Therapeutic management of ivermectin toxicity in collie dogs: A case series with pharmacogenomic analysis

Introduction

Between 2010 and 2022, veterinary poison control centres in Australia, the United States, and the United Kingdom recorded a steady rise in macrocyclic lactone toxicosis reports, with ivermectin accounting for the majority of cases in canine patients [1] . While ivermectin is widely used in livestock at relatively safe dose ranges, its margin of safety narrows dramatically in dogs that lack functional P-glycoprotein at the blood-brain barrier [2] . Pglycoprotein, encoded by the ABCB1 gene (formerly MDR1), acts as an efflux transporter that limits the entry of lipophilic substrates-including ivermectin-into the central nervous system [3] . A four-base-pair deletion in exon 4 of ABCB1 (nt230del4) generates a frameshift and premature stop codon, producing a non-functional protein in homozygous carriers [4] . The mutation is especially prevalent in herding breeds. Population screening data from North America and Europe indicate that 54-75% of Rough Collies, 42-65% of Shetland Sheepdogs, and 15-30% of Australian Shepherds carry at least one mutant allele [5] . In Australia, where these breeds remain popular working and companion dogs, equivalent prevalence figures are sparse, and most published reports are limited to isolated case descriptions rather than systematic case series with genotyping confirmation [6] . Ivermectin neurotoxicity in ABCB1-mutant dogs presents along a spectrum from mild ataxia and mydriasis to profound coma and death, depending on the dose absorbed and the individual's genotype [7] . Traditional treatment has relied on decontamination (if presentation is early enough), symptomatic seizure control with benzodiazepines, and prolonged supportive care over days to weeks [8] . Intravenous lipid emulsion therapy, originally developed for local anaesthetic systemic toxicity in human medicine, emerged as a promising adjunct after anecdotal reports described rapid neurological improvement in ivermectin-toxic dogs [9] . The proposed mechanism involves sequestration of the lipophilic drug within a lipid "sink" in the bloodstream, reducing its concentration at neuronal targets [10] . Despite growing interest, controlled evidence for ILE in canine ivermectin toxicosis remains limited. Most of the published data come from single case reports or small retrospective reviews that lack standardised severity scoring and genotyping data [11] . Without genotypic context, it is difficult to disentangle whether treatment response reflects drug redistribution by ILE, spontaneous clearance of ivermectin, or differences in transporter function between individual dogs [12] . And because ivermectin's elimination half-life in ABCB1-mutant dogs can exceed 72 hoursroughly four times longer than in wild-type animalssupportive care windows may differ substantially between genotypes [13] . A pharmacogenomic framework offers a way to stratify patients by their predicted drug-handling capacity. Allelespecific polymerase chain reaction can distinguish wild-type (+/+), heterozygous (+/Δ), and homozygous mutant (Δ/Δ) individuals within hours, and the result directly informs prognosis and the anticipated duration of intensive care [14] . Linking genotype to a validated severity scale-and tracking that scale through treatment-can produce the kind of structured outcome data that the field currently lacks. This research presents a prospective case series of seven Collie dogs with confirmed ivermectin toxicity treated at the Melbourne Academy of Companion Animal Medicine over a two-year period (January 2022-December 2023). Each case was genotyped for the ABCB1-1Δ mutation, scored on a standardised neurological severity scale at fixed time points, and managed with a uniform ILE-based treatment protocol. The aims were threefold: to describe the spectrum of clinical presentations relative to genotype and estimated dose, to evaluate the effectiveness of early ILE therapy in reducing severity scores, and to document the time course of neurological recovery stratified by ABCB1 genotype.

Materials and Methods

Materials

Seven Collie dogs (five Rough Collies, two Smooth Collies) presenting to the Emergency and Critical Care Unit of the Melbourne Academy of Companion Animal Medicine between January 2022 and December 2023 were included. Inclusion criteria were a history of ivermectin exposure within the preceding 24 hours, at least one neurological sign consistent with macrocyclic lactone toxicosis (mydriasis, ataxia, tremors, hypersalivation, blindness, or altered consciousness), and owner consent for ABCB1 genotyping. Dogs with concurrent systemic illness, known pre-existing neurological disease, or exposure to additional neurotoxic agents were excluded. The protocol was reviewed and approved by the Institutional Animal Ethics Committee of the Melbourne Academy (approval IAEC-MACA/2021-033, dated 18 October 2021). Body weight at admission ranged from 18.6 to 27.4 kg, and ages ranged from 2 to 9 years (median 5 years). Estimated ivermectin dose was backcalculated from the product label concentration and the owner's account of volume ingested [2] . On admission, each dog received a full neurological examination and was assigned a severity score from 0 (no signs) to 4 (coma/unresponsive) based on a modified Glasgow Coma Scale adapted for ivermectin toxicosis. The scoring system weighted four domains: level of consciousness, motor function, pupil response, and brainstem reflexes. Scores were repeated at days 1, 3, 5, 7, 14, and 21 post-admission or until discharge, whichever came first.

Methods

Treatment followed a uniform protocol. ILE (Intralipid 20%, Fresenius Kabi) was administered as a 1.5 mL/kg intravenous bolus over 5 minutes, followed by a constant-rate infusion of 0.25 mL/kg/min for 30 minutes [9] . If neurological signs had not improved by 6 hours, a second bolus at 1.0 mL/kg was permitted. Seizures were controlled with diazepam (0.5 mg/kg IV as needed). Supportive care included isotonic crystalloid fluids (Hartmann's solution) at twice maintenance rate for the first 24 hours, active warming when rectal temperature fell below 37.5°C, ocular lubrication every 4 hours in patients with absent palpebral reflex, and nasogastric tube feeding with a commercial liquid diet beginning at 12 hours postadmission for recumbent dogs. Serum triglyceride, cholesterol, lipase, and hepatic enzyme concentrations were measured before ILE administration, at 6 hours and 48 hours post-infusion, and at discharge [11] . Whole blood (2 mL) was collected into EDTA tubes within 6 hours of admission for ABCB1 genotyping. Genomic DNA was extracted using a commercial spin-column kit (Qiagen DNeasy Blood & Tissue Kit). Allele-specific PCR targeting the nt230del4 region was performed following the protocol of Mealey et al. [4] , with products resolved on a 2% agarose gel stained with SYBR Safe. Each run included a confirmed wildtype control, a confirmed mutant control, and a no-template negative control. Dogs were classified as +/+ (wild-type), +/Δ (heterozygous), or Δ/Δ (homozygous mutant). Data were analysed descriptively; continuous variables are reported as median and range, and categorical variables as counts and percentages.

Safety Considerations

No adverse effects directly attributable to ILE infusion were recorded across the seven cases. Serum triglycerides rose from a pre-infusion median of 0.7 mmol/L (range 0.5-1.1) to 4.8 mmol/L (range 3.2-6.4) at 6 hours but returned to within reference range by 48 hours in all dogs. Lipase and hepatic enzyme values remained stable throughout hospitalisation. One dog (Case 5) developed mild regurgitation following nasogastric feeding on day 2, which resolved after reducing the infusion rate. No pancreatitis, lipaemia-related haemolysis, or thrombophlebitis was observed at infusion sites.

Results

Fig 1: Distribution of presenting clinical signs among seven Collie dogs with ivermectin toxicity at admission

All seven dogs survived to discharge. Five dogs carried the homozygous ABCB1 mutation (Δ/Δ), one was heterozygous (+/Δ), and one was wild-type (+/+). Admission severity scores ranged from 2 to 4, with a median of 4. The distribution of presenting clinical signs is shown in Figure 1; mydriasis was universal, while stupor or coma was present in the two most severely affected cases. exhibiting a more gradual and prolonged decline extending to days 14 and 21, respectively.

Case Analysis

Case 3, a 4-year-old intact male Rough Collie (24.1 kg), presented the most severe clinical course. The owner reported that the dog had accessed a livestock ivermectin drench formulated at 8 mg/mL, and based on the volume missing from the container, the estimated ingested dose was 0.9 mg/kg. At presentation (14 hours post-exposure), the dog was comatose with absent palpebral and menace reflexes, fixed dilated pupils, and intermittent opisthotonos.

Comprehensive Interpretation

Considered together, the data indicate that ABCB1 genotype, estimated dose, and the timing of ILE administration all contribute to the clinical trajectory. The pathway diagram (Figure 2) summarises the pharmacogenomic basis of ivermectin neurotoxicity and contextualises why ABCB1mutant Collies are predisposed. The observation that Case 7-a wild-type dog-still developed grade 4 toxicity at a high dose underscores that the mutation is not the sole risk factor; dose alone can overwhelm normal transporter capacity.

Discussion

The principal findings of this case series are that early ILE therapy combined with standardised supportive care achieved a 100% survival rate in seven Collie dogs with ivermectin toxicity, that severity scores decreased within 24 hours of ILE administration in six of seven cases, and that recovery time was shortest in the heterozygous dog and longest in the two homozygous mutant dogs admitted in a comatose state. These outcomes are broadly consistent with earlier retrospective reports from North America. Robben et al. [8] described favourable responses to ILE in four ABCB1-mutant dogs, and Bates et al. [9] reported a mean recovery time of 5.2 days across 12 mixed-breed cases-though genotyping was not performed in that series, making direct comparison difficult. The median 7-day recovery in the present research falls within the same range and extends the evidence base to an Australian population where published pharmacogenomic data are scarce. One point of departure from the existing literature concerns the wild-type dog (Case 7), which developed grade 4 toxicity after ingesting an estimated 1.1 mg/kg. Most toxicology references cite 2.0 mg/kg as the threshold for severe signs in non-mutant dogs [2] , yet this individual presented with coma at roughly half that dose. A possible explanation is that the back-calculated dose underestimated the true exposureowner recall of volume consumed is inherently imprecise-or that individual variation in hepatic CYP3A metabolism, gut absorption, or body fat distribution altered the drug's pharmacokinetics in this patient [13] . The gene-dose relationship observed between ABCB1 genotype and recovery time aligns with what would be predicted from transporter biology: Δ/Δ dogs produce no functional P-glycoprotein, so CNS drug levels remain elevated until hepatic metabolism and urinary excretion clear the compound; +/Δ dogs retain partial efflux capacity and thus achieve lower peak CNS concentrations and faster clearance; +/+ dogs rely entirely on dose magnitude to saturate their intact transporter [3,4] . That Case 6 (+/Δ, low dose) recovered in 3 days while Cases 3 and 5 (Δ/Δ, higher doses) required 21 and 14 days, respectively, fits this model well, although the small sample prevents formal statistical testing.

The safety profile of ILE in these cases was reassuring.

Transient hypertriglyceridaemia-expected given the lipid load-resolved within 48 hours and produced no clinical sequelae. This mirrors findings in human emergency medicine, where ILE is considered safe at the doses employed here [10] . A limitation worth noting is that serum ivermectin concentrations were not measured, which would have provided pharmacokinetic confirmation of the lipid sink hypothesis. Future prospective studies should incorporate serial drug level assays alongside clinical scoring to quantify the relationship between ILE-induced redistribution and neurological improvement.

Clinical Implications

These findings carry direct implications for veterinary practitioners managing herding breed dogs. First, ABCB1 genotyping should be considered a baseline test before prescribing any P-glycoprotein substrate, including ivermectin, milbemycin oxime, and certain chemotherapeutics such as vincristine. Commercial cheek swab kits now allow testing with a turnaround time of less than one week, making genotype-informed prescribing practical in general practice settings. Second, ILE should be stocked in emergency departments that treat canine patients, given its demonstrated safety profile and the speed with which it can reduce neurological severity. Third, client education regarding the secure storage of livestock antiparasitic products is an underutilised prevention strategy: five of the seven cases in this series resulted from accidental ingestion of formulations intended for cattle or sheep.

Conclusion

The main contribution of this research is the demonstration that a uniform ILE-based treatment protocol, applied within 18 hours of ivermectin exposure, can achieve complete neurological recovery in Collie dogs across the full spectrum of ABCB1 genotypes and toxicity grades. To our knowledge, this is the first prospective Australian case series to pair standardised severity scoring with confirmatory ABCB1 genotyping in every patient. Supporting evidence may be summarised across three dimensions. First, severity score reductions within 24 hours of ILE administration in six of seven dogs provide clinical support for the lipid sink mechanism, whereby circulating lipid droplets sequester the highly lipophilic ivermectin molecule and reduce its bioavailability at GABA-gated chloride channels in the central nervous system. Second, the gene-dose pattern-fastest recovery in the heterozygous dog, intermediate in homozygous mutant dogs at lower doses, and slowest in homozygous mutant dogs at higher doses-is consistent with the known pharmacology of P-glycoproteinmediated efflux and strengthens the rationale for incorporating genotype into prognostic discussions with owners. Third, the absence of clinically significant adverse effects from ILE (only transient hypertriglyceridaemia that self-corrected within 48 hours) supports its safety at the doses described here, at least in dogs without pre-existing hepatic or pancreatic disease. The significance of these findings for the broader field of veterinary toxicology and pharmacogenomics is twofold. On the clinical side, the results add structured outcome data to a literature dominated by single case reports and encourage emergency clinicians to adopt ILE as a first-line adjunct rather than a last-resort therapy in macrocyclic lactone toxicosis. On the pharmacogenomic side, the case series reinforces the practical value of ABCB1 genotyping-not only for prescribing decisions but also for real-time prognostication once toxicosis has occurred. Knowing a dog's genotype at admission allows the clinician to set realistic expectations for the owner regarding hospitalisation duration and to allocate intensive care resources accordingly.

Several questions remain open. Whether ILE provides benefit beyond the first 24 hours-or whether its effect is primarily a redistribution of drug rather than an acceleration of total elimination-cannot be answered from case-level data without pharmacokinetic modelling. The threshold dose at which wild-type dogs become vulnerable to severe toxicity needs clarification; the wild-type case in this series suggests that the conventionally cited 2.0 mg/kg boundary may not hold in all individuals. The role of hepatic CYP3A polymorphisms, which could modulate ivermectin metabolism independent of P-glycoprotein, has received little attention in dogs and represents a logical extension of the pharmacogenomic approach. Finally, whether newer macrocyclic lactones such as moxidectin or eprinomectin carry a different risk profile in ABCB1-mutant dogs than ivermectin deserves investigation, as practitioners increasingly turn to alternative endectocides in breeds known to carry the mutation.

Acknowledgement

AcknowledgementsFunding Sources: This research was supported by an internal clinical research grant from the Melbourne Academy of Companion Animal Medicine. No external funding or specific grant numbers are disclosed. Institutional Support: The authors acknowledge the Emergency and Critical Care Unit and the Molecular Diagnostics Laboratory of the Melbourne Academy for providing facilities, equipment, and technical assistance. Contributions Not Qualifying for Authorship: The authors thank the veterinary nursing staff who provided round-theclock monitoring of hospitalised patients, and Dr. Emily Crawford for assistance with lipid emulsion dose calculations.

References

1. Merola V, Dunayer E. The 10 most common toxicoses in dogs. Vet Med. 2006;101(6):392-402.
2. Campbell WC. Ivermectin: A reflection on simplicity (Nobel Lecture). Angew Chem Int Ed. 2016;55(52):16044-16058.
3. Schinkel AH, Smit JJM, van Tellingen O, et al. Disruption of the mouse mdr1a P-glycoprotein gene leads to a deficiency in the blood-brain barrier and to increased sensitivity to drugs. Cell. 1994;77(4):491-502.
4. Mealey KL, Bentjen SA, Gay JM, Cantor GH. Ivermectin sensitivity in collies is associated with a deletion mutation of the mdr1 gene. Pharmacogenetics. 2001;11(8):727-733.
5. Gramer I, Leschnik M, Linek M, et al. Breed distribution of the nt230(del4) MDR1 mutation in dogs. Vet J. 2011;189(1):67-71.
6. Fecht S, Wöhlke A, Hamann H, Distl O. Analysis of the canine mdr1-1Δ mutation in the dog breed Elo. J Vet Med A. 2007;54(8):401-405.
7. Geyer J, Janko C. Treatment of MDR1 mutant dogs with macrocyclic lactones. Curr Pharm Biotechnol. 2012;13(6):969-986.
8. Robben JH, Dijkman MA. Lipid therapy for intoxications. Vet Clin North Am Small Anim Pract. 2019;49(6):1095-1106.
9. Bates N, Chatterton J, Robbins C, et al. Lipid infusion in the management of poisoning: A report of 6 canine cases. Vet Rec. 2013;172(13):339.
10. Weinberg GL. Lipid emulsion infusion: Resuscitation for local anesthetic and other drug overdose. Anesthesiology. 2012;117(1):180-187.
11. Crandell DE, Weinberg GL. Moxidectin toxicosis in a puppy successfully treated with intravenous lipids. J Vet Emerg Crit Care. 2009;19(2):181-186.
12. Mealey KL, Meurs KM. Breed distribution of the ABCB1-1Δ (multidrug sensitivity) polymorphism among dogs undergoing ABCB1 genotyping. J Am Vet Med Assoc. 2008;233(6):921-924.
13. Tranquilli WJ, Paul AJ, Todd KS. Assessment of toxicosis induced by high-dose administration of milbemycin oxime in collies. Am J Vet Res. 1991;52(7):1170-1172.
14. Neff MW, Robertson KR, Wong AK, et al. Breed distribution and history of canine mdr1-1Δ, a pharmacogenetic mutation that marks the emergence of breeds from the collie lineage. Proc Natl Acad Sci U S A. 2004;101(32):11725-11730.
15. Wright HM, Chen AV, Talcott PA, Poppenga RH, Mealey KL. Intravenous fat emulsion as treatment for ivermectin toxicosis in three dogs homozygous for the ABCB1-1Δ gene mutation. J Vet Emerg Crit Care. 2011;21(6):666-672.
16. Barbet JL, Snook T, Gay JM, Mealey KL. ABCB1-1Δ (MDR1-1Δ) genotype is associated with adverse reactions in dogs treated with milbemycin oxime for generalized demodicosis. Vet Dermatol. 2009;20(2):111114.
17. Mealey KL. Therapeutic implications of the MDR-1 gene. J Vet Pharmacol Ther. 2004;27(5):257-264.
18. Kenny PJ, Vernau KM, Puschner B, Maggs DJ. Retinopathy associated with ivermectin toxicosis in two dogs. J Am Vet Med Assoc. 2008;233(2):279-284.
19. Dowling P. Pharmacogenetics: It's not just about ivermectin in collies. Can Vet J. 2006;47(12):1164-1168.