Page 1 - Levetiracetam (Keppra®)

Monitoring Recommendations Clinical Pharmacology Laboratory
Auburn University
(Updated May 2012)

General Comments:

Why a population “therapeutic range” may be irrelevant to your patient: The therapeutic range of anticonvulsants is a
population statistic and as such, should not be used as an indication of therapeutic failure if your patient has not responded despite
concentrations being in a population “therapeutic range”. Rather, therapeutic drug monitoring (TDM) should be used to establish
your patient’s therapeutic range. The population therapeutic range can be helpful in that if your patient is at the maximum end of
the range and has not responded, a further increase is less likely to be helpful. Also, if your patient has responded at the very low
end of the therapeutic range, a dose increase is not necessarily indicated, but the patient may be more at risk of failure than a
patient whose concentrations are closer to the median of the therapeutic range

For antimicrobials, the therapeutic range depends on the target organism ‘sminimum inhibitory concentration. As such, providing
the target organism and its MIC is helpful for recommendations.

Why a “check up” is the least important reason for monitoring: In general, we recommend monitoring in the following
situations: Start-up: to establish your patient’s therapeutic range at baseline (steady-state) in a responding patient. This is critical
information in the patient who has a change in clinical signs (see “Whats’ Up”); Follow-up: to re-establish your patient’s baseline
after a dose adjustment, a diet change (eg, bromide), the addition of a drug (eg, phenobarbital, cyclosporine) and thus the potential
for drug interactions, changes in the manufacturer of a human generic drug (eg, cyclosporine, zonisamide) etc; Check-up: routine
monitoring at pre-set intervals (eg, 3 months, 6 months, yearly) with the interval depending on the drug being tested and the risk
the patient for therapeutic failure (ie, the more at risk and the more damaging the sequelae of failure, the more frequent the interval
should be); and “What’s up”: a change in clinical signs indicating therapeutic failure (seizures, loss of remission of
immunomodulation) or evidence of toxicity in a previously well controlled patient, or in a patient that has failed to respond as
therapy is initiated and further changes in the dose are being scrutinized.

In general, do NOT use serum separator tubes. The silicon gel may bind drug resulting in a falsely lower concentration.

Peak or trough or what difference does it make?
a. For drugs with a long half-life compared to the dosing interval, because little drug is eliminated during a dosing interval,
the time the sample is collected is not important since concentrations do not change much during the dosing interval.
Bromide, phenobarbital and zonisamide are examples, although phenobarbital and zonisamide half-life may get short
enough that timing is important. For this reason, for anticonvulsants, we always recommend a trough (just before the next
dose) sample so you can assess the lowest concentrations that occur during a dosing interval.
b. For drugs with a short half-life compared to the dosing interval, much to almost all drug will be eliminated during the
dosing interval. As such, no single sample can predict drug concentrations at any other time in the dosing interval. The
question then is peak versus trough or both? Both is indicated if a half-life is to be calculated or if it is important to know
how high and low concentrations occur. If only a single sample can be collected, which one depends on the drug. For
anticonvulsants with a short half-life, because we are interested in the lowest concentration that occurs during a dosing
interval and a trough is recommended. For cyclosporine, in humans, a peak concentration has been associated with the
best prediction of overall exposure to drug during a 12 hour dosing interval and ideally, if only a single sample can be
collected, it would be peak. However, in veterinary medicine, the timing of drug concentrations collected as part of
scientific clinical research reports generally was not given and as such, guidance is problematic. If a 24 hr dosing interval
is used for cyclosporine, the half-life may be so short that drug is not longer detectable at 24 hrs. As such, if both a peak
and trough can not be collected, a peak is preferred. Consistency is important: the same time should be used if comparison
across time is desired. If the patient has not responded well, at that time, both a peak and trough should be considered. For
antibiotics, in general both a peak and trough (or second sample 2 2 to 3 half-lives later) is important to design a dosing
regimen.
c. Calculating a half-life: Determining the patient’s half-life for a drug might be important and for such drugs, both a peak
and trough sample should be collected, with the trough always being before the next dose. The exception of a trough
sample is if the half-life is so short that no drug is likely to be present, then the second sample should be collected two to
three half-lives after the first sample (eg, aminoglycoside antimicrobials). A half-life can be calculated based on the slope
of the line resulting from the two points: half-life (hr) = 0.693/kel where kel = (ln[peak/trough])/time lapsed between
peak and trough (ln=natural log).

ANTICONVULSANTS

For information on anticonvulsants not listed below, please call. In general, for drugs with short half-lives, a peak and trough is
recommended.

Bromide (potassium or sodium salt) is characterized by a half-life of approximately 21 days in dogs and 14 days in cats;
variability among animals is likely to be marked. Bromide is influenced by chloride intake (more chloride intake results in more
rapid elimination). As such, during a 12 or 24 hr dosing interval, little drug is eliminated between doses and drug concentrations