48 Reviews / Epilepsy Research 68 (2006) 19–94
Table 3
Proposed stratification of elderly for drug studies
Elderly healthy Studies for registration
Elderly with multiple Determine isoenzyme profile
medical problems
Determine renal clearance profile
Frail elderly Absorption studies
Side effects studies
6. Conclusions
Elderly should be included in clinical trials of AEDs
for efficacy. For registration, the most reasonable route
would be to include a sufficient number of elderly, espe-
cially those in the 85+ years of age who also meet
the entry criteria of younger adults. Those would be in
the sub category of elderly healthy except for epilepsy
(Table 3). However, because drug-drug interactions
may be a major concern, new AEDs should be evalu-
ated for the potential of drug interactions with all of the
categories of drugs that may be used. These, however,
can be studied independent of registration studies. For
drugs primarily cleared by the kidneys, an assessment
of renal clearance should be done.
The frail elderly pose a different and more diffi-
cult population. Specific studies evaluating absorption,
clearance, and side effects should be done. Measure-
ment of AEDs during studies in elderly populations
is crucial. Elderly may have substantial differences in
absorption and clearance. Doses in elderly may need to
be lower to avoid side effects. Also, therapeutic effects
may be observed at lower doses as well.
Acknowledgement
Supported in part by NIH 2P50NS16308.
Reference
Birnbaum, A., Hardie, N.A., Leppik, I.E., Conway, J.M., Bowers,
S.E., Lackner, T., Graves, N.M., 2003. Variability of total pheny-
toin serum concentrations within elderly nursing home residents.
Neurology 60, 555–559.
Brodie, M.J., Overstall, P.W., Giorgi, L., 1999. Multicentre, double-
blind, randomized comparison between lamotrigine and carba-
mazepine in elderly patients with newly diagnosed epilepsy.
Epilepsy Res. 37, 81–87.
Epilepsy Foundation of America, 1999. Epilepsy, a Report to the
Nation. Landover, MD.
Garrard, J., Cloyd, J.C., Gross, C., Hardie, N., Thomas, L.W., Lack-
ner, T.E., Graves, N.M., Leppik, I.E., 2000. Factors associated
with antiepileptic drug use among nursing home elderly. J.
Gerontol. 55, 384–392.
Garrard, J., Harms, S., Hardie, N., Eberly, L.E., Nitz, N., Bland, P.,
Gross, C.R., Leppik, I.E., 2003. Antiepileptic drug use in nursing
home admissions. Ann. Neurol. 54, 75–85.
Greenblatt, D.J., 1979. Reduced serum albumin concentrations in the
elderly: a report from the Boston Collaborative Drug Surveillance
Program. J. Am. Geriatr. Soc. 27, 20–22.
Hauser, W.A., 1997. Epidemiology of seizures in the elderly. In:
Rowan, A.J., Ramsay, RE. (Eds.), Seizures and Epilepsy in the
Elderly. Butterworth-Heinemann, Boston, MA, pp. 7–20.
Hauser, W.A., Hesdorffer, D.C. (Eds.), 1990. Epilepsy: Frequency,
Causes and Consequences. Demos Publications, New York, NY,
pp. 1–51.
Leppik, I.E., Bergey, G.K., Ramsay, R.E., Gidal, B.E., Birnbaum,
A.K., Elliott, M.B., 2004. Advances in antiepileptic treatments. A
rational basis for selecting drugs for older patients with epilepsy.
Geriatrics 59, 14–18.
Leppik, I.E., Cloyd, J.C., Sawchuk, R.J., Pepin, S.M., 1979. Com-
pliance and variability of plasma phenytoin levels in epileptic
patients. Ther. Drug Mon. 1, 475–483.
Rowan, A.J., Ramay, R.E., Collins, J.F., Pryor, F., Boardman, K.D.,
Uthman, B.M., Spitz, M., Frederick, T., Towne, A., Carter, G.S.,
Marks, W., Felicetta, J., Tomyanovich, M.L., The VA Coopera-
tive Study 428 Group, 2005. New onset geriatric epilepsy, A ran-
domized study of gabapentin, lamotrigine, and carbamazepine.
Neurology 64, 1868–1873.
US Bureau of the Census, 2004.
Wynne, H.A., Cope, L.H., Mutch, E., Rawlins, M.D., Woodhouse,
K.W., James, O.F., 1989. The effect of age on liver volume and
apparent liver blood flow in healthy man. Hepatology 9, 297–301.
Woodhouse, K.W., Wynne, H.A., 1988. Hepatic drug metabolism
and aging. Br. Med. Bull. 15, 287–296.
doi:10.1016/j.eplepsyres.2005.09.018
CLINICAL TRIALS: SPECIAL ISSUES
This section addresses a number of very important
issues in AED development. Preclinical animal test-
ing of AEDs, as described by Dr. Rogawski in section
1, is incapable of determining whether a drug will have
an important clinical effect in human epilepsy. There-
fore, there is a great need for early studies that will
confirm such an effect in a short time period, expos-
ing as few patients as possible. Such studies are called
“proof of principle trials”, and are discussed by Dr.
Schmidt. Dr. Privitera discusses an issue that is rarely
addressed: Regulatory approval of new AEDs is gained
after thorough evaluation of safety and efficacy in a
relatively narrow population. After approval there is
 Reviews / Epilepsy Research 68 (2006) 19–94
a need both for further exploration of drug effects in
other populations, as well for comparative trials. But,
who will pay for these trials? He compares the oppor-
tunities and issues related to NIH vs privately funded
studies. Another topic that is addressed in this section
relates to evidence-based guidelines. These guidelines
are becoming more prevalent, and are having a greater
impact on clinical care. Dr. French discusses the guide-
line process, and explores how choice of study design
can impact guideline conclusions. In the final section,
Dr. Pippinger emphasizes how the availability of ther-
apeutic drug monitoring has enhanced the efficacy and
safety of antiepileptic drugs. He suggests that availabil-
ity of monitoring should continue to be explored during
the pre and post-regulatory development process.
Proof of principle studies
Bernd Schmidt
Abstract
Proof of Principle studies are an early stage of clin-
ical drug development when a compound has shown
potential in animal models and early safety testing. This
step of proof-of-principle (PoP) or proof-of-concept
(PoC) often links between Phase-I and dose rang-
ing Phase-II studies. These small-scale studies are
designed to detect a signal that the drug is active on
a pathophysiologically relevant mechanism, as well as
preliminary evidence of efficacy in a clinically rele-
vant endpoint. Sponsors use these studies to estimate
whether their compound might have clinically signifi-
cant efficacy in other diseases states as well as epilepsy
(e.g., migraine, neuropathic pain, anxiety, depression).
Endpoints other than seizure frequency often reveal
special characteristics of the drug. A structured dose
escalation design can reveal dose-dependent effects
and adverse effects, crossover studies can demonstrate
change, presurgical studies can define efficacy, interic-
tal discharges and photosensitivity models can explain
changes in seizure features, transcranial magnetic stim-
ulation evaluates hyperexcitability. PoP studies allow
exploration of a wide range of potential therapeutic
areas beyond epilepsy as part of an integrated CNS
development plan.
49
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2. PoP study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3. Structured dose-escalation . . . . . . . . . . . . . . 50
4. Crossover-studies . . . . . . . . . . . . . . . . . . . . . . 50
5. Presurgical model . . . . . . . . . . . . . . . . . . . . . . 50
6. Examples of PoP models . . . . . . . . . . . . . . . . 51
6.1. Interictal discharges . . . . . . . . . . . . . . 51
6.2. Transcranial magnetic stimulation
(TMS) . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.3. Photosensitive epilepsy . . . . . . . . . . . 51
7. Selection of a PoP design . . . . . . . . . . . . . . . 52
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
1. Introduction
1.1. In recent years, drug development has been
focused on early identification of the viability of can-
didate molecules for full clinical development to a
commercially competitive product. The important step
of proof-of-principle (PoP) or proof-of-concept (PoC)
studies shifts decisions for continuing development to
early, less costly stages of development, thereby reduc-
ing the cost of failures. PoP/PoC studies are defined as
tools to detect a signal that the drug in development is
active on a pathophysiologically relevant mechanism,
as well as preliminary evidence of efficacy in a clini-
cally relevant endpoint. Such studies may include proof
of bioavailability in humans as well as an indication of
tolerability and safety. PoP studies are small, brief, sci-
entifically rather than regulatory driven designs. These
studies often link between phase-I and dose ranging
phase-II studies.
1.2. PoP studies are valuable hurdles that can
add value to a drug to attract development partners,
potential investors, or licensees for molecules under
review by big pharma and biotech companies. With the
increasing importance of PoP as critical for decision
making, requirements for an optimised lead compound
to be tested in a PoP study are upgraded as well. Strong
support of the approached indication has to be avail-
able in a whole range of suitable in-vitro and animal
models, with dose responses investigated. Human effi-
cacious concentrations and pharmacokinetic behaviour
of the compound has to be reasonably predicted, and
50 Reviews / Epilepsy Research 68 (2006) 19–94
metabolic profiling done. Toxicology liabilities and a
potential therapeutic window has been determined pre-
clinically. Single and possibly multiple dose toxicology
studies including maximal tolerated doses have been
determined in healthy volunteers, to support short term
application in a suitable PoP human population. An
often overlooked requirement is that formulations that
achieve efficacious concentrations without too much
variability in plasma and the central nervous system
compartment are available and can be manufactured to
good manufacturing standards.
2. PoP study
In many disease areas, CNS included, surrogate
markers are used for PoP to provide an efficacy signal.
However, evidence-based misunderstandings happen
frequently when biomarkers or surrogates are mixed
up with or are misinterpreted for clinical endpoints.
A surrogate endpoint is expected to predict clinical
benefit (or harm or lack of benefit), based on epidemio-
logic, therapeutic, pathophysiologic or other scientific
evidence. In epilepsy PoP studies, electrophysiologi-
cal parameters are used in some paradigms, although
clinical seizures are the main endpoint.
2.1. There is no general recipe of how to do a PoP
study, or how extensive or limited a PoP exploration has
to be. All this depends on the features known for the
compound so far, and on the target profile intended for
the later drug. For potential antiepileptic drugs (AEDs),
a handful of models have been used. They can be
grouped into those using seizure counts as clinical end-
point and those models with an electrophysiological
outcome measure. For PoP studies evaluating seizure
counts chronic toxicology studies are required to allow
for longer observation periods. When using surrogates
single dose applications may be sufficient to achieve a
result.
3. Structured dose-escalation
In this design, a low number of pharmacoresis-
tant patients with very frequent partial seizures, often
many seizures per day, receive the experimental AED
in increasing doses adjunctive to their stable base-
line regimen. Time of dose increments is dependent
on the pharmacokinetic parameters, but weekly dose
increases until the individual maximum tolerated dose
is achieved are usual. Earlier AED development pro-
grams used a completely open approach, more recent
ones generally have been randomized and blinded, with
some patients on placebo. There is a high likelihood for
the dose-escalation design to be either false negative
or false positive, due to a number of reasons, includ-
ing the atypical patient population, regression-to-the
mean phenomena, natural fluctuation of seizure activ-
ity, clustering, underestimation of pharmacokinetic
and potentially pharmacodynamic drug-drug interac-
tions, missing out on a suitable titration scheme, rel-
atively short term exposure and the low power of
a small scale study, even when a control group is
introduced.
3.1. When structured dose escalation is used within
a randomised, adjunctive, placebo-controlled study
to investigate maximum tolerated dose, pharmacoki-
netic and drug–drug interactions in patients are impor-
tant. When patients respond, the added experimental
AED are slowly withdrawn from one or more of their
baseline drugs (“staircase design”) the value for PoP
becomes a bit better, provided the number of patients
included is adequate and the duration of exposure suf-
ficient for the baseline seizure frequency (Sachdeo et
al., 2001).
4. Crossover-studies
With a number of earlier AED developments,
crossover trials versus placebo have been used to
demonstrate preliminary evidence of efficacy, mostly
for patients with frequent partial seizures as adjunctive
to a baseline treatment. The potential for well know
carry-over effects, short duration of exposure, missing
the optimal dose and titration schedule, as well as sta-
tistical issues with the trial analysis make go-/no-go
decisions based on this type of PoP risky (Loiseau et
al., 1990).
5. Presurgical model
Within the framework of video-EEG monitoring,
in order to identify one or more epilepsy foci and
to explore suitability for surgery in partial epilepsy,
 Reviews / Epilepsy Research 68 (2006) 19–94
in-patients are rapidly withdrawn from their baseline
AEDs. Then, patients are randomised to receive either
the experimental AED or placebo over a period of up
to 10 days, or until predefined exit criteria have been
reached. Difference in time-to-exit and percentage of
seizure free patients at the end of the trial are outcome
measures. This type of design has been criticized for
ethical and medical reasons, particularly when done
with a larger N for the purpose of a trial intended for
approval of a monotherapy indication. However, within
the context of PoP, accepting a lower power by includ-
ing fewer patients, not exceeding the period absolutely
necessary to complete the presurgical workup and rigid
exit criteria, the presurgical design is still worth consid-
ering. It provides a setting for monotherapy application
of the novel AED, while investigating partial seizures
by video-observation, and is short in duration. How-
ever, for the experimental anticonvulsant, ideally no
titration period should be needed and effective levels
achieved rapidly. A positive outcome is indicative of an
early robust effect on withdrawal type partial seizures,
but an effect of a potential AED building up slower
over time may be overlooked. The standard design
has been fine-tuned by using “time until Kth seizure”
as a more sensitive analysis and to shorten the total
trial duration and by doing so limiting patient’s expo-
sure to an ineffective treatment. Another variation has
suggested, using TISA (therapeutic intensive seizure
analysis), would reduce the exposure time to 3 days,
provided therapeutic drug levels, orally or parenterally,
can be safely achieved within that window. (Stefan et
al., 2004; Pledger and Sahlroot, 1993).
6. Examples of PoP models
6.1. Interictal discharges
Patients with frequent inter-ictal EEG activity
receive the experimental AED in a crossover fashion,
either versus baseline recordings only, or also versus
placebo application. Patients with both focal or gener-
alized inter-ictal spike-waves can be included. Eval-
uation is done by comparing spike counts obtained
in matching time intervals of EEG recording between
baseline or control days, versus drug application, done
either as single dose or after chronic dosing. Downsides
of this model include heterogeneity of the patient pop-
51
ulation presenting with frequent inter-ictal activity, the
high and rapid intra-individual and diurnal variability,
dependency of stable observation during recordings,
and the fact that a plasma level that cures spikes may not
translate into effectiveness on seizure counts (Binnie,
2001).
6.2. Transcranial magnetic stimulation (TMS)
TMS has been investigated for PoP purposes within
early AED development not only for therapeutic use in
epilepsy, but also as a safe and easy tool. TMS induces
a state of hyperexitability of cortical interneurons that
can potentially be reversed by acute or chronic activity
of AEDs. The parameters measured after either sin-
gle or paired-pulses are applied over the motor cortex
include minimal threshold for motor response in the
target muscle, motor evoked potential amplitude, the
cortical silent period and the intracortical inhibition.
Nearly all older and newer AEDs have been studied.
A gross pattern discriminating gaba-ergic drugs from
those working on excitatory ion channels has emerged.
The pattern of TMS evoked responses is highly incon-
sistent, and may even within a particular AED be time
and dose dependent. Given the fact that AEDs mostly
have a variety of pharmacological activities, use of
TMS is rather a way to look into potential mechanisms
of influencing cortical excitation, than a predictive tool
useful for PoP of a novel AED, than would translate
into clinical antiseizure activity (Theodore, 2003).
6.3. Photosensitive epilepsy
In about 5% of epileptic patients, mostly associated
with generalized epilepsies, the phenomenon of photo-
sensitivity occurs. It has a strong genetic component, a
higher incidence in females and is commonly found in
patients under the age of 18. Photosensitivity is defined
as a generalized epileptiform reaction on intermittent
photic stimulation (IPS), outlasting the visual stimulus
train, on the EEG. For the purpose of the model, provo-
cation of paroxysmal discharges by IPS, not of seizures,
are important. The outcome measure is the number
of standard visual stimulation frequencies (in Hertz)
that the patient is sensitive to, between their lower and
upper thresholds (photo-paroxysmal response, PPR).
The PPR itself is a dimensionless parameter ranging
from 0–14, zero meaning complete abolition of reac-
52 Reviews / Epilepsy Research 68 (2006) 19–94
tivity on any stimulation frequency. Advantages for use
of the photosensitivity model in PoP include possi-
bility to perform single dose studies without a need
to wait for chronic toxicology and prolonged multi-
ple dose phase-I results, the possibility to correlate a
pharmacodynamic response with the pharmacokinetic
behavior and to get a preliminary indication of effective
plasma/dose ranges. Crossovers with various doses,
versus placebo, as well as a parallel groups and adaptive
designs are feasible.
6.3.1. The test procedure requires experienced and
sophisticated centers, but not many exist around the
world. If done as a multicenter-trial, all participants
have to follow exactly the same standardized IPS tech-
niques and procedures, with a comparable patient pop-
ulation. Preferably a skilled central reader would need
to follow exactly the established EEG criteria for PPR
in order to avoid misinterpretation of the results with
non-specific phenomena.
Disadvantages include difficulties in recruiting suit-
able patients (e.g. institutional review boards are some-
times reluctant to approve trials in patients under 18
years of age, or patients do not like the stimulation
procedure), exclusion of co-medication if a potential
interaction profile is not known, patients with unstable
baseline PPR ranges and when EEGs are finally anal-
ysed confusion of true PPR with non-pathological or
ill defined reactions to IPS, or photomyoclonus.
Despite an ongoing debate how close the model is
to epilepsy and if it constitutes a reliable indicator of
broad-spectrum features of the experimental AED, over
the past three decades the photosensitive paradigm has
been proven useful as a tool for PoP. For a number of
novel AEDs that made a successful entry on the market,
this model has been a reliable early indicator of phar-
macodynamic activity. False positives are not reported,
for obvious reasons negative studies are frequently not
published (Kasteleijn-Nolst Trenite et al., 1996).
7. Selection of a PoP design
Predicting clinical effectiveness is difficult, even
with a broad spectrum AED that has ideal pharmacoki-
netic and metabolic features. Fortunately, therapeutic
intervention at a specific target is often relevant to more
than one disease area. Therefore, PoP studies allow
exploration of a wide range of potential therapeutic
areas beyond epilepsy as part of an integrated CNS
development plan. Alternatively, there are many exam-
ples of a clinical path for initial development as an AED
an early market entry for a CNS-active compound, fol-
lowed by additional indications.
Reference
Binnie, C., 2001. Proof of principle trials: EEG surrogate endpoints.
Epilepsy Res. 45, 7–11.
Kasteleijn-Nolst Trenite, D.G., et al., 1996. Photosensitive epilepsy:
a model to study the effects of antiepileptic drugs. Epilepsy Res.
25, 225–230.
Loiseau, P., et al., 1990. A randomised, double-blind, placebo-
controlled, cross-over, add-on trial of lamotrigine in patients
with treatment resistant partial seizures. Epilepsy Res. 7,
136–145.
Pledger, G.W., Sahlroot, J.T., 1993. Alternative analyses for
antiepileptic drug trials, in New Antiepileptic Drug Develop-
ment. Epilepsy Res. Suppl., 10.
Sachdeo, R.C., et al., 2001. An early determination of drug–drug
interactions in epileptic patients. Epilepsia 42 (Suppl. 7),
289–299.
Stefan, H., Wang, Y., Pauli, E., Schmidt, B., 2004. A new approach
in anti-epileptic drug evaluation. Eur. J. Neurol. 11, 467–473.
Theodore, W.H., 2003. Transcranial magnetic stimulation in
epilepsy. Epilepsy Curr. 3, 191–197.
doi:10.1016/j.eplepsyres.2005.09.019
Large clinical trials in epilepsy: Funding by the NIH
versus pharmaceutical industry
Michael Privitera
Abstract
Different approaches to clinical trials are taken by
government-sponsored (NIH, VA, MRC) and industry-
sponsored clinical trials. Each sector has a different
perspective and funding capacity. Sponsorship often
depends on the question. Initial safety and efficacy
studies for regulatory purposes usually are industry-
driven. Studies of drug use as monotherapy, direct com-
parisons among drugs, and comparisons between cate-
gories of drugs often are more important to prescribers
and payors, thereby requiring government funding. The
challenge to clinician-investigators is to now find ways
to fund comparison trials to determine evidence-based
guidelines so patients receive the best treatment both