The National Antimicrobial Resistance Monitoring System (NARMS) is a national public health surveillance system that monitors enteric bacteria and select animal pathogens to determine if they are resistant to antimicrobials used in human and veterinary medicine. NARMS is a collaboration of agencies within the U.S. Department of Health and Human Services (HHS) (FDA and CDC) and the U.S. Department of Agriculture (USDA) (FSIS, APHIS, and ARS).
The NARMS program helps promote and protect public health by providing information on the levels of bacterial resistance in the food chain and in our pets, whether resistance is increasing or decreasing, and the impact of interventions designed to limit the spread of resistance. NARMS data are used by FDA to inform its policy development, decision making, and other activities focused on preserving the effectiveness of antimicrobials for humans and animals. The CDC and FSIS may use NARMS information to support foodborne illness outbreak investigations on a case-by-case basis.
What has changed?
Resistance in animal pathogens
The previous 2016–2017 Integrated Summary featured, for the first time, pathogen data from companion animals and food-producing animals, collected through a pilot study with FDA's Veterinary Laboratory Investigation and Response Network (Vet-LIRN). The 2018 Integrated Summary features dog pathogen data from both Vet-LIRN and the USDA APHIS National Animal Health Laboratory Network (NAHLN) Antimicrobial Resistance Pilot Project.
In 2018, NARMS began using a new antimicrobial susceptibility testing panel (CMV4AGP) for Enterococcus which has the following changes:
- Expansion of dilution ranges for gentamicin and streptomycin dilution to 16–1024 µg/mL and 64–2048 µg/mL, respectively
- Expansion of the dilution ranges for chloramphenicol and ciprofloxacin to 2–64 µg/mL and 0.12–16 µg/mL, respectively
- Reduction of the dilution range for nitrofurantoin to 2–32 µg/mL
- Addition of avilamycin, an orthosomycin drug, with a dilution range of 0.25-32 µg/mL
- Replacement of penicillin with ampicillin
- Removal of kanamycin, lincomycin, and tylosin
Additionally, the Clinical and Laboratory Standards Institute (CLSI) updated breakpoints for daptomycin for all Enterococcus in its M100-S30 document published January 2020.1 The new breakpoints for daptomycin were applied to all Enterococcus data collected prior to and during 2018. The new breakpoint defines resistance to daptomycin as ≥ 8 µg/mL for all enterococcal species.
New retail meat sites
In 2018, NARMS retail meat testing expanded to southern California and North Carolina for a total of 23 sampling sites.
NARMS collects a large number of bacterial isolates from 15 distinct human, animal, and food sources (see Table 1 below). This summary report focuses on antimicrobial resistance to drug classes that are most important to human medicine (generally, first- or second-line treatments), multidrug-resistance, and specific drug resistance profiles of epidemiological importance. It is important to note that the changes highlighted in this summary may be influenced by changes in animal health, antimicrobial use, environmental influences, animal and food production practices, human behavior (such as travel and food-product choices), sampling and laboratory methodologies, year-to-year variations in serotype distributions, and other factors.
Salmonella and Campylobacter are the leading bacterial causes of foodborne illness in the United States. Nontyphoidal serotypes of Salmonella enterica (henceforth referred to as Salmonella) and Campylobacter can be present in the intestinal tracts of a wide range of animals including wildlife, livestock, and domestic pets. Salmonella and Campylobacter exposure occurs primarily through the consumption of contaminated foods. Salmonella is estimated to cause over 1.35 million illnesses and 420 deaths, and Campylobacter is estimated to cause over 1.5 million illnesses and 240 deaths each year (1). A small fraction of Salmonella and Campylobacter infections result in long-term sequelae such as reactive arthritis and Guillain-Barré syndrome.
Generic Escherichia coli and Enterococcus are also tested for antimicrobial susceptibility. These bacteria are cultured from animal samples at slaughter and from retail meats. Generic E. coli and Enterococcus are used by NARMS as indicator organisms to detect both emerging resistance patterns and specific resistance genes that could potentially be transferred to pathogenic bacteria.
For more information about these four bacterial organisms, visit the NARMS webpage.
This summary represents a consolidated view of the data generated from the four NARMS sample sources:
- Human clinical isolates
- Food-producing animal isolates from cecal (intestinal) samples at slaughter
- Samples routinely collected at inspected establishments as part of FSIS verification testing, and
- Raw meats (chicken, ground turkey, ground beef, and pork chops) collected at retail outlets in 20 states (23 sites).
The NARMS samples were collected from the sources listed in Table 1 below.
Table 1: Data sources for 2018 NARMS Integrated Summary
* Previously used terms include - Pathogen Reduction/Hazard Analysis Critical Control Point (PR/HACCP) testing or Routine testing
The majority (81%) of Salmonella from humans were not resistant to any of the antimicrobials tested. In humans, the overall level of resistance remains unchanged from 2017, and is in line with other data from 2006-2016, where 76-85% of Salmonella tested were susceptible to all antimicrobials tested. However, important resistance threats to human and veterinary health remain. Resistance to the most important clinical agents is summarized below.
- Although ceftriaxone resistance in Salmonella from humans remained low, this represented the fourth straight year of continuous increase, with levels rising from 2.4% in 2014 to 3.4% in 2018. For most food and animal Salmonella isolates, ceftriaxone resistance declined or remained stable since 2014. In contrast, ceftriaxone resistance increased in isolates from product verification chicken samples and turkey cecal samples. The blaCMY2 gene accounted for the majority of ceftriaxone resistance genes in all sources except retail chickens, chicken cecal samples, and retail ground turkey. In those sources, blaCTX-M-65 represented >65% of the ceftriaxone resistance genes. Similar to previous years’ findings, Infantis (carrying blaCTX-M-65) was the predominant ceftriaxone-resistant serotype among Salmonella from humans, all chicken sources, retail turkey meat, and product verification turkey samples.
- The percentage of Salmonella isolates from humans and poultry sources with decreased susceptibility to ciprofloxacin (DSC; minimum inhibitory concentration (MIC) ≥0.12 µg/mL) continued to increase. In humans, Salmonella with DSC increased from 7.5% in 2017 to 8.7% in 2018. Between 2017 and 2018, Salmonella with DSC increased from 9% to 18% in isolates from retail chicken meat, 14% to 20% in isolates from product verification chicken samples, and 18% to 26% in isolates from chicken ceca. Increases were also observed among retail turkey meat (4.6% - 9.2%). Similar to the 2016–2017 data, a substantial proportion (~40%) of DSC isolates from humans was serotype Enteritidis. Historically, DSC in serotype Enteritidis has been associated with international travel (2). In 2018, the second most abundant Salmonella serotype with DSC was Infantis. The increase in DSC among poultry isolates was primarily due to the increase in multidrug resistance (MDR; defined as resistance to three or more antimicrobial drug classes) in serotype Infantis (which carried gyrA mutations) along with a slight increase in DSC among Enteritidis from product verification chicken samples (from 3.2% in 2017 to 5.7% in 2018). Transmissible quinolone resistance genes qnrB, qnrS, aac(6’)-lb-cr, and oqxA continued to be identified among DSC isolates from humans, retail meats, and animal sources. They were most common in isolates from pigs (market swine and sows), where they were found in over 60% of DSC swine isolates.
- Since publication of the 2017 data, further testing showed that resistance to azithromycin2 was present in 0.3% of human isolates (not the previously reported 1.1%). A re-examination of 2017 cecal data also revealed three resistant isolates, one from beef cattle and two from market swine. Azithromycin resistance is rare, appearing in only 0.8% of human isolates tested, with serotype Newport comprising the majority (52%) of the isolates. Azithromycin resistant Salmonella were also recovered from retail and food animals in 2018, but it was rare, appearing in 1/19 (5%) retail pork isolates, 1/264 (<1%) turkey product isolates, 3/513 (< 1%) market swine isolates, and 2/214 (1%) beef cattle isolates. The mph genes were found in many of these resistant isolates while erm(42) was found in one beef cattle isolate. Some of the azithromycin resistant Salmonella Newport isolates from humans were linked to an outbreak attributed to consumption of contaminated beef and dairy products (3).
- Salmonella isolates from humans, retail meats, and animals did not show any resistance to carbapenem (routine meropenem testing began in 2016).
Multidrug Resistance (MDR; defined as resistance to three or more antimicrobial drug classes)
- In isolates collected from humans, MDR Salmonella has remained constant at around 10% for the last 11 years. In 2018, the most common MDR serotype in human isolates was I 4,,12:i:- accounting for 23% of MDR Salmonella. Approximately 92% of MDR I 4,,12:i:- isolates had combined resistance to ampicillin, streptomycin, sulfisoxazole, and tetracycline (ASSuT).
- MDR Salmonella recovered from product verification chicken samples showed an increase from 18% in 2017 to 22% in 2018. The percentage of Salmonella recovered from chicken cecal samples that were MDR increased from 25% in 2017 to 32% in 2018. Similarly, MDR Salmonella recovered from retail chicken also increased from 17% in 2017 to 20% in 2018. All increases were largely driven by the rise in MDR Salmonella Infantis isolates.
- MDR continued to either decline or remain stable in Salmonella from turkey sources between 2017 and 2018. The majority of MDR isolates from turkey samples were serotypes Infantis, Reading, and I 4,,12:i:-.
- There were thirteen human isolates, 7 isolates from cattle, 3 isolates from swine, and 1 chicken isolate that were extremely resistant (defined as resistant to eight or more antimicrobial classes). Nine of the twenty-four isolates were serotype Dublin, six were Typhimurium, and four were Heidelberg. The remaining five isolates were of serotypes Agona, Enteritidis, I 4,,12:i:-, Kentucky, and Mbandaka.
- No isolates exhibited decreased susceptibility to at least three of the four clinically important antimicrobial agents used for the treatment of complicated or invasive salmonellosis (ceftriaxone, ciprofloxacin, azithromycin, meropenem)
- MDR has decreased among serotypes Heidelberg and Reading.
MDR Salmonella Infantis
The recent emergence of MDR Salmonella Infantis has driven many important changes in Salmonella trends. The MDR Salmonella Infantis strain has supplanted other leading serotypes in U.S. poultry since it was first identified in that source type in 2014. This strain shows decreased susceptibility to fluoroquinolones due to a gyrA mutation and contains an MDR plasmid that carries up to ten resistance genes, which sometimes includes blaCTX-M-65. This plasmid carries genes conferring resistance to cephalosporins, tetracycline, aminoglycosides, chloramphenicol, and sulfonamides. NARMS scientists identified this strain of Salmonella Infantis in retail poultry, food animal cecal samples, and humans (4).
Whereas the initial human cases were associated with international travel, particularly to and from South America, later cases were mostly acquired domestically (5). This strain has dramatically increased in poultry products since 2016. There is also an increased proportion of this strain among isolates causing human illnesses and those found in food animal cecal isolates from poultry. We believe the Salmonella Infantis phenomenon reflects the international spread of a clone that emerged in one region, like Salmonella Typhimurium DT104 did in the 1990s, and spread to other countries (6,7). While it carries resistance to drugs not used in U.S. poultry production, this strain of Salmonella Infantis has gained a foothold in domestic poultry systems and can be found even in food producing animals raised without antibiotics. This highlights the importance of global cooperation in combating antimicrobial resistance.
- The proportion of macrolide-resistant (azithromycin or erythromycin) C. jejuni isolates from humans and chickens (ceca, production, and retail) remained at ≤ 3%. However, between 2017 and 2018, macrolide resistance among C. coli isolates from humans increased from 7% to 13%. Macrolide resistance among C. coli isolates from both routine and cecal chicken samples decreased from 8% to 4%. Among cecal samples, macrolide resistance was found most often in C. coli isolates from market swine (31%).
- The proportion of ciprofloxacin-resistant Campylobacter isolates from humans did not change for C. jejuni (28% in 2017 and 29% in 2018) or C. coli (39% in 2017 to 40% in 2018). In 2018, there was no significant change in ciprofloxacin resistance among C. jejuni and C. coli isolated from chickens compared with the previous year. Among all animal sources of Campylobacter tested in 2018, beef cattle cecal samples yielded the highest levels of ciprofloxacin resistance (63% in C. coli).
- Ceftriaxone-resistant E. coli in market swine cecal samples, increased to 8.6% (from 6.3% in 2017). Additionally, ceftriaxone resistant E. coli from retail pork chop samples remained at the level observed in 2017 (4-5%).
- Azithromycin resistance3 was detected in three E. coli isolates from retail pork chop samples, although only one carried a corresponding gene [mph(A)]. All azithromycin resistant cecal isolates were from market swine and sows, most carrying mph genes and some also having msr(E).
- While quinolone resistance was ≤ 8% among E. coli from food animals and retail foods, there has been a steady rise in DSC among E. coli from chicken cecal samples, market swine, and retail pork since 2016. Seventy-one percent (64/90) of food animal isolates with DSC were associated with gyr or par mutations. Seventeen isolates contained qnr genes. In retail foods, 13/30 DSC isolates were sequenced. Six of the thirteen contained qnr resistance genes, and the remaining seven had at least one mutation in the gyrA gene.
- MDR E. coli either decreased or remained stable in all food animal sources except for sows, where it increased from 16% to 20% from 2017 to 2018.
- No E. coli isolate from retail meats or animal ceca showed elevated MICs (≥ 2 µg/mL) to meropenem.
- Since testing began in 2013, erythromycin resistance has increased from 2.1% to 11.4% among predominant enterococcal species in beef cattle ceca. The top three Enterococcus species that account for the greatest proportion of macrolide resistance were Enterococcus hirae, E. durans and E. gallinarum.
- Chloramphenicol resistance, which had substantially risen among E. faecalis from market swine ceca and retail pork from 2013 to 2017 decreased in 2018 to 23% and 2.5%, respectively. Chloramphenicol resistance in sow ceca declined from 21% to 8.5% during the same period and increased to 11.8% in 2018.
- Since cecal testing began in 2013, there has been a decrease in gentamicin-resistant E. faecalis among turkeys (39% to 22%), chickens (44% to 9.7%) and sows (17% to 5.9%). There has also been a decrease in gentamicin-resistant E. faecalis from retail chicken meat (24% to 8%) during the same time period, and a recent decrease in gentamicin-resistant E. faecalis from retail turkey (31% in 2017 to 18% in 2018).
Although isolates were not tested for susceptibility to colistin, the genomes of all Salmonella and selected E. coli isolates were examined for the presence of transmissible colistin resistance genes (mcr-1 through mcr-9). In general, the U.S. finds fewer isolates positive for mcr genes than in countries where colistin is or has been used in food animal production (8,9). A comprehensive search of genome sequences in NCBI revealed the mcr-1 gene in 14 Salmonella isolates and two pathogenic E. coli (O111 and O145) isolates collected from humans. All but two patients reported having traveled internationally before their illnesses began. An additional Salmonella isolate was found to contain the mcr-3.1 gene, but the travel history of the patient is unknown. All isolates underwent whole genome sequencing at state health departments. In NARMS, there were no Salmonella nor E. coli isolates from food animals or retail meats that harbored mcr-1 through mcr-8 resistance genes. However, mcr-9.1 (10) was found in thirty-two Salmonella isolates from humans, in five E. coli and two Salmonella from retail meat sources, and in twenty-one Salmonella isolates from food animal ceca. Further work showed that the mcr-9 gene does not confer clinical resistance (11).
Antimicrobial Resistance in Animal Pathogens
Antimicrobial resistance (AMR) of bacterial pathogens is an emerging public health threat to people and animals because the ability to treat infections is compromised. Traditionally, surveillance programs in the U.S. have focused on collecting data from food animals, foods, and people. In March of 2015, the President of the U.S. released The National Action Plan for Combating Antibiotic-Resistant Bacteria (CARB), with the primary purpose to guide activities and actions by the government, public heath, healthcare, and veterinary partners to address the AMR threat. The National Action Plan laid out five main goals and charged the FDA’s Veterinary Laboratory Investigation and Response Network (Vet-LIRN) and USDA’s National Animal Health Laboratory Network (NAHLN) with enhancing their efforts to identify emerging resistance in animal pathogens with the goal of increasing antimicrobial stewardship. As of 2017 and 2018 (respectively), Vet-LIRN and NAHLN are collecting data on the antimicrobial susceptibility of clinically relevant bacterial isolates from different animal hosts, including companion animal species.
The primary goal of both projects is to monitor AMR profiles in animal pathogens routinely isolated by veterinary clinics and diagnostic laboratories across the U.S. By developing a centralized data collection and reporting process across all of these laboratories, data can be monitored for trends in antimicrobial resistance phenotypes and genotypes to identify new or emerging resistance profiles, to help monitor the continued usefulness of antibiotics over time, and to provide information back to our stakeholders regarding these trends.
In 2018, Vet-LIRN continued a pilot project with twenty “source” laboratories to conduct antimicrobial susceptibility testing (AST) of S. pseudintermedius and E. coli from dogs and Salmonella enterica spp. from any animal host. Concurrently, NAHLN initiated the first year of the NAHLN AMR pilot project, covering the period January 1, 2018 to December 19, 2018. Nineteen laboratories provided AST data to the NAHLN. Microbiological data from four livestock species (cattle, swine, poultry and horses), and two companion animal species (dogs and cats) were included. Bacterial isolates surveyed were E. coli (across all animal species), Salmonella enterica spp. (across all species), Mannheimia haemolytica (from cattle), and Staphylococcus intermedius group (from dogs and cats). Both networks followed CLSI AST testing methods. To view the Joint FDA and USDA Report on Antimicrobial Resistance visit: 2018 Animal Pathogen AMR Data.
The NARMS collaboration is possible only because of the dedicated effort of many people across Federal, State, and academic institutions. A complete list of NARMS partners can be found in the supplemental material.
- CDC. Antibiotic Resistance Threats in the United States, 2019. Atlanta, GA: U.S. Department of Health and Human Services, CDC; 2019.
- O’Donnell AT, et al. Quinolone-Resistant Salmonella enterica Serotype Enteritidis Infections Associated with International Travel. Clin Infect Dis. 2014 Nov 1:59(9):e139-141.
- Plumb ID, et al. Outbreak of Salmonella Newport Infections with Decreased Susceptibility to Azithromycin Linked to Beef Obtained in the United States and Soft Cheese Obtained in Mexico - United States, 2018-2019. MMWR. Morbidity and mortality weekly report, 2019 Aug 23: 68(33): 713–717.
- Tate H, et al. Comparative Analysis of Extended-Spectrum-β-Lactamase CTX-M-65-Producing Salmonella Enterica Serovar Infantis Isolates from Humans, Food Animals, and Retail Chickens in the United States. Antimicrob Agents Chemother. 2017 Jun 27; 61(7):e00488-17.
- Brown AC, et al. CTX_M-65 Extended-Spectrum β-Lactamase-Producing Salmonella Enterica Serotype Infantis, United States. Emerg Infect Dis. 2018 Dec; 24(12):2284-2291.
- Threlfall EJ, et al. Spread of Resistance from Food Animals to Man-The UK Experience. Acta Vet Scan Suppl. 2000; 93:63-8; discussion 68-74.
- Poppe C, et al. Salmonella Typhimurium DT104: A Virulent and Drug-Resistant Pathogen. Can Vet J. 1998 Sep;39(9):559-65.
- Sun J, et al. Towards Understanding MCR-like Colistin Resistance. Trend Microbiol. 2018 Sep;26(9):794-808.
- Nang SC, et al. The Rise and Spread of mcr Plasmid-Mediated Polymyxin Resistance. Crit Rev Microbiol. 2019 Mar;45(2):131-161.
- Carroll, LM et al. Identification of Novel Mobilized Colistin Resistance Gene mcr-9 in a Multidrug-Resistant, Colistin-Susceptible Salmonella enterica Serotype Typhimurium Isolate. MBio. 2019 May 7; 10(3).
- Tyson GH, et al. The mcr-9 Gene of Salmonella and Escherichia coli Is Not Associated with Colistin Resistance in the United States. Antimicrob Agents Chemother. 2020 Jul 22;64(8)e00573-20.
1 CLSI. Performance Standards for Antimicrobial Susceptibility Testing. 30th ed. CLSI supplement M100. Wayne, PA: Clinical and Laboratory Standards Institute; 2020.
2 CLSI breakpoints for azithromycin are only established for Salmonella ser. Typhi. Interpretive criteria for Salmonella ser. Typhi are based on MIC distribution data and limited clinical data. The azithromycin interpretive standards used for Salmonella serotypes other than ser. Typhi and for E. coli are NARMS-established breakpoints for resistance monitoring and should not be used to predict clinical efficacy.
3 CLSI breakpoints for azithromycin are only established for Salmonella ser. Typhi. Interpretive criteria for Salmonella ser. Typhi are based on MIC distribution data and limited clinical data. The azithromycin interpretive standards used for Salmonella serotypes other than ser. Typhi and for E. coli are NARMS-established breakpoints for resistance monitoring and should not be used to predict clinical efficacy.