Catmanager saw this AVMA news item the other day: “How to submit samples, report cases related to adulterated pet food.”

The story explains that “the American Association of Veterinary Laboratory Diagnosticians, in consultation with the Food and Drug Administration, was refining a working definition” of what consitutes a “case” of food-recall-related illness. As part of their effort, the AAVLD has set up an online survey (available on the AAVLD Web site and conducted by Michigan State University and the University of Guelph on behalf of the AAVLD) to collect incidents that diagnostic laboratories or veterinarians suspect are related to the food recall.

The AAVLD will share data with the FDA and will present an analysis of the data at the October 2007 AAVLD meeting.

Catmanager was excited to hear about the AAVLD survey, but after checking it out I was somewhat confused.

What is the purpose of the survey? Veterinarians have already been encouraged to report incidents to the FDA, their state veterinarian, and (if they are members) to VIN. Now they need to submit to another organization? Already taxed for time, why would they contribute to AAVLD?

According to the AVMA, the primary goal of the survey

is to distinguish true cases of nephropathy unique to this recall, hopefully resulting in a set of criteria defining a true case.

Other survey objectives are to characterize the spectrum of lesions; the temporal and geographic distributions of the suspected intoxications; the species, breeds, and ages of affected animals; and when possible, the brands, lot numbers, and UPC numbers of pet food involved in the toxic exposure, and results of chemical analyses.

Catmanager interprets this to mean that the AAVLD isn’t in the business of counting cases. They aren’t trying to create an authoritative tally (that’s up to the FDA). Instead, they’re trying to figure out how one should go about counting cases. In an e-mail responding to my questions, survey coordinator Dr. Wilson Rumbeiha seemed to confirm this interpretation.

The main objective is to define a true case of pet-food-induced nephrotoxicity derived from several criteria including history of ingestion of contaminated pet foods, documentation of renal failure by chemistry assays and urinalysis, histopathologic signs consistent with pe-food-induced nephrotoxicity and chemistry tests for markers of contaminants found in food, tissues and/or urine. Everyday animals fall sick or die of acute renal faliure and so our job is to come up with criteria that decide what is in and what is out.

Who should be submitting data to the survey? The survey is publicly accessible and doesn’t require registration or passwords. According to Dr. Rumbeiha, the openness of the survey “is meant to encourage participation rather than discourage.” However, some of the

information asked for in the survey is not what you would expect regular veterinarians to have. That information would be gathered at diagnostic laboratories or selfstanding commercial labs like IDEXX. Hence we expect that only terminal diagnosticians will complete the survey as primary practitioners will not have the information handy.

Looking over the survey questions, catmanager suspects primary practitioners could answer all of the questions (in fact, many of the questions about patient history will be more easily answered by the primary practitioner unless they provide extremely detailed histories to the labs) but might not be able to fully answer question 9, which asks about crystal composition. The average practitioner will need to rely on labs to confirm the presence of melamine in the pet’s food or tissues, for example.

Still, I wondered about the openness of the survey. I can without difficulty imagine pet owners attempting to complete the survey. Would the researchers screen out data submitted by laypersons? Dr. Rumbeiha indicated that they will, and he pointed to the unique animal ID (such as a case number assigned by a lab) and contact information asked by the survey. Dr. Rumbeiha and three other pathologists will contact “each and every case entered” using that contact info. Any unauthenticated cases will be discarded from the survey.

Will the survey, which was released on April 4, be updated to include the foods subsequently added to the recall? Dr. Rumbeiha said no, they’ll rely on people using the spot for “other” in the list of foods. This holds open the possibility that the survey could identify new foods in need of recall, although it’s not certain whether that information would be identified soon enough to have a practical effect.

When can we expect to see results of the survey? After one or two months, the survey will close so that the team contacting the cases, reviewing slides, and checking the data has enough time to complete their work before the fall presentation to the AAVLD. Because the AAVLD commissioned the survey, they get to decide when and how they’ll release the data to the public. Dr. Rumbeiha wasn’t sure when that might be. He also wasn’t sure when or how the data would be shared with the FDA, but he did say they wouldn’t be sharing data in real time.

Our job is to define criteria of what is a real case of pet food poisoning and what is not. . . . the real work will follow the survey where pathologists will come up with these criteria defining what is in and what is out. That requires time and good scientific review.

So, check with your diagnostic lab to see whether they’re submitting reports to the AAVLD survey. If they’re not, consider submitting them yourself. (Pet owners, please save yourself and the researchers conducting the survey some time and don’t try to submit data yourself. You might, however, want to check that your veterinarian is aware of survey—many veterinarians aren’t members of AAVLD, and the AVMA news story announcing the survey won’t be published in print until the May 1 issue of JAVMA.)

As Pet Connection noted last week, this survey could serve as a model for a future reporting mechanism. One thing that struck catmanager was how simple setting up the survey appeared to be. It’s hosted at, which is one of many online survey companies that make it surprisingly easy—and inexpensive—to create your own survey. Of course, someone has to review the data (which presumes people have taken the time to submit data), and as Dr. Rumbeiha noted in his e-mail to me, the real work doesn’t start until the data are all in.

Still, from a technological standpoint, setting up a national adverse-events database for veterinary medicine should be doable with existing, commericially available software. Finding money for the real work of reviewing all the data is where the political will is needed.