For the last year or so, we have been working on a book about pet food, What Pets Eat, to be published by Harcourt late in 2009. One of the pleasures of a long-term project like this is the time to follow digressions wherever they lead. Last year’s (2007) massive pet food recall  was so much of a diversion that it resulted in a spin-off publication—Pet Food Politics:Chihuahua in the Coal Mine  —scheduled for release this September by University of California Press.
The recall made us especially curious about the role of melamine, the substance responsible for kidney failure in pets eating food that supposedly contained wheat gluten. Just a glance at the chemical structure of melamine shows that it is high in nitrogen, a nutrient usually obtained from protein. This made us suspect that melamine must have been added deliberately to boost the apparent amount of protein in wheat gluten, because methods that measure the amount of protein in animal feed count nitrogen, not protein itself. Our suspicions were confirmed. The toxic “wheat gluten” turned out to be wheat flour laced with melamine.
But why would melamine harm cats and dogs? A quick search for studies of melamine toxicity turned up several performed on rats and mice, but just one on dogs (which dated from 1945).These studies gave the impression that melamine was not very toxic except at extremely high doses. Furthermore, the kidneys of animals who had eaten the contaminated food contained odd crystals that did not look like crystals of melamine.These turned out to be formed from complexes of melamine and one of its by-products, cyanuric acid. Even so, researchers and federal officials were puzzled. They had not heard of associations of melamine or cyanuric acid with kidney failure.
Really? Our Internet search turned up a brief and not particularly informative abstract of a 1960s study on melamine toxicity in sheep.We thought we needed to look at the entire paper, and found it and others in old bound journals in the Cornell library. These decades-old studies demonstrated that melamine is quite toxic, and causes kidney-related symptoms in animals at doses nearly identical to those reported in contaminated pet food. The studies were designed to test the idea that, because melamine nitrogen is far less expensive than protein nitrogen, melamine might have two useful purposes: (1) as an honest feed additive for ruminant animals, whose microorganisms can convert nonprotein nitrogen to amino acids, and (2) as a dishonest adulterant that makes feed test as though it contains more protein than it really does.
By following other leads,we also found references to relevant studies from the early 1980s in Italian journals that we had to request through interlibrary loan. These showed that melamine was so frequently used for fraudulent purposes in the 1980s that Italian scientists developed a test for it. They used the test to show that melaminna had been used to adulterate more than half the samples of fish meal they examined.
Most of the early information about melamine toxicity came from attempts to use it as a drug or nutrient. In the 1940s, investigators explored its potential as a canine diuretic. To follow what comes next, pay attention to the size of the melamine dose in milligrams per kilogram (mg/kg) of body weight, and recall that a kilogram is 2.2 pounds. The 1945 study showed that at a dose of about 120 mg/kg, dogs excreted crystals in their urine but otherwise did fine. In the 1960s, investigators used cyanuric acid (which is 32 percent nitrogen) to feed ruminant animals, and observed no problems even at high doses.Wouldn’t melamine (66.6 percent nitrogen) work even better?
A South African scientist fed daily doses of about 250 mg/kg to sheep, but most animals refused food and lost weight, and some of them died. Another South African investigator gave melamine doses to a single sheep, starting with 2,600 mg/kg. At such high doses, the sheep died within a few days from kidney damage, and the investigators could see crystals of melamine hanging from the animal’s prepuce. Lower doses of melamine caused sheep to stop eating, especially if their water intake was restricted. These studies suggested that a dose of about 250 mg/kg kills some— but not all—sheep over time.So, by 1968, melamine was known to induce kidney damage when fed to sheep over prolonged periods. Ten years later, American investigators tested melamine in cattle and found that a dose of about 100 mg/kg caused four out of six steers to refuse feed.
We think these studies are highly relevant to the pet food situation. The FDA reported that melamine could have accounted for 10 percent of the weight of the false wheat gluten, and the false wheat gluten could have accounted for as much as 10 percent of the weight of the pet food. If so, 100 grams—about 3 ounces—of pet food could have contained as much as a gram (1,000 milligrams) of melamine, and an average cat or small dog could have eaten an amount close to the 250 mg/kg level that proved toxic to sheep in the 1960s.
Recently, investigators from Georgia and the University of California, Davis, have shown that much smaller amounts of melamine can form crystals if cyanuric acid is also present. In cats, doses as low as 32 mg/kg each of melamine and cyanuric acid caused crystallization and kidney blockage.
We were surprised that neither the university or FDA veterinarians involved in the melamine investigations knew about this earlier work, but we think we can guess why. Papers in international journals are not readily accessible on the Internet, and the old animal feed literature is not likely to be studied these days.We had to discover the papers the old-fashioned way, by going to the library in person, sifting through reference lists, following up leads that sometimes required interlibrary loans and pursuing the reference trail back to its origins.We had the interest and time to pursue these questions. For the veterinarians and FDA officials caught up in the heat of the recall, a trip to the library might have seemed like a luxury they could ill afford.
We can’t say whether earlier suspicion of melamine would have hastened the recall or improved veterinary care of the sick dogs and cats.But we can say that the old experiments on animal feeding are well worth reading, that it’s best to read entire papers and not just their abstracts, and that libraries still have much to offer that the Internet cannot.
This article is based on Nestle M., Nesheim M.C. Additional information on melamine in pet food [letter]. JAMVA 231(2007):1647.
Illustration by Heather Horton