What an adverse event report is, and is not

An adverse event report tells FDA's Center for Veterinary Medicine (CVM) that something unwanted was observed in connection with an animal drug. FDA holds 1,357,245 of them, going back to 1987. They are the raw material of drug monitoring after approval, and they are easy to misread.

Who files them

Drug companies must send FDA every report they receive about their approved animal drugs. Most reports start as a phone call from an owner or a veterinary clinic to the company, which writes up the case and submits it. Owners and veterinarians can also report directly to FDA. Nobody checks a report against medical records before it enters the database.

What a report contains

• Every drug the animal was given around the time, not only one under suspicion. A dog on a heartworm preventive, a flea product and a pain medication appears under the ingredients of all three.
• The signs observed, coded by a trained person into VeDDRA dictionary terms such as "Emesis" or "Lethargy".
• The animal: species, breed, sex, age and weight, when given.
• The outcome when the report was written: recovered, ongoing, died, euthanized or unknown.
• The type of report. 65% are filed as health events. 23% are wholly or partly about a product not working as expected (a dog on a preventive that still tested positive for heartworm, for example), and 12% are product or packaging defects, often with no animal involved.

FDA's public files leave out the written description of what happened, anything identifying the owner or clinic, and the brand and manufacturer names, which FDA masks. This site therefore counts by active ingredient and does not attribute any figure to a brand; see why you will not find brand names in the statistics.

What a report is not

It is not a finding. In FDA's words, the information "is as reported to the FDA’s CVM, and the agency has not necessarily determined if the product(s) in question was/were the actual cause of the events reported." An old dog with kidney disease that dies a week after a routine injection generates a report listing death; the report does not settle whether the injection played any part.

It is not a measure of frequency. FDA notes that reports "represent a small percentage of total usage numbers of a product" and that adverse events are under-reported. How many doses were sold is not public, so no rate per animal treated can be worked out.

It is not a basis for comparing products. FDA says so directly: the data cannot be used "to estimate the risk associated with the product or compare one drug or device product with another."

What FDA does with them

CVM reviewers read the full reports, including the parts the public does not see. They weigh the timing, the amount given, other drugs, the animal's history and whether signs stopped when the drug was stopped. Patterns that hold up can lead to label changes, letters to veterinarians or other action. The database is, in FDA's description, an early warning system. Counting reports, which is all this site does, is the first and crudest step of that work.

How to use these pages

To see what has been reported, how much, in which animals, and whether volume is changing. To prepare questions for a veterinarian. Not to decide on your own whether an animal should start, stop or switch a medication; this site is not veterinary advice.