, E coli, and Enterobacter aerogenes, but rarely in Salmonella from the United States.^{9- 14,31- 33}

AmpC β-lactamases have been recently described among E coli isolates from calves with diarrhea in the United States³⁴; although treatment histories were not described, the authors suggest that the emergence of AmpC-mediated resistance may have been a consequence of use of ceftiofur in the calves. Our demonstration of cross-resistance between ceftiofur and ceftriaxone supports a similar conclusion. Ceftiofur is the only cephalosporin approved for systemic use in food-producing animals in the United States. This antimicrobial was approved for therapeutic use in cattle by injection in the United States in 1988, and is now also approved for similar use in pigs, sheep, chicken, and turkeys. Ceftiofur use other than labelled indications is also permitted; for example, ceftiofur is frequently injected into chicken eggs prior to hatching on a flock-wide basis for disease prevention.

Ceftriaxone is a standard therapy for serious Salmonella infections in children. Although most persons with Salmonella infection do not require antimicrobial therapy, it can be lifesaving for persons with invasive disease. Unfortunately, resistance to ceftriaxone probably will continue to increase in Salmonella. To mitigate potential rapid dissemination, there is a need to control selective pressure in human and particularly veterinary medicine by reducing the misuse and overuse of extended-spectrum cephalosporins. Efforts to reduce the transmission of Salmonella to humans through the food supply, such as the new regulations for safe processing of meat and poultry to decrease the prevalence of Salmonella in these products and enhanced food safety education programs, also need to be emphasized.

Studies are needed to further determine the association between use of ceftiofur in food-producing animals and the emergence and dissemination of ceftriaxone resistance, particularly resistance caused by an ampC resistance gene. Because there is cross-resistance between ceftiofur and ceftriaxone, and food-producing animals are the source of most domestically acquired Salmonella infections, the use of ceftiofur in food-producing animals apparently is contributing to ceftriaxone-resistant Salmonella, which is transmitted to consumers through the food supply.

To protect the public's health, acceptable thresholds for ceftriaxone resistance in Salmonella should be established that are consistent with the framework proposed in December 1998 by the US Food and Drug Administration for evaluating the human safety of antimicrobial agents in food-producing animals.³⁵ Limitations on expanded-spectrum cephalosporin use in food-producing animals may be indicated, particularly if such resistance increases.

Recently available data indicate that in 1999, 28 (1.9%) of 1499 Salmonella isolates submitted to the National Antimicrobial Resistance Monitoring System had ceftriaxone MIC of 16 µm/mL or higher, indicating decreased susceptibility to ceftriaxone. Full-range MIC testing by broth microdilution, epidemiologic investigation, and molecular characterization is pending.