An 11-month-old miniature Shetland Pony colt weighing 26 kg was presented to the University of Zurich Equine Hospital for evaluation of acute onset of generalized seizures, followed by unconsciousness and recumbency, 34 hours after a massive (>25-fold) overdose of deworming paste1 (140 mg ivermectin and 1.05 g praziquantel). Instead of the labeled quantities of 0.2 mg ivermectin/kg and 1.5 mg praziquantel/kg, the foal had received the entire tube, equivalent to 5.4 mg ivermectin/kg and 40.4 mg praziquantel/kg.
At presentation, the colt was in a stuporous condition but responded to pain stimuli including jugular venous catheter placement. Heart rate was 26/min, respiratory rate was 20/min, and rectal temperature was 34.5°C (94.1°F). The peripheral pulses were very weak and not detectable by palpation at all superficial arteries. Both jugular veins were poorly filled and capillary refill time was prolonged at approximately 3 seconds. The distal limbs and ears were cold to the touch. Palpebral and pupillary reflexes as well as menace responses were absent. In contrast, the corneal reflex was preserved and both eyes showed spontaneous horizontal nystagmus. Muscular tone of the tongue was decreased, whereas the anal reflex appeared normal.
The clinical signs were most consistent with ivermectin intoxication. Ivermectin overdoses trigger severe neurologic signs by opening γ-aminobutyric acid (GABA)-gated chloride channels, which in turn causes membrane hyperpolarization and blockade of neuronal impulses.[1, 2] In adult horses, no signs of intoxication are observed up to an ivermectin dosage of 1.8 mg/kg, whereas ingestion of 2 mg/kg has been shown to cause lethargy, ataxia, and visual impairment. The praziquantel overdose was considered less relevant in this foal, because praziquantel has a wide margin of safety, and the present exposure would result in only mild neurological and gastrointestinal signs.
Analyses of blood samples at admission identified a low blood glucose concentration2 (3.4 mmol/L; normal, 3.6–6.1 mmol/L), normal plasma protein concentration3 (70 g/L; normal, 57–80 g/L) with low albumin3 (11 g/L; normal, 22–37 g/L) and increased globulin3 concentrations (59 g/L; normal, 27–50 g/L), a total bilirubin concentration3 of 9 μmol/L (normal, 9–39 μmol/L), increased γ-glutamyl transferase activity3 (GGT, 53 IU/L; normal, 5–24 IU/L), and low total calcium concentration3 (2.40 mmol/L; normal, 2.88–3.55 mmol/L). Blood urea nitrogen and creatinine concentrations, as well as plasma sodium, potassium, and chloride concentrations,3 were within reference ranges. The blood lactate concentration4 was 2.3 mmol/L (normal, <3 mmol/L). A CBC5 disclosed normocytic, normochromic anemia (hematocrit, 19.3%; normal, 30–42%; hemoglobin concentration, 7.0 g/dL; normal, 10.8–14.9 g/dL) with increased white blood cell count (WBC, 11.2 × 103/μL; normal, 4.7–8.2 × 103/μL).
Treatment was initiated with warmed lactated Ringer's solution (LRS)6 supplemented with glucose7 to a concentration of 2.5%, administered through a jugular venous catheter at a rate of 60 mL/kg/h. After 10 minutes, the blood glucose concentration had reached 8.3 mmol/L and treatment was continued with LRS without glucose at the same rate. After 1 hour, the fluid was changed to a maintenance crystalloid solution8 containing 1.5% glucose and the infusion rate was decreased to 4 mL/kg/h. By that time, peripheral pulse quality, jugular vein filling, and capillary refill time had improved, and the patient urinated. The colt was placed under a heat lamp, on a forced-air heating blanket, and the extremities were bandaged. To improve ventilation and oxygenation of arterial blood, the colt was kept in sternal recumbency and supplemented with 6 L/min humidified oxygen via nasal insufflation. Based on its benefical effects in a previous moxidectin overdose in a foal, the partial inverse benzodiazepine receptor agonist sarmazenil9 was administered at a dosage of 0.04 mg/kg IV q2h, in an attempt to antagonize the effect of ivermectin on GABA receptors.
Despite treatment, the colt remained stuporous 12 hours after admission. The heart rate had increased to 48/min, the respiratory rate was 18/min, and the rectal temperature had increased to 37.0°C (98.6°F). Venous blood gas analysis10 yielded a pH of 7.40 (normal, 7.34–7.43), a PvCO2 of 50 mmHg (normal, 38–48 mmHg), a PvO2 of 38 mmHg (normal, 37–56 mmHg), a HCO3− concentration of 28.7 mmol/L (normal, 22–29 mmol/L), and a base excess of 3.7 mmol/L (normal, 0–6 mmol/L). The blood glucose2 and lactate4 concentrations were 7.6 and 4.5 mmol/L, respectively.
Over time, blood lactate concentration increased, reached a peak of 7.5 mmol/L about 17 hours after admission, and decreased continuously thereafter to reach concentrations below 3 mmol/L 28 hours after admission. After initial treatment, blood glucose concentration remained between 5.1 and 7.6 mmol/L over the first 24 hours after admission.
Twenty-four hours after admission, the colt still was recumbent and unconscious. No improvement of its neurologic status was evident. A CBC11 now yielded a WBC of 16.4 × 103/μL with 79.5% segmented neutrophils, 0.5% monocytes, and 20.0% lymphocytes. On plasma biochemistry analysis,12 the GGT activity was 50 U/L (normal, 6–31 U/L), the glutamate dehydrogenase activity 18.1 U/L (normal, 0.5–2.2 U/L), the lactate dehydrogenase activity 1053 IU/L (normal, 369–822 IU/L), the sorbitol dehydrogenase activity 133.4 IU/L (normal, 0.1–7.6 IU/L), and the bile acid concentration 19.1 μmol/L (normal, 5–15 μmol/L). All other variables were similar to those observed at admission. Venous blood gas analysis still was unremarkable. Blood lactate4 and glucose2 concentrations were 2.0 and 5.8 mmol/L, respectively. A broad-spectrum antibiotic (cefquinome,13 1 mg/kg q12h) and flunixin meglumine14 (2.2 mg/kg q12h) were added to the therapeutic regimen. Fluid therapy was continued with maintenance fluids8 administered at a rate of 2.5 mL/kg/h IV.
Thirty-seven hours after presentation (71 hours after ivermectin exposure), no neurological improvement was evident. As a consequence, treatment with sarmazenil was replaced by an IV lipid emulsion containing 20% soybean oil in water.15 Because no recommendations on the use of lipid emulsions for the treatment of lipophilic drug toxicities in horses were available, the dosage was based on previously reported cases in small animals[4, 5] and on current protocols for human patients. First, a bolus of 1.5 mL/kg was injected IV through the jugular catheter, followed by 0.25 mL/kg/min over 30 minutes, until reaching a total volume of 234 mL. During lipid administration and for 2 hours thereafter, the colt was closely monitored. Physical examination findings (heart rate, respiratory rate, rectal temperature, capillary refill time, neurologic status) as well as ECG recordings16 and noninvasive oscillometric measurements of systemic blood pressures16 obtained every 10 minutes remained unchanged. Twenty minutes after completion of the lipid infusion, the colt showed slight improvement of the pupillary light reflexes in both eyes and the nystagmus stopped. However, it remained recumbent and unconscious.
Because no further clinical improvement was evident subsequently, IV lipid treatment was repeated at the same dose 19 hours later (ie, 56 hours after admission and 90 hours after ivermectin exposure). Again, heart rate, respiratory rate, rectal temperature, capillary refill time, ECG recordings, and systemic blood pressure measurements did not show any remarkable changes during and after treatment. However, during the 30-minute infusion, neurologic status improved steadily and the colt regained consciousness. Immediately after the 2nd lipid treatment, the colt was able to stay in sternal recumbency and to hold its head in a normal position. The colt also started showing moderate appetite and, with return of the swallowing reflex, was able to ingest small amounts of hay. The pupillary light reflexes still were sluggish and the menace response absent. However, 1 hour later, the colt tried to stand up and, with minimal help, was able to maintain a standing position for 3 minutes. Four hours after the second lipid infusion, the colt stood up independently and remained in a standing position for 7 hours. At that time, with the exception of a dysfunctional menace response and suspected cortical blindness, the foal appeared neurologically normal.
Heparinized and EDTA-supplemented blood samples17 were collected before and repeatedly after the first and second IV lipid treatments for subsequent measurement of plasma triglyceride and ivermectin concentrations (see below). The samples were centrifuged at ambient temperature and plasma was harvested, transferred into cryovials,18 and immediately frozen and stored at −20°C until batch-wise analysis 6 weeks after collection.
Serum biochemistry was repeated after the lipid treatments. Blood ammonia and bile acid concentrations12 were 60 μmol/L (normal, <63 μmol/L) and 15.5 μmol/L (normal, 5–15 μmol/L), respectively, 18 hours after the 1st lipid treatment. Twenty-four hours after the 2nd lipid treatment, blood ammonia concentration had decreased to 24.1 μmol/L and bile acid concentration had increased to 36.6 μmol/L. All other biochemical variables did not change markedly compared with the results obtained 24 hours after admission.
Fluid therapy with the crystalloid maintenance solution8 (2.5 mL/kg/h) was continued for 5 days, and the antibiotic and anti-inflammatory treatments were extended for 3 days. Eyesight and menace responses recovered completely within 5 days after the 2nd lipid infusion. There was no relapse of neurologic signs, and the colt was discharged 8 days after the second IV lipid treatment, at which time the owners assessed the foal's behavior as completely normal. Three weeks later, the colt was presented to the hospital for reevaluation. At that time, no abnormal clinical signs were detected and laboratory variables had further improved.
This case indicates that IV administration of lipid emulsions may be effective for treatment of avermectin overdoses in equids. Both after the 1st and the 2nd IV lipid treatment, the foal appeared to show sudden and marked improvement in neurologic deficits and state of consciousness.
Triglyceride concentrations12 were monitored using plasma from heparinized blood samples. Concomitantly, ivermectin and the internal standard abamectin19 were measured in EDTA-supplemented plasma by solid reversed-phase extraction20 followed by high-performance liquid chromatography (HPLC) coupled to tandem mass spectrometry (MS/MS). First, the analytes were separated using an UltiMate 3000 HPLC system21 equipped with a C18 column.22 A gradient elution was carried out using 5 mM ammonium formate buffer (pH 3) and acetonitrile/formic acid at a flow rate of 0.5 mL/min. The final detection took place in an AB Sciex 5500 Q-Trap tandem mass spectrometer23 by electrospray ionization, multiple reaction monitoring (MRM), and enhanced product ion scans using information-dependent acquisition. Three MRM transitions for ivermectin (892/569, 892/307, 892/551) and 2 MRM transitions for the internal standard (890/305, 890/567) were used.
After therapeutic administration of ivermectin to horses, maximal plasma concentrations of the drug have been reported in the range of 21.4–82.3 ng/mL.[7, 8] In contrast, ivermectin concentrations measured in the plasma of this overdosed pony attained a peak of 1,930 ng/mL (Fig 1). After each lipid treatment, the transient rise in triglyceride concentrations coincided with considerably increased plasma ivermectin concentrations. Subsequently, the drug concentration steadily decreased to reach concentrations of 83 ng/mL at the day of discharge (8 days after the 2nd lipid infusion) and 4 ng/mL 3 weeks later. The parallel course of triglyceride and ivermectin plasma concentrations (Fig 1), in conjunction with the concurrent clinical improvements, allowed us to address the question of how lipid infusions resolve the adverse consequences of drug overdoses. Previous publications already described the use of IV lipid emulsions to manage life-threatening toxicities caused by lipophilic agents unresponsive to other treatments in humans[6, 9, 10] and animals.[4, 5, 11] This therapeutic approach was first described in an experimental model in 1998, but the exact mechanism of action is still a subject of debate. In the case of overdosed local anesthetics, IV lipids have been postulated to either beneficially influence cellular metabolism or exert a direct inotropic effect on the myocardium. In the presence of lipophilic substances that do not cause cardiotoxicity (eg, ivermectin), an alternative scenario known as the “lipid sink” hypothesis postulates that lipophilic drugs are effectively transferred from the central nervous system into the lipid fraction of the vascular compartment.[11, 12] The systemically circulating ivermectin then is prone to metabolism mainly in the liver, or to direct secretion into the gut followed by excretion in the feces, thus accelerating its overall elimination.[1, 2, 14] Here, we were able to confirm the lipid sink hypothesis in a clinical case by demonstrating that the blood concentration of a lipophilic drug indeed increases substantially after IV triglyceride infusion (Fig 1).
The normally low toxicity of ivermectin or other avermectins in mammals is because of their exclusion from the central nervous system. Although highly lipophilic molecules readily diffuse across biological membranes, their penetration into the central nervous system is counteracted by a P-glycoprotein pump, encoded by the multidrug resistance (MDR1) gene, which is expressed at the blood-brain barrier in the luminal membrane of capillary endothelial cells. As a consequence, incomplete expression of P-glycoprotein in neonates[16, 17] or foals under 4 months[3, 18] as well as homozygous mutations of the MDR1 gene may lead to accumulation of lipophilic drugs in the brain. Alternatively, a massive overdose in an otherwise healthy animal may have a comparable outcome by saturating the excretory function of the P-glycoprotein pump.[20, 21] Because no specific antidote exists, the treatment of avermectin overdoses is mostly supportive. Sarmazenil may counteract the action of avermectins by down-regulating chloride conductance, although it did not appear to be beneficial at the dosage and dosing intervals used in the present case.
Conversely, the lipid infusion appeared to dramatically improve the clinical status of the pony. To our knowledge, this is the 1st case study demonstrating that IV administration of a lipid emulsion may be effective for the treatment of avermectin overdoses in horses. However, despite potential benefits in the treatment of lipophilic drug overdoses, IV administration of high doses of lipid emulsions is not without risk of adverse effects.[22, 23] For example, lipid emulsions are known to cause pancreatitis when used for parenteral nutrition in humans, and increased serum amylase activity has been reported after lipid treatment for local anesthetic toxicity in an adult human. In horses, IV administration of phospholipid emulsions has been shown to result in hemolysis with minimal changes in hematocrit, but no signs of pancreatitis.[24, 25] In the present case, no immediate clinical adverse effects were noted during and immediately after lipid infusion and no hemolysis was evident. However, transient increases in blood ammonia and bile acid concentrations occurred and may have been related to lipid treatment, indicating that the functional capacity of the liver might be transiently overwhelmed by the massive lipid load. This was consistent with a previous study on the use of a phospholipid emulsion in horses, which also identified a marked increase in bile acid concentrations after treatment. However, this effect was only transient and no clinical signs of acute liver failure occurred.
In conclusion, this case demonstrates that lipid emulsions should be considered for treatment of avermectin overdoses in horses and foals and might also be used in cases of intoxication with other lipophilic drugs (eg, local anesthetics). The data presented here supports the “lipid sink” theory, explaining the potential mechanism of action of lipid emulsions for treatment of lipophilic drug overdoses. Possible adverse reactions, including liver dysfunction and mild to moderate hemolysis, must be considered when using high doses of lipids IV. Therefore, we recommend that hematocrit, plasma hemoglobin concentration, and integrity and function of the liver be monitored when using lipid emulsions in horses. Additional studies will be required to determine the optimal dosage, efficacy, and full spectrum of adverse reactions of IV lipids for treatment of drug overdoses in horses.