Seizure & anticonvulsant FAQ

Mark Troxel, DVM, DACVIM (Neurology)
Massachusetts Veterinary Referral Hospital, Woburn, MA

Management of seizures is an art. There is no “correct way” to treat seizures, and neurologists frequently have differences of opinion regarding anticonvulsants. The goals of this article are to answer many of the frequently asked questions about seizure treatment in dogs and cats and to provide information regarding the more commonly used anticonvulsants.

Please note that some of the information below is my personal view on seizures and their treatment. I have tried to identify where the information below is based on opinion.

What tests should I run?

The minimum database is a CBC, biochemical profile, and urinalysis. Thyroid hormone levels should be determined in patients older than five years of age. There is no proven cause-and-effect relationship between hypothyroidism and seizures, but there are many anecdotal reports suggesting that hypothyroidism causes seizures or at least makes them more difficult to control. Phenobarbital administration can lower both total T4 and free T4 leading to a subsequent increase in TSH, making it difficult to determine if a patient receiving phenobarbital is truly hypothyroid. Therefore, thyroid hormone levels should be obtained in older patients before starting phenobarbital.

Bile acids testing should be performed in all young to middle-age patients to rule out a portosystemic shunt (PSS) and to obtain a baseline before starting phenobarbital. It is important to remember that the CBC/biochemical profile can be normal even in patients with PSS, so a bile acids test should still be performed. I’ve diagnosed many middle-aged patients with PSS that present for seizures or other neurological signs that have normal liver values. Also, remember that small and toy breed dogs frequently have microvascular dysplasia (MVD), which can cause mild bile acid elevations (usually less than 50). As a result, bile acids testing should be performed before starting liver-metabolized anticonvulsants so that any post-treatment bile acids increase can be interpreted in light of pre-treatment levels.

Blood lead level should be obtained if the owners’ house was built before 1978, especially if there’s a reasonable risk of lead exposure (e.g., flaking paint, recent remodeling) or if there are CBC changes consistent with lead exposure (e.g., nucleated red blood cells, basophilic stippling).

If the minimum database is normal, then the next step would be advanced imaging (MRI > CT) and cerebrospinal fluid (CSF) analysis. Advanced imaging is recommended for any patient younger than one year of age or older than five years of age, if seizures are refractory to standard anticonvulsants, or if there are any neurological signs between seizures.

When should I start treatment?

There is no hard and fast rule. Frequent seizures require treatment, but how often is too frequent? Chronic, repetitive seizures cause microscopic changes that make it more likely to have seizures (“kindling phenomenon”). We also need to keep in mind our goal for seizure control. I shoot for no more than one seizure every two to three months. As a result, I typically don’t start anticonvulsants unless there is more than one seizure every three to four months. An anticonvulsant should be started immediately for any unprovoked status epilepticus or cluster seizures, if there is a trend toward increasing frequency, or if there is a known structural cause (e.g., brain tumor) that is causing seizures. Some clinicians advocate starting an anticonvulsant for any forebrain disorder that could cause seizures even if the patient is not currently having seizures. I do not start an anticonvulsant until the patient has had a seizure because of the cost of treatment, adverse effects, and need for monitoring, but it’s not wrong to start preemptive anticonvulsants, especially if there’s a high likelihood for seizures. I counsel all owners that the patient is at risk of seizures and often send them home with a rectal Valium kit to use if the patient starts having seizures and needs emergency treatment before starting a daily anticonvulsant.

Client education is extremely important before starting an anticonvulsant. I often talk to clients about how treatment is a team effort between the veterinarian and family members. It is important for owners to understand that anticonvulsants may be needed for the remainder of their pet’s life and that it is critical to give the anticonvulsants as directed. Clients must also be informed of the cost of treatment and therapeutic monitoring. If they do not buy into the treatment plan, failure is almost certain.

How many seizures are acceptable to be considered controlled?

Again, there is no hard and fast rule. My goal is always to eliminate all seizure activity, but that is uncommon, and not always realistic or practical. Treatment involves achieving a balance between seizure control and side effects of medication(s). For dogs and cats with idiopathic epilepsy (i.e., no underlying structural disease that makes seizure control more difficult), I typically try to reduce seizure frequency to no more than one short seizure every two to three months. A recent study that examined quality of life and lifespan of seizure patients on phenobarbital and/or potassium bromide found that most owners considered one seizure every three months to be adequate seizure control with a good quality of life.1

What is the probability of treatment success?

Approximately 75-80% of dogs with idiopathic epilepsy can be controlled with one or two anticonvulsants. The other 20-25% of dogs are more difficult to control and are on three or more anticonvulsants. These statistics are similar to human medicine. It is imperative that our patients to have a good quality of life, but it’s equally important for owners to have a good quality of life with their pets. As a result, a balance between seizure control and side effects must be achieved with the owner’s help. Ultimately, it’s often the owners that guide me on how aggressive I am in trying to control the seizures. In some patients with very frequent seizure activity or significant side effects, I am happy to get to one seizure per month, as long as the client is happy.

Which anticonvulsant should I use?

The standard first-line anticonvulsants for dogs remains phenobarbital and potassium bromide (KBr). For cats, the first anticonvulsant is phenobarbital. My choice is based on many factors, including the age of the patients, suspected underlying cause of seizures, and length of treatment that will be required. I typically reach for phenobarbital in cases where seizures are frequent and/or severe due to its rapid onset of action. It’s also my first line drug for older patients with a structural brain disease (usually at a lower dose to prevent excessive sedation). I’ve been using KBr as the first anticonvulsant less and less because it’s less effective than phenobarbital for monotherapy according to a recent study,2 because it has a long time delay to reach steady state, and because, in my experience, I see more weakness and ataxia compared to phenobarbital. If I use KBr as a monotherapy, then it’s only if the seizure frequency is less often than one seizure per month. If seizures are more frequent, and I need to use KBr for some reason, then I load the patient before starting the maintenance dose of KBr. Phenobarbital requires monitoring every six months, which significantly increases the annual cost of treatment compared to other anticonvulsants. Also, many owners are aware of the possibility of hepatotoxicity and are reluctant to use the medication, but the incidence of hepatotoxicity is low (less than 2% of cases) and can often be reversed if detected early and treated appropriately.

Some of the newer anticonvulsants are now available as generics, are more cost effective for owners, and have fewer short term side effects (long-term adverse reactions have not been fully elucidated). As a result, many veterinarians are now using zonisamide and levetiracetam (generic version of Keppra) more often as the first anticonvulsant for some patients. However, their efficacy as a single agent has not been fully established and long-term adverse effects are not fully known.

When and how should I monitor anticonvulsants?

When to monitor anticonvulsants depends on the time needed to reach steady-state levels. For single-pass elimination medications, steady state is achieved at about five times the elimination half-life. For phenobarbital, the elimination half-life is approximately 48-72 hours so a phenobarbital level should be checked two weeks after starting the medication or increasing the dose. I perform a CBC at the first two-week phenobarbital level to screen for early blood dyscrasias. A CBC, liver profile, phenobarbital blood level and, ideally, bile acids testing should be performed every six months.

Potassium bromide has a long elimination half-life (dogs about 25 days, cats about 10 days) so steady state levels is reached until 3-4 months in dogs and 6 weeks in cats.

Steady state is reached by two weeks for zonisamide and levetiracetam so a serum level can be checked at that time. That being said, many neurologists don’t check levels very often. We don’t truly know the therapeutic levels for either of these medications so it can be seen as a waste of the clients money if we don’t truly know how to monitor these patients.

Therapeutic monitoring should be performed in the following circumstances:

  • When steady state levels have been reached after starting or changing the dose of an anticonvulsant.
  • Serum anticonvulsant levels can be checked immediately after completing a loading regimen.
  • When there are signs of dose-related toxicity.
  • When seizures are not well controlled, to help determine whether an additional anticonvulsant should be started.
  • Every 6-12 months depending on the anticonvulsant for routine monitoring.

Practical tips:

  • Ideally, a trough level should be obtained, but this is not practical for many clients due to the timing of medication administration vs. appointment availability. A study showed that the time of day a serum phenobarbital level was drawn had little effect on the decision to alter dose in approximately 90% of patients.3
  • Peak and trough anticonvulsant levels may help estimate the half-life and determine whether a medication should be given more often.
  • “Therapeutic levels” are often approximations based on retrospective studies or small pharmacokinetic studies. If the patient is well-controlled, even at a “sub-therapeutic” level, then no change in dose is needed. In other words, treat the patient, not the lab test results.
  • For phenobarbital, I never go above 35 μg/mL because there’s increased risk of hepatotoxicity with chronic blood levels higher than 35.4 I try to get the level to 30-35 μg/mL before adding another anticonvulsant.
  • Fasted samples are preferred – lipemia can interfere with accurate measurement.
    Do not use serum separator tubes as the silicone can bind anticonvulsants leading to an artificially low value.
  • Some neurologists do not perform serum bromide levels. The test can be expensive, and signs of toxicity from too high a blood level (sedation, ataxia, weakness) are easily observed. If a patient on KBr suddenly shows these signs after the initial period of adjustment, then KBr dose should be decreased by about 25%.
  • If there is a question as to whether KBr is causing signs of weakness/ataxia vs. a new problem (e.g. disk disease), the KBr can usually be safely stopped completely for 3-5 days. The serum level usually won’t drop enough to trigger seizures because of its long half-life, but may lead to clinical improvement if the weakness/ataxia is due to KBr.
  • Since zonisamide and levetiracetam levels are expensive and “therapeutic levels” are currently approximations based on human levels or small veterinary pharmacokinetic studies, they often are performed only if there is a concern for toxicity or the patient is poorly controlled.

When should I add a second (or third, or fourth) anticonvulsant?

In general, you should try to maximize the medication(s) the patient is currently receiving before adding another medication. Be sure that the target blood level or medication dose has been maximized first. When adding another medication, make one change at a time. Just because drug A doesn’t seem to be “working” doesn’t mean that you should add drug B and stop drug A. It may be that drug A has helped reduce the frequency/severity, but isn’t fully controlling seizures so you need to add another medication. Make one change at a time, otherwise it’s difficult to know what is working and what isn’t. Reassess the patient to be sure that the correct dose was calculated and prescribed and there is client compliance.

When can I stop anticonvulsants?

In human medicine, patients who are seizure-free for two years often remain so after discontinuation of treatment, especially if the length of time prior to starting anticonvulsant and the duration of time to reach control of seizures is short. To my knowledge, there is only one veterinary study that has evaluated anticonvulsant withdrawal in dogs with idiopathic epilepsy.5 The study enrolled 138 patients and the authors found the following:

“In 11 cases, the therapy had been stopped after the dogs had become seizure free for a median time of 1 year. Reasons for AED withdrawal were appearance or fear of adverse side effects, financial aspects, and the idea that the medication could be unnecessary. Following AED withdrawal, four of these dogs remained seizure free, seven dogs suffered from seizure recurrence, of which only three dogs could regain seizure freedom after resuming AED therapy.”

In general, I won’t consider discontinuation of anticonvulsants for patients with idiopathic epilepsy until the patient is seizure-free for at least two years. If there was a structural or inciting cause of the seizures that has been cured or is in clinical remission, then I will consider starting to taper the medication(s) six months after the last seizure. Anticonvulsants should be tapered slowly, ideally over the span of several months, to prevent rebound seizures that can be worse than the original seizures.


  1. Chang Y, Mellor DJ, Anderson TJ. Idiopathic epilepsy in dogs: owners’ perspectives on management with phenobarbitone and/or potassium bromide. J Sm Anim Pract 2006;47:574-81.
  2. Boothe DM, Dewey C, Carpenter DM. Comparison of phenobarbital with bromide as a first-choice antiepileptic drug for treatment of epilepsy in dogs. J Am Vet Med Assoc 2012;240:1073-83.
  3. Levitski RE, Trepanier LA. Effect of timing of blood collection on serum phenobarbital concentrations in dogs with epilepsy. J Am Vet Med Assoc 2000;217:200-4.
  4. Dayrell-Hart B, Steinberg SA, VanWinkle TJ, et al. Hepatoxicity of phenobarbital in dogs: 18 cases (1985–1989). J Am Vet Med Assoc 1991;199:1060–6.
  5. Gesell FK, Hoppe S, Löscher W, et al. Antiepileptic drug withdrawal in dogs with epilepsy. Front Vet Sci 2015;2:23. doi: 10.3389/fvets.2015.00023