Unmasking the toxic culprit(s) in pet-food recalls

WHAT IS YOUR INTERPRETATION AND DIAGNOSIS?

BACKGROUND: On March 16, 2007, Menu Foods Inc., a Canadian manufacturer of wet pet food, issued a recall because of concerns about adverse effects of some of their products on kidney functions of cats and dogs.

The recall was unprecedented in scope, eventually involving an estimated 1,000 products manufactured by Menu Foods and other large pet-food companies.

The food was manufactured by Menu Foods between Dec. 3, 2006, and March 6, 2007. On Feb. 22 and 28, 2007, Menu learned of the illness of two cats that ate some of their food. About the same date, it received a separate consumer complaint about the death of a cat that consumed the food.

On March 6 and 7, a company that Menu Foods retained to perform routine palatability tests informed the company of the death of one cat and the euthanasia of two of 20 other cats in a routine palatability study.

On March 9, four more cats assigned to the first palatability panel were euthanized because of severe illness. Two cats assigned to a different palatability panel also were euthanized.

About the same time, food-related acute to subacute renal failure was diagnosed in dogs and cats in veterinary clinics across the country. Menu Foods contacted the Food and Drug Administration (FDA) on March 15, 2007, and issued a voluntary pet-food recall the next day.

Aminopterin, an antagonist of folic acid at one time used to treat cancer, was initially thought to be the toxin in the recalled pet food. However, this finding was not confirmed by the FDA. Therefore, testing for other food-related contaminants continued. Menu Foods identified seven different ingredients, including wheat gluten, common to all incidents. Tests did not reveal any contaminants in the ingredients. However, melamine was eventually detected in the wheat gluten. The FDA confirmed this finding.

Later it was learned that melamine originating in China was added to wheat gluten. Contaminated wheat gluten subsequently was identified in rice protein concentrate. Then, the FDA learned that samples of the contaminated wheat gluten and rice protein were mislabeled. Both products actually were wheat flour contaminated with melamine and melamine-related products.

Melamine is an industrial chemical of no known nutritional value to dogs or cats. In mice and rats, it causes urolithiasis. Chronic irritation of the urothelium by uroliths induces urothelial hyperplasia and subsequent urinary bladder neoplasia in these rodents.

Melamine has been reported to cause diuresis in rats and dogs, crystalluria in mice, rats and dogs and fatal uremia characterized primarily by crystalluria in sheep.

Apparently melamine has been marketed as a fertilizer because of its high nitrogen content. In context of pet foods, because protein concentration in ingredients of food is measured by analysis of total nitrogen content, it has been alleged that melamine was intentionally added to increase the apparent protein content.

Melamine has a relatively high safety margin, and it is unlikely that melamine itself directly caused renal failure in these cats.

The FDA later announced that, in addition to melamine, cyanuric acid (also originating in China and containing a relatively high level of non-protein nitrogen) was added to pet foods.

Cyanuric acid is structurally related to melamine. It is sometimes used as a stabilizer in outdoor swimming pools and hot tubs to minimize the decomposition of hypochlorous acid by light. Unfortunately, a paucity of data is available about the toxicity of cyanuric acid in mammals. Sodium cyanurate fed subchronically to mice and rats caused uroliths, indicating poor solubility.

The evidence suggests that a combination of chemicals (melamine, cyan-uric acid, possibly others) formed insoluble crystals in the kidneys of these unfortunate pets, with subsequent physical damage to the renal tubules. Cornell University's Animal Health Diagnostic Center identified melamine in:

1) pet food

2) urine from cats that ate the contaminated food

3) in the kidney of one cat participating in Menu Food's palatability studies.

The University of Guelph's Laboratory Services Division induced a crystal-line complex of melamine-cyanuric acid in vitro by adding melamine and cyanuric acid to cat urine. Further analysis of the crystals revealed that they were composed of approximately 70 percent cyanuric acid and 30 percent melamine and were extremely insoluble.

Two other melamine-related substances (ammelide and ammeline) were identified. Whether these metabolites also are nephrotoxic is being evaluated. Additional in vivo studies revealed that melamine alone and cyanuric acid alone were not acutely nephrotoxic. However, as demonstrated by investigators at the University of California-Davis, the combination of these two industrial chemicals was extremely nephrotoxic, especially to cats.

The March 2007 pet food-associated renal failure epidemic in North America is similar to a March 2004 epidemic of pet food–associated renal failure that occurred in Asia, in which an estimated 6,000 dogs and fewer cats developed nephrotoxic renal failure after ingesting a certain brand of pet food.

Compare this data to that from Banfield, The Pet Hospital, which suggest that an estimated 39,000 dogs and cats developed renal failure in the 2007 North-American food-borne epidemic.

Minnesota Urolith Center input

The Asian epidemic (2004) and the North American (2007) epidemic corre-spond in many respects to the sudden recognition by our staff of uroliths received beginning in October 2002.

These uroliths and crystals were similar (if not identical) in appearance to the crystals and uroliths later described by several investigators (Figures 2 to 8).

The infrared spectrum of the uroliths did not match any of the known minerals in our database. We consulted with several laboratories in the United States; none was familiar with the substance in question.

Figure 1: Graph illustrating number of uroliths containing uric acid monohydrate (UAM). These uroliths were analyzed at the Minnesota Urolith Center.

Therefore, we sent samples to colleagues in Germany for analysis. They were evaluated via infrared spectro-scopy and solid-state NMR spectro-scopy; the analysis revealed uric acid monohydrate. This mineral type had been recently (2005) reported in humans. Since 2002, we have received uroliths containing the minerals of interest (uric acid monohydrate) retrieved from 520 dogs and cats living in Asia (Figure 1).

Figure 2: Photomicrograph of uric acid monohydrate crystals from a dog illustrating unique spherical yellow-brown crystals with radial striations.

During the period from 2006 to Aug. 11 this year, we have received 16 uroliths from the United States containing uric acid monohydrate (12 dogs and 4 cats: Figure 1).

Figure 3: Photomicrograph of uric acid monohydrate crystals from the dog described in Figure 2 illustrating characteristic crystals.

Staff at the Characterization Facility, University of Minnesota, compared our previously unknown samples to references for uric acid monohydrate obtained from the Cambridge Crystallographic Data Center (<www.ccdc.cam.ac.uk>), utilizing X-ray microdiffraction and electron dispersive spectroscopy.

Figure 4: Photomicrograph of amber-colored uric acid monohydrate crystals from the dog described in Figure 2, illustrating abundant crystals. These crystals appear similar to those reported by other investigators.

The results were consistent with references for uric acid monohydrate. Comparison of our samples by X-ray diffraction and infrared spectro-scopy with references for melamine and cyanuric acid did not match. From November 2002 to date, we have been reporting uroliths with these characteristics as uric acid monohydrate. (Figures 1 to 8)

Figure 5: Photograph of a nephrolith composed of uric acid monohydrate crystals enmeshed in large quantities of matrix. The nephrolith was soft and gooey. This nephrolith was obtained at necropsy from a 13-year-old, mixed breed, male neutered dog.

Greenish-yellow nephroliths were detected in some of the dogs involved in the Asian epidemic (Figures 5, 7, 8).

Figure 6: Photomicrograph of an H&E stained section of kidney removed from a 12-year-old mixed breed dog with acute renal failure. Note the characteristic appearance of the crystals. The crystals were highly birefringent when examined by polarizing light microscopy.

Additionally, crystals were detected in the renal tubules (Figure 5). The crystals in the kidneys of animals involved in the Asian episode were green-yellow in color, circular in shape and had striations radiating from their centers. They were identical to the crystals observed in the kidneys of animals involved in the 2007 pet-food associated epidemic and similar (if not identical) to our samples (Figures 2,3,4).

Figure 7: This nephrolith is composed of calcium oxalate with prominent surface crystals composed of uric acid monohydrate. This nephrolith was obtained from a 7-year-old, male Shih Tzu.

The Asian endemic of pet food-associ-ated renal failure was attributed to contamination of raw materials with mycotoxins (especially ochratoxin and/or citrinin) in a pet-food manufacturing plant in Thailand. Viewed in retrospect, this diagnosis is questionable.

Figure 8: Photograph of uroliths removed from a dog. Uroliths are comprised of an outer layer of magnesium ammonium phosphate (struvite) with an inner core of brown-yellow uric acid monohydrate

Armed Forces Institute of Pathology weighs in

Renal tissue from three dogs was submitted to the Armed Forces Institute of Pathology, Washington, D.C. Two of the cases were submitted by IDEXX Veterinary Services, West Sacra-mento, Calif., for light microscopic examination and characterization of the chemical composition of crystals observed previously in hematoxylin and eosin (H&E)-stained sections. The third case had been submitted from the Animal Technology Institute Taiwan, Division of Animal Medicine, Chunan, Miaoli, Taiwan, as a Wednesday Slide Conference case submission for the 2004-2005 academic year, and was retrieved retrospectively.

» Case No. 1 was a 3-year-old female Parson Russell Terrier that developed acute renal failure after eating canned food on the Menu Foods recall list. There was no known prior history of renal disease or ethylene glycol exposure. Clinical findings included azotemia and hyperphosphatemia. A CBC revealed that neutrophilic leuko-

cytosis was present. Liver enzyme values were within normal limits. The patient died. Necropsy findings included a perforated gastric ulcer, hyperemia and bleeding from the gastric wall, and green crystals in the renal pelves.

» Case No. 2 was a 3-year-old spayed female Bernese Mountain Dog that developed acute renal failure after eating canned food on the Menu Foods recall list. There was no known prior history of renal disease or exposure to ethylene glycol. Clinical findings included azotemia, hyperphosphatemia, hyperkalemia, decreased TCO2 and increased anion gap consistent with acute renal failure and concurrent metabolic acidosis. Because of the severity of the illness, the dog was euthanitized. The kidneys were collected at necropsy.

» Case No. 3 was a 1-year-old male mixed-breed dog that developed renal failure after eating a commercial dog food for several months. Macroscopic findings included prominent white powderlike deposits in the kidney. Clinical findings included marked azotemia consistent with renal failure, and neutrophilic leukocytosis.

Light microscopic morphology and histochemistry: In dog No. 1, approximately 90 percent of the renal tubules were either pale, swollen and vacuolated or were occasionally hyper-eosinophilic and shrunken with nuclear pyknosis or karyolysis on the HE-stained slide.

The basement membranes of tubules in the inner cortex were thickened and basophilic. There were numerous crystals evenly distributed throughout the cortex and medulla within renal tubules and collecting ducts. The crystals demonstrated bright birefringence when viewed under polarized light.

Approximately 75 percent of the crystals were round, pale brown and appeared to have a rough surface as a result of smaller crystalline structures being arranged radially and more randomly within the entire birefringent crystal. Some of these crystal structures were arranged in concentric circles. Occasionally, the centers of these concentric circles were empty.

The crystals were present within renal tubular epithelial cells and in the lumens of tubules, where they filled and in some areas distended the tubules. The crystals measured up to approximately 80 micrometers in diameter.

Birefringent crystals were present in the renal calyces. A second type of crystal comprising approximately 25 percent of the birefringent crystals was present within tubule lumens. Some of these appeared to be in the walls of blood vessels, in addition to the lumens of renal tubules.

Low numbers of lymphocytes and plasma cells were observed in the renal interstitium. Similar lesions and crystals were observed on HE-stained sections in dogs No. 2 and 3. Tubular basement membrane mineralization and basophilic non-birefringent particles were also present in dog No. 3 but were absent in dog No. 2.

Similar to dog No. 1, the crystals in dogs No. 2 and 3 were almost exclusively within renal tubules. In dog No. 2, occasional crystals within the renal interstitium elicited a granulomatous inflammatory response with multinucleate giant cells. In dogs No. 2 and 3, the rough, pale-brown crystals measured up to approximately 100 micrometers in diameter. There were multifocal aggregates of many lymphocytes, plasma cells and fewer neutrophils within the interstitium of dog No. 2. Prominent interstitial fibrosis was present in dog No. 3.

» In all three cases, Oil Red O72h-stain demonstrated variable degrees of the pale brown, rough-textured crystals, indicating a plastic or lipid origin. This stain was negative for the smoother platelike crystals. Application of 23 Oil Red O stain for at 24 and 48 hours yielded similar results.

In dogs No. 1 and 3, Alizarin Red S (pH 4.1-4.3) did not stain either birefringent crystal type. However, it did stain the basement membranes that were basophilic on HE section. This finding was consistent with the deposition of a calcium salt other than calcium oxalate, likely calcium phosphate.

The rough, pale brown birefringent crystals demonstrated the IR spectral characteristics of melamine-containing material. The infrared spectra of the pale-brown crystals did not match those of pure melamine or pure cyanuric acid, but were very similar to spectra obtained at the University of Guelph of a melamine-cyanuric acid complex.

To produce the same complex, they mixed aqueous solutions of melamine and cyanuric acid in varying relative amounts. When the solutions were mixed, spontaneous precipitation of crystals caused the solutions to appear cloudy, but the addition of urea (or formalin) to the solution significantly enhanced the formation of precipitates.

The recovered crystals were found to exhibit infrared spectra similar to those obtained from the kidney tissue of the three dogs.

The spectra were the same as those reported at the University of Guelph for urinary crystals from cats and melamine-cyanuric acid crystals formed in vitro in cat urine

Under scanning electron micro-scopy with energy dispersive X-ray analysis, melamine crystals demonstrated peaks for carbon, nitrogen and oxygen consistent with melamine-containing crystals. The presence of a nitrogen peak in the melamine-containing crystals reflected concentrations of nitrogen in excess of those seen in normal tissue protein.

» Comparison was made of urolith samples submitted from 2002 to date by Asian colleagues, with urolith samples submitted from 2006 to date from colleagues in the United States. These uroliths were evaluated by X-ray crystallography and electron dispersive spectroscopy. Representative samples from each time period were found to be identical in composition. We agree with the previous reports by Brown and colleagues and Thompson and colleagues that the pet food-associated renal failure outbreaks in 2004 and 2007 share clinical, histologic and toxicologic findings. This evidence indicates that they share the same cause.

» Melamine has a relatively high safety margin, and it is unlikely that melamine itself directly caused renal failure in these dogs. The evidence suggests that a combination of chemicals (melamine, cyanuric acid, possibly others) formed insoluble crystals in the kidneys of these unfortunate dogs with subsequent physical damage to renal tubules. SEM/EDXA revealed characteristics of melamine-containing crystals, including a peak for nitrogen, which is not seen when examining tissue protein under normal operating conditions. This finding highlights the high nitrogen content of both melamine and to a lesser extent cyanuric acid.

What's next?

We are contacting other colleagues with interest in this phenomenon to resolve the identity of the type of mineral in the uroliths and kidney tissue. Are they uric acid monohydrate? Are they metabolites of melamine and cyan-uric acid? Or are they something else?

Carl A. Osborne, DVM, PhD

Lisa K. Ulrich, CVT

Lori A. Koehler, CVT

Jody P. Lulich, DVM, PhD

Minnesota Urolith Center, University of Minnesota

REFERENCE: Schubert G, Reck G, Jaancke H and colleagues. Uric acid monohydrate-A new urinary calculus phase. Urology Research 2005, 33: 231-2238. Others available on request.

Author's Note: Because of an educational gift from Hill's Pet Nutrition Inc., the Minnesota Urolith Center is able to perform quantitative urolith analysis at no charge. A urolith analysis submission form may be obtained by faxing your request to (612) 624-0751 or visiting our Web site: www.cvm.umn.edu. Click the link to department and centers to find Minnesota Urolith Center. Click on the icon labeled "How to submit samples."

Acknowledgements: Parts of this work were carried out in the Institute of Technology Characterization Facility, University of Minnesota, which receives partial support from NSF through the NNIN program.