Materials and Methods

Criteria for case selection—An electronic computer database^a from an animal poison control center^b based in Minneapolis was searched to identify dogs ingesting zinc phosphide between November 2004 and July 2009. Inclusion criteria included a witnessed or suspected exposure to a zinc phosphide product or ingestion of an animal previously poisoned by zinc phosphide (defined as relay toxicosis); suspected exposures were included if packaging was destroyed, missing, or extensively altered. Whenever possible, the active ingredient was identified by the Environmental Protection Agency registration number, which is federally mandated to be on the label. Exclusion criteria included any noncanine species, poor product identification, and incomplete medical records.

Abstraction of data from medical records—Records were reviewed for location of the caller; breed, sex, age, and body weight of the dog; time since exposure; development or lack of clinical signs; clinical signs categorized into body systems affected (including gastrointestinal system, CNS, respiratory system, and cardiovascular system); treatments performed; and veterinary care (characterized as either an outpatient or in-hospital basis). In patients that were hospitalized, the duration of hospitalization, when available, was recorded. Follow-up with pet owners or hospital staff was performed by phone call at the time of this study and was designed to confirm outcome and treatment performed.

Statistical analysis—For parametric data, the mean and SD were reported, and for nonparametric data, median values and ranges were reported. A Wilcoxon-Mann-Whitney test was performed to determine significant differences between survivors and nonsurvivors regarding age, duration of hospitalization, and time since exposure to initiation of the call. A Fischer exact test was used to determine whether sex significantly affected survival rate. All analyses were performed with a commercially available statistical software package.^c Values of P < 0.05 were considered significant.

Results

During the defined study period, 375 cases were identified; of these, 362 dogs met the inclusion criteria of presumptive toxicosis from exposure to zinc phosphide. The remaining 13 cases were excluded because they were in noncanine species (including 8 cats, 2 horses, and 1 each of a cow, goat, and pig). Ten dogs were excluded from sex determination data because no sex was identified either on the initial call or by the veterinarian. Twenty dogs were excluded from statistics involving clinical signs because an accurate determination of their signs could not be made. Sixty-six dogs were excluded from survivability data because follow-up was judged to be incomplete or inaccurate.

In 247 of 362 (68.2%) dogs, the active ingredient was confirmed as zinc phosphide on the basis of the Environmental Protection Agency registration number. Identification of the active ingredient in the remaining 115 dogs was made by product name and active ingredients listed on the packaging.

Call initiation—In 72.1% (261/362) of the cases, the call was initiated by the animal owner. The remaining calls were initiated by other professionals, including veterinarians (14.4% [n = 52]), veterinary staff (6.9% [25]), or pesticide control officers and other miscellaneous individuals (6.6% [24]). Calls originated from 43 states as well as Canada and Puerto Rico.

Population distribution—There was no significant effect of breed (P = 0.55), sex (P = 0.67), or weight (P = 0.73) on overall survival rate; however, age was associated with survival rate (P = 0.01). Overall, 60 specific breeds, in addition to mixed-breed dogs, were represented. Mixed-breed dogs (25.4% [92/362 dogs]) and Labrador Retrievers (14.9% [54/362 dogs]) were the most commonly exposed. Other specified breeds included Golden Retrievers (5% [n = 18]), Chihuahuas (3.9% [14]), Dachshunds (3.6% [13]), German Shepherd Dogs (2.5% [9]), Beagles (2.2% [8]), Shih Tzus (1.9% [7]), and German Shorthair Pointers (1.9% [7]). The remaining breeds had < 5 representatives, and there were 14 unspecified breeds. One hundred seventy (48.3%) dogs were male, 182 (51.7%) were female, and in 10 cases, the sex was not recorded (2.8%). For the 359 for which it was recorded, the median weight was 15.9 kg (35 lb) and it ranged from 0.68 to 60.45 kg (1.5 to 133 lb). Overall, survivors were significantly older (median age, 2 years; range, 3 weeks to 14 years) than nonsurvivors (median age, 6 months; range, 4 months to 2 years).

Outcome—Overall, the survival rate associated with zinc phosphide toxicosis was 98.3% (291/296 dogs). A total of 5 dogs were nonsurvivors; 1 of these was found deceased by the pet owner, with no observed clinical signs prior to death, and of the remaining 4 cases, 2 dogs died during hospitalization and 2 dogs were euthanized.

The time from exposure to initiation of a phone call to the animal poison control center^b was recorded in 91.4% (331/362) of cases. In survivors (n = 291), the median time to initiation of a phone call was 1 hour (range, 2 minutes to 90 days). In nonsurvivors (n = 5), the median time since exposure to call initiation was 18 hours (range, 6 to 24 hours). A shorter time interval for call to the animal poison control center for assistance was found to have a significant (P = 0.009) effect on survival rate.

Clinical signs—Accurate information regarding development of clinical signs was available in 94.5% (342/362) of cases. Over half the dogs (58.8% [201/342]) did not develop clinical signs or have any clinical signs reported, and specific clinical signs were reported for the remaining 41.2% (141/342) of dogs. There were 180 total clinical signs seen in the 141 dogs for which they were reported, with some dogs having developed > 1 category of clinical signs. Of the dogs that developed clinical signs, 76.6% (108/141) had signs in a single body system, 18.4% (26/141) had clinical signs in 2 body systems, and 5% (7/141) had clinical signs in 3 body systems. Clinical signs involving the gastrointestinal tract, including vomiting, diarrhea, anorexia, ptyalism, and abdominal distension, were the most commonly reported type of clinical sign (66.7% [n = 120/180 reported signs]). Generalized malaise accounted for 17.8% (32/180) of the clinical signs, and CNS signs accounted for 8.9% (16/180). The most commonly reported CNS signs included altered mentation, unusual behavior, ataxia, tremors, and seizures. Respiratory signs accounted for 3.3% (6/180) of the clinical signs and included tachypnea, increased respiratory effort, coughing, sneezing, and evidence of pulmonary edema (on the basis of radiographic findings; n = 1 dog). Cardiovascular signs, including tachycardia, arrhythmias, and hypovolemic shock, developed in only 1.7% (3/180) of patients. Two dogs had miscellaneous clinical signs, including facial swelling and stiffness. Four of the 5 dogs that did not survive developed signs of gastrointestinal tract disease, and 2 developed neurologic signs, 1 developed respiratory signs, and 1 developed cardiovascular shock. Sudden death was the only description reported for 1 dog.

In addition to direct product ingestion, owners reported suspected relay toxicosis in 11 cases. Three of the 11 dogs developed clinical signs after ingesting a rodent presumably killed by zinc phosphide. Gastrointestinal disturbances occurred in 1 dog, generalized malaise occurred in another, and both gastrointestinal signs and malaise occurred in the third dog. The remaining 8 dogs suspected of having relay toxicosis did not develop any clinical signs. Confirmation of relay toxicosis was hindered by lack of consumed carcasses and advanced diagnostics.

Hospitalization and treatments—Based on follow-up phone calls, the type of veterinary care received was established in 90% (326/362) of cases. Approximately 72% (234/326) of dogs received primary veterinary care. Of these, 51.3% were hospitalized (120/234) for a median of 15.0 hours (range, 5 to 80 hours). Over a third of the dogs (36.3% [85/234]) were treated on an outpatient basis. The remaining 29 dogs received veterinary care, but an accurate determination of outpatient treatment versus hospitalization could not be made. In 107 affected dogs, emesis was induced as a means of decontamination, with 21.5% (23/107) of dogs treated by the pet owner on the basis of recommendations from a veterinarian or an animal poison control center; the remaining 78.5% (84/107) had emesis induced by a veterinarian. Gastric lavage was performed in 20.1% (47/234) of dogs, and activated charcoal was administered in 44.9% (105/234).

Of the 234 dogs receiving veterinary care, patients received the following treatments: IV fluids (19.7% [46/234]), antacids (16.7% [39/234]), antiemetics (6% [14/234]), antimicrobials (3% [7/234]), anticonvulsants (1.7% [4/234]), sucralfate (1.7% [4/234]), vitamin K₁ (1.7% [4/234]), corticosteroids (0.85% [2/234]), and oxygen (0.43% [1/234]).

Four nonsurvivors were hospitalized for treatment. Only 2 received decontamination procedures that included gastric lavage (2 dogs) and activated charcoal administration (1). Similar treatment modalities, including IV fluid therapy (3 nonsurvivors), antiemetics (1), anticonvulsants (1), and oxygen supplementation (1), were performed.

Discussion

Phosphide rodenticides are highly effective products designed to kill mice, rats, gophers, and other small rodents.^2,5 Aluminum phosphide is used primarily as a fumigant to protect grain in storage bins, ship holds, and railroad cars, but zinc phosphide is more commonly used in residential and agricultural settings as a bait.³ Zinc phosphide was first registered by the Environmental Protection Agency in 1947. It is classified, on the basis of animal studies, in Toxicity Category I, the highest of 4 categories for both oral and inhalation routes of exposure.¹ Although zinc phosphide can exert a toxic effect from either ingestion or inhalation exposures in human beings, only toxicosis resulting from ingestion has been reported for dogs and horses.^2,6,7

The mechanism of action of phosphide rodenticides is not completely understood. It is widely accepted that they exert their toxic effects by liberation of phosphine gas, produced by hydrolysis of zinc phosphide in a moist or acid environment.^3,7 Zinc phosphide is resistant to hydrolysis above a pH of 4.^3,5

Liberated phosphine gas is rapidly absorbed across gastric mucosa and distributed systemically where it exerts its toxic effect.³ How these effects occur remains unknown, but postulated mechanisms include inhibition of cytochrome C oxidase with mitochondrial dysfunction and interruption of cellular respiration, inhibition of serum acetyl cholinesterase activity, formation of reactive hydroxyl radicals, and inhibition of catalase and peroxidase resulting in lipid peroxidation.^3,7 Organs with high metabolic rates (eg, the brain, heart, and liver) are affected first, followed by the kidneys, should the animal survive the initial 24 to 30 hours.³ In addition, the production of phosphine gas within the stomach may lead to gastric or abdominal distension, pain, secondary bloat, and the potential for gastric dilatation-volvulus.^8–11

In addition to generating phosphine gas, the zinc phosphide compound itself plays a role in development of toxicosis. Zinc, as a metal, is associated with protracted vomiting, and the phosphide salt is considered corrosive and, as such, a direct irritant to the gastrointestinal tract.² Vomiting, anorexia, hematemesis, and melena can all develop after ingestion. In general, elemental zinc toxicosis is not expected with ingestion of zinc phosphide rodenticides.

Confirmation of zinc phosphide exposure can be made on the basis of gas chromatography-mass spectrometry or a Dräger detector tube test,^d which can be performed by a diagnostic laboratory. The Dräger detection tube test has been validated by use of canine stomach contents and vomitus.⁴ Gastric contents or postmortem samples can be analyzed for the presence of zinc (or aluminum) phosphide by use of gas chromatography-mass spectrometry. Additionally, tissue samples (of liver, stomach contents, and blood) can be evaluated for reaction with silver nitrate or silver diethyldithiocarbamate to identify exposure.^12,13 Despite the availability of laboratory diagnostic tests, no confirmatory testing was performed for any of the affected dogs in the present study.

In the present study, median age of survivors was significantly different from that of nonsurvivors. The 5 nonsurvivors were young dogs and may have been more severely affected owing to several factors, including dietary indiscretion, poor owner monitoring, or altered metabolic systems.

In this study, the median time from exposure to call initiation by the pet owner or health-care professional was found to have a significant effect on survival rate and indicates that early decontamination procedures are an important part of treatment in this toxicosis. Nonsurvivors had a longer period prior to call initiation, which may indicate that early and appropriate veterinary care is associated with improved outcome in dogs that develop clinical signs.

Clinical signs associated with zinc phosphide toxicosis are reported to occur within 15 minutes to 4 hours after ingestion but may be delayed by 12 to 18 hours.^2,5,14,15 Clinical signs identified in the present study were similar to signs previously described in human and veterinary literature.^{3,6,9,10,14,15} Early and protracted vomiting, followed by abdominal pain and distension, are signs commonly reported for many species.^2,6,9 Lethargy, coma, seizures, and sudden death are also common features of zinc phosphide toxicosis in both humans and animals. Human beings tend to have more severe systemic signs, in particular multisystem organ dysfunction syndrome; this may be associated with the ingestion of larger doses, which is often seen with toxicosis following suicide attempts.⁵ Clinicopathologic abnormalities reported for human beings include methemoglobinemia, Heinz body formation, hemolysis,¹ azotemia, increased liver enzymes (including alanine transaminase, aspartate aminotransferase, and total bilirubin), and electrolyte abnormalities (including hypokalemia and hypomagnesemia).^3,5,16 Other testing may reveal decreased cholinesterase activity, increased myocardial troponin, metabolic acidosis, and hypoxemia.^3,17,18

Clinical signs involving the gastrointestinal tract were the most commonly reported clinical sign in the present study (66.7%). The emetic action of zinc is well documented and the likely source of early and protracted vomiting associated with ingestion. Both the phosphide salt and phosphine gas have been implicated in the development of corrosive gastroenteritis leading to further emesis, abdominal pain and distension, and ulceration with hemorrhage.^2,6

Generalized malaise was a frequently reported (17.8%) clinical sign. Given the equivocal nature of this clinical sign, it is difficult to determine the exact relation to the mechanism of action for phosphine; however, because phosphine is rapidly absorbed and well distributed, this may lead to an overall ill or lethargic state. Human beings exposed secondarily to phosphine gas have reported signs consistent with malaise, including fatigue and headaches.⁹ Those with direct zinc phosphide–induced poisoning have an early onset of nausea, abdominal pain, and odd sensations such as chills and tremors.^5,9

Approximately 8.9% of the clinical signs were categorized as neurologic. The incidence of neurologic signs was much lower in this study, compared with that found in previous studies^2,6,7 reported in the veterinary literature. Stowe et al⁶ reported 5 of 8 affected dogs with confirmed phosphide ingestion developing CNS-related signs prior to death. Fatalities secondary to phosphide toxicosis have been reported to occur secondary to uncontrolled seizures.^2,7,11

Respiratory signs were reported to comprise only 3.3% of total clinical signs seen in dogs of the present study. Respiratory injury is likely due to capillary damage, which results in pulmonary edema formation and cellular hypoxia.^2,3 Postmortem findings in both humans and animals document the development of pulmonary edema secondary to phosphide toxicosis.^5,7,11 Respiratory signs should be considered a serious sequela to exposure and supportive care administered, including oxygen therapy, judicious fluid therapy, and appropriate monitoring.

Cardiovascular signs were rarely seen and comprised only 1.7% of the total clinical signs observed. Venous congestion and capillary damage are likely the underlying causes, and they have been documented on postmortem examinations in both humans and animals.^3,5,7 One study¹⁰ showed a correlation between the degree of abdominal organ congestion and severity of hypotension. Affected animals with evidence of cardiovascular compromise should be monitored closely, including blood pressure monitoring, and treated with fluid therapy and careful vasopressor support as needed.

Oddly, 58.8% of the dogs exposed in the present study did not develop any clinical signs. This could have been due to several factors, including exposure amount and variable exposure conditions. The retrospective nature of the present study did not allow a reliable determination of exposure amount in these dogs. Owners and veterinarians, when queried, were often unable to provide accurate estimates of ingested amounts. In addition, the amount retrieved by veterinarian-induced emesis or gastric lavage was often dissimilar to that reported by the owner. The acidity of the gastric environment at the time of ingestion likely also played an important role in the development of clinical signs. The toxic dose of zinc phosphide varies considerably depending on the presence of gastric contents and gastric pH. The LD₅₀ in dogs is reported to be 20 to 40 mg/kg (9 to 18 mg/lb).^2,4,15 However, dogs ingesting bait on an empty stomach have reportedly survived ingesting doses as high as 300 mg/kg (136 mg/lb).² Food or stomach contents increase gastric acid secretion and phosphine gas production and can result in the development of clinical signs. These 2 variables, along with other unidentifiable factors such as age and moisture content of the bait, may explain why there were a substantial number of dogs that did not develop clinical signs. Baits placed in moist environments or baits otherwise becoming wet rapidly undergo hydrolysis, decreasing their toxic potential. More importantly, veterinary professionals should not advise pet owners to feed their exposed pet any home remedies (eg, milk or bread) with zinc phosphide exposure because food or stomach contents increase gastric acid secretion and secondary phosphine gas production.

Dogs have been poisoned by feeding on the dead carcasses of rodents and other animals poisoned by zinc phosphide.^2,6 In this study, relay toxicosis was suspected in 11 dogs, with 3 dogs developing clinical signs associated with toxicosis. As zinc phosphide is a nonspecific rodenticide, it may cause harm or death in a number of targeted and nontargeted species.¹ In the present study, we were unable to confirm relay toxicosis, as no dead and previously poisoned animals were available for analysis. Because 3 dogs developed clinical signs potentially associated with toxicosis, the possibility of clinical signs developing secondary to consumption of a dead and decaying animal (ie, gastrointestinal upset) rather than primarily because of zinc phosphide rodenticides toxicosis cannot be excluded.

Death in animals has been reported to occur within 3 to 48 hours after zinc phosphide ingestion.^2,6,7 In the present study, the 5 nonsurvivors were euthanized or died within 48 hours after zinc phosphide ingestion. Two hospitalized dogs died from respiratory failure and uncontrolled seizures at 26 and 30 hours, respectively. Two additional dogs were euthanized secondary to uncontrolled seizures and cardiovascular shock at 36 and 48 hours, respectively. The remaining dog was found deceased by the owners within 14 hours after exposure.

In the present study, most dogs (71.8%) received some form of veterinary care, with just over half of these dogs hospitalized and one-third treated on an outpatient basis. Specific treatments varied among veterinarians but generally included some type of decontamination procedure. This remains a mainstay treatment for toxicoses in veterinary medicine, as there is no antidote for zinc phosphide toxicosis. Various types of decontamination procedures were performed, including emesis induction, activated charcoal administration, and gastric lavage. In the present study, emesis was induced by the veterinarian in most affected dogs. Although some patients also spontaneously vomit because of the direct gastric irritant effect of zinc phosphide, appropriate decontamination procedures should still include reducing the acidity of the gastric lumen, as this has a substantial effect on phosphine gas production. Procedures designed to neutralize the production of phosphine gas included the administration of a liquid antacid (ie, magnesium or aluminum hydroxide and calcium carbonate) or administration of a 5% solution of sodium bicarbonate prior to emesis or prior to gastric lavage.^2,5 As for decontamination with activated charcoal, there is limited evidence that it decreases toxic effects of zinc phosphide; however, until there is proof otherwise, it is still currently a recommended treatment.

The need for accurate product identification was highlighted in the present study. Many active ingredients are used in rodenticides currently on the market. The most common exposures in the United States are those to long-acting anticoagulant rodenticides, bromethalin, cholecalciferol, strychnine, and zinc phosphide.⁸ Phytonadione (vitamin K₁) was administered to 4 dogs in the present study. In 1 situation, this was warranted because of exposure of a combination of both zinc phosphide and bromadiolone (a long-acting anticoagulant rodenticide). However, in the remaining 3 dogs, vitamin K₁ treatment was used for presumptive long-acting anticoagulant rodenticide exposure without confirmation of the active ingredient or recognition of the mechanism of action of zinc phosphide.

The vomitus associated with zinc phosphide rodenticide toxicosis is often characterized as having a rotten fish or acetylene gas smell; this odor may also be present on the patient's breath and should raise suspicion that zinc phosphide toxicosis may have occurred. However, some types of zinc phosphide may also be odorless, so the lack of a distinct smell does not necessarily rule out toxicosis. The source of this smell is currently unknown but has been associated with the presence of phosphine gas, an inert ingredient, or a contaminant.^19,20 Owners should be warned about the risk of phosphine gas exposure during transit to the veterinary hospital because of risks of human toxicosis secondary to inhalational exposure. Adequate ventilation (eg, driving with the car windows down or inducing emesis outside or in a well-ventilated area) is imperative with decontamination of zinc phosphide ingestions. There are limited data as to the likelihood or seriousness of this type of human exposure. In the present study, there is at least 1 case where a pet owner developed shortness of breath that required emergency care after the dog vomited in the car on the drive to the veterinary hospital. Regardless of the type of decontamination procedure used, preventative measures should always be taken to limit human exposure.

The limitations of the present study include limitations commonly encountered in a retrospective study. One such limitation was the inability to confirm the exact amount of zinc phosphide ingested, as this information was often unavailable or unable to be obtained from the pet owner. Another limitation of the present study was that there were no advanced diagnostic tests performed in any of these cases to confirm or quantify zinc phosphide ingestion. Although case management and medical treatments were recommended to the veterinary professionals calling the animal poison control center, the final clinical decision for advanced diagnostics was left to the pet owner and the treating veterinarian. Additionally, follow-up information could not be obtained on all patients, further limiting data collection. Finally, as there were only 5 nonsurvivors in the study, it was difficult to make statistical comparisons between the survivors and nonsurvivors because of a lack of statistical power.

Although a previous report⁶ introduced the possibility of severe clinical signs and the potential for fatality associated with zinc phosphide exposures, the present study demonstrates that overall outcome is actually quite good with appropriate and early intervention. The positive outcome in the present study is likely due to a number of factors, including awareness of pet owners and veterinary staff of the dangers associated with rodenticide ingestions, availability of animal poison helplines to assist both pet owners and veterinary staff, and the use of early and appropriate decontamination procedures. Many dogs were hospitalized for continued monitoring and treatment, which likely contributed as well to an improved outcome. Zinc phosphide ingestion is a serious exposure, but accurate product identification and prompt veterinary intervention help ensure a positive outcome.