Anticoagulant rodenticide toxicity in dogs: A retrospective study of 349 confirmed cases in Saskatchewan

Abstract

Objective

To evaluate the signalment and clinical, laboratory, treatment, and outcome features of dogs diagnosed with anticoagulant rodenticide (AR) intoxication in Saskatchewan.

Animals

We studied 349 dogs.

Procedure

Medical records from the Veterinary Medical Centre (Saskatoon, Saskatchewan) between 1999 and 2022 were reviewed. Cases were included if they met at least 1 of the following criteria: owner witnessed the dog ingesting an AR; AR was seen in the vomitus when emesis was induced; the dog had clinical signs of coagulopathy, with elevation of PT ± aPTT that normalized after vitamin K1 therapy, in the presence of appropriate clinical and paraclinical data and the absence of other causes of hypocoagulable state determined by the primary clinician.

Results

Fifty-three percent of cases were seen between July and October. Most dogs (61%) came from an urban setting. Ninety-two percent of dogs ingested a 2nd-generation AR and the most frequent toxin was bromadiolone. Clinical signs were reported in 30% of AR intoxications and included lethargy (86%), dyspnea (55%), and evidence of external hemorrhage (44%). The most common site of hemorrhage was the pleural space, accounting for 43% of hemorrhage sites. Consumptive thrombocytopenia was reported in 24% of dogs with evidence of AR-induced hemorrhage, with moderate (platelet count < 60 K/μL) and marked (< 30 K/μL) thrombocytopenia in 7/12 and 2/12 dogs, respectively. Blood products were administered to 84% of dogs with AR-induced hemorrhage; the most common product administered was fresh frozen plasma (56% of cases). Among dogs with AR-induced hemorrhage, those that received blood products were more likely to survive to discharge (81%) compared to those that did not (19%) (P = 0.017). Eighty-six percent of dogs with AR-induced hemorrhage survived to discharge.

Conclusion and clinical relevance

The pleural space was the most common site of hemorrhage. Moderate thrombocytopenia was a common finding. Eighty-six percent of dogs with AR-induced hemorrhage survived to discharge.

Introduction

Anticoagulant rodenticides (ARs) are pest-control products that represent a common cause of intoxication in dogs (1–4). In recent years, legislation has been passed in North America to control and limit the use of 2nd-generation ARs (SGARs) yet, despite these actions, ARs remain one of the most commonly reported toxicoses in companion animals. Intoxication in dogs typically results from accidental ingestion of rodent baits and, in rare cases, from malicious poisoning (1,3,4). Secondary or relay intoxication of dogs due to ingestion of intoxicated rodents or non-target animals (such as birds and invertebrates) is also possible but uncommon (1,5).

Anticoagulant rodenticides induce coagulopathy and subsequent hemorrhage by impairing hepatic synthesis of certain coagulation factors. Anticoagulants are generally well-absorbed through the intestinal epithelium within a few hours (6). In addition, absorption through the respiratory tract and skin has also been reported; however, is less frequently relevant in dogs (6). In blood, warfarin is almost entirely proteinbound, primarily to albumin. After intestinal absorption, ARs travel to the liver via the portal vein system or chylomicrons, and are metabolized by the liver largely through the action of cytochrome P450 (6). Both 1st-generation ARs (FGARs) and SGARs accumulate rapidly in the liver until the microsomal binding sites are saturated (6). Anticoagulant rodenticides inhibit the hepatic enzyme vitamin K1 epoxide reductase (1). This enzyme is required for the recycling of vitamin K1 into its active form, which is a necessary cofactor for the γ-carboxylation of glutamic acid residues on clotting factors II, VII, IX, and X (7). Anticoagulant rodenticides therefore cause depletion of the active form of vitamin K1, leading to a progressive decrease in plasma concentrations of functional forms of vitamin K1-dependant clotting factors, and eventually leading to a hypocoagulable state and hemorrhage (8). Anticoagulant rodenticides then exit the liver via the hepatic vein (where they can be measured in circulation) and are eliminated through urine or bile. In the case of biliary elimination, some ARs are subject to enterohepatic recirculation and can remain in the liver tissues for weeks, even after successful treatment (1,5).

Anticoagulant rodenticides can be divided into 2 subcategories based on their potency and activity: FGARs and SGARs. Warfarin was the first AR developed and was widely used until rodents began to develop resistance. Warfarin is derived from dicoumarol, a naturally occurring chemical found in some plants, including sweet clover (Melilotus spp.) (9). Resistance to warfarin prompted the development of SGARs, sometimes referred to as superwarfarins (1,3,5). These superwarfarins (i.e., brodifacoum, bromadiolone, difethialone) are more potent and longer-acting than FGARs. This is attributed to their greater affinity for vitamin K1 epoxide reductase, ability to disrupt the vitamin K1 epoxide cycle at more than one location, hepatic accumulation, and unusually long biological half-lives due to high lipid solubility and enterohepatic recirculation. For example, the plasma elimination half-lives of the SGARs bromadiolone and brodifacoum are 6 d, compared to warfarin’s half-life of only 14 h (3,10). The toxic dose and median lethal dose depend on the AR product, but SGARs have median lethal doses significantly lower than those of FGARs and generally require only a single feeding to result in the death of the target species (5).

Clinical signs of AR toxicosis usually develop 2 to 5 d after exposure, depending on the dose and type (FGAR versus SGAR) of AR product ingested (1,3,11,12). The impaired coagulation can be assessed by measurement of blood coagulation parameters including prothrombin time (PT) and activated partial thromboplastin time (aPTT). Among the 4 vitamin K1-dependant clotting factors (Factors II, V, VII, IX), Factor VII has the shortest half-life, and so an increase in PT is expected before an increase in aPTT (1,3). Therefore, PT testing is considered more sensitive than aPTT testing for early detection of AR intoxication (1,3). Advanced and confirmation testing are sometimes needed, especially in medicolegal cases, since results are generally not received in time to affect treatment decisions. Antemortem confirmatory tests include measurement of AR in stomach contents or serum/blood, whereas postmortem samples also include liver and kidneys for liquid chromatography (1).

Treatment of dogs with AR intoxication involves the administration of vitamin K1. The appropriate length of treatment required depends on the dose ingested (most often unknown) and, more importantly, the generation of the AR ingested. Treatment with vitamin K1 is normally given for 2 to 4 wk, depending on the specific AR (1,3). This highlights the importance of recognizing, if possible, which form of rodenticide was ingested.

To the authors’ knowledge, only a few studies on dogs diagnosed with AR intoxication are available in the literature, and these reported comparatively small numbers of cases (8,13–15). The purpose of this study was to evaluate the signalment, epidemiology, clinical and clinicopathologic features, medical management, and outcomes of a large number of cases of AR intoxication (N = 349) presented to the Veterinary Medicine Centre (VMC), Western College of Veterinary Medicine (Saskatoon, Saskatchewan), over a 23-year period. Our aim is that the information presented in this report will aid clinicians in the diagnostic evaluation and treatment of suspected AR intoxications.

Materials and methods

Database

Medical records from the VMC from January 1999 to December 2022 were reviewed for confirmed cases of dogs with AR intoxication. Cases were included in the study if they met at least 1 of the following criteria: i) the owner witnessed the dog ingesting an AR; ii) a formed product matching common characteristics of AR products was present in the vomitus when emesis was induced; or iii) the dog had clinical signs of coagulopathy in which an elevated PT ± aPTT was documented, along with evidence of normalization of PT/aPTT measurements following vitamin K1 therapy, in the presence of appropriate clinical and paraclinical data and the absence of other causes of hypocoagulable state identified by the primary clinician.

Signalment and epidemiologic data were extracted from the medical record when the information was available. These included breed, age, sex, weight, urban versus rural setting, year and month when AR intoxication occurred, rodenticide active ingredient, and generation of AR. Sites of hemorrhage were recorded based on the clinical signs; physical examination findings at presentation (the hair coat was not routinely clipped); diagnostic imaging studies performed [thoracic and abdominal radiographs, abdominal- (AFAST) and thoracic- (TFAST) focused assessment with sonography for trauma (FASTVet, Spicewood, Texas, USA)]; any interventional procedures performed (e.g., thoracocentesis, abdominocentesis); and necropsy findings, if intoxication was fatal or the owner requested euthanasia. The hospitalization time, if any, was also recorded.

The PT and aPTT testing were completed with 1 of 2 machines, depending on whether samples were processed within normal business hours (MLA Electra 750 photo-optical plasma coagulation timing instrument; Beckman Coulter, Mississauga, Ontario: RR 10 to 13 sec for PT and 9.6 to 45 sec for aPTT), or outside of normal business hours (patient-side Idexx Coag Dx Analyzer; Idexx, Markham, Ontario: RR 11 to 17 sec for PT and 72 to 102 sec for aPTT). For the complete blood (cell) count (CBC) data, the Cell Dyn 3500 (Abbott Laboratories, Mississauga, Ontario) automated hematology analyzer was used until the end of 2013, and the ADVIA 2120i (Siemens Healthineers, Oakville, Ontario) was used thereafter. As part of the CBC, blood samples from dogs classified as anemic were smeared and stained with new methylene blue dye, to calculate the percentage of reticulocytes present and to further characterize the anemia as regenerative or non-regenerative. In all cases, the automated platelet count was also confirmed as adequate by a clinical pathologist or clinical pathology technician based on manual evaluation of a blood smear.

Blood serum biochemistries were measured using a Hitachi 911 chemistry analyzer (Roche Diagnostics, Laval, Quebec) until mid-2011, and thereafter using a Cobas C311 (Roche Diagnostics). If full serum biochemical analysis was not undertaken, the total protein (TP) concentration was measured from plasma using a refractometer, as part of an emergency panel on admission, along with the dog’s packed cell volume (PCV), blood glucose, and Azostix (Siemens Medical Solutions, Oakville, Ontario) blood urea nitrogen.

The treatment regimen for each animal was also reviewed, and data were extracted from the record, including induction of emesis, administration of activated charcoal or intravenous fluids, symptomatic treatments, blood product transfusions, and vitamin K1 supplementation. The dose, route, and duration of administration of vitamin K1 prescribed in hospital and prescribed at discharge were recorded. In addition, the times to recheck at the VMC, if any, for retesting of PT, aPTT, and other blood parameters were noted. The dog’s outcome (alive, euthanized, deceased) was recorded, and necropsy reports, when available, were reviewed for any additional information to confirm AR intoxication.

Statistics

Qualitative descriptive data were presented as percentage, mean ± standard deviation, and range. For nonparametric contingency tables, the Fisher exact test was used. A P-value of < 0.05 was considered statistically significant. Statistical analyses were completed, and graphs were produced, using a commercial software package (GraphPad Prism 9 version 9.5.1; GraphPad Software, La Jolla, California, USA).

Results

Study population epidemiology and rodenticide products

A total of 349 confirmed cases of dogs with AR intoxication were identified between January 1999 and December 2022, with a mean of 15 ± 9 cases/y (range: 4 to 35/y) (Figure 1 A). Most cases were seen between July and October, with 185/349 cases (53%) seen during this period (Figure 1 B). No sex predisposition was noticed: 161/349 dogs (46%) were males (69 intact and 92 castrated) and 188/349 (54%) were females (70 intact and 118 spayed). Age of dog at presentation was available for 342 dogs and median age was 42 mo (3.5 y) ± 42 mo (range: 1 mo to 17 y). Weight was available for 331 dogs and median weight was 16.6 ± 13 kg (range: 1.1 to 66.2 kg). There were 103 dogs (30%) in the small-breed category (< 10 kg), 126 dogs (36%) in the medium-breed dog category (10 to 25 kg), 86 dogs (25%) in the large-breed category (25.1 to 45 kg), and 11 dogs (3%) in the giant-breed category (> 45.1 kg). The most common breeds affected were Labrador retriever (66/349, 19%), German shepherd (29/349, 8%), and border collie (22/349, 6%), followed by golden retriever (5%), shih tzu (5%), beagle (4%), cocker spaniel (3%), Pomeranian (3%), and 60 other breeds with frequency < 3% for each.

Figure 1.

A — Distribution of yearly confirmed cases of anticoagulant rodenticide (ACR) intoxication in dogs between 1999 and 2022. B — Number of confirmed cases of ACR intoxication by month between 1999 and 2022.

Dog and owner lived in a rural setting (acreage or a farm) in 136/349 cases (39%) and in an urban setting in 213/349 cases (61%). In 240/349 cases (69%), the owner claimed to have witnessed their dog ingesting the rodenticide. The type of AR was recorded and provided by the owner in 157/349 cases (45%): 13 dogs (8%) ingested warfarin, an FGAR; the remaining 144 dogs ingested an SGAR, including bromadiolone (n = 51), brodifacoum (n = 40), difethialone (n = 20), diphacinone (n = 16), chlorophacinone (n = 15), and indandione (n = 2) (Figure 2).

Figure 2.

Distribution of active ingredient in 157 of 349 dogs with confirmed anticoagulant rodenticide intoxication.

Clinical signs and sites of hemorrhage

Clinical signs were reported in 104/349 (30%) of AR intoxications and included lethargy (n = 89), dyspnea (n = 57), evidence of external hemorrhage (ocular, oral, nasal, cutaneous or subcutaneous, vulvar or penile hemorrhages; melena/hematochezia; or hematemesis) (n = 46), anorexia (n = 26), cough (n = 23), vomiting (n = 19), diarrhea (n = 6), and neurological signs (seizures, paraparesis) (n = 3).

As part of the diagnostic evaluation, radiographs were obtained for 58/104 dogs with clinical signs of AR intoxication (thoracic radiographs in 54 dogs, abdominal radiographs in 19). No radiographs were obtained for animals that were asymptomatic. Point-of-care TFAST and AFAST became available in 2015 and were done in 33/104 dogs with clinical signs of AR intoxication and in only 7/245 asymptomatic dogs.

Sites of hemorrhage were reported in 81 cases (Figure 3). The most common site of hemorrhage was the pleural space (n = 35); thoracocentesis was performed in 12/35 cases and confirmed hemothorax in all cases. The other locations of hemorrhage were cutaneous or subcutaneous (ecchymosis) (n = 21); respiratory tract, including in the lungs (n = 19), based on radiographic changes or necropsy findings, epistaxis (n = 13), tracheal submucosal hemorrhage (n = 1) (identified on necropsy); oral cavity (n = 19), with evidence of bleeding from the tongue or gingiva; gastrointestinal tract, with melena (n = 15); peritoneal cavity (n = 11); mediastinum (n = 8); conjunctiva and sclera (n = 7); and urogenital tract, with hematuria (n = 6) or bleeding from the vulva or penis (n = 4). Suspected CNS involvement was also reported in 6 dogs but was not confirmed by advanced diagnostic imaging. Two dogs had acute onset of seizures, and another dog had extensor rigidity of the front limbs and paralysis of the back limbs, which was suspicious for multifocal hemorrhage into the cerebellum and spinal cord. Anecdotal sites of hemorrhage included eyelid hematoma (n = 1). In this case, the dog was bleeding from a surgical wound on the upper eyelid following removal of a benign mass 2 d before admission for AR intoxication. One dog also had hemarthrosis in multiple joints, confirmed by arthrocentesis; and another had thymic hemorrhages identified on necropsy. Of note, no dog had petechial hemorrhage despite severe thrombocytopenia in some cases, though hair coats were not clipped and petechia could have been missed. However, all dogs with cutaneous or subcutaneous hemorrhages had ecchymosis. Pericardial effusion was also not reported in any dog, but postmortem examination revealed multifocal hemorrhages in the endocardium and on the outer layer of the parietal pericardium in 1 dog.

Figure 3.

Site of hemorrhage for the 81 cases of dogs with anticoagulant rodenticide-induced hemorrhage.

Coagulation times

Of the 349 dogs with confirmed AR intoxication, PT was measured in 111 dogs and aPTT was measured in 107. Of 111 dogs with available PT measurements, 30 had increased PT on the MLA analyzer (including 19 above the limit of detection: > 60 s) and 50 had increased PT on the Idexx analyzer (including 29 above the limit of detection: > 100 s). Of 107 dogs with available aPTT measurements, 22 had increased aPTT on the MLA analyzer (including 19 above the limit of detection: > 60 s) and 44 had increased aPTT on the Idexx analyzer (including 15 above the limit of detection: > 350 s) (Table 1).

Table 1.

Coagulation times (prothrombin time and activated partial thromboplastin time) and results of hematological analysis in dogs with anticoagulant rodenticide intoxication.

Parameters	Results	Reference intervals (RI)
On presentation
PT (n = 111)
MLA (n = 51)	• 19 dogs > UDL	10 to 13 s
	• 30 dogs had increased PT
POC (n = 62)	• 29 dogs > UDL	11 to 17 s
	• 50 dogs had increased PT
aPTT (n = 107)
MLA (n = 48)	• 19 dogs > UDL	9.6 to 45 s
	• 22 dogs had increased aPTT
POC (n = 61)	• 15 dogs > UDL	72 to 102 s
	• 44 dogs had increased aPTT
PCV (n = 73)	• 35 ± 14%	36 to 56%
Hematocrit (n = 65)	• 0.30 ± 0.11 L/L	0.36 to 0.56 L/L
RBC count (n = 65)	• 4.61 ± 1.8 × 1012/L	5.2 to 8.2 × 1012/L
Hemoglobin (n = 65)	• 108 ± 40 g/L	133 to 197 g/L
Total proteins (n = 73)	• 6.1 ± 1.3 g/dL	5.6 to 7.4 g/dL
Platelet count (n = 65)	• 147 K ± 101 K/μL	117 K to 418 K/μL
Platelet count in dogs with	• 120 K ± 83 K/μL
AR-induced hemorrhages (n = 49)	• 66 K ± 39 K/μL in the 12 dogs with thrombocytopenia	117 K to 418 K/μL
Follow-up after treatment
PT (n = 67), MLA	• Mean ± SD: 8.0 ± 1.3 s	10 to 13 s
	• All values were within RI
aPTT (n = 46), MLA	• Mean ± SD: 13.7 ± 11.6 s	9.6 to 45 s
	• All values were within RI

aPTT — Activated partial thromboplastin time; AR — Anticoagulant rodenticide; MLA — MLA Electra 750 photo-optical plasma coagulation timing instrument (Beckman Coulter, Mississauga, Ontario); PCV — Packed cell volume; POC — Patient-side Idexx Coag Dx Analyzer (Idexx, Markham, Ontario); PT — Prothrombin time; RBC — Red blood cell; SD — Standard deviation; UDL — Upper detection limit.

Of the 81 dogs with clinical signs of hemorrhage, PT was measured in 62 dogs and aPTT was measured in 60. Of 62 dogs with available PT measurements, 27 had increased PT on the MLA analyzer (including 18 above the limit of detection) and 28 had increased PT on the Idexx analyzer (including 24 above the limit of detection). The remaining dogs with clinical signs of hemorrhage but normal PT had received vitamin K1 before testing (usually by the referring veterinarian). Of 60 dogs with available aPTT measurements, 21 had increased aPTT on the MLA analyzer (including 17 above the limit of detection) and 26 had increased aPTT on the Idexx analyzer (including 11 above the limit of detection).

Of 303 dogs that received vitamin K supplementation, follow-up PT and aPTT results were available for 67 and 46 dogs, respectively, and were measured 48 to 72 h after discontinuation of oral vitamin K supplementation. On average, PT and aPTT were rechecked 21.9 ± 13 d after initial presentation. All recheck PT/aPTT measurements were with the MLA analyzer and none of the values were increased. The mean values at follow-up were 8.0 ± 1.3 s for PT (reference range: 10 to 13 s) and 13.7 ± 11.6 s for aPTT (reference range: 9.6 to 45 s).

Hematological analysis

Of the 73 emergency panels obtained, the mean PCV was 35 ± 14% and the mean TP was 6.1 ± 1.3 g/dL. A CBC was done in 65/349 dogs with AR intoxication and in 58/104 dogs with clinical signs of AR intoxication on initial examination. Based on the 65 CBCs, mean hematocrit was 0.30 ± 0.11 L/L (reference interval: 0.36 to 0.56 L/L), mean RBC count was 4.61 ± 1.8 × 10¹²/L (reference interval: 5.2 to 8.2 × 10¹²/L), and mean hemoglobin was 108 ± 40 g/L (reference interval: 133 to 197 g/L) (Table 1). Anemia, defined as PCV < 36% or hematocrit < 0.36 g/L, was present in 62 dogs out of 73 that had either a CBC or a PCV/TP measured. Based on reticulocyte count, the anemia was considered regenerative in 40/62 anemic dogs and non-regenerative in 22/62 anemic dogs, suggesting acute blood loss and a pre-regenerative state for the latter category. Of the 62 dogs with anemia, mean TP was 4.6 ± 1.5 g/L (reference interval: 5.5 to 7.1 g/L), consistent with blood loss.

Of the 81 dogs with evidence of hemorrhage, 49 dogs had a CBC analysis, with thrombocytopenia reported and confirmed on blood smear review in 12 cases (24%) and normal or increased platelet levels in the remaining 37 cases (76%). The mean platelet count was 120 236 ± 83 251/μL (range: 12 to 467 K/μL) in these 81 dogs, and 65 750 ± 39 269/μL (range: 12 to 126 K/μL) in the 12 dogs with thrombocytopenia, with 7/12 dogs having a platelet count < 60 K/μL and 2/12 dogs having a platelet count < 30 K/μL (1 dog had a platelet count of 12 K/μL and the other, a count of 20 K/μL) (Table 1). Of 46 dogs with anemia and with a CBC reported, 12 dogs had thrombocytopenia and 34 had normal-to-increased platelet counts. However, all dogs with thrombocytopenia were anemic.

Treatment

Among the 349 cases with confirmed AR intoxication, emesis was attempted with intravenous apomorphine in 189 dogs and with oral hydrogen peroxide in 61 dogs. Activated charcoal was given to 163 dogs as part of decontamination treatment on admission. Intravenous fluid therapy was administered to 73 dogs, and 34 dogs received oxygen therapy. Blood products were administered to 68/81 dogs (84%) with evidence of hemorrhage: 38 received fresh frozen plasma transfusions, 25 received fresh whole blood transfusions, 11 received packed RBC transfusions, and 8 received stored whole blood transfusions (these 8 dogs all had anemia with a mean PCV of 29 ± 9%). No blood products were administered to any dog not having clinical signs. Among dogs with evidence of hemorrhage that survived to discharge, the percentage of dogs that received blood products (81%) was significantly higher than that of dogs that did not receive blood products (19%) (P = 0.017).

In 223/349 confirmed AR intoxication cases (64%), vitamin K1 supplementation was started in hospital, at a mean dose of 3.6 ± 1.3 mg/kg (range: 1.2 to 5.5 mg/kg), administered SC in 187 dogs, PO in 34 dogs, and IM in 2 dogs. Vitamin K1 supplementation was started in hospital for 70/81 dogs (70%) with AR-induced hemorrhage. At the time of discharge, oral vitamin K1 supplementation was prescribed for 310/349 dogs (89%), and 68/81 dogs (84%) with AR-induced hemorrhages (the remaining 13 dogs were euthanized or died before discharge). The mean dose of oral vitamin K supplementation was 2.6 ± 0.9 mg/kg. The frequency of oral administration of vitamin K1 was available for all cases and was q24h (n = 70), q12h (n = 235), and q8h (n = 5). The mean vitamin K1 treatment duration at home for all AR intoxications was 25.8 ± 7.5 d (means of 20 d for FGARs and 27 d for SGARs). Of 81 dogs with evidence of hemorrhage, combination treatment with intravenous fluid therapy, blood product transfusion, and vitamin K1 supplementation was reported in 58 dogs (72%).

Outcome

Of 104 dogs with clinical signs of AR toxicity, 78 dogs were hospitalized, with a mean length of hospitalization of 3.0 ± 1.5 d (range: 1 to 8 d). Of the 245 asymptomatic dogs, only 12 were hospitalized, with a mean length of hospitalization of 1.8 ± 0.6 d.

Of the 104 cases with clinical signs of AR intoxication on presentation, 91 dogs survived to discharge and 13 did not survive (7 dogs were euthanized). Of the 81 cases of AR intoxication with hemorrhage, 70 dogs (86%) survived to discharge and 13 dogs (14%) did not survive (7 dogs were euthanized). All dogs without clinical signs on admission survived to discharge.

Discussion

To the authors’ knowledge, this is the first study evaluating the signalment and clinical, laboratory, treatment, and outcome features of a large population of dogs with AR intoxication in western Canada. Cases of AR intoxication in dogs are frequently seen at our institution, possibly because of the intensive nature of agriculture in our geographic region and the large rodent populations associated with our climate. Most cases (53%) were seen between mid-summer and early fall (July to October), when dogs have increased access to barns, farms, lake cottages, and even parks or wildlife areas.

The active AR ingredient was unknown in 55% of the cases in this study. When the active ingredient was known, SGARs accounted for 92% of cases, with bromadiolone, brodifacoum, and difethialone being the most common toxins. This first reflects that consumers have moved away from buying FGARs such as warfarin and preferably purchase SGARs that are more toxic and more efficient at killing rodents (1,3). It may also reflect a change in which products are available for sale through hardware and pest extermination supply stores. A similar overrepresentation of SGARs has been documented in terrestrial birds of prey from western Canada with AR intoxication (16). In that study, brodifacoum, bromadiolone, and difethialone were detected in 58%, 54%, and 34% of raptor livers, respectively (16).

The clinical signs of AR intoxication vary depending on the site of hemorrhage. In this study, hemorrhage occurred at multiple sites, including the pleural space, peritoneal cavity, oral cavity, cutaneous or subcutaneous tissues, lungs, gastrointestinal tract, upper respiratory tract, mediastinum, conjunctiva and sclera, urogenital tract, CNS, eyelids, and joints. The most common site of documented hemorrhage in our study was the pleural space, accounting for 43% of hemorrhage sites (1,3,17). In a recent study, hemothorax was also the most prevalent single site of hemorrhage, and about 1/2 of the dogs had evidence of intracavitary hemorrhage (2). However, AR intoxication may cause hemorrhage at any site within the body, as seen in our study and supported by other retrospective studies, case reports, and case series, which reported hematometra (18), gastric wall hemorrhage (19), ureteral hemorrhage (20), extradural hematoma (21), tracheal mucosal hemorrhage (22), and others. Interestingly, conjunctival and scleral hemorrhages were not uncommon in the current study and were seen in 7/81 dogs with AR-induced hemorrhage. A previous study similarly reported 6 dogs with suspected or confirmed AR toxicity that presented with predominantly ocular manifestations, including subconjunctival hemorrhage, exophthalmos, and orbital pain (23). Interestingly, no pericardial effusion was reported in the present study, though the pericardium is a reported site of hemorrhage in cases of AR toxicity (2).

Thrombocytopenia was documented after blood smear review in 12 cases (24%) of the 49 cases of AR intoxication with hemorrhage that had a CBC, and 2 dogs had marked thrombocytopenia (< 30 K/μL). In another study, thrombocytopenia was documented in 8/11 dogs with AR-induced hemorrhage and was also marked in 2 cases (< 30 K/μL) (13). In both studies, petechial hemorrhage was not noted in any case. This highlights that AR intoxication should be part of the differential diagnosis for dogs with hemorrhage accompanied by mild-to-severe thrombocytopenia, and especially should be distinguished from disseminated intravascular coagulation.

Treatment of dogs with AR intoxication involves administration of vitamin K1 and transfusion therapy to provide active clotting factors, and in some cases, RBCs, if moderate-to-severe anemia is evident (1,3). Blood products were administered in 84% of dogs with evidence of hemorrhage; a recent study involving 62 dogs with hemorrhage secondary to AR intoxication similarly reported 77% of dogs received a transfusion (2). Among dogs with evidence of hemorrhage that survived to discharge, the percentage of dogs that received blood products (81%) was significantly higher than that of dogs that did not receive blood products (19%). Similar results were reported for another recent study, where dogs that received blood transfusions had a significantly higher rate of survival to discharge (64%) versus dogs that did not receive blood products (36%) (2). Of the 81 cases of AR intoxication with hemorrhage, 70 dogs (86%) survived to discharge and 13 dogs (14%) did not survive; these results are similar to those in previous reports with survival-to-discharge rates of 87% (2) and 83% (8). However, of the 13 dogs that did not survive, 7 dogs were euthanized shortly after admission, and financial constraints were reported in 6/7 cases. This may have decreased the survival rate of dogs with AR-induced hemorrhage seen in our institution.

Limitations of this study include the retrospective nature of the study design, and thus the lack of standardization in diagnosis and treatment. Due to the broad time frame of the study (1999 to 2022), single blood analyzers were not used for all samples; however, reference ranges for the analyzer used at the time of presentation have been used to avoid any overinterpretation of the data. One of the inclusion criteria was that the owner witnessed the dog ingesting a formed product matching common characteristics of AR products in the vomitus when emesis was induced; however, no laboratory verifications were done to confirm these products were ARs. Another inclusion criterion was that the dog had clinical signs of coagulopathy in which an elevated PT ± aPTT was documented, along with evidence of normalization of PT/aPTT measurements following vitamin K1 therapy. The authors cannot exclude the possibility that the increase followed by normalization on PT/PTT was secondary to liver failure; however, this is unlikely given the absence of other clinical and paraclinical data (i.e., hyperbilirubinemia) to support hepatic insufficiency in these cases.

This study highlights the wide spectrum of clinical signs and sites of hemorrhage in dogs with AR intoxication, despite the classical association of AR intoxication with cavitary hemorrhages. The most common site of hemorrhage was the pleural space, accounting for 43% of the sites of hemorrhage. Thrombocytopenia was not uncommon and was reported in 24% of dogs. Blood products were administered to 84% of dogs with AR-induced hemorrhages, most commonly fresh frozen plasma (56% of cases). The prognosis for AR intoxication is good, with 86% of dogs with AR-induced hemorrhages surviving to discharge. CVJ