Compounding errors in 2 dogs receiving anticonvulsants

Abstract

Two cases that involve drug compounding errors are described. One dog exhibited increased seizure activity due to a compounded, flavored phenobarbital solution that deteriorated before the expiration date provided by the compounder. The other dog developed clinical signs of hyperkalemia and bromine toxicity following a 5-fold compounding error in the concentration of potassium bromide (KBr).

Drug compounding is the mixing of 2 or more ingredients of which at least one is a drug or pharmacologically active component to create a final product in an appropriate form for dosing (1). Drug compounding is used in veterinary medicine to dilute medicine for smaller patients, to flavor or reformulate medication, for in-clinic mixing of drugs such as in a surgical premix, and to obtain drugs that are otherwise unavailable such as potassium bromide (KBr) and cisapride (2,3). The compounded product is not a generic product and does not have a licensed trade name. There is no obligatory, independent third-party oversight to ensure quality assurance as is required for licensed pharmaceuticals and the scriptor (veterinarian) assumes responsibility for the efficacy and the quality assurance of a compounded product (4).

The following 2 cases of compounding errors were identified by the therapeutic drug monitoring service offered by Diagnostic Services at the Atlantic Veterinary College (AVC), University of Prince Edward Island. Clinical cases involving poor stability of compounded flavored phenobarbital or simple calculation errors in KBr concentration have not been previously described.

Case description

Case 1

The first case involved an 800 g, 1.5-year-old castrated male hydrocephalic Chihuahua dog. The dog had a previously implanted intracranial shunt and the hydrocephalus was in a stabilized, non-progressive state. The Chihuahua was started at 1.25 mg/kg body weight (BW) phenobarbital BID, PO for seizures secondary to the hydrocephalus. This dose was almost immediately increased to 2.5 mg/kg BW BID to achieve seizure control. This is a typical dose; phenobarbital maintenance therapy is often initiated at 1 to 2 mg/kg BW, BID, PO and may reach 2 to 8 mg/kg BW or more BID in order to control seizures yet stay within the therapeutic range (5). Seizure control was excellent for ~3 to 4 mo before the seizure frequency started to increase. At this time, the trough phenobarbital concentration was low at 22 μmol/L (phenobarbital therapeutic range: 54 to 190 μmol/L), as measured by a Cobas c 501 Chemistry Analyzer (Roche/Hitachi Phenobarbital Cobas; Roche Diagnostics GmbH, Indianapolis, Indiana, USA). The dosage was increased by 50% to 3.75 mg/kg BW, BID, PO resulting in a new steady state trough concentration of 44 μmol/L. Because of the continued low concentration, the dosage was further increased to 6.25 mg/kg, BW, BID, PO which was sufficient for seizure control. The new serum concentration wasn’t tested until 3 mo later, when the seizure frequency and severity started to escalate again. At that time, the steady state trough concentration of phenobarbital had decreased to 38 μmol/L (lower than it had been at 3.75 mg/kg BW, BID) despite the dose increase to 6.25 mg/kg BW, BID. This concentration was also surprisingly low for the phenobarbital dosage the animal was receiving.

Before increasing the dosage further, it was important to ensure there were no underlying errors. There were multiple rule outs for the low concentration of phenobarbital: 1) laboratory error, 2) miscommunication with the owner or lack of owner compliance, 3) rapid phenobarbital metabolism — phenobarbital metabolism differs markedly between dogs and while the half-life averages ~48 h, it can vary from < 20 h to > 100 h (5,6), 4) phenobarbital induction of cytochrome P450 enzymes increasing its own metabolism and causing the serum concentration of phenobarbital to decrease slightly, and 5) the drug itself could have been at fault. The veterinarian had prescribed a compounded product because the patient was very small and refused to take any solution that wasn’t specially flavored.

Diagnostic Services repeated the measurement, confirming the results and ruling out laboratory error. The dog was owned by the referring veterinarian who felt that compliance had been excellent. The veterinarian submitted trough and peak serum samples to Diagnostic Services for determination of the phenobarbital half-life in this dog. The half-life was 15 h, indicating a rapid metabolism and that TID therapy would be beneficial. A sample of the compounded phenobarbital solution was submitted for analysis. The phenobarbital concentration was ~2.5 mg/mL, which was half of the 5.0 mg/mL indicated on the label. The testing was done in duplicate to confirm the results; however, the method used was not validated specifically for use in therapeutic solutions of phenobarbital. This suggested that the compounded solution had either been mixed incorrectly or was unstable. The solution was ~4 to 5 wk old and the compounding pharmacist had stated that the product had a shelf-life of 3 mo.

The solution was replaced by a newly compounded product that was not tested for its concentration; seizure control improved immediately and the new steady state trough phenobarbital measurement increased to 76 μmol/L despite continuing with the same dosage regimen. The compounded product was subsequently renewed every 3 wk and 6.25 mg/kg BW, BID continued to provide excellent seizure control for the following 18 mo after which the dog died of unrelated causes.

Case 2

The second case involved a castrated male 9-year-old Shetland sheepdog that had been treated with KBr and phenobarbital for idiopathic epilepsy for 6 y prior to presentation for the compounding error. Therapeutic levels of both anticonvulsants were measured annually and had been stable for > 2 y. The dog was receiving 27 mg/kg BW, SID of KBr (normal dosage of KBr is 20 to 60 mg/kg BW SID or divided BID) and 4 mg/kg BW of phenobarbital divided BID PO at the time of the compounding error.

Potassium bromide is only available as a compounded drug. The owner refilled the KBr prescription at a local pharmacy employing a compounding pharmacist. Five days later, the dog fell down the stairs and appeared to have a subdued demeanor, sore neck, and delayed proprioception in both forelimbs. Within a few days his signs progressed to depression, inability to walk, mild abdominal discomfort, and bradycardia (heart rate 60 beats/min). Cervical and abdominal radiographs were unremarkable. The complete blood profile showed a mild, normocytic, normochromic, non regenerative anemia compatible with decreased red cell production due to chronic disease or a recent episode of hemolysis or blood loss with no time to respond. An in-clinic serum biochemical analysis showed an elevated potassium (K⁺) [8.4 mmol/L, reference interval (RI): 3.7 to 5.8 mmol/L] and chlorine (Cl⁻) (123 mmol/L, RI: 106 to 120 mmol/L), and normal sodium (Na⁺) (150 mmol/L, RI: 138 to 160 mmol/L). The alkaline phosphatase, amylase, and creatinine were mildly elevated. Hyperkalemia can occur with hypoadrenocorticism, renal insufficiency or urinary tract obstruction, metabolic acidosis, tissue necrosis, rhadomyolysis, repeated drainage of thoracic chylous effusions, peritoneal effusion or increased K⁺ intake (7). Hyperchloremia can occur with salt toxicity, hypertonic saline, decreased renal excretion of Na or Cl, hyperaldosteronism, hyperchloremic metabolic acidosis, gastrointestinal or renal loss of bicarbonate, chronic respiratory alkalosis, excessive water loss such as with an osmotic diarrhea or renal loss, pure water loss or decreased water intake or interference from Br producing a spurious hyperchloremia (7). The urinalysis indicated moderate proteinuria and a urine specific gravity of 1.020. The protein/creatinine ratio was increased to 2.6 (normal < 1). Aggressive therapy with IV 0.9% NaCl resulted in improved mental status and heart rate (120 beats/min).

A normal ACTH stimulation test ruled out hypoadrenocorticism. The serum KBr concentration was elevated at 45 mmol/L despite having been maintained at the same dose for more than 2 y and normally having a serum concentration of ~26 mmol/L (normal therapeutic range for KBr is 12.5 to 31 mmol/L) for dogs receiving phenobarbital and KBr simultaneously, as measured using an F1022Br Bromide Selectrode in conjunction with a PHM-24 pH meter (8,9), (Radiometer; Denmark at Diagnostic Services, AVC). Failure of owner compliance or miscommunication was considered unlikely as the owner had been administering the drug for several years correctly but the owner was interviewed to rule out noncompliance. A sample of the compounded drug was sent to Diagnostic Services. The KBr concentration in the solution was 225 mg/mL instead of the 200 mg/5 mL that had been prescribed and written on the label. The laboratory rechecked the concentration to confirm the results; however, the methodology used was not validated for use in any matrix other than serum; therefore, an error in the laboratory results cannot be ruled out. The presenting signs, hyperchloremia, hyperkalemia, and subsequent response to therapy, however, were all compatible with the laboratory results. The dog was given supportive treatment of 0.9% NaCl IV and the KBr was discontinued for 7 d at which time the serum KBr was found to have returned to 24.8 mmol/L. The regular maintenance therapy was reinstituted using newly compounded KBr.

This dog had clinical signs of K⁺ and Br⁺ toxicity due to a 5-fold compounding error. Hyperkalemia can cause bradycardia, cardiac arrhythmias, muscle weakness, depression, and anorexia (10). Bromide toxicosis can cause a spurious hyperchloremia (11) and is associated clinically with depression, lethargy, decreased proprioception and hyporeflexia, stupor, erythematous dermatitis, conjunctivitis, nausea, anorexia, and weight loss (11,12). Pancreatitis has also been reported in dogs treated with KBr and phenobarbital (13). Disorientation and ataxia due to the bromide toxicosis likely resulted in the initial fall down the stairs. The mild abdominal discomfort exhibited throughout may have been due to constipation, gastrointestinal irritation, or pancreatitis, which have all been reported secondary to KBr toxicity. The pancreatic lipase immunoreactivity was later determined to be elevated (653 μg/L, RI: 0 to 200 μg/L), supportive of pancreatitis.

The KBr was continued at 27 mg/kg BW, SID, PO and the phenobarbital was decreased to 15 mg BID, PO. The clinical response and subsequent serum KBr concentrations continued to be monitored at regular intervals. Four months following the episode, the dog was presented to his veterinarian for decreased appetite, weight loss, and evidence of cognitive dysfunction. Following extensive testing, he was diagnosed with a protein-losing nephropathy, decreased liver function, hepatic nodular hyperplasia, and a steroid hepatopathy. The dog was euthanized within a year of being presented for KBr toxicity. The dog had appeared healthy prior to the intoxication. It cannot be determined if any of his subsequent health problems were due to the toxicity event or whether they were coincidental.

Discussion

Original, patented drugs undergo careful scrutiny by Health Canada to assure efficacy, safety, and quality before being given a Drug Identification Number (DIN) indicating the right to mass produce and market a drug in Canada (2). The drug company, packagers, labelers, distributors, and all other individuals involved in producing and selling a patented drug must have an Establishment License, abide by Good Manufacturing Practices (GMP), and comply with other applicable regulations (1).

Unlike patented and generic products, compounded drugs have no DIN, are not tested for bioequivalence, and are not produced under the guidance of GMP (14). Compounded drugs are produced in accordance with Health Canada Policy Guidelines regarding compounding (1,15). These guidelines are intended to ensure that patients receive safe and efficacious products, that there are no violative food residues, and that compounding doesn’t circumvent the regular process of drug application, certification, and production (16).

The Health Canada Guidelines on compounding are explicit (1). Compounding must be done by a licensed pharmacist or veterinarian. A compounded drug can only be produced for an animal with a valid veterinary-client-patient relationship on an as needed basis. An animal must be suffering or its health threatened without treatment. Cost is not an acceptable reason for compounding except in extreme circumstances. There cannot be an equivalent veterinary or human product already available. The compounded drug must be safe, efficacious, and within 10% of its labeled dose. Compounded products are usually made from a Health Canada approved product, preferably veterinary, but if none is available then from a human drug (1). Unlike the situation in the United States, it is legal to use a bulk active product ingredient (API) in Canada if no approved product exists, such as with KBr and cisapride (1,15). Owners must give informed consent and records are required for the documentation of all compounded product use (1).

There are potential pitfalls when using compounded products. Reformulating a product may change its stability, potency, or bioavailability. Compounders use stability studies to give an expiration date if possible. If there are no data available, then the compounder should use 25% or less of the “beyond use” date of the original product (3,17). In the USA, compounders may quote 6 mo for solid products and 14 d for aqueous solutions (3). In Case 1, Health Canada approved phenobarbital tablets were altered by crushing and putting them into solution and then flavoring. Altering the formulation can change a drug significantly. Omeprazole suspensions were found to give a significantly lower serum concentration and poorer efficacy compared with the patented paste form of omeprazole (Gastrogard), possibly because the suspensions were more acidic and omeprazole is less stable in acidic solutions (18). Alteration of the pH of some drugs by 1 unit can change stability by a factor of 10 (19). Flavoring can also alter the stability of a drug. Orbifloxacin was found to be 61% of its original strength by day 4 when the flavoring lixotinic was added, but stable when a variety of other flavorings were used (20). Addition of water to put a drug into solution can cause hydrolysis of some drugs (19). Reformulating a drug into a transdermal form by combining it with a penetration enhancer such as pluronic lecithin organogel is an attractive option for some feline patients but it is not possible for many drugs (21). Whereas transdermal methimazole is efficacious when compounded correctly and associated with less gastric irritation than oral methimazole products (22), morphine, dexamethasone, and fluoxetine are poorly absorbed in transdermal products (23–25).

Compounding services are offered by compounding companies and human pharmacies (26). Compounding pharmacists may assume that veterinarians understand that compounded drugs are not tested for efficacy, bioavailability, stability, or safety. Veterinarians may assume that a product being offered by a compounder must be safe and efficacious (26). According to Canadian Compounding Guidelines, a pharmacist compounds a product according to the request or prescription of the veterinarian and the veterinarian then takes responsibility for the safety and efficacy (1). If the compounded drug is being used in a performance animal, the veterinarian must provide the new withdrawal time. The Global Food Animal Residue Avoidance Databank (gFERAD), which can be used to obtain new withdrawal times for drugs used in an extralabel manner, does not provide withdrawal times for compounded products.

The Canadian Veterinary Medical Association (CVMA) has compounding guidelines on its Web site (4). These guidelines provide a scripting hierarchy. Veterinarians should choose a drug based on the least level of risk to both the patient and public (4). When possible, an approved veterinary drug with a DIN and label instructions should be used. If none is available, then an approved veterinary drug with a DIN can be used extra-label. When no veterinary drug is available, then an approved human drug with a DIN is used extra-label. When neither of these is possible, then a compounded drug that contains an approved veterinary drug is used (extra-label). If this isn’t available, then a compounded approved human drug with a DIN is used. A compounded drug containing an active pharmaceutical ingredient (as opposed to a drug with a DIN) is only used if none of the previous choices are available (4). The CVMA compounding guidelines have been included in some provincial veterinary association guidelines and can be used for guidance during investigations into allegations of professional misconduct in regards to the use of compounded drugs (27). The CVMA recommends that veterinarians only use compounded drugs when necessary and use a pharmacist (or company) with expertise in compounding. Veterinarians should feel comfortable in asking for pertinent information such as: the compounding credentials of the compounder, the active ingredient and its source, tests for quality assurance, the method used to formulate the new product, and validation of effectiveness (for example, if they are making a transdermal formulation, is there any proof that it works). The veterinarian should also have predetermined and practical clinical assessment parameters for efficacy and toxicity before dispensing the drug (4). Monitoring of serum concentrations of the drug (if there are validated methods to do so) in addition to clinical response is especially helpful, given that the products have not necessarily been evaluated for efficacy or pharmacodynamic/pharmacokinetic characteristics.

There must be informed consent when prescribing a compounded product. The veterinarian must inform the owner of the risks of using compounded drugs and what they should watch for with regards to toxicity or drug failure. The owners must understand that the drug is not approved, has not gone through a rigorous government approval process, and that the efficacy is not necessarily known (27). The College of Veterinarians of Ontario includes sample consent forms on its Web site (27).

Adverse reactions to compounded drugs should be reported to both Health Canada’s Veterinary Drug Directorate and the compounding pharmacist (4). Products that are imitations of approved products or bulk batches of compounded medication for resale to clients should not be prescribed as, by definition, these are pirated products (4,28).

The veterinarians in both of these cases complied with compounding guidelines and suggested practices. In the first case, the Chihuahua dog required a compounded product due to its small size and taste issues; in the second case, KBr is only available in a compounded form. Both veterinarians provided prescriptions to local reputable pharmacists advertising expertise in compounding. In both cases, the compounded pharmaceuticals were drugs commonly measured for therapeutic drug monitoring which allowed prompt analysis when questions arose about efficacy or safety. When prescribing compounded drugs, veterinarians should remember that the products are made on an individual basis, are more susceptible to human error, and that they, as the scriptors, take responsibility. CVJ