When dogs started getting sick after eating jerky treats made in China, the U.S. Food and Drug Administration (FDA) wasn’t the only government agency to hear from affected pet owners.

The New York State Division of Food Safety and Inspection also received consumer complaints. In response, it collected samples of potentially tainted jerky and submitted them to the state Food Laboratory, a sister operation within the Department of Agriculture and Markets.

That laboratory, after years of investigation, eventually detected unapproved antibiotics in major brands of chicken jerky sold by national chain retailers. The discovery led in January to the first manufacturer recalls stemming from the protracted search for pathogens and toxicants in the pet treats.

The search isn’t over, though. The FDA does not consider the antibiotic contamination to be responsible for the sometimes fatal kidney disorder and other illnesses in dogs associated with consumption of jerky treats.

At the same time, the discovery highlights differences in the ways individual laboratories tackle the inquiry and identifies a variable that may be critical to their results: namely, how they define a test sample.

Dan Rice, director of the New York State Food Laboratory, said in recent interviews with the VIN News Service by telephone and email that in the course of their investigation, his lab decided to treat a single piece of jerky — one average serving size — as a test sample.

That decision appears to have been instrumental to the discovery of antibiotic residues because the residues showed up inconsistently among the samples. “What we found in our lab, if you did individual chicken jerky treats, some were quite high. But another chicken breast sample within the same package would be quite low,” Rice said.

An alternative to examining one piece at a time is to test homogenized samples. Homogenate is derived through a variety of methods, such as by grinding single strips or parts of strips and combining the grounds from multiple pieces. Sampling homogenate is a common approach used by many labs, Rice said, including his. “We typically start with homogenates and then during the investigation process, may focus in on individual components of a sample (i.e., a single jerky treat),” he said.

Homogenizing may affect test results by diluting any contaminant present, particularly if the food portions that went into the homogenate were unevenly contaminated.  

The FDA, which has been investigating the jerky puzzle since 2007, is among the laboratories that use homogenized samples. “It is true in general that the FDA’s definition of a sample is sometimes different from the state definition of a sample,” Rice said. “The main differences relate to subsampling. For example, say with fresh produce, what I consider 10 samples is considered one sample by FDA. There are national efforts to standardize the definition of a sample among state and federal programs,” he added.

Siobhan DeLancey, a spokeswoman for the FDA Office of Foods and Veterinary Medicine, said that FDA labs contracted to assist in the jerky investigation do test individual pieces but also sometimes test composites of three pieces.

“It depends on the test, how much of the sample we have (some owners only have a half-piece remaining in the bag) and how much is needed for a particular analysis,” DeLancey said.

In a statement posted to its website on Jan. 9, the FDA described the New York Department of Agriculture and Markets (NYSDAM) testing methodology as “different and reportedly more sensitive than currently validated and approved regulatory methods.”

DeLancey said the agency was not referring to the test samples but another aspect of testing methodology. However, she was unable to elaborate.

“Everybody does things slightly differently, and also, a lot of the tests are proprietary,” she said, adding that the FDA, too, “tends to closely guard” its methods. 

She added: "FDA is seeking to improve its ability to detect antibiotics and we are therefore eager to validate NYSDAM’s method, adapt it to our equipment, and begin using it to support regulatory action when such action is warranted."

Results from the New York Food Laboratory analysis provided to the VIN News Service in response to a Freedom of Information Law request show the lab detected five antibiotics: sulfaclozine, sulfaquinoxiline, tilmicosin, trimethroprim and enrofloxacin. Levels range from a few parts per billion to, in one instance, 1,596 ppb.

Why the levels were so variable is unknown. One theory, Rice said, is that the chicken in a given package comes from multiple sources. “So that one bag probably represents multiple sources of chicken,” he offered. “Maybe some producers are overusing antibiotics and some aren’t.”

Treat manufacturers characterized the findings as “trace.” Rice said some measurements were trace but not all. “We also found a few at what is considered very high levels,” he said.

Four of the five antibiotics detected are not approved in the United States for use in poultry. The exception, sulfaquinoxaline, is FDA-approved for treatment of coccidiosis and acute fowl cholera in chickens and turkeys. The amount of sulfaquinoxaline detected by the New York lab exceeded the regulatory tolerance of 0.1 ppm (100 ppb) in at least one sample.

Citing the antibiotic residues problem, between Jan. 9 and Jan. 25, five companies recalled assorted brands of chicken treats for dogs: Milo’s Kitchen Chicken Jerky and Kitchen Grillers sold by Del Monte; Waggin’ Train and Canyon Creek Ranch sold by Nestle-Purina; Publix chicken tenders; IMS Trading Corp. Cadet brand chicken jerky treats; and Hartz Mountain Chicken Chew and Oinkies Pig Skin Twists wrapped with chicken.

Those were the first manufacturer recalls of jerky treats related to the search for a mystery agent since the FDA began receiving reports of illness in dogs in 2006. When reports intensified in 2007, Wal-Mart and other stores pulled suspect product off their shelves, although jerky makers hadn’t issued recalls. Since then, some jerky treats have been recalled for conventional contaminants such as Salmonella, but those episodes are unrelated to the bigger jerky puzzle.

Rice said the New York laboratory is continuing its investigation into the pet treats.  

The lab conducts pet food and livestock feed testing primarily in support of state regulatory programs but has a cooperative agreement as well with the FDA Center for Veterinary Medicine to look at specific hazards in pet food and animal feed, Rice said.

However, its greater emphasis is on human food safety. The state laboratory analyzes some 25,000 samples of human comestibles each year, Rice said, compared with 600 to 700 samples of livestock and pet food.

The 42-person laboratory staff recently moved into a new facility. Soon, they expect to receive a sophisticated piece of equipment known as Time of Flight that can delve down to the molecular level of a sample to better identify what harmful substances may be present.

“It’s the closest you can get to ... a magic instrument,” Rice said.