Resistance changes mixed across bug and drug combos

The percentage of Salmonella isolates found in ground turkey that displayed resistance to nalidixic acid dropped from 8.1 percent in 2002 to 2.6 percent in 2007.

Similarly, the percentage of isolates resistant to ceftiofur dropped from 8.1 percent to 5.3 percent in that period.

In contrast, the percentage of Salmonella isolates in ground turkey with resistance to ampicillin increased from 16.2 percent to 42.6 percent in that time.

These are just a few examples of the changes in antimicrobial resistance prevalence across permutations of bacterial isolates, antimicrobials, and meat samples included in recently released data from the Food and Drug Administration's arm of the National Antimicrobial Resistance Monitoring System and included in the FDA's summary of the NARMS Retail Meat Annual Report.

The Department of Agriculture and the Centers for Disease Control and Prevention had not, at press time, published their agencies' 2007 NARMS reports, which are used with the FDA data for surveillance of antimicrobial resistance in the United States. The last available collaborative annual report is the NARMS 2006 Executive Report.

David G. White, PhD, director of the office of research at the FDA Center for Veterinary Medicine, noted the recently released data from the FDA indicate the percentage of Enterococcus isolates found in pork chops that displayed resistance to quinupristin-dalfopristin dropped from 27 percent in 2002 to 2 percent in 2007. However, these figures do not include E faecalis, which the FDA considers to be intrinsically resistant to the antimicrobial and which accounted for most isolates.

But, he said, percentages of Enterococcus isolates from ground turkey samples displaying resistance to some aminoglycosides increased in the same period. The prevalence of gentamicin resistance rose from 20.4 percent in 2002 to 34 percent in 2007, and the prevalence of kanamycin resistance rose from 28.9 percent to 41.6 percent.

Enterococcus data reflect ongoing debate

Dr. White acknowledged some concern that the use of virginiamycin in food-producing animals could select for resistance to quinupristin-dalfopristin in certain bacteria. Both are streptogramin antimicrobials.

In a November 2004 draft risk assessment regarding streptogramin resistance in E faecium, however, the FDA said: "It is difficult to assess the extent of transfer of streptogramin resistance from virginiamycin-exposed E faecium to E faecium found in human infections based on the available data." The report includes an analysis of two possible scenarios in which 10 percent or 100 percent of resistance to the drug class was attributed to the use of virginiamycin in food animals. Under these scenarios, the risk in any given year of a person becoming infected with E faecium resistant to streptogramin antimicrobials as a result of food animal use ranged from 0.7 to 140 chances in 100 million.

The AVMA has disputed suggestions that a ban on virginiamycin in food animals would benefit human health. Dr. Lyle P. Vogel, then assistant executive vice president of the AVMA, testified before a Senate committee in June 2008 that the prevalence of quinupristin-dalfopristin resistance among Enterococcus isolates from humans in Denmark is 10 times the prevalence among isolates from people in the United States, despite Denmark's ban on the use of virginiamycin in food animals.

Dr. White said decreased use of an antimicrobial has been associated with decreased resistance, but NARMS does not currently obtain the types of usage data needed to determine what led to the drop in resistance to quinupristin-dalfopristin (trade name Synercid) among Enterococcus isolates from pork chops. Streptogramins may be used less often in swine production than in previous years, for example, but other factors, such as improved biosecurity, could also be impacting resistance development.

"The pork industry has a very good quality assurance program, and that actually shows up in our E(scherichia) coli prevalence (data)," Dr. White said. E coli is less common in pork chops than in other retail meat samples, he said.

The NARMS data indicate the percentage of Enterococcus isolates in beef—excluding E faecalis isolates—displaying quinupristin-dalfopristin resistance also dropped from 46.2 percent in 2002 to 6.2 percent in 2007. Resistance prevalence dropped by a smaller amount among isolates from ground turkey, for which prevalence decreased from 79.6 percent to 73.5 percent, and among isolates from chicken, for which prevalence decreased from about 56.3 percent to about 54.6 percent.

Data provide opportunities for analysis

Dr. White said it is also unclear why the percentage of Enterococcus isolates from ground turkey resistant to gentamicin and kanamycin has increased so substantially, and multiple years of data are needed to evaluate trends.

He said the number of combinations of bacteria, drugs, and meats has so far led the FDA to stick with a data-driven report, but the agency could have more-interpretive reports within the next few years. For now, he hopes veterinarians and others involved in public health will examine the data for substantial trends.

In gathering the 2007 NARMS retail meat data, officials at FoodNet laboratories in California, Colorado, Connecticut, Georgia, Minnesota, New Mexico, New York, Oregon, and Tennessee collected about 4,300 meat samples from randomly selected chain grocery stores near the laboratories. Officials at the FoodNet facilities cultured meat samples for Campylobacter and nontyphoidal Salmonella organisms, and sites in Tennessee, Georgia, and Oregon also cultured the samples for E coli and Enterococcus organisms.

Samples that tested positive for those bacteria were sent to the FDA-CVM for serotype or species confirmation.

More data needed to complete the picture

Dr. White said comparing the FDA's retail meat data with data from the CDC regarding antimicrobial resistance in humans and with data from the USDA regarding antimicrobial resistance in food animals will provide a more complete picture.

"The retail (information) is interesting, but it needs to be compared with the human data and what we see in the actual food animals as well," Dr. White said.

Felicita Medalla, MD, a CDC epidemiologist with NARMS, said, "We are part of that three-federal-agency collaboration, and to get the complete picture, we need the data from all three arms of NARMS. And we work very closely with each other."

The USDA's data were expected to be available by the end of October, and the CDC's data by early 2010.

The FDA's NARMS data are available at www.fda.gov. Under the "Animal & Veterinary" tab, click on "Antimicrobial Resistance."