Abstract
Eight out of a litter of 13 puppies were either born dead or died within 48 hours of birth. Three puppies that died shortly after birth were necropsied. Two puppies had hemorrhage in the thoracic and peritoneal cavities, intestinal serosa, and meninges. The third puppy was smaller than the other two puppies but did not have detectable hemorrhage. Brodifacoum, a second-generation coumarin anticoagulant, was detected in livers from the two puppies with hemorrhage. The dam did not have clinical signs of coagulopathy before or subsequent to whelping. The owners were confident that the dog had not been exposed to rodenticide for at least 4 weeks before whelping. A presumptive diagnosis of in utero brodifacoum toxicity was made. To the authors' knowledge this is the first time a second-generation coumarin anticoagulant has been detected in the liver of a newborn animal. This case is also unique because the dam was unaffected, suggesting that fetuses are more susceptible to brodifacoum toxicity than adult animals.
A 1-year-old Great Pyrenees bitch that was housed at a breeding kennel gave birth to 13 puppies after 63 days of gestation. Whelping was observed and was considered normal. However, two puppies were born dead, three died within 6 hours, and three died between 6 and 48 hours of birth. Four of the six puppies that died postpartum had variable hemoptysis or epistaxis before death. Two puppies were reported to have died suddenly without previous clinical signs of disease. The birth weight of the eight puppies that were born dead or died was 280 ± 107 g (mean ± 1 SD), whereas the birth weight of the five puppies that survived was 502 ± 57 g (mean ± 1 SD).
The three puppies that were born alive but died within 6 hours of birth were necropsied. The abdominal cavities of two puppies (puppy Nos. 1 and 2) each contained approximately 40 ml of uncoagulated blood. Both puppies also had approximately 15 ml of uncoagulated blood within the thoracic cavity. However, one (puppy No. 3) did not have visible hemorrhage. Representative tissue samples from all three puppies were fixed in 10% buffered formalin and routinely processed for histologic examination. Histologic examination of tissues from puppy Nos. 1 and 2 revealed microscopic hemorrhage within the intestinal serosa and meninges. No significant lesions were observed by histologic examination of tissues from puppy No. 3. No histologic evidence of any infectious process was present in tissues from any of the puppies. Aerobic cultures of lung, liver, small intestine, and kidney from all three puppies did not yield pathogenic bacteria. Direct fluorescent antibody (American Bioresearch, Seivierville, TN) examination of kidney, lung, and liver did not reveal canine herpes virus antigen in any of the puppies.
Liver samples from all three puppies were analyzed by high performance liquid chromatography (HPLC) for warfarin, bromadiolone, and brodifacoum. 17 Briefly, liver samples were extracted with HPLC grade acetonitrile (Fisher Scientific, Suwanee, GA). This was then evaporated under nitrogen and the residue redissolved in 0.5 ml of methanol-water (4:1). Solutions (0.1 mg/ml) of the three anticoagulant standards (Chem Service, Inc, West Chester, PA) were prepared in HPLC grade acetonitrile. The mobile phase consisted of 33% acetic acid (1.5%, pH 4.7), 66% acetonitrile, and 1% dibutylamine at a flow rate of 2 ml/minute. HPLC was performed using a C-18 reverse phase column (Aquasil, 250 mm × 4.6 mm × 5 μ, Keystone Scientific, Inc., Bellefonte, PA) and detection with UV light and fluorescence (excitation wavelength at 285 nm, emission wavelength at 390 nm) (Thermoquest Corporation, Riviera Beach, FL). Detection limits were 1 ppm for warfarin, 10 ppb for bromadiolone, and 5 ppb for brodifacoum.
Brodifacoum was detected in samples of liver from puppy No. 1 (0.63 ppm) and puppy No. 2 (0.23 ppm). Brodifacoum was not detected in liver samples from puppy No. 3. Both ingestion by the puppies and transplacental transfer were considered as possible routes by which the puppies were exposed to the anticoagulant. Because coumarins are not excreted in milk, 11 ingestion of the anticoagulant could only have been from the environment of the whelping box or from the skin of the bitch during nursing. However, the owners were confident that the whelping box that the dogs were housed in was regularly cleaned and that rodent poisons were not used anywhere in the breeding facility. Additionally, because the shortest half-life of the vitamin K–dependent coagulation factors in adult dogs is 6.2 hours (factor VII), 19 it was considered unlikely that ingestion of brodifacoum would result in a fatal coagulopathy within the first 6 hours of life. Although the dog had been housed at the breeding kennel for the last 4 weeks of gestation, the owners considered that rodenticide exposure was possible before this. Therefore, placental transfer was considered the likely source of the anticoagulant, and a presumptive diagnosis of in utero brodifacoum toxicosis was made. It was most likely that the dam ingested brodifacoum during the first 3 weeks of gestation. No subsequent cases of rodenticide toxicity have been observed within this breeding kennel.
Brodifacoum is a coumarin anticoagulant. All coumarins block the vitamin K 2,3-epoxide reductase enzyme. 6 This prevents cycling between the inactive oxidized and the active reduced form of vitamin K in the liver. 6 Reduced vitamin K is essential for the carboxylation (and thus activation) of the clotting factors II, VII, IX, and X, 6 and a coagulopathy results once the existing factors are depleted. Coumarin anticoagulants are divided into two generations. Warfarin is an example of a first-generation compound. Although developed as a rodenticide, warfarin is also used as an anticoagulant in human and, less commonly, veterinary medicine. Second-generation coumarins were developed in the 1970s in response to the emergence of warfarin-resistant rats. 6 Common examples of these chemicals include brodifacoum, bromadiolone, and difenacoum. Because second-generation coumarins have half-lives up to 10 times longer than warfarin, 4 single doses are lethal to rodents. The increased half-life is caused by the addition of a tetrahydronaphthyl side chain that increases the affinity of the molecule to the vitamin K 2,3-epoxide reductase enzyme. 4
Poisoning of both animals and humans by coumarin derivatives has been reported. 5 , 19 In the United States, brodifacoum is the most commonly implicated chemical in cases of rodenticide toxicity. 5 In animals, common clinical signs of toxicosis include dyspnea, lethargy, coughing, hemoptysis, and pallor. 19 The serum half-life of brodifacoum in adult dogs is 6 ± 4 days, 22 however, persistence of brodifacoum in liver samples has not been defined. In human fetuses, immaturity of fetal hepatic biotransformation enzyme systems and divergent protein-binding properties markedly slow warfarin elimination when compared with adults. 1 Because brodifacoum and warfarin have similar chemical structures, persistence of brodifacoum in fetal livers is expected to be longer than in adults.
The majority of the information regarding coumarin anticoagulants in pregnancy is derived from the use of warfarin to prevent pulmonary embolism in pregnant women. 10 , 11 Although an effective anticoagulant, warfarin was found to readily cross the placenta and administration during pregnancy decreased mean birth weight 18 and resulted in abortion, stillbirth, or neonatal death in 17% 10 to 36% 18 of patients. In humans, although not in other animals, warfarin is also teratogenic resulting in nasal hypoplasia, stippled epiphyses, and central nervous system defects. 10 One case of canine fetal warfarin toxicity has been reported. 9 In this case, the dog was observed to ingest warfarin 4 days before whelping. Vitamin K therapy normalized the prothrombin time in the dam but not in the full-term puppies that were all either born dead or died because of massive hemorrhage within the first day of life. 9
There are few reports of second-generation coumarin anticoagulant poisoning during pregnancy. Coagulopathy caused by brodifacoum was observed in a dog 10 days after mating. 13 In this case, treatment with vitamin K throughout gestation resulted in resolution of clinical signs and the puppies were born clinically normal. Brodifacoum ingestion resulted in multiple cases of toxicity in sheep and goats in Egypt. 8 Death of adult animals as well as abortions and stillbirths were reported. One woman developed oropharyngeal hemorrhage, back pain, and hematuria after deliberately ingesting brodifacoum during the 21st week of pregnancy. 23 Vitamin K therapy resolved the coagulopathy, and the patient subsequently delivered a full-term clinically normal baby.
To the authors' knowledge, this is the first time a second-generation coumarin has been detected in the liver of a neonate, presumably after crossing the placenta. Warfarin readily crosses the placenta, 11 and considering the similar molecular structure of brodifacoum it is likely that this molecule is also able to pass transplacentally.
The present case is also unique because no symptoms of bleeding were observed in the bitch during gestation. In a report of four fatal brodifacoum toxicities in adult dogs, hepatic brodifacoum concentrations ranged from 1.2 to 13 ppm. 7 This compares with hepatic brodifacoum concentrations of 0.63 and 0.23 ppm in puppy Nos. 1 and 2, respectively. Although the dam was asymptomatic, it appears that the quantity of rodenticide that she ingested was sufficient to decrease the activation of vitamin K–dependent coagulation factors in the fetus. The puppies then developed fatal hemorrhage as a result of trauma during birth and the neonatal period. This suggests that fetuses are more susceptible to brodifacoum toxicity than adult animals.
There are three possible mechanisms to explain the increased susceptibility of fetuses and neonatal animals to maternal ingestion of vitamin K antagonist anticoagulants. First, transplacental movement of anticoagulant will inhibit vitamin K 2,3-epoxide reductase enzyme activity and reduce hepatic vitamin K cycling more in fetuses than in adult animals. 21 This is because adults have an alternative pathway to reduce inactive vitamin K. This pathway is considered the antidotal pathway to coumarin anticoagulants and is catalyzed by pyridine nucleotide–dependent enzymes that are markedly less sensitive to coumarin inactivation. 21 Studies in fetal rats show that the pyridine nucleotide–dependent pathway does not start functioning until late in gestation (day 17 of 21). 21
Second, fetuses are more susceptible to maternal coumarin toxicity because of in utero vitamin K metabolism. The placenta is the only source of vitamin K for the fetus. 12 However, because of the lipophilic nature of the vitamin, placental transfer of vitamin K is slow. 20 If maternal plasma vitamin K concentrations remain normal throughout gestation, adequate vitamin K crosses the placenta to enable normal coagulation factor production. However, if maternal plasma vitamin K concentrations are decreased during pregnancy, fetal vitamin K deficiency can result. 16 Furthermore, because the quantity of vitamin K able to cross the placenta is limited by the slow placental transfer rate, the fetus cannot compensate for periods of prolonged maternal deficiency, even if maternal vitamin K concentrations return to normal. 16 Hypovitaminosis K has been reported in babies of patients suffering from diseases that cause vitamin K malabsorption. 16 In a study of such cases, two of three patients did not show other clinical evidence of vitamin K deficiency. 16 In the present case, maternal brodifacoum ingestion may have asymptomatically reduced maternal plasma vitamin K concentrations during gestation. This may then have prevented transplacental transfer of sufficient vitamin K to enable normal fetal coagulation factor production. The detection of brodifacoum within fetal liver samples may then have been an indicator of maternal hypovitaminosis K rather than the direct cause of the coagulopathy.
Third, puppies are born with concentrations of coagulation factors that are approximately half those observed in adult dogs. 15 This results in an activated partial thromboplastin time in neonatal puppies that is approximately 60% longer than in adults. 15 Because neonates are inherently mildly deficient in coagulation factors, it is likely that any further inhibition of coagulation factor production will have a greater impact in a neonatal puppy than in an adult dog.
Except for one, all the puppies that died had birth weights lower than average for this breed. In humans, reduced birth weight has been reported with fetal vitamin K deficiency caused by transplacental warfarin toxicosis, 18 hereditary fetal vitamin K deficiency, 3 and maternal hypovitaminosis K. 16 Although the mechanism for this reduced weight is unclear, it may be caused by inadequate vitamin K–mediated carboxylation of bone proteins. 14 Surprisingly, low birth weights were not reported in hereditary vitamin K–deficient Rambouillet lambs. 2
Five out of the litter of 13 puppies survived and are reported to be clinically normal. The mean weight of these five puppies was significantly greater than the mean weight of the eight puppies that died. The origin of this variation in weight and survival is unclear. However, it is speculated that the puppies that survived may have had greater placental development. Although this may have resulted in a higher initial transfer of brodifacoum, a larger placental surface area may have enabled more placental transfer of vitamin K. This could then allow these five puppies to develop to a greater size and also to produce adequate activated coagulation factors before birth.
Brodifacoum was not detected in samples of liver from puppy No. 3. This puppy did not have gross or histologic evidence of hemorrhage, and it is probable that this puppy did not die of brodifacoum toxicosis. Puppy No. 3 was the smallest of the three puppies necropsied (198 g compared with puppy Nos. 1 and 2 that both weighed 340 g). It is hypothesized that the small size of this puppy was due to inadequate placental development. If the surface area of the placenta was small, it is possible that only undetectable levels of brodifacoum crossed the placenta of this puppy.
Serum samples from the dam were not available for brodifacoum testing in this case. However, considering that the serum half-life of brodifacoum in adult dogs is 6 ± 4 days 22 and that serum concentrations of brodifacoum are much lower than liver concentrations, 7 it is unlikely that brodifacoum would be detectable in samples taken from the dam 4 weeks after subclinical ingestion of rodenticide.
In conclusion, this case provides circumstantial evidence that brodifacoum can cross the canine placenta and cause neonatal hemorrhage. Maternal ingestions of coumarin anticoagulants may cause fetal coagulopathy by blocking fetal hepatic vitamin K cycling or decreasing maternal plasma vitamin K concentrations and subsequent placental transfer of vitamin K. Regardless of the mechanism, it appears that the amount of brodifacoum required to be ingested to cause fetal coagulopathy is less than the amount required to cause coagulopathy in an adult dog. Therefore, brodifacoum toxicity in a pregnant animal should be treated aggressively even if maternal coagulopathy is not observed.