Barriers to Drug Distribution

The blood–brain barrier and blood–cerebrospinal fluid (CSF) barrier form a permeability boundary, rendering the brain a sanctuary by limiting penetration of macromolecules. They also act as selective barriers to inflow and outflow of small, charged molecules, including drugs, neurotransmitters, and their precursors or metabolites. The blood–brain barrier is absent in selected locations, thus allowing the brain to monitor chemicals. These areas include but are not limited to the area postrema, median eminence of hypothalamus, and the posterior pituitary and pineal glands. Cerebral ischemia and inflammation modify the integrity and function of the barriers. The blood–brain barrier comprises blood capillaries that differ in structure compared to other tissues in that they lack fenestrae and are joined by tight junctions similar to the epithelium of other tissues. Microvessels comprise about 95% of the total surface area of the barrier. Astrocyte foot processes encapsulate the capillaries, forming the tight junctions (Figure 27-1). Because pinocytosis also is absent, compounds can enter the brain only by passive diffusion, limiting penetration to small, lipid-soluble, non-ionized molecules. In addition to the physical barrier, function mechanisms exist to preclude central nervous system (CNS) penetration. These include degradative enzymes, which target not only chemicals but also many peptides, and high concentrations of efflux transporters. These include adenosine triphosphate (ATP)–binding cassette proteins such as P-glycoprotein (among the most abundant and most well characterized) and others, resulting in a combined broad substrate specificity that targets many drugs.² The substrate specificity of the P-glycoprotein transporter is large; its role in breed susceptibility to CNS adverse drug reactions is discussed in Chapters 2 and 3. Interactions involving P-glycoprotein (or other efflux transporters) are among the drug interactions described at the level of the blood–brain barrier.³ In the dog this protein is regulated by only one gene (ABCB1).⁴

Figure 27-1 Diagrammatic representation of the blood brain (left) and blood–cerebrospinal fluid (right) barriers, structures that limit compound movement from blood into the interstitial fluids of brain parenchyma. The endothelium of the capillaries of the blood–brain barrier lack fenestrae, are joined by tight junctions, and are encapsulated by astrocyte foot processes. Only small, lipid-soluble, non-ionized molecules in the blood stream are likely to move through the endothelium into the brain. The blood–CSF barrier is formed from the epithelium of the choroid plexus, which forms tight junctions. The plexus dips into a double-celled structure made from folding of ependymals cells on themselves. (See text for other mechanisms by which drugs might be excluded from each of these sites.)

The blood–CSF barrier prevents compounds moving from interstitial fluid into brain parenchyma.⁵ This barrier is located in the epithelium of the choroid plexus, limiting movement of compounds from the blood into the CSF. The choroid plexus consists of the highly vascularized pia mater and cells with microvilli on the CSF side. The plexus dips into pockets formed by folding of ependymal cells onto themselves, forming a double layer structure between the dura and pia (i.e., the arachnoid membrane). This double layer contains the subarachnoid space. Although the capillaries of the choroid plexus are fenestrated, the adjacent choroidal epithelial cells form tight junctions, precluding movement of most macromolecules. Like the blood–brain barrier, the blood–CSF barrier contains organic acid transporters that preclude penetration of many therapeutic agents.⁶ The role of efflux pumps and P-glycoprotein in particular is discussed later.

The formidable barrier presented to chemical penetration of the brain contributes to therapeutic failure when treating dysfunction of the CNS. A number of strategies have been described to enhance drug delivery to the brain.⁶ These include manipulation of the drugs such that they are chemically more likely to penetrate the barrier, administration of prodrugs, the combination of drugs with compounds that decrease drug efflux, the design of vectors or other carriers that are able to penetrate, and temporary disruption of the barriers themselves. The use of nanotechnology offers potentially viable options.⁷ Finally, alternative routes of delivery include intraventricular, intrathecal, olfactory, and direct interstitial routes. None thus far has proved consistently safe or relevant to all therapeutic agents.

Amino Acids

Glycine receptors are located primarily in the brainstem and spinal cord, where they are inhibitory in nature.

Glutamate and aspartate occur in very high concentrations and are powerful excitatory signals throughout the brain. Glutamate receptors are either ligand-gated ionotropic receptor channels or G protein coupled. Ligand-gated ion channels include N-methyl-d-aspartate (NMDA) and non-NMDA receptors (alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid [AMPA] and kainic acid [KA]).^1,8,9 The NMDA receptors, derivatives of aspartic acid, are pH sensitive, and response to endogenous signals, including zinc ions, selected neurosteroids, arachidonic acid metabolites, redox regions, and selected polyamines. Although NDMA receptors are involved in normal synaptic transmission, they appear to be more closely involved with induction of synaptic plasticity. This includes long-term potentiation, which reflects an increase in the magnitude of the postsynaptic response to presynaptic stimulation of a specific strength. The activation of NMDA receptors requires simultaneous firing of at least two neurons, including binding of glutamate released from the synapse and simultaneous postsynaptic depolarization from different neurons. High concentrations of glutamate can cause neuronal death. The mechanism partially, but not solely, reflects excessive activation of NMDA or AMPA/KA receptors and calcium influx. For example, ischemia or hypoglycemia may be associated with massive glutamate release; when coupled with impaired cellular reuptake, cell death may occur. Depletion of Na⁺ and K⁺ and increased extracellular Zn2⁺ may play a role in activation of necrotic or apoptotic cascades.¹ The role of these receptors in pain is discussed in Chapter 28.

Monoamines

Acetylcholine is an excitatory neurotransmitter for Renshaw interneurons and other CNS cells. Actions reflect interactions between nicotinic and muscarinic receptors.

Several catecholamines act as neurotransmitters in the CNS. Dopamine comprises more than 50% of CNS catecholamine and is found in high concentrations is selected areas. Its effect depends on the subreceptor targeted, with five thus far identified in the brain (see Table 27-1). Epinephrine is present only in limited areas, and its role has not yet been identified. However, norepinephrine is found throughout the brain but in very large concentrations in the hypothalamus, certain zones in the limbic system, and other regions. All three receptor types (alpha 1, alpha 2, and beta) and subtypes are located in the brain. The nonclassical electrophysiologic synaptic effects of norepinephrine vary with the state but are considered “enabling.” Although excitatory in nature, the effects of norepinephrine (and serotonin) can be anticonvulsant, rather than proconvulsant.¹⁰

Biogenic Amines

Serotonin is one of two biogenic amines active in the CNS. Thus far 14 receptor (5-hydroxytryptamine [5-HT]) types or subtypes are found in CNS nuclei that are located in or adjacent to the midline regions of the pons and upper brainstem. Up to five different 5-HT₁ receptor subtypes inhibit adenylyl cyclase or regulate K⁺ and Ca²⁺ channels. 5-HT₂ receptor subtypes (three) are linked to G proteins and phospholipase C. 5-HT_2C may modulate CSF production. 5-HT₃ receptors are located in the area postrema and solitary tract nucleus, where they cause emesis and antinociception. Members of the 5-HT₄ through 5-HT₇ receptor types have not yet been studied in the CNS.

Histamine is the second biogenic amine that serves as a neurotransmitter. Four subtypes have been described in the CNS, all acting through G proteins. Most neurons that synthesize histamine are located in the ventral posterior hypothalamus and extend throughout the entire CNS through ascending and descending tracts.¹

Synaptic Vesicle Protein

The importance of the synaptic vesicle protein 2 (SV2) emerged as part of the research into the mechanism of action of levetiracetam. Levetiracetam, but not other anticonvulsants, binds to synaptic vesicle (SV) protein through the brain in a steroselective, concentration-dependent manner.¹¹ However, ethosuximide, pentobarbital, and pentylenetetrazole compete with levetiracetam for binding. The distribution of the protein does not follow any pattern consistent with classical receptors, and the protein is located in synaptic vesicled (SVs). Three isoforms exist (A, B, C), with SV2A the most widely distributed. Knockout mice without SV2A die within 3 weeks as a result of spontaneous seizures. The role of SV2A is not clear, but it may modulate exocytosis that precedes neurotransmitter release; disruption of activity appears to result in calcium accumulation and increased neurotransmitter release.¹¹

The Action Potential

The normal resting membrane potential (RMP) of the neuronal cell is 70 mV. The electrical difference across the cell membrane is maintained by a Na⁺, K⁺-ATPase pump. Depolarization and the generation of an action potential occur when the RMP becomes sufficiently positive to reach threshold. As with other membranes, the RMP of a neuron is determined by the concentration of negative and positive ions across the membrane. Important ions include Na⁺, K⁺, Ca²⁺, and Cl^–. The concentration reflects ion fluxes and thus permeability of the cell membrane to the ions. Fluxes resulting in an increase in positive ions inside the cell relative to the outside hypopolarize the RMP, bringing it closer to threshold and subsequent depolarization. The tendency of a neuron to depolarize reflects, in part, the sum total effect of neurotransmitters (NTs) that interact with the cell membrane. Inhibitory NTs such as GABA render the RMP more negative and less susceptible to depolarization (see Figure 27-2).¹² The principal postsynaptic receptor for GABA is the GABA_A receptor; activation increases inflow of Cl- ions into the cell, hyperpolarizing the neuron. Glycine and serine are also inhibitory neurotransmitters. Excitatory neurotransmitters, such as acetylcholine and glutamate, elevate the RMP to a more positive status, rendering it more susceptible to reaching the threshold necessary for depolarization.¹²

Mechanism of Action

Phenobarbital sodium (see Figure 27-4) specifically depresses the motor centers of the cerebral cortex, thus enhancing anticonvulsant properties. Electroshock experiments with cats and other species established phenobarbital as one of the most potent anticonvulsants available. It has the widest spectrum of activity in different convulsive seizure patterns. Many AED have been synthesized as structural variants of phenobarbital. For example, primidone is a close congener of phenobarbital.

Phenobarbital is the most effective anticonvulsant to delay progressive intensification of seizure activity that may accompany epilepsy. Phenobarbital both increases the seizure threshold required for seizure discharge and decreases the spread of discharge to surrounding neurons; its primary mechanism is enhancement of responsiveness to the inhibitory postsynaptic effects of GABA. Interaction of GABA opens a chloride channel, resulting in higher intracellular concentrations of chloride and hyperpolarization of the RMP (see Figure 27-3). Phenobarbital also, however, inhibits glutamate activity and probably calcium fluxes across the neuronal membrane. A study that compared CSF concentrations of GABA and glutamate in epileptic Labrador Retrievers versus non–Labrador Retrievers and in non-epileptic Beagles also compared concentrations in those epileptic dogs treated with phenobarbital and those not treated.²² Within the non–Labrador Retriever phenobarbital-treated group, GABA, glutamate, and aspartate CSF concentrations were lower compared with those of their untreated counterparts. This difference was not detected in the Labrador Retrievers treated versus untreated groups; however, only 14 Labrador Retrievers were receiving phenobarbital, which may have limited the power to detect a significant difference.

Phenobarbital might be considered a broad-spectrum anticonvulsant. Despite introduction of new antiepileptic drugs, phenobarbital has generally been the anticonvulsant of choice for the cat and dog.¹⁷ It is effective in all types of epileptic seizures observed in cats and dogs.

Disposition

As a weak acid (pK_a 7.3), phenobarbital is well absorbed after oral administration, although peak plasma concentrations may not be reached for 4 to 6 hours after administration. The absorption half-life in dogs is 1.27 ± 0.21 hours.⁴⁰ With an elimination half-life of 0.26 ± 0.18, about 6.4 hours may be required for near complete absorption of phenobarbital from the gastrointestinal tract. Absorption is 88% to 95% complete.⁴¹ Phenobarbital is 45% bound to serum protein in dogs.⁴² The volume of distribution of phenobarbital in dogs is 0.7 ± 0.15 L/kg and 0.7 ± 0.4 in cats.⁴³ Approximately 16 days (8 to 15.5 days) of multiple dosing is necessary to attain steady-state serum concentrations in the normal (uninduced) animal; however, induction may cause this period to be shorter after a change in the dosing regimen of an animal already receiving phenobarbital. Maintenance doses of 5.5 mg/kg daily (in one to three equal doses) administered orally are required to reach an average serum concentration of 20 μg/mL, according to one study.⁴⁴

Through microsomal enzyme action, phenobarbital is metabolized by oxidative hydroxylation to form hydroxyphenobarbital. This metabolite has a weak anticonvulsant activity that does not contribute significantly to the action of phenobarbital. Hydroxyphenobarbital is rapidly eliminated from blood by conjugation with glucuronide and excretion in urine of the dog. Up to 25% of the parent drug is eliminated renally in dogs. Individual variability in the rate of phenobarbital elimination is marked owing to differences in hepatic metabolism. Half-life varies not only between and within species but also in the same animal, in part because of induction (discussed later).

Breed differences may exist in phenobarbital clearance. Frey and Löscher⁴² reported a half-life of phenobarbital at 64 ± 15 hours in largely mongrel dogs, but 32 ± 4.8 in Beagles; clearance was 7.0 ± 1.3 and 13 ± 1.7 mL/kg/min, respectively. Alkalinization of urine (pH >7.5) accelerates excretion of unaltered phenobarbital by reducing back-diffusion (tubular reabsorption) through drug ionization.⁴⁵ Fukunaga and coworkers⁴⁶ demonstrated that a twice-daily fed diet containing either 0.6 gof potassium citrate (pH 7.8 to 8.2) or 0.2 g of ammonium chloride (pH approximating 5.9 to 6.5) increased urine clearance of a single dose of phenobarbital compared with a water vehicle (pH 6.8 to 7.5) in Beagles (n = 5; female). However, the effect did not significantly (statistically nor clinically)impact peak plasma drug concentrations (μg/mL: 3.25, 2.96, and 2.97 μg/mL for water placebo, ammonium chloride, or potassium citrate, respectively) or area under the curve. Yet, half-life was affected (52, 59.6, and 51.1 hours, respectively), and as such peak concentrations might be slightly affected if the drug is dosed to steady state.

Phenobarbital is a potent inducer of hepatic drug-metabolizing enzymes and is capable of increasing the rate of clearance of other drugs metabolized by the liver as well as increasing its own rate of metabolism.⁴⁷ In the dog phenobarbital (2 mg/kg) administered orally 3 times a day for 5 days results in an elimination half-life between 37 and 75 hours, with a mean elimination half-life of 53 ± 15 hours.⁴⁴ After a single 5-mg/kg intravenous dose, clearance is 5.6 to 6.6 mL/kg per hour, and elimination half-life is 92.6 ± 23.7 hours in dogs⁴⁰ and 47 ± 3 hours after oral administration (5 mg/kg) in cats. The effects of multiple doses of phenobarbital were documented by Ravis and coworkers.⁴⁸ After 90 days of treatment (5.5 mg/kg), mean elimination half-life decreased from 88.7 ± 19.6 hours to 47.5 ± 10.7 hours in dogs. In cats 21 days of oral phenobarbital at 5 mg/kg resulted in a half-life of 43 ± 3 hours.⁴³

Preparations

Phenobarbital is available as either an oral or an injectable preparation. Oral tablets contain 0.25, 0.50, or 1 grain (15, 30, and 65 mg, respectively) phenobarbital. An elixir is also available (4 mg/mL) for treatment of very small animals. The injectable form is intended for intravenous use but can be given intramuscularly and rectally. Under the 1970 Controlled Substances Act, phenobarbital is classified as a Schedule IV drug.

Clinical Use

Phenobarbital is reportedly effective in 60% to 80% of canine patients suffering from epilepsy if serum concentrations of the drug are maintained within the recommended therapeutic ranges of 20 to 45 μg/mL.⁴⁹ This is consistent with the early work of Farnbach,⁵⁰ who reported that the concentrations of phenobarbital necessary to control seizures (control defined as interictal interval prolonged at least twice that of the pretreatment period). Concentrations considered effective (n = 22) ranged from extremes of 6.5 to 81.3 μg/mL, with others ranging from 14.2 to 46 μg/mL (mean 23 μg/mL); however, concentrations not associated with control (n=20) ranged from 4.7 to 42 μg/mL. Efficacy might be higher if therapeutic failure is not considered until the maximum end of the range (i.e., 35 to 45 μg/mL) has been reached (Figure 27-6.). The author has reported 85% efficacy in eradication of seizures in spontaneous canine epilepsy with phenobarbital concentrations at 25.4 ± 5.7 μg/mL with such an approach.²⁸ In this same study, phenobarbital was found to be superior to bromide for first choice control of seizures in dogs (see the discussion of bromide). Phenobarbital has been used successfully to treat hypersialosis.⁵¹

Figure 27-6 Algorithm for modification of anticonvulsant dosing regimens. The goal of monitoring in a stepwise fashion is to achieve adequate control of seizures with the lowest plasma drug concentration possible. This approach minimizes side effects and drug interactions and may allow for the least adaptation by the brain. The highest concentration is generally not initially chosen, particularly for drugs associated with side effects. Combination therapy generally should not be pursued until monotherapy fails (i.e., seizures remain unacceptable despite drug concentrations at the maximum end of the therapeutic range, or in the presence of unacceptable side effects). If initial combination therapy fails, a third drug may be necessary. If the patient responds to combination therapy and if seizure history warrants doing so, the drug that is least effective or associated with the most side effects might be gradually discontinued. Monitoring during withdrawal might be prudent to determine the minimum effective concentration necessary to control seizures should the patient develop breakthrough seizures. Should grogginess emerge with the addition of a second or third anticonvulsant drug, a 25% decrease in the drug most sedative and with the shortest half-life should be considered.

Patients should not considered refractory to phenobarbital therapy until plasma concentrations reach 35 to 40 μg/mL unless unacceptable side effects persist at lower concentrations. However, because of the potential for phenobarbital-induced hepatotoxicity, the author often recommends adjuvant therapy once phenobarbital concentrations exceed 25 μg/mL in an uncontrolled dog. The dose necessary to achieve target concentrations may vary dramatically among dogs and in the same dog across time because of differences or changes in drug disposition. Changes are likely to reflect, in part, induction of drug-metabolizing enzymes; monitoring should occur 1 to 2 months after baseline is established to detect the possible impact of induction. Measurement of both a peak and trough sample can detect a short half-life; for patients in which phenobarbital half-life is 24 hours or less, the same or a slightly higher (e.g., 25% increase) total dose divided into 8-hour rather than 12-hour intervals may minimize fluctuation of plasma drug concentrations. Clearly, once-daily dosing should be avoided in the face of induction that decreases half-life to 36 hours or less. Phenobarbital elimination half-lives of 9 to 12 hours have been documented in the author’s laboratory in some dogs that have been receiving phenobarbital for several months.

Phenobarbital can be administered intravenously for acute control of seizures, although a lag time (i.e., 20 to 30 minutes) may be observed before control of seizures. An intravenous loading dose of 12 mg/kg is designed to achieve therapeutic concentrations (20 μg/mL) immediately; the dose can be decreased proportionately (on the basis of serum phenobarbital concentrations) if the patient is currently receiving phenobarbital. Alternatively, the calculated dose can be given in four to six equal hourly doses. Collection of a postload monitoring sample would be prudent once seizures are controlled; this sample can provide a target for maintenance therapy. A second sample 2 or 3 days later should be used to assess the oral maintenance dose. If used in combination with intravenous diazepam (for its more immediate effects) to prolong the control of seizures, the phenobarbital dose might be administered intramuscularly to avoid respiratory and cardiovascular depression. Once seizures are controlled with phenobarbital alone, monitoring should establish the target concentration for chronic therapy in the patient.

Limited information is available regarding the use of phenobarbital in cats. Phenobarbital administered at 6 mg/kg daily was useful in suppressing experimentally induced hippocampal generalized seizures in cats; drug concentrations ranged between 15 and 25 μg/mL. After-discharge was totally suppressed in 33% of cats at 35 to 50 μg/mL (12 mg/kg daily). However, phenobarbital was less effective for amygdaloid-kindled seizures.⁵²

Drug Interactions

Hepatic microsomal enzyme activity, especially mixed-function oxidase (cytochrome P450 [CYP]) induction, is accelerated by phenobarbital.^53-56 Enzyme induction by phenobarbital appears to be dose related.⁵⁷ Long-acting barbiturates are better inducers of microsomal enzyme activity than are short-acting compounds. For example, compared on a molar basis, phenobarbital has been described as the most potent enzyme stimulatory agent known,⁵⁸ whereas pentobarbital and thiopental sodium are less potent inducers of microsomal enzyme activity, in part because duration of therapy dose, and thus impact, is limited. Enzyme induction may take weeks to months and may occur with each dose increase. Induction has been documented in dogs^59-62 but apparently not in cats. Although it may occur occur in the latter species, saturation of drug met. Phenobarbital was associated with a twofold increase in CYP protein and increased clearance of the hepatic metabolism marker antipyrine in Beagles treated with 5 mg/kg orally every 12 hours for 35 days. Among the CYP isoenzymes characterized by an increase in activity were CYP1A, CYP2B, CYP2C, and CYP3A; in contrast, CYP2D did not appear to increase.⁶³

Induction does not immediately resolve once a drug is discontinued: For example, once phenobarbital initiates induction, up to 7 months may be required for complete resolution in the dog. Fukunaga and coworkers⁶⁴ demonstrated that antipyrine clearance was increased in response to phenobarbital at 4 weeks but not 6 weeks in dogs receiving 5 mg/kg of phenobarbital twice daily. Clearance was increased by close to 60% during phenobarbital therapy. Interestingly, volume of distribution increased almost threefold, and half-life increased, as did mean residence time, despite the change in clearance. The investigators did not include a control group that did not receive phenobarbital, complicating interpretation. Clinically, as per the author’s therapeutic drug monitoring laboratory, however, induction often decreases the elimination half-life and thus duration of therapeutic response to other drugs metabolized by those enzymes induced by phenobarbital (see Chapter 2). In the dog, experimental documentation includes digoxin and digitoxin^44,65,66 and thiopental.⁶⁷ However, induction does not appear to affect thiopentone (thiopental) dose requirements in the dog.⁶⁸ For drugs in which toxic metabolites are produced, the induced liver will produce more metabolites; therefore potentially hepatotoxic drugs should not be used with phenobarbital. N-Acetylcysteine should be used in cases of drug-induced acute hepatopathy, regardless of the cause (exceptions might be made in the presence of hepatic encephalopathy). Longer term support might include s-adenosyl methionine (SAMe) or milk thistle as well as other therapies.

Phenobarbital therapy (2.5 mg/kg orally every 12 hours for 30 days) was associated with a decrease in the peak concentrations of benzodiazepines (diazepam/oxazepam) after single-dose administration of 2 mg/kg either intravenously (about 10%, from 5963 to 5565 ng/mL) or rectally (about 60% decrease, from 630 to 270 ng/mL). However, the target associated with seizure control (150 ng/mL) was achieved in five of six dogs within 8 minutes of rectal administration.⁶⁹

Side Effects

Treatment of Acute Phenobarbital Toxicosis

Treatment of acute phenobarbital toxicosis is generally supportive. Dogs apparently can tolerate marked acute overdosing (in the author’s experience, concentrations approximately 150 μg/mL) without persistent effects, once the drug has been discontinued. Sedation is the most detrimental effect. Artificial respiration with oxygen should be administered to prevent respiratory arrest induced. Although less effective than oxygen, doxapram or other analeptic drugs might stimulate the respiratory center. Alkalinization of the urine accelerates renal excretion of phenobarbital because of ion trapping in the urine, although this is not easily accomplished.⁴⁵ Activated charcoal effectively accelerates the body clearance of phenobarbital.⁸² When charcoal is administered in the human, the biological half-life of phenobarbital is decreased from 110 ± 8 to 45 ± 6 hours; it increases the total body clearance of phenobarbital from 4.4 ± 0.2 to 12.0 ± 1.6 mL/kg per hour.⁸² Assuming drug metabolizing enzymes are not saturated (threshold not known in the dog or cat), the time needed for phenobarbital concentrations to decline such that sedation is no longer present depends on the drug concentrations (i.e., the extent of overdose) and the elimination half-life of the drug in the patient. For example, at 150 μg/mL, approximately three half-lives must elapse before concentrations drop below 35 μg/mL; for a 72-hour half-life, 9 days must elapse. Both the magnitude and half-life (assuming saturation has not occurred) of phenobarbital can be monitored in an overdose patient; a sample should be collected on presentation and again 24 hours later. Timing of sample collections must accompany the samples if a half-life is to be calculated. A follow-up sample 2 or 3 days later can confirm that the original half-life was correct (i.e., saturation has not occurred). In cases indicative of acute hepatopathy, N-acetylcysteine should be administered, as for acetaminophen toxicosis.

Side Effects

All side effects noted for phenobarbital are caused by primidone. Primidone may induce nystagmus, nausea, drowsiness, and ataxia. According to Schwartz-Porsche et al.,¹⁷ polydipsia is more common in dogs treated with primidone. In humans it is recommended that therapeutic plasma concentrations of primidone and its metabolite phenobarbital not exceed 15 and 30 μg/mL, respectively. Megaloblastic anemia is one of the more serious adverse effects of primidone in humans. Pruritis characterized by alopecia, scaling, ulceration, pigmentation, and fissuring has been recorded in dogs.⁸⁶

In the dog primidone induces progressive hepatic injury, as manifested by increases in liver enzyme values.⁸⁹ In a clinical study, signs of liver toxicity were reported in 14 of 20 dogs.¹⁷ Hepatic cirrhosis associated with primidone and phenobarbital after 7 years of use has been reported in a dog.⁹⁰ Primidone-induced dermatitis has been reported in a dog.⁸⁸

Drug Interactions

Drug interactions previously described for phenobarbital also occur with primidone. Primidone should not be used concurrently with chloramphenicol, which is a potent inhibitor of the microsomal enzyme system. Severe CNS depression and inappetence occur in the dog after concurrent use of these drugs.⁹¹

Mechanism of Action

Benzodiazepines enhance the inhibitory effects of GABA in both the brain and spinal cord (see Figure 27-2). Thus they not only decrease seizure spread but also block arousal and centrally depress spinal reflexes (see Chapter 26 for additional discussion). In dogs, tolerance to the anticonvulsant activity of diazepam develops rapidly, within 1 week, and as such, diazepam is not an effective anticonvulsant for chronic therapy in dogs. Intravenous diazepam is, however, the drug of choice for the treatment of SE in both dogs and cats, in part because of its mechanism bt also because it rapidly crosses the blood–brain barrier into the CSF.

Disposition

Diazepam is the prototype benzodiazepine used in small animals. The drug is well absorbed after oral administration but undergoes rapid and extensive hepatic metabolism. Although only 1% to 3% of diazepam is orally bioavailable, 74% to 100% of the drug and all active metabolites are available.⁴² Diazepam is generally administered intravenously but can also be administered intramuscularly, although absorption is not predictable using this route. In human pediatric and canine patients, diazepam has been administered rectally.^95,96

The metabolites of diazepam (nordiazepam [desmethyldiazepam] and oxazepam) are active (see Figure 27-4), although less so (25% to 33%) than the parent compound. 3-Hydroxydiazepam also is a metabolite, although its anticonvulsant activity is not known.⁹⁶ Although less potent than the parent compound, the half-lives of the metabolites are slightly longer than that of diazepam (3.6 hours for desmethyldiazepam and 5.2 hours for oxazepam, compared with 15 minutes for diazepam).^94,96 After oral administration diazepam undergoes rapid first-pass metabolism, with diazepam representing only 1% to 3% of the total (parent and metabolite) area under the curve, compared with 7% to 21% after intravenous administration. Oral bioavailability of all compounds approximates 74% to 100%. After oral administration metabolite concentration surpasses that of the parent compound.⁹⁶ Using a crossover study in both mixed-breed dogs (n = 6) and Beagles (n = 4), intravenous (0.5 mg/kg) or rectal (2 mg/kg to mixed-breed dogs and 0.5 mg/kg to Beagles) administration, and high-performance liquid chromatography to detect either diazepam and its active metabolites, desmethyldiazepam (nordiazepam) and oxazepam, Papich and Alcorn⁹⁵ demonstrated that the sum total of metabolites will surpass the parent compound. Löscher and Frey⁹⁷ reported that the sum bioavailabililty (%) of parent compound and metabolites ranged from 74 to 100 (median 86). Although rectal bioavailaibity of diazepam was only 7.4 ± 5.9% and 2.7 ± 3.2 % for the high and low dose (or mixed breed and Beagles), respectively, bioavailability of diazepam and its two metabolites increased to 79.9 ± 20.7 and 66.0 ± 23.8 for the high dose and low dose, respectively. After intravenous administration, diazepam was characterized by an elimination half-life of approximately of 15 minutes, compared with approximately 2.5 hours disappearance half-life for desmethyldiazepam and 3.8 hours for oxazepam. Peak concentrations of both diazepam and desmethyldiazepam after intravenous administration (0.5 mg/kg) approximated 800 ng/mL; oxazepam peaked at approximately 350 ng/mL. However, whereas diazepam concentrations were below 100 ng/mL within 45 minutes of administration, concentrations were above 300 ng/mL (adjusting for 30% potency of diazepam) for desmethyldiazepam at 8 hours and oxazepam at 1.5 hours. Rectal administration of 2 mg/kg diazepam yielded peak diazepam concentrations of only 75 ng/mL; this compared to peak desmethyldiazepam and oxazepam concentrations of approximately 1600 and 550 ng/mL, respectively. Both desmethyldiazepam and oxazepam concentrations remained above 300 ng/mL at study end (8 hours). As such, rectal administration of diazepam can be expected to yield effective anticonvulsant activity, with concentrations remaining above the minimum recommended for about 8 hours for desmethyldiazepam.⁹⁵

Diazepam has also been studied in dogs (n=6) after intranasal administration (0.5 mg/kg) either as drops or a commercially available atomizer. Bioavailability approximated 40%. Of the two routes, the atomizer was well tolerated. ⁹⁸ Another study diazepam administered IN (0.5 mg/kg; drug measured as benzodiazepines) in dogs (n = 6, crossover design) reported mean peak plasma concentration of benzodiazepine was 448 ± 41 ng/mL at 4.5 minutes, compared with 1316 ± 216 at 3 minutes after intravenous administration. Intranasal bioavailability of benzodiazepine was 80 ± 9%. Plasma drug concentrations exceeded the recommended anticonvulsant therapeutic concentration (300 ng/mL).⁹⁹

The generation of active metabolites complicates the utility of therapeutic monitoring for benzodiazepines as a guide to therapy because anticonvulsant activity is not necessarily correlated with concentration of the parent compound. For diazepam all metabolites and parent drugs should be measured. Methods based on polarized immunofluorescence appear to correlate with total activity relatively well but are likely to underestimate concentrations by 50% or more;^94,95 the amount of underestimation does not appear to be predictable but probably reflects failure to detect oxazepam.

Benzodiazepines enter the CSF rapidly, with peak concentrations usually occurring within 15 minutes of intravenous administration. The more lipophilic drugs enter CSF most rapidly. CSF to plasma drug ratios generally are less than unity, although concentrations may stay higher in CSF compared to plasma, prolonging the duration nof effect.¹⁰⁰ Both diazepam and desmethyldiazepam rapidly pass into the CSF.^96,101

Diazepam and nordiazepam have been studied in cats after intravenous administration of 5, 10, and 20 mg/kg of diazepam and 5 and 10 mg/kg of nordiazepam. Elimination of both drugs was linear over the range of doses covered. Total body clearance of diazepam (4.72 ± 2.45 mL/min/kg) was sixfold greater than that of nordiazepam (0.85 ± 0.25 mL/min/kg). Approximately 50% of an administered dose of diazepam was biotransformed to nordiazepam in the cat.¹⁰²

Clorazepate is metabolized in the stomach to its active metabolite, nordiazepam (desmethyl diazepam), which is also a major, although less efficacious, metabolite of diazepam. After oral administration of 2 mg/kg, clorazepate reaches peak benzodiazepine concentration of 446 to 1542 ng/mL in dogs; mean residence time was 8.5 hours. After multiple administration, mean residence time was significantly longer (approximately 12 hours).¹⁰³ Scherkl and coworkers¹⁰¹ described the consequences of long-term use of clorazepate in dogs. After 2 hours of clorazepate infusion, only nordiazepam was detected in CSF. The elimination half-life of nordiazepam ranged from 7.2 to 12 hours and did not change with 6 weeks of oral therapy. Only nordiazepam and oxazepam were in plasma. Oral doses of 2 mg/kg were associated with ataxia and difficulty standing; these signs resolved 3 to 4 days after treatment. Concentrations of 0.180 to 0.780 μg/mL were associated with decreased seizure threshold. Tolerance was demonstrated in two of six dogs, with some, but not total, loss of anticonvulsant activity. Withdrawal also was demonstrated by manifestation of generalized tonic-clonic seizures in two of six dogs rapidly withdrawn from therapy; death occurred in one of the dogs. The remaining four dogs exhibited no signs of withdrawal.

Clorazepate anecdotally is an alternative anticonvulsant in cats (3.75 to 7.5 mg/cat orally every 8 to 12 hours). However, because it is an active metabolite of diazepam in other species, its role in diazepam-induced hepatotoxicity cannot be ruled out and caution is probably indicated with its use in cats.

Lorazepam is a benzodiazepine derivative used as an anxiolytic in human medicine. It has been studied after intravenous and rectal administration in dogs,¹⁰⁴ although its role in the management of seizures has not been identified. After IV administration of 0.2 mg/kg IV, concentrations remained above 190 ng/mL for 20 min in two of three dogs, and above 30 ng/mL for 90 min in all dogs. However, lorazepam was not detectable after rectal administration of 1 mg/kg, presumably due to high first pass metabolism. Lorazepam also has been studied after intranasal administration. Lorezpam is absorbed almost 3 times as rapidly after intranasal administration compared with oral administration, yielding a relative bioavailability that approximates 2.5 times that after oral administration.¹⁰⁵ Mariami reported in abstract form that the IN administration of 0.2 mg/kg in dogs yielded peak plasma lorazepam concentrations of 106 ng/mL compared with 165 ng/mL following intravenous administration. Concentrations of lorazepam achieved 30 ng/mL (effective anticonvulsant target in humans) by 9 minutes after IN administration in six of six dogs and remained at or above the target for 60 minutes in three of six dogs. Midazolam also has been studied after IN administration with IN aadminsitration exceeding oral administration by 2.5 fold.¹⁰⁶

Clonazepam (Klonopin) is a benzodiazepine derivative and chemically is 5–(o–chlorophenyl)-1,3-dihydro-7-nitro-2H-1,4-benzodiazepine-2-one (see Figure 27-4). It is more potent than diazepam and is used only in the emergency treatment of SE in the dog.⁴² Clonazepam is given orally or intravenously (0.05-0.2 mg/kg; the intravenous preparation is not available in the United States). Accumulation occurs with continued administration. Tolerance develops within days to weeks after administration, however, as a result of hepatic enzyme induction. Consequently, clonazepam, like diazepam, is unsatisfactory for long-term control of epilepsy in dogs. However, it might be considered for long-term use in cats (0.016 mg/kg/day; target concentration 70 ng/mL). The relationship between clonazepam and liver disease has not been addressed in cats. The disposition of clonazepam has been described in the dog (see Table 27-4).^107,108

Preparations

Diazepam is available as both an intravenous and an oral preparation; clorazepate is available as an oral preparation and a sustained-release preparation.¹⁰⁹ It is classified as a Schedule IV drug under the 1970 Controlled Substances Act. Clonazepam is available as an intravenous preparation (not in the United States) and as oral tablets.

Safety

Sedation is the most common direct side effect of the benzodiazepines. Adverse effects (sedation, ataxia, increased appetite, and in some cases hyperactivity) are likely to occur if concentrations reach 500 ng/mL. An 8-hour rather than a 12-hour dosing interval may be indicated to avoid both toxic and side effects. Care must be taken not to discontinue benzodiazepines abruptly because of the potential for SE.¹¹⁰

Diazepam may cause hepatotoxicity in cats.¹¹¹ Reports have focused on cats receiving diazepam as an appetite stimulant rather than as an anticonvulsant. Manifestations include vomiting, depression, jaundice, lethargy, and acute death. Clinical laboratory tests associated with toxicity include increased serum alanine transaminase, aspartate transferase, and alkaline phosphatase activities and increased bilirubin. Toxicity does not appear to be associated with the dose or duration. Toxicity has not been experimentally induced, suggesting that the reaction is idiosyncratic (i.e., nonpredictable).

Physical dependency toward benzodiapenes has been described in dogs, with the risk and severity of signs of withdrawal dependent on dose and drug. Studies that demonstrate dependency often are based on precipitation of clinical signs after administration of a benzodiazepine antagonist; the severity of withdrawal also appears to be dependent on the dose of flumazenil. Both diazepam and lorazepam can be associated with withdrawal in dogs, although only high doses (60 and 1000 mg/kg per day, respectively, for diazaepam and lorazepam) have been studied. Clinical signs included, but were not limited to, tremor, “hot foot walking,” rigidity, and decreased food intake. Lorazepam withdrawal appears to be much less intense but has a shorter latency to onset than the diazepam in manifestation of abstinence syndrome. For diazepam, withdrawal signs include bi-phasic clonic and tonic–clonic convulsions, appearing at 24 and 48 hours. Diazapam syndrome was lethal in two dogs studied experimentally. Administration of diazepam can reduce but not obliterate the major signs of the diazepam withdrawal. The syndrome was worsened by the administration of a benzodiazepine antagonist, although tonic–clonic seizures were not precipitated¹¹² A follow-up study described withdrawal after oral administration of increasing doses of diazepam administered every 8 hours for 5 to 6 weeks: 0.05625, 0.225, 0.5625, 4.5, 9, or 36 mg/kg daily every 8 hours. Abstinence was precipitated by administration of a benzodiazepine antagonist, flumazenil, at 0.66, 2, 6, 18, 36, and 72 mg/kg, but included a placebo. Withdrawal intensity increased proportionately with the dose of diazepam and the dose of flumazenil. The pattern of withdrawal signs varied with dogs receiving high doses of diazepam more sensitive to flumazenil than those receiving lower doses. Seizure activity occurred only in dogs receiving 9 and 36 mg/kg/day of diazepam. Withdrawal signs increased linearly with plasma and brain concentrations of diazepam, oxazepam, and nordiazepam but not with free concentrations of diazepam alone.¹¹³

Although desmethyldiazepam (nordiazepam) has been theorized to be the metabolite causing physical dependence in dogs receiving diazepam (in part because the area under the curve for nordiazepam is much greater than that for diazepam),¹¹⁴ follow-up studies using benzodiazepines whose area under the curve is not surpassed by nordiazepam suggest otherwise.^115,116

The ability of flumazenil to induce clinical signs of withdrawal, including clonic seizures, in dogs, is well described.¹¹⁴ Flumazenil precipitated withdrawal scores were higher in dogs receiving diazepam (and halazepam, which is converted to nordiazepam) compared with nordiazepam, although clonic seizures were greater in the diazepam- and nordiazepam-dependent dogs, compared with halazepam-dependent dogs. The magnitude of withdrawal does not appear to reflect a predominance of one parent compound or metabolite but rather a combined effect.¹¹⁵ Oral flunitrazepam at 7.6 mg/kg per day caused more severe clinical signs of precipitated withdrawal, compared with diazepam (at 6 to 9 mg/kg every 6 hours) administered in four equally divided doses. However, precipitated signs persisted longer with diazepam than flunitrazepam, leading investigators to conclude that diazepam and flunitrazepam were equivalent in their ability to produce induced signs of withdrawal. Difference in kinetics, including protein-binding, among drugs and metabolites, as well as receptor interactions are likely to play a role in differences in response. For example, the estimated free plasma concentration of flunitrazepam and its metabolites was equal to or greater than that of diazepam and its metabolites; further, affinity of flunitrazepam for the benzodiazepine receptor is greater than that of either diazepam, nordiazepam, or oxazepam.¹¹⁶ The disposition of flumazenil in dogs has been described after oral dosing. Rapid absorption is followed by rapid elimination with concentrations being essentially nondetectable in 4 hours.¹¹⁷ Interestingly, benzodiazepine withdrawal induced by flumazenil was described in one study to occur particularly as plasma concentrations markedly decreased.¹¹⁷

Drug Interactions

Clinically important drug interactions resulting from chronic diazepam therapy have not been reported. Despite reports to the contrary,¹¹⁸ interactions between clorazepate and phenobarbital appear to confound therapy. In our studies clorazepate consistently increases phenobarbital concentrations in patients that have been receiving long-term phenobarbital therapy if doses of clorazepate exceed 1 mg/kg every 8 to 12 hours. The increases are usually evident by the first month of therapy but may take longer. It may be necessary to decrease phenobarbital doses. Yet clorazepate concentrations tend to decrease across time despite no change in dose, presumably because of the inductive effects of phenobarbital. As clorazepate concentrations decrease, phenobarbital concentrations may also decrease, which may put an epileptic at risk, particularly if the dose of phenobarbital has been reduced.

Therapeutic Use

As in humans, efficacy, including rapidity of onset, coupled with lack of toxicity, leads to intravenous diazepam as the first drug of choice for control of SE in humans.⁴⁵ Likewise, diazepam is the drug of choice for emergency treatment of SE in dogs and cats.⁴² However, because of its short half-life, diazepam is likely to have to be repeated once or twice during the first 2 hours of stabilization.⁴⁹ Not surprisingly, various methods of administration have been recommended in an attempt to maintain effective diazepam concentrations necessary to maintain efficacy. Intravenous doses vary, beginning at 5 to 20 mg or 0.5 to 1 mg/kg.⁴² The dose can be repeated in 1 to 2 minutes if necessary; giving the same dose intramuscularly may provide longer therapeutic concentrations. If a response has not occurred after the second dose of the drug, an intravenous infusion can be implemented (note that the infusion line should first be flushed with the diazepam solution to allow diazepam binding to the polyvinyl). Alternative anticonvulsants might be considered in animals failing to respond (discussed later). Patients that respond to the first or second dosages of diazepam are carefully monitored, and if SE returns within 2 to 4 hours after the initial treatment, the diazepam regimen is repeated. Intravenous diazepam may be replaced with other benzodiazepines, such as clonazepam (0.05 to 0.2 mg/kg), toward which tolerance develops more slowly. For cats an intravenous dose (5 to 10 mg) generally is given to effect. The dose may need to be repeated; up to 20 mg may be necessary. High doses should be injected slowly. Diazepam (0.5 mg/kg using 5 mg/mL solution) has been used successfully by rectal administration for home control of cluster seizures.¹¹⁹

For chronic therapy rapid metabolism coupled with rapid development of tolerance precludes long-term treatment with diazepam and several other benzodiazepines in dogs.²⁹ However, compared with diazepam, tolerance does not appear to develop as readily to the anticonvulsant effects of clorazepate in dogs. Clorazepate has been studied when added to phenobarbital in dogs still seizuring despite phenobarbital concentrations that exceeded 32 μg/ml (author, unpublished data). Seizures were eradicated in about 35% of animals and reduced 50% or more in another 15%. However, its use in combination with phenobarbital was complicated by drug interactions (including marked fluctuations in nordiazepam disappearance half-life), and an apparent increased risk of hepatotoxicity. Drug interactions between phenobarbital and clorazepate may necessitate dose modification in response to changes in drug concentrations. The short half-life (which might be documented with sequential peak and trough plasma drug concentration measurements) may increase the risk of seizures if drug concentrations decrease below effective concentrations in the patient. The therapeutic range of benzodiazepines (including metabolites) in dogs has been extrapolated from human studies and does not reflect combination therapy. For efficacy, trough concentrations should be maintained at least above 100 ng/mL, although monitoring should establish each patient’s range. Dogs receiving phenobarbital are likely to become groggy when peak clorazepate concentrations exceed 300 to 500 ng/mL (varying with the patient).

For chronic control in cats, diazepam can be given orally in doses of 2 to 5 mg 2 or 3 times daily; diazepam may be increased or decreased in increments of 2 mg with or without phenobarbital. Whereas distribution of diazepam in the cat brain reflects blood flow, nordiazepam accumulates to high concentrations throughout all brain matter, which may be the reason for its efficacy with long-term use.¹²⁰

Mechanism of Action

Bromides mechanism of action is not completely understood.¹²¹ The original proposed mechanism—replacement of negatively charged chloride with bromide in the neuron—is unlikely because bromide is less negatively charged than chloride.¹²² In vitro, bromide enhances GABA-activated currents, leading the authors to conclude that bromide potentiates inhibitory postsynaptic potentials of GABA. Regardless of the mechanism, the anticonvulsant effects of bromide correlate with plasma concentration.¹²³

Disposition

The pharmacokinetics of bromide have been described using either the sodium or potassium salt. In a small study, the half-life in dogs was reported as 21 to 24 days after oral administration, suggesting steady-state concentrations being achieved only at 2.5 to 3 months.¹²⁴ Distribution is to extracellular fluid,¹²⁵ yet sufficient quantities penetrate the CNS. Bromide is eliminated slowly (presumably as a result of marked reabsorption) in the kidney. Bromide appears to compete with chloride for renal elimination, with elimination appearing to decrease proportionately to the amount of chloride in the diet.^126,127

Bromide (20 mg/kg bromide; approximately 26 mg/kg sodium bromide) has been studied in Beagles as the sodium salt after intravenous (n = 4) and oral (n = 4) administration, but only at the daily maintenance dose (which is about to the loading dose).¹²⁸ Dogs received a diet containing 0.4% chloride on a dry-matter basis. Elimination half-life after intravenous administration was 39 ± 10 days and after oral administration, 46 ± 9 days; clearance (reported in ml/kg/day) was 9 ± 3.9, and apparent volume of distribution (Vd area) was 0.45 ± 0.07 L/kg. Oral bioavailability (calculated from mean area under the curve from intravenous and oral groups) was 46%. A study by the same investigators¹²⁷ in Beagles measured changes in bromide (sodium; 14 mg/kg orally) disposition in response to dietary chloride content (0.2, 0.4, and 1.3%, dry-matter basis). Mean apparent elimination half-life decreased from 69 ± 22 days (0.2%) to 24 ± 7 days (1.3%). Predicted maximum drug concentrations were quite variable within treatment groups but were profoundly influenced by dietary chloride content, ranging from 1.95 ± 1.1 mg/mL at 0.2% chloride, 1.27 ± 0.36 mg/mL at 0.4% chloride, to a low of 0.3 ± 0.15 mg/mL at 1.3% dietary chloride.

A study of potassium bromide (20 mg/kg per day bromide or approximately 30 mg/kg potassium bromide), mixed as a 20 mg/mL solution in canned dog food, in Beagles (n = 6) receiving 0.55 to 0.72% dietary chloride reported a median bromide concentration at 115 days of 2.45 mg/mL (low of 1.8 and high of 2.7 mg/mL); the time to steady state was not provided, but concentrations were within 90% of steady-state concentrations by 60 days.¹²⁹ The median elimination half-life, estimated from time to reach steady state, was 15 days (low of 12, high of 20 days). Renal clearance (based on 24-hour urine collection) ranged from 6 to 12 .6 mL/kg per day (median 8.2). The median ratio of bromide in CSF to serum at day 9 was 0.63; this improved to 0.86 (0.82 to 0.99) by day 120. An increase in dose sufficient to generate serum bromide concentrations 3.4 mg/mL (median; does not provide but designed to target 4 mg/mL) resulted in an increase in the CSF: serum ratio to 0.86, suggesting a greater risk than anticipated of CNS side effects. Whereas neurologic side effects were not described at the dose of 60 mg/kg per day, caudal paresis and ataxia developed in two of six dogs. Electrodiagnostic changes were mild even at the high dose; within individual dogs, latency shifts were mild and either progressively increased over time or appeared with doses designed to target 4 mg/mL.

Bromide was studied (report in progress by the author) after rectal administration in normal hound dogs after a loading dose (600 mg/kg) of sterile potassium bromide solution (concentration 250 mg/mL) over 24 hours, either by 6 intrarectal boluses of 100 mg/kg, each every 4 hours, or constant-rate infusion (CRI) throughout the 24-hour period. The average peak serum bromide concentration after intrarectal loading was 0.91 mg/mL (range: 0.81 to 1.11 mg/mL), compared with 1.10 mg/mL after intravenous loading (range: 0.89 to 1.22 mg/mL). The average peak serum K⁺ concentration during potassium bromide loading was 5.80 mEq/L (range: 5.5 to 6.8 mEq/L) and 6.14 mEq/L (range: 5.6 to 7.7 mEq/L) for intrarectal and intravenous loading, respectively (normal range: 3.5 to 5.5 mEq/L). The mean half-life (t_1-2) of intrarectally administered bromide was 488.8 hours, or 20.4 days. A side effect of both intrarectal and intravenous potassium bromide loading was mild sedation, and of intrarectal loading was transient (24 to 48 hours) diarrhea. Bioavailability (F) of intrarectally administered bromide for the 24-hour loading period was calculated to be 107% for the five dogs. In two dogs overall bioavailability (F) of intrarectally administered bromide was calculated to be 57.7%. Results of this study indicate that potassium bromide is well absorbed rectally and that a 24-hour intrarectal loading protocol is safe to administer in normal dogs.

Bromide has been studied in cats after oral administration of the potassium salt at the lower end of the canine dose. After a dose of 30 mg/kg, maximum serum bromide concentration was 1.1 ± 0.2 mg/mL at 8 weeks. Mean disappearance half-life was 1.6 ± 0.2 weeks. Steady state was achieved at a mean of 5.3 ± 1.1 weeks.^129a However, the use of bromide in cats is discouraged due to safety reasons.^129a

Preparations and Sources

Bromide is available as a potassium or sodium salt. Triple bromide salt preparations (Na, K, and NH₄) also may be available through some pharmacies. The accompanying cation does not appear to alter efficacy, although sodium bromide is more difficult to solubilize in water than is potassium bromide. In addition, because potassium weighs less than sodium, 1 g of sodium bromide contains more bromide (78% bromide, or 780 mg) than does 1 g of potassium bromide (67% bromide, or 670 mg). Thus the amount of sodium bromide used to make a solution should be less (211 mg/mL) than potassium bromide (250 mg/mL) to achieve equivalent amounts of bromide in the solution. At the time of publication, a commercial product of potassium bromide is being marketed as a chewable tablet and solution. The product contains B vitamins as well. This product has not been approved by the Food and Drug Administration. It is being marketed under the “generally recognized as safe” category reserved for food of food ingredients; however, this status does not apply to drugs It is important to recognize that the product has not undergone premarket evaluation by any regulatory agency for safety, efficacy, or quality and although its use is not necessarily discouraged, clinicians should be aware that this is a manufactured-compounded product.

Potassium bromide can be compounded by veterinarians or pharmacists on an individual-need basis. It can be purchased through chemical companies (request medicinal or ACS grade [Curtin Matheson]). Chemical companies have refused to sell bromide for medicinal purposes without an investigational new animal drug application (INADA). This application is no longer necessary, however, because the Food and Drug Administration (Division of Drug Compliance) will grant regulatory discretion.

Bromide can be compounded in a syrup or water solution or administered in a gelatin capsule. A number of compounding pharmacies now offer this service, although the compounded solution can be much more costly when compounded by a pharmacist compared with the cost of the ingredients. The solution can be compounded by the clinician with minimal equipment, particularly if monitoring is used to document maintenance of drug concentrations with compounded preparations. If purchased (1-kg bottle) from a chemical company, potassium bromide can be weighed and divided into 250-g (four equal) packets. A liter bottle of distilled water can be purchased from any grocery store. Before the bromide is mixed, a line should be drawn at the 1-L volume mark and approximately 50% of the water removed and set aside. One of the four 250-g packets of bromide is added to the bottle and the bottle shaken well to dissolve the bromide (this may require aggressive shaking and the addition of more water). Once the bromide is dissolved, the volume should be returned to 1 L with either water or a flavored syrup. The final solution will approximate 250 mg/mL. The solution should be stable for at least 6 months, and refrigeration should minimize microbial growth in the solution. Refrigerating the solution may cause the salt to crystallize; warming the solution should cause the drug to redissolve. Note that while this solution can be dispensed from a practice, it cannot legally be sold to others.

Side Effects

Adverse reactions to bromide tend to be dose dependent. They appear to be related to the anticonvulsant or other centrally acting effects of the drug and predominantly affect the CNS (e.g., ataxia, grogginess). This is supported by the recent retrospective report of Rossmeisl and Inzana,¹³⁰ which described signs of “bromism” in epileptic dogs (n = 31). Clinical signs included ataxia and stupor, with one dog being comatose. Other clinical signs included bilateral mydriasis; head pressing was evident in two dogs. Mean bromide concentration at time of admission was 3.7 ± 0.3 mg/mL compared with the most recent monitoring result before admission (1.9 ± 0.3 mg/mL) and compared with a control group of epileptic dogs without bromism at 1.7 ± 0.1 mg/mL. Of the 31 dogs, 26 dogs also were receiving phenobarbital at a mean concentration of 31 ± 12 μg/mL; control group phenobarbital was 26 ± 1 μg/mL. Causes for the increase in bromide were attributed to reduction in renal function, dosing error, shifting from potassium to sodium bromide, and exposure to bromide in drinking water or alternative medications. Changes in diets, the most common reason for increased bromide in the author’s therapeutic drug monitoring laboratory, was not addressed. Diuresis was implemented in affected dogs (n = 27); four of these dogs had breakthrough seizures. Furosemide therapy was initiated in five dogs; however, it is not clear what impact furosemide has on bromide clearance.

Up to 3 months may be required for accommodation to the sedative effects of bromide. This reflects in part the time necessary to reach steady-state concentrations. Gradual reduction of phenobarbital in 25% increments may resolve some of the side effects and may be preferred to saline diuresis in those patients for which nonthreatening clinical signs of bromism emerge. Alternatively, the bromide dose can be decreased by 25%, although 1 to 2 weeks may elapse before a response is seen. Sodium bromide administered as a loading dose sufficient to achieve steady-state concentrations has been reported to be well tolerated in dogs.¹³¹

Fluids containing sodium chloride can be used to treat acute bromide toxicity; monitoring should occur immediately after saline treatment to establish a new baseline. Note that diuresis may increase the risk of breakthrough seizures and SE. Diuresis should be reserved for animals with profound sedation for which a decrease in phenobarbital (if on combination therapy) or bromide will not provide a sufficiently rapid response. Pruritic skin lesions may occur, particularly in patients that had pruritic disorders before starting therapy. A short period of glucocorticoid therapy may control pruritis. Hyperactivity is an occasional side effect and may or may not be dose dependent. Like other anticonvulsants, bromide tends to increase the appetite of dogs. Vomition is not uncommon and appears to reflect the hypertonicity of the salt and direct gastric irritation. Solutions appear to be better tolerated than capsules, although this may vary. Dividing the daily dose into smaller, more frequent doses or feeding before or with medication may decrease gastrointestinal side effects. Sodium bromide may be more tolerated than other bromide salts.

Pancreatitis has been associated with the use of bromide when combined with phenobarbital. A retrospective report found 10% of dogs receiving a combination of potassium bromide and phenobarbital developed clinical signs consistent with pancreatitis, compared with 0.3% of dogs receiving phenobarbital alone.⁸⁰ Serum pancreatic lipase immunoreactivity was found to be increased in the serum of dogs receiving either bromide or phenobarbital, alone or in combination, indicating that the risk may occur with either drug. ⁸¹

Although cats tolerated 8 weeks of oral potassium bromide dosing (experimentally) at 30 mg/kg, clinical patients do not appear to tolerate the drug as well. However, in a retrospective study, bromide (n = 4) or bromide and phenobarbital (n = 3) was associated with eradication of seizures in 7 of 15 cats (bromide dosage range, 1 to 1.6 mg/kg [0.45 to 0.73 mg/lb]); however, bromide administration was associated with adverse effects in 8 of 16 cats. Coughing developed in six of these cats, leading to euthanasia in one cat and discontinuation of bromide administration in two cats. Although somewhat effective in seizure control, the incidence of adverse effects may not warrant routine use of bromide for control of seizures in cats.

Because it is renally eliminated, bromide does not negatively interact with other anticonvulsants metabolized by the liver. However, dietary (or therapeutic) chloride will compete with bromide for renal excretion and can shorten bromide half-life, probably because the kidneys will selectively resorb chloride.¹²⁷ In addition, laboratory assays may not be able to distinguish among anions; for some, bromide may artificially increase serum chloride measurements.¹³²

Therapeutic Use

Bromide has been used to treat seizures in dogs for over 15 years. Initially, it was used as an adjunct to phenobarbital therapy, an indication that persists.^133,134 However, it is increasingly being used as sole therapy.²⁸ Therapeutic ranges vary, with initial studies indicating 0.7 to 2 mg/mL.¹³⁴ More recently, dogs (n = 122; from 1992-1996) with major motor epilepsy were studied retrospectively to determine the therapeutic range of bromide, either alone or in combination with phenobarbital or primidone. Bromide successfully controlled seizures (≥50% reduction in seizure frequency) in 72% of animals. Phenobarbital or primidone could be discontinued in 19% of animals receiving the combination. For dogs that continued to receive both drugs, 45% remained controlled with a phenobarbital dose below 20 μg/mL. The concentration of bromide necessary to control seizures when given as sole therapy was higher (1.9 mg/mL) than when combined with either barbiturate (1.6 mg/mL). The authors concluded that the range for bromide depends on whether bromide will provide sole anticonvulstant support (0.8 to 2.4 mg/mL) or is combined with a barbiturate (0.88 to 3 mg/mL). If phenobarbital could not be discontinued, the range necessary for control was 9 to 36 μg/mL.¹³⁵ A distinction in therapeutic ranges as sole or combination therapy of bromide may not be necessary. Concentrations may fall outside or within a range of 1 to 3 mg/mL (based on the gold chloride method of detection) as needed to control the patient’s seizures, regardless of concomitant therapy. Because the side effects of bromide apparently are not associated with cellular toxicity (unlike phenobarbital), the therapeutic range can be surpassed in animals, if tolerated and if necessary to control seizures. Drug concentrations can be increased beyond this concentration if there is no evidence of toxicity. Some animals will present with side effects at concentrations in the lower end of the range, particularly if a loading dose is administered. Therefore it is generally wise to avoid a loading dose if seizure history and drug safety (e.g., phenobarbital) is supportive. Bromide generally should be given once or twice daily, depending on the animal’s gastric tolerance of the drug. Twice-daily administration of bromide offers no benefit over once-daily administration, with the potential exception of decreased gastric irritation. Indeed, an animal can miss several days of dosing with minimal impact on serum drug concentrations. Missed doses in such instances should be given when convenient. On the other hand, if bromide is better tolerated when broken into much smaller, more frequent doses, no disadvantages are apparent with this approach, other than inconvenience to the owner.

In humans bromide has been used to treat intractable seizures in pediatric patients.^81,136,137 Bromide’s initial introduction into veterinary medicine was as an add-on anticonvulsant, particularly combined with phenobarbital, for control of refractory seizures in dogs or in patients that have developed liver disease in association with phenobarbital therapy. Increasingly, however, bromide is proving efficacious as a first-choice antiepileptic drug for dogs. For example, bromide has been used as the sole drug for patients whose seizure history is limited to mild seizure episodes.¹³³ Because of its long half-life, bromide may be the drug of choice for dogs whose owners are noncompliant because the plasma drug concentrations are not easily manipulated. The author studied bromide as an add-on anticonvulsant in dogs with refractory epilepsy (n = 20) in which seizures continued despite phenobarbital concentations at 32 μg/ml or more. Bromide eradicated seizures in 60% of the patients and reduced seizure frequency by 50% or more in another 20%.

Two choices should be considered when starting bromide therapy: administration of a maintenance dose, which allows the body to gradually adapt to the presence of the drug, or administration of a loading dose. The maintenance dose approach is indicated when possible to minimize the advent of side effects. A targeted concentration should be selected (lower concentrations for milder seizures; higher concentrations may be necessary for severe seizures). In general, 30 mg/kg per day orally will achieve 1 mg/mL at steady state. For every 0.5 mg/mL increment above 1 mg/mL, another 15 mg/kg per day should be administered. To proactively monitor when using the maintenance dose approach to steady state, one sample should be collected at 3 weeks, or halfway to steady state. Doubling the measured concentrations provides an estimate of steady-state concentrations. However, because kinetics vary among animals, a steady-state sample should be collected to confirm baseline concentrations at 3 months.

The second option for beginning bromide therapy is implementation of loading dose. This approach is recommended in patients for which lack of seizure control can be life threatening (including the risk of euthanasia by a distraught owner), or in patients for whom serum phenobarbital concentrations must be rapidly decreased because of hepatotoxicity or bone marrow suppression. Both loading dose and maintenance dose must be established; the loading dose is designed to target steady-state concentrations, but the patient will not be at steady state until being on the same dose for 2 to 3 months. Thus as the loading dose is discontinued and the maintenance dose is begun, a new time to steady state begins. Drug concentrations may increase or decrease compared with the loading dose as steady-state concentrations are reached. Monitoring can ensure that the maintenance dose will maintain what the loading dose achieved. However, this requires a sample collected after loading (within 1 or 2 days) to document what loading achieved, and another sample one elimination half-life after the maintenance dose is started (about 3 weeks). This time is chosen because if concentrations are not maintained by the loading dose, the majority of change (50%) will occur in the first elimination half-life. If the two samples are not within about 10 to 15% of one another (i.e., postload and 3-week maintenance), the maintenance dose should be increased or decreased accordingly. Generally, an increase at the 3-week point is not problematic (and often is desirable), but a decrease is undesirable. If not curtailed by an increase in the maintenance dose, as steady state is reached, the decrease may be sufficient to lead to seizures 2 to 3 months after the loading dose successfully controlled seizures.

A loading dose should reflect all the doses that would have been given as steady state is reached, less drug eliminated by clearance. Thus the loading dose of bromide is substantially higher than the maintenance dose. To target the lower end of the therapeutic range (1 mg/mL), 450 mg/kg should be given as a loading dose. Thereafter, for every 0.5 mg/mL desired increase in bromide concentrations, the loading dose should be increased by approximately 225 mg/kg. Because animals may not tolerate the loading dose as a single oral dose, it generally is divided into 10 equal doses given twice daily for 5 days. However, note that the maintenance dose should be added to the loading dose if the targeted concentration is to be reached. For example, to target 1.5 mg/mL with a 5-day loading period, the patient should receive 675 mg/kg split over a 5-day period, plus a maintenance dose of 45 mg/kg/day. Thus for 6 days, the patient should receive approximately 135 mg/kg. On day 6 the maintenance dose of 45 mg/kg is begun and a postload sample is collected (assume the loading dose achieved 1.3 mg/mL). A second sample is collected 3 weeks later. Allowing for some variability in analytic error, then the maintenance dose can be adjusted in proportion to the difference between the postload and 3 week sample. For example, if the postload was 1.3 mg/ml, and a 3-week concentration is 0.8 mg/mL, the maintenance dose may need to be increased by about 25% to 30%). If a patient already is receiving bromide and an increase in concentrations is desired (e.g., 0.5 mg/mL), the maintenance dose can be increased 15 mg/kg/day. A third sample can be checked at 3 weeks to make sure drug concentrations are moving in the right direction. If a rapid response is desired in a patient already receiving bromide, a “mini” loading dose of 225 to 250 mg/kg total can be given along with an increase in maintenance dose.

The use of bromide as an add-on anticonvulsant to phenobarbital also is discussed later in this chapter. Monitoring in such patients is important if the goal of the additional bromide is to decrease or discontinue phenobarbital. If the goal is to replace phenobarbital, the bromide target should approximate the current phenobarbital concentration in the patient. For example, if the patient has phenobarbital concentrations of 20 μg/mL, (midtherapeutic range), then the bromide dosing regimen might approximate the midtherapeutic range (i.e., about 1.5 to 2 mg/mL). To prevent excessive grogginess, bromide should be initiated as a maintenance rather than loading dose. At least 3 weeks of dosing (i.e., halfway to steady state) should occur, ideally, to allow a halfway to steady-state sample that confirms the bromide dose is on target before phenobarbital is decreased. However, if the patient becomes unacceptably groggy at any point, assuming the seizure history is appropriate, phenobarbital rather than bromide should be decreased by 25% (not only is bromide safer, response will be more rapid for phenobarbital because its half-life is shorter). If all remains well as the new phenobarbital dose reaches steady state (about 2 weeks), assuming the patient has been challenged by a seizure (thus waiting at least one seizure interval is ideal), then an additional 25% decline in phenobarbital can be considered. Ideally, phenobarbital concentrations are measured at each step. This process can be repeated until phenobarbital is eradicated or the patient seizures again. A much more rapid approach will be necessary in animals that develop a life-threatening adversity to phenobarbital (including bone marrow dyscrasias). In such cases a loading dose is indicated, with immediate (6- to 7-day) and 3-week postload samples being critical to ensure maintenance of bromide concentrations in the targeted range. If the loading dose option is taken, phenobarbital probably can be decreased by 25% to 50% as little as 2 to 3 days into the loading dose. Ideally, phenobarbital concentrations should not be decreased completely unless the patient has a bromide concentration of 1.5 mg/mL. For some patients phenobarbital can be completely discontinued once bromide is begun. For others, a combination may be necessary to control seizures. Unfortunately, for a small percentage of animals, seizures may continue even if both drugs are at the maximum end of the therapeutic range or side effects are untenable.

For life-threatening seizures that have not responded to phenobarbital, bromide also can be given rectally as a loading dose (see the discussion of disposition). Efficacy and safety of bromide versus phenobarbital has been compared in spontaneously epileptic dogs (n=46)²⁸ using a parallel, randomized double-blinded study design. Enrollment was based on seizure history, physical and neurologic examinations, and clinical pathology. Dogs were loaded over a 7-day period to achieve the minimum end of the therapeutic range of the assigned drug phenobarbital (3.5 mg/kg), or bromide (15 mg/kg) was administered every 12 hours. Data (clinical pathology and drug concentrations) were measured at baseline and at 30-day intervals for 6 months. All but three patients completed the study. Seizures initially worsened in three dogs on bromide but not in any phenobarbital-treated patient. Compared with baseline, mean seizure number, frequency, and severity were reduced at 6 months for both drugs; seizure duration was shorter for phenobarbital but not bromide. Seizure activity was eradicated in a greater percentage of phenobarbital-treated patients (85%) compared with bromide-treated patients (65%), and the seizure duration decreased more for those on phenobarbital than for those on bromide, but successful control (at least 50% reduction in seizure number with no unacceptable adverse drug reactions) tended (p = 0.06) to be better only for those in the phenobarbital group. For patients in which seizures were eradicated, mean phenobarbital concentration was 25.4 ± 5.4 (95% confidence interval of 23 to 28 μg/mL) (see Table 27-4) but ranged from 12 μg/mL to 34 μg/mL (dose was 4.1 ± 1.1 mg/kg). For bromide mean concentration associated with eradication of seizures was 1.8 ± 0.5 mg/mL at a range of 0.9 to 3.3 mg/mL (dose of 31 ± 11 mg/kg). Both drugs caused abnormal behaviors. Weight increased by 10% in both groups. Changes in clinical pathology were limited to increased (but within normal) serum alkaline phosphatase and decreased (but within normal) serum albumin at 6 months for phenobarbital compared with baseline and compared with bromide at 6 months. Side effects at 1 month were greater for phenobarbital vs bromide and ataxia (55% vs 22%), polyuria (35% vs 13%), but greater for bromide vs phenobarbital for polyphagia (43% vs 30%), vomiting (57% vs 20%), and hyperactivity (43% vs 35%). At 6 months, most clinical signs had resolved to less than 15% of animals in either group with the exception of vomiting that continued in 21% of dogs in the bromide group (0% in the phenobarbital group). One dog failed phenobarbital therapy because of neutropenia; two dogs failed bromide due to vomiting. This study suggests that both phenobarbital and bromide are reasonable first choices for control of epilepsy in dogs, although phenobarbital may provide better control. Side effects can be expected to be greater in bromide following chronic dosing.

Although bromide has been used in the cat successfully, adverse effects (consistent with bronchial asthma) are sufficiently common that an alternative anticonvulsant should be considered.^129a While administration of bromide (n = 4) or bromide and phenobarbital (n = 3) was associated with eradication of seizures in 7 of 15 cats (serum bromide concentration range, 1.0 to 1.6 mg/mL); bromide coughing developed in 6 cats, leading to euthanasia in 1 cat and discontinuation of bromide administration in 2 cats.

Pharmacologic Activity

Phenytoin produces a stabilizing effect on synaptic junctions that ordinarily allow nerve impulses to be readily transmitted at lower thresholds. Consequently, the level of synaptic excitability that permits impulses to be transmitted easily is reduced or stabilized or both. This effect appears to be associated with active extrusion of Na⁺ from neurons and decreased posttetanic potentiation or spread of nerve impulses to adjacent neurons. Phenytoin may reduce calcium movement across cell membranes. Further, phenytoin may inhibit activation of protein phosphorylation by the calcium–calmodulin complex.¹⁸¹ Phosphorylation and norepinephrine release in neurons requires calmodulin. Reduction in spread of the “burst” activity associated with epilepsy prevents genesis of the cortical seizure. Phenytoin stabilizes hyperexcitable neurons without causing general depression of the CNS.⁴⁵

Disposition

Absorption of phenytoin is erratic after intramuscular administration. This may be related to crystallization of the drug at the injection site because of alteration in pH by tissues.⁴⁵ Administration of phenytoin by the intramuscular route is not advised because considerable necrosis and sloughing at the injection site occur.¹⁷⁸ Absorption of the drug from the gastrointestinal tract of the dog is rapid but erratic and incomplete.¹⁷⁹ Bioavailability of phenytoin from the tablet formulation averages 36% in the dog.¹⁸⁰ Poor oral absorption and differences in product bioavailability contribute to the difficulty in achieving effective serum levels of phenytoin. The generic preparations of phenytoin should not be used.¹⁸² Dosing at 30 mg/kg thrice daily for 3 days resulted in concentrations of less than 8.5 μg/mL.¹⁷⁹ Dosing at 10 mg/kg thrice daily for 5 to 8 days yielded concentrations less than 2.5 μg/mL.

At therapeutic concentrations (10 to 20 μg/mL), phenytoin is highly bound (75% to 85%) to plasma proteins in animals and humans.¹⁸³ The high degree of phenytoin binding predisposes this acidic drug to interaction with other drugs by a displacing effect at protein (albumin) binding sites, although the relevance of this in the presence of increased clearance is not known. In uremic patients decreased plasma protein binding is associated with accelerated renal clearance or elimination of the drug. Phenytoin readily crosses the placenta.¹⁸⁴ High concentrations of phenytoin are attained in the maternal liver and maternal and fetal hearts. The brain (ostensibly the primary target organ) contains nearly the lowest concentration of the drug.

Phenytoin is metabolized via CYP2C9 and to a lesser degree CYP2C19¹⁸⁵ into metahydroxyphenytoin or parahydroxyphenytoin, which are conjugated with glucuronic acid. In humans approximately 60% to 75% of the daily dose of phenytoin is excreted in the glucuronide form;⁴⁵ the dog also converts a high percentage of phenytoin into this form. In addition, diphenylhydantoic acid, a minor metabolite in some laboratory animals, and dihydrodiol are formed in dogs. Interestingly, high concentrations of diphenylhydantoic acid are found in cat urine. The dihydrodiol metabolite is probably involved in formation of catechol metabolites; these are also formed in most animals.¹⁸⁶ Epoxide metabolites are also speculated to be formed in humans. Because phenytoin is not very soluble in water, little of the unmetabolized drug is excreted in urine.

In the dog, despite relatively large single daily oral doses (50 mg/kg), the plasma concentration of the drug is low, perhaps reflecting the short half-life of 6 to 7.8 hours¹⁸⁷ compared with 4 to 6 hours after an intramuscular injection (50 mg/kg) as well. The shorter half-life of the second study probably reflects induction after multiple dosing;¹⁸⁸ the half-life of phenytoin in the dog dramatically decreases after 7 to 9 days of treatment.^179,180. A 2.5-hour half-life was predicted for phenytoin after 14 days of therapy compared with 3.3 hours with single dosing. Other studies have confirmed a short half-life for phenytoin in the dog: After a single intravenous dose (15 mg/kg), a value of 4.5 hours was obtained by Sanders and coworkers,¹⁸⁹ and a half-life of 3.65 hours was determined by Pedersoli and coworkers¹⁹⁰ after an intravenous bolus of 11 mg/kg. The clearance of phenytoin has been described as dose dependent in dogs.¹⁷⁹ More recently, using a parallel design, the disposition of phenytoin was described in Beagles (n=5) after 5 days of intravenuos administration of 12 mg/kg (see Table 27-5).¹⁹¹

After oral administration (10 mg/kg) in the cat, plasma half-life of phenytoin was 24 to 108 hours,^188,192 contributing to the prolonged effect of phenytoin observed in the cat more than in other species. This may reflect decreased glucuronide conjugation.

Clinical Use

In humans clinical therapeutic effects and intoxication are related to the blood concentration of phenytoin. A reduction in the number of seizures occurs when phenytoin blood concentrations exceed 10 μg/mL. Recommended therapeutic doses of phenytoin administered orally every 8 hours for control of seizure disorders in the dog indicate considerable variation: 6.6 to 11 mg/kg,¹⁷⁸ 11 mg/kg,⁴⁹ and 35 mg/kg.¹⁷⁹ However, oral administration of phenytoin (4.4 and 11 mg/kg) every 8 hours fails to reach serum concentrations of 10 μg/mL in the dog. Indeed, serum phenytoin after either single or repeated oral doses of 10 mg/kg does not exceed a concentration of 2 μg/mL.¹⁷⁹ To achieve a serum concentration of approximately 10 μg/mL phenytoin, it appears that an oral dose of at least 35 mg/kg given thrice daily is necessary for the adult dog.^49,179 According to Pedersoli and coworkers,¹⁹⁰ an oral dosage schedule of 20 mg/kg every 8 hours of the phenytoin microcrystalline suspension should be sufficient to reach a serum concentration of 10 μg/mL or higher. This dose, however, will maintain a plasma therapeutic level for only the first 2 or 3 days of treatment.¹⁸⁰ Farnbach⁵⁰ found that only 3 of 77 epileptic dogs receiving phenytoin were controlled; concentrations in all dogs ranged from 0.2 to 13 μg/mL, with a median of 2.3 μg/mL, despite doses ranging from 3 to 129 mg/kg per day (generally divided into twice-daily doses).

Drug Interactions

The combined use of phenytoin and phenobarbital or primidone may lead to increased formation of epoxide metabolites in animals. This could possibly result in cholestatic hepatic injury similar to that reported in three dogs.³⁶ The effect may also reflect induction by phenytoin. Phenytoin must be considered a potent inducer of the hepatic microsomal enzyme system in the dog.¹⁸⁰ Between 7 and 9 days after administration of phenytoin, its half-life may be reduced from 5.5 to 1.3 hours. In contrast, the half-life after oral administration in humans averages 22 hours, with a range of 7 to 42 hours. This contrasts to humans, in which phenytoin has a moderate inducing ability in the human. The differences between the two species may help explain the efficacy of phenytoin for control of epileptic seizures in humans compared to dogs. Further, although the combined use of phenytoin and phenobarbital is considered optimal therapy in humans,⁵⁵ use of both drugs is controversial because both induce hepatic microsomal enzyme activity. This combined induction complicates successful therapy because therapeutic concentrations are difficult to achieve and the increased production of potentially toxic metabolites.¹⁷⁸ Metabolism of a number of chemicals or drugs is enhanced by phenytoin. These include digitoxin, dexamethasone, and cortisol.¹⁹³ In humans the half-life of theophylline decreases by 50% with an increase in clearance of twofold.¹⁹⁴

Phenytoin metabolism in humans is inhibited by other drugs prolonging their effect. Examples include dicumarol, chloramphenicol, phenylbutazone, and the phenothiazines. In vitro inhibition has been demonstrated by diazepam and propoxyphene hydrochloride. In the dog the serum half-life of intravenous phenytoin increased from 3 to 15 hours when administered with chloramphenicol, and signs of phenytoin toxicosis reversed within 24 hours after cessation of chloramphenicol treatment.¹⁹⁵ Phenylbutazone is also known to increase plasma concentration as a result of inhibition.⁴⁵ Despite predominant renal elimination, vigabatrin, a drug similar to gabapentin, deceased phenytoin clearance by 31% and increased phenytoin half-life by 45% in Beagles.¹⁹¹ In contrast, gabapentin had no effect.

Serum phenytoin concentration declines in the presence of folic acid therapy, presumably because the hydroxylase enzyme that metabolizes phenytoin is folate dependent. Phenytoin may prolong the prothrombin time.¹⁹⁶ Blood coagulation defects similar to that induced by vitamin K deficiency can occur in neonates exposed to phenytoin in utero. The coagulation defect can be reversed by treatment with vitamin K.

Blood Concentrations and Associated Toxicity

In humans mild signs of intoxication such as nystagmus develop with blood levels of 20 μg/mL; patients with levels over 40 μg/mL have marked nystagmus and are uncoordinated and lethargic. Blood levels in the dog would probably have to increase a comparable 100% to 400% as in humans before serious signs of intoxication develop.

Hepatitis, jaundice, and death after clinical use of phenytoin have been reported in one dog that initially received primidone (500 mg daily).¹⁹⁷ Toxic hepatopathy and intrahepatic cholestasis associated with phenytoin administration in combination with phenobarbital or primidone (or both) have been reported in three dogs.^36,89 Two distinct forms of hepatotoxicity have been ascribed to reflect the sequelae of anticonvulsant phenytoin in dogs.³⁶ The first (type B reaction) is characterized by clinical signs after extended treatment at lower than recommended doses and may result from an (unpredictable) idiosyncratic reaction. Indications are that histologic changes with this form will progress from chronic hepatitis to cirrhosis. The second form is more frequently characterized by intrahepatic cholestasis and is associated with a poor prognosis. This form of liver disease has been associated with high doses of phenytoin in combination with primidone or phenobarbital and may represent an intrinsic hepatotoxicity. Induction of enzymes is likely to increase formation of toxic metabolites, which may contribute to hepatotoxicity. Hepatotoxicity resulting from phenytoin is more likely if the drug is used in combination therapy with either primidone or phenobarbital. Toxicity may be related to generation of toxic metabolites. Two forms of toxicity appear to occur with phenytoin therapy: a dose-independent chronic hepatitis that may progress to cirrhosis and that appears to be reversible after discontinuation of the drug early in the disease and a dose-dependent intrahepatic cholestasis that is accompanied by a poor prognosis.

Other Side Effects

The CNS side effects of phenytoin in the dog are moderate in part because of its action but also because it is rapidly metabolized.⁴⁹ Transient incoordination and oversedation may occasionally occur after administration of phenytoin. A moderate degree of polyphagia, polydipsia, and polyuria may be seen in animals medicated with this drug. Sialosis, weight loss, and vomiting have been reported after the use of phenytoin in the cat. Inhibition of release of antidiuretic hormone accounts for the polyuria that develops after administration of phenytoin. Insulin secretion also is inhibited.⁴⁵

Displacement of T₄ by highly protein-bound drugs (e.g., phenytoin) from T₄-binding globulin increases T₄ concentrations, induction of hepatic drug-metabolizing enzymes results in increased clearance of both T₄ and T₃, and increased conversion of T₄ to T₃ by peripheral tissues further decreases T₄. The latter mechanism has been postulated as the reason that T₃ concentrations may remain normal despite increased T₃ clearance¹⁹⁸ in patients receiving phenytoin. Clinical signs of hypothyroidism may not be apparent in such cases. Note, however, that both serum T₄ and free T₄ may be decreased in some patients receiving anticonvulsants.

Fosphenytoin is a phosphate ester prodrug of phenytoin, with the latter being released after dephosphorylation by tissue (liver, red blood cells, and others) phosphatases.¹⁹⁹ Enzymatic conversion half-life is about 3 minutes in dogs, with equimolar concentrations of phenytoin emerging with fosphenytoin. Similar antiarrhythmic (antiseizure was not reported) effects occurred for fosphenytoin and phenytoin in dogs. Side effects are similar to phenytoin, although local irritation was less with intramuscular administration of fosphenytoin compared with phenytoin. No information could be found regarding use of fosphenytoin in dogs.