FULL-LENGTH ORIGINAL RESEARCH

A randomized, double-blind comparison of antiepileptic

drug treatment in the elderly with new-onset focal epilepsy
*1Konrad J. Werhahn, †‡Eugen Trinka, †‡Judith Dobesberger, ‡Iris Unterberger, §Petra Baum,
¶Maria Deckert-Schmitz, #Tobias Kniess, **Bettina Schmitz, *Viviane Bernedo, ††Christian
Ruckes, ††Anne Ehrlich, and ‡‡2G€unter Kr€
amer

Epilepsia, **(*):1–10, 2015

doi: 10.1111/epi.12926

SUMMARY
Objective: To compare the effectiveness of controlled-released carbamazepine (CR-
CBZ) to levetiracetam (LEV) and to lamotrigine (LTG) in elderly patients with newly
diagnosed focal epilepsy.
Methods: Randomized, double-blind, parallel-group trial conducted between January
2007 and August 2011, in 47 ambulatory or hospital sites in Germany, Austria, or Swit-
zerland. Eligible participants were aged ≥60, had new-onset epilepsy, had no acute ill-
ness as the cause of their seizures, and had no contraindication to the drugs in the trial.
Patients were randomized 1:1:1 to CR-CBZ, LTG, or LEV. Doses were up-titrated for
6 weeks and could be maintained or adjusted depending on seizure relapse or tolera-
bility over an additional period of 52 weeks. Primary outcome was the retention to
treatment at week 58; secondary measures related to seizure and adverse event fre-
quency.
Results: Of 361 randomized patients, 359 were included (CR-CBZ n = 121, LTG
n = 117, LEV n = 122) in the modified intent-to-treat population (mean age [range]
71.4 [60–95] years). At week 58, the retention rate for LEV was significantly higher
than for CR-CBZ (61.5% vs. 45.8%, p = 0.02), and similar to LTG (55.6%). Seizure free-
Konrad J. Werhahn is dom rates at weeks 30 and 58 were not different across the groups. Twice as many
a professor of patients receiving CR-CBZ discontinued due to adverse events or death compared to
neurology and those in the LEV group (32.2% vs. 17.2%; odds ratio 2.28, 95% confidence interval [CI]
epileptologist working 1.25–4.19, p = 0.007), whereas discontinuation was intermediate for LTG (26.3%).
for UCB Pharma. Median daily doses of completers (n = 195) were CR-CBZ 380.0 mg/day (333.0–384.0),
LTG 95 mg/day (94.0–97.0), and LEV 950 mg/day (940.0–985.0).
Significance: In the initial monotherapy of focal epilepsy in the elderly, 1-year reten-
tion to LEV was higher compared to CR-CBZ due to better tolerability. Retention of
LTG was intermediate and close to LEV, but did not differ significantly from either
comparators. NCT00438451, www.clinicaltrials.gov.
KEY WORDS: Elderly, Monotherapy, Antiepileptic drugs, Levetiracetam, Carbamaz-
epine.

Accepted December 24, 2014.

*Department of Neurology, University Medical Center of the Johannes Gutenberg University, Mainz, Germany; †Department of Neurology,
Paracelsus Medical University, Salzburg, Austria; ‡Department of Neurology, Medical University Innsbruck, Innsbruck, Austria; §Department of
Neurology, University of Leipzig, Leipzig, Germany; ¶Department of Neurology, Harlaching Hospital, Munich, Germany; #Clinic of Neurology, Rh€on-
Klinikum, Bad Neustadt, Germany; **Department of Neurology, Vivantes Humboldt Klinikum, Berlin, Germany; ††Interdisciplinary Center for
Clinical Trials, University Medical Center of the Johannes Gutenberg University, Mainz, Germany; and ‡‡Swiss Epilepsy Center, Zurich, Switzerland
1
UCB Pharma, Brussels, Belgium.
2
Neurozentrum Bellevue, Zurich, Switzerland.
Address correspondence to Prof. Dr. Konrad J. Werhahn, Salierstr. 8, 40545 D€usseldorf, Germany. E-mail: [email protected]
Wiley Periodicals, Inc.
© 2015 International League Against Epilepsy

1
2
K. J. Werhahn et al.

The elderly represent the most rapidly growing age group valproate within 4 weeks of screening to avoid an interac-
within the global population, and this demographic group tion due a reduced hepatic clearance in the lamotrigine arm.
also has the highest incidence of epilepsy.1 Half of all newly Other exclusion criteria were renal insufficiency (glomeru-
occurring epileptic seizures are expected to occur in patients lar filtration rate <50 ml/min), increased liver enzymes or
aged 60 and older in 2020.2 The estimated incidence of trea- bilirubin, dementia, history of drug or alcohol abuse, a psy-
ted epilepsy in the elderly ranges from 110.5 in Iceland to chiatric condition requiring legal guardianship, or life
2,400 per 100,000 for patients ≥65 years using Medicare in expectancy <1 year due to malignant disease.
the United States and varies depending on whether the
patient independent and living in the community or in nurs- Study design
ing homes.3,4 As a group, elderly patients with epilepsy face This was a randomized, double-blind, active compara-
several treatment challenges. They are likely to have comor- tor, multicenter, parallel-group trial conducted over
bid conditions, which increases the potential for drug–drug 58 weeks (www.clinicaltrials.gov: NCT00438451; EudraCT
interactions.5 Furthermore, age-related metabolic changes Number: 2005-003324-19). It comprised a 6-week titration
and reduced drug clearance combined with increased phar- followed by a 52-week maintenance period. Patients were
macodynamic sensitivity further decrease the therapeutic randomized 1:1:1 to controlled-released carbamazepine
window and lower tolerance to treatment.6 (CR-CBZ, Carbamazepine Sandoz Retard 200 mg tablets
Antiepileptic drugs (AEDs) are used by 2% of commu- (Sandoz International GmbH, Holzkirchen, Germany)
nity-dwelling and by 4–11% of the institutionalized became unavailable during the study and were with regu-
elderly.7–11 Carbamazepine and gabapentin are the most latory approval replaced by bioequivalent Tegretol Retard
commonly used AEDs in Europe,10,12 and phenytoin and by Novartis, Basel, Switzerland), lamotrigine (LTG;
gabapentin are the most commonly used in the United Lamotrigine 1A Pharma, 1A Pharma GmbH, Oberhaching,
States.13 At the time the present study was designed, carba- Germany), or levetiracetam (LEV, Keppra; UCB Pharma,
mazepine was a standard choice for elderly patients in the Brussels, Belgium). A randomization list for each center
European Union; however, as cytochrome P450 inducers, (random permuted blocks) was prepared by the Interdisci-
carbamazepine and phenytoin interact with many plinary Centre for Clinical Trials (IZKS), Mainz,
drugs.12,14,15 Of the newer drugs, lamotrigine was included Germany. The pharmacy of the University Hospital Mainz
because of previous conflicting observations.16–18 In random- encapsulated the study drugs and labeled the blinded med-
ized trials, lamotrigine showed superiority in patient retention ication including the randomization number. The
versus immediate-release carbamazepine16,17 but not versus randomization number was assigned to the subjects after
controlled-released carbamazepine.18 Observational data in enrollment according to the numeric order of inclusion.
elderly patients suggested that levetiracetam may be effec- After inclusion, study medication was uptitrated over
tive19; however, a comparative trial has not been conducted. 6 weeks to initial target doses of 400 mg/day CR-CBZ,
Herein, we report the findings of a multicenter, double- 100 mg/day LTG, or 1,000 mg/day LEV, divided in two
blind, randomized trial evaluating controlled-released equal doses/day. This duration of titration was chosen
carbamazepine, lamotrigine, and levetiracetam in the initial mainly to adjust to the LTG arm. Study medication con-
anticonvulsive therapy of elderly patients with newly diag- sisted of capsules each containing 100 mg of CR-CBZ or
nosed focal epilepsy. 25 mg of LTG or 250 mg of LEV. During titration,
patients received one capsule in the evening in the first
14 days, two capsules (one in the morning and one in the
Methods evening) in weeks 3 and 4, three capsules in week 5 (one
Study population in the morning and two in the evening), and four capsules
Based on the United Nations (UN) definition of the from week 6 onward (two in the morning and two in the
elderly,20 patients aged ≥60 years with new-onset focal epi- evening). Doses were selected based on previous studies
lepsy in whom treatment was deemed necessary were in the elderly.16,17,23,24 During maintenance (weeks 7–58),
recruited at 47 sites in Austria, Germany, and Switzerland patients took between 2 and 12 capsules per day, each cap-
(according to the new International League Against sule containing either 100 mg CR-CBZ, 25 mg LTG, or
Epilepsy [ILAE] definition21,22 for epilepsy: two spontane- 250 mg LEV. After reaching target doses, dose adjust-
ous seizures with one occurring in the last 6 months, or one ments were allowed according to tolerability and seizure
seizure plus an epileptiform discharge shown by electroen- control in steps of one capsule per week to doses between
cephalography or an enduring relevant lesion shown by 200 and 1,200 mg for CR-CBZ, 50 and 300 mg for LTG,
imaging. Patients were included if they had no or <4 weeks and 500 and 3,000 mg for LEV to mimic clinical practice.
previous AED treatment. Patients were excluded if they had Patients who did not tolerate a study drug within this dose
symptomatic seizures occurring <2 weeks after the onset of range or whose seizures recurred despite the maximum
acute cerebral insults or had been previously treated with dose had to be withdrawn. Patients continued until week

Epilepsia, **(*):1–10, 2015

doi: 10.1111/epi.12926
4
K. J. Werhahn et al.

ized to CR-CBZ was found not to have epilepsy and was Differences between CR-CBZ and LTG (p = 0.15), or
therefore excluded from the ITT population. LTG and LEV (p = 0.36) were not significant (Fig. 2).
The mean (standard deviation [SD]) age of patients at Logistic regression analysis of the primary outcome indi-
baseline was 71.4 (7.2) years, range 60–95 years; median cated that retention rate at week 58 was influenced by the
(Q1, Q3) age was 71.0 (66.0, 76.0) years. Baseline demo- number of concomitant diseases (odds ratio [OR] 0.92, 95%
graphic and epilepsy characteristics were similar across the CI 0.86–0.99, p = 0.02), but not by country or center. The
treatment groups (Table 1). influence of treatment did not reach statistical significance
(p = 0.06). The odds of remaining on treatment were 83.8%
Primary outcome measure higher for LEV compared with CR-CBZ (OR 1.84, 95% CI
The retention rate at week 58 was 45.8% (95% CI 36.9– 1.1–3.1) and 16.9% higher for LEV versus LTG (OR 1.17,
54.7%) for CR-CBZ, 55.6% (95% CI 46.6–64.6%) for 95% CI 0.69–1.98). The odds of remaining on CR-CBZ
LTG, and 61.5% (95% CI 52.8–70.1%) for LEV. Retention were 36.4% lower than LTG (OR 0.64, 95% CI .38–1.07).
rates were significantly different for all groups (p = 0.048) The primary analysis was confirmed by an analysis of time
and between the CR-CBZ and LEV groups (p = 0.02). to discontinuation.

Figure 1.
Patient disposition. *At the end of the study, all patients were offered further open-label treatment, either with their current drug or an
alternative antiepileptic drug of their choice.
Epilepsia ILAE

Epilepsia, **(*):1–10, 2015

doi: 10.1111/epi.12926
4
K. J. Werhahn et al.

ized to CR-CBZ was found not to have epilepsy and was Differences between CR-CBZ and LTG (p = 0.15), or
therefore excluded from the ITT population. LTG and LEV (p = 0.36) were not significant (Fig. 2).
The mean (standard deviation [SD]) age of patients at Logistic regression analysis of the primary outcome indi-
baseline was 71.4 (7.2) years, range 60–95 years; median cated that retention rate at week 58 was influenced by the
(Q1, Q3) age was 71.0 (66.0, 76.0) years. Baseline demo- number of concomitant diseases (odds ratio [OR] 0.92, 95%
graphic and epilepsy characteristics were similar across the CI 0.86–0.99, p = 0.02), but not by country or center. The
treatment groups (Table 1). influence of treatment did not reach statistical significance
(p = 0.06). The odds of remaining on treatment were 83.8%
Primary outcome measure higher for LEV compared with CR-CBZ (OR 1.84, 95% CI
The retention rate at week 58 was 45.8% (95% CI 36.9– 1.1–3.1) and 16.9% higher for LEV versus LTG (OR 1.17,
54.7%) for CR-CBZ, 55.6% (95% CI 46.6–64.6%) for 95% CI 0.69–1.98). The odds of remaining on CR-CBZ
LTG, and 61.5% (95% CI 52.8–70.1%) for LEV. Retention were 36.4% lower than LTG (OR 0.64, 95% CI .38–1.07).
rates were significantly different for all groups (p = 0.048) The primary analysis was confirmed by an analysis of time
and between the CR-CBZ and LEV groups (p = 0.02). to discontinuation.

Figure 1.
Patient disposition. *At the end of the study, all patients were offered further open-label treatment, either with their current drug or an
alternative antiepileptic drug of their choice.
Epilepsia ILAE

Epilepsia, **(*):1–10, 2015

doi: 10.1111/epi.12926
 5
Comparison of Epilepsy Drugs in Elderly

Table 1. Patient characteristics

Overall (n = 359) Carbamazepine (n = 120) Lamotrigine (n = 117) Levetiracetam (n = 122) p-Valuea
Age, years 71.4  7.2 71.7  6.7 70.7  7.4 71.8  7.5
Median (Q1, Q3) 71.0 (66.0, 76.0) 71.0 (66.5, 77.0) 70.0 (65.0, 76.0) 71.0 (66.0, 76.0) 0.450
Female sex, no. (%) 144 (40.1) 55 (45.8) 48 (41.0) 41 (33.6) 0.148
Body mass index, kg/m2 26.7  3.9 27.0  4.2 26.5  4.0 26.7  3.5 0.584
N seizures at randomization
Median 2.0 2.0 2.0 2.0 0.177
N seizures, no. (%)
1b 151 (42.1) 55 (45
8) 45 (38
5) 51 (41
8) –
2 91 (25.3) 26 (21
7) 37 (31
6) 28 (23
0) –
3–5 71 (19.8) 18 (15
0) 21 (17
9) 32 (26
2) –
>5 38 (10.6) 18 (15
0) 11 (9
4) 9 (7
4) –
Aura, no. (%) 72 (20.1) 28 (23.3) 21 (17.9) 23 (18.9) 0.557
Etiology, no. (%)
Vascular 240 (65.9) 84 (70.0) 74 (61.7) 82 (66.1) –
Neoplasm 14 (3.8) 5 (4.2) 6 (5.0) 3 (2.4)
Inflammatory 4 (1.1) 1 (0.8) 3 (2.5) 0 (0)
Neurodegenerative 4 (1.1) 0 (0) 1 (0.8) 3 (2.4)
Infectious 2 (0.5) 1 (0.8) 1 (0.8) 0 (0)
Metabolic 1 (0.3) 0 (0) 1 (0.8) 0 (0)
Other 34 (9.3) 10 (8.3) 12 (10.0) 12 (9.7)
Unknown 65 (17.9) 19 (15.8) 22 (18.3) 24 (19
4)
Epileptiform EEG, no. (%) 101 (28.6) 37 (30.8) 26 (23.2) 38 (31.4) 0.318
Abnormal MRI, no. (%) 184 (80.3) 63 (82.9) 62 (77.5) 59 (80.8) 0.844
N drugs at study start 5.1  3.0 5.6  3.2 5.3  3.2 4.6  2.7 –
Concomitant diseases 5.8  3.6 5.8  3.5 6.4  3.8 5.3  3.3 –
Modified intent-to-treat population, n = 359.
CCT, cranial computed tomography; EEG, electroencephalography; MRI, magnetic resonance imaging.
Plus–minus values are means  standard deviation.
a
Between all treatment groups.
b
Plus epileptiform EEG or lesion on MRI or CCT imaging.

Figure 2.
Kaplan-Meier plot of time to
discontinuation (intent-to-treat
population). Censored values are
indicated by crosses.
Epilepsia ILAE

Secondary outcome measures 38.5%, LEV 42.6%; p = 0.33). Time to first seizure after
There were no significant differences in seizure freedom randomization (Fig. S1) (p = 0.50) or after titration
rates at week 30 (CR-CBZ 39.2%, LTG 48.7%, LEV (p = 0.78) was similar between treatment arms. In patients
48.4%; p = 0.25) or at week 58 (CR-CBZ 33.3%, LTG who completed the trial (n = 195), the absolute number of

Epilepsia, **(*):1–10, 2015

doi: 10.1111/epi.12926
6
K. J. Werhahn et al.

seizures during the 52-week maintenance phase was not dif- 110/117 (94.0%), and LEV 108/122 (88.5%). Most AEs
ferent between groups (CR-CBZ 131 in 55, LTG 130 in 65, were mild to moderate in severity, with the most frequently
and LEV 168 in 75 patients; p = 0.92) as was the number of reported for all treatments being dizziness, fatigue, and
seizure-free days. Log-rank comparison of the time to first headache (Tables 2 and S3). A full list of AEs reported by
related AE during titration (Fig. 3) did not reveal a signifi- patients in any treatment group by MedDRA system organ
cant difference (p = 0.0617), although it was numerically class is given in Table S4. AEs occurred earlier with CR-
shorter for CR-CBZ compared to LEV and LTG (median CBZ (18.0 days) than with LTG (27.0 days) or LEV
[95% CI] in days: CBZ-CR 18 [15–31], LTG 27 [15–43], (24.0 days) (p = 0.06 between all groups). Overall, 962/
and LEV 24 [15–43]). 2,208 AEs (43.6%) were considered to be drug related (CR-
CBZ 52.7%, LTG 39.2%, LEV 38.6%). Discontinuations
Subgroup analysis due to AEs (Table 3) were higher in the CR-CBZ group (39/
Logistic regression analysis indicated that there was no 121; 32.2%, 95% CI 23.9–40.5), compared with LTG (31/
statistically significant difference in treatment effects 118; 26.3%, 95% CI 18.4–34.2) or LEV (21/122; 17.2%,
within the previous epileptic seizure categories between 95% CI 10.5–23.9, p < 0.01 for CR-CBZ vs. LEV), the
treatment arms (p = 0.51). In three of four subgroups, by most frequent being allergic skin reaction (3.3%), fatigue,
number of seizures at presentation before study inclusion, and dizziness (each 2.5%) for CR-CBZ; fatigue (4.3%),
patients on LEV had higher retention rates than those on dizziness (1.7%), and renal failure (0.9%) for LTG; and
CR-CBZ or LTG. This was most apparent for the subgroup fatigue, renal failure, and epileptic seizures (each 2.5%) for
of patients with one seizure and pathologic electroencepha- LEV (for AEs most frequently leading to discontinuation in
lography (EEG) study or an enduring and relevant lesion ≥2% see Table S5). Serious AEs occurred with similar fre-
shown by imaging (CR-CBZ 32.7%, LTG 51.1%, LEV quency in patients on CR-CBZ (36/121; 29.8%) compared
60.8%, p = 0.013; Table S2) and not explained by the num- to those on LTG (32/117; 27.4%) or LEV (33/122; 27.0%).
ber of concomitant diseases, which was similar between Drug-related serious AEs, however, were more frequent
previous seizure subgroups. with CR-CBZ (15/121; 12.4%) than with LTG (3/117;
2.6%, OR 5.38, 95% CI 1.51–19.10, p = 0.09) or LEV
Safety and tolerability (5/122; 4.1%, OR 3.31, 95% CI 1.16–9.42, p = 0.025).
Median (Q1–Q3) daily exposures (safety set n = 360) to Based on MedDra system organ class, psychiatric AEs
study drugs during the trial were below the initial target occurred in 23.6% of patients in the safety set (CR-CBZ
doses: CR-CBZ 280.0 mg/day (114.0–380.0), LTG 20.7%, LTG 26.5%, and LEV 23.8%). The most common
93.8 mg/day (58.3–95.8), and LEV 937.5 mg/day (412.5– psychiatric AEs (MedDRA Preferred Terms) were depres-
955.0). Median daily doses in patients who completed the sion (5.0%), confusional state (4.1%), and sleep disorder
trial (n = 195) were CR-CBZ 380.0 mg/day (333.0–384.0), (4.1%) in patients on CR-CBZ; sleep disorder (7.7%), anxi-
LTG 95 mg/day (94.0–97.0), and LEV 950 mg/day (940.0– ety (4.3%), and depression (3.4%) for LTG; and agitation
985.0) and were not different from those who were seizure- (4.1%), depression (4.1%), and sleep disorder (3.3%) for
free at visit 6 at week 58 at the end of the trial (n = 137). LEV (see Table S4).
The incidence of patients reporting AEs was similar There were 14 deaths during the study; four in the
between treatment groups: CR-CBZ 108/121 (89.3%), LTG CR-CBZ group and five each in the LTG and LEV groups.

Figure 3.
Kaplan-Meier plot of time to first
related adverse event (intent-to-treat
population). Censored values are
indicated by crosses. V0, visit 0
(inclusion and randomization).
Epilepsia ILAE

Epilepsia, **(*):1–10, 2015

doi: 10.1111/epi.12926
 7
Comparison of Epilepsy Drugs in Elderly

Table 2. Adverse events reported by ≥10% of patients in any treatment group

Overall (n = 360) Carbamazepine (n = 121) Lamotrigine (n = 117) Levetiracetam (n = 122)
Adverse event No. of patients (%)
Nervous system disorders
Dizziness 105 (29.2) 33 (27.3) 36 (30.8) 36 (29.5)
Headache 89 (24.7) 29 (24.0) 29 (24.8) 31 (25.4)
Epilepsy 48 (13.3) 11 (9.1) 15 (12.8) 22 (18.0)a,b
Convulsion 25 (6.9) 8 (6.6) 12 (10.3)d 5 (4.1)
General disorders
Fatigue 100 (27.8) 46 (38.0)a,c 23 (19.7) 31 (25.4)b
Gastrointestinal disorders
Nausea 48 (13.3) 20 (16.5)a 17 (14.5)b 11 (9.0)
Diarrhea 39 (10.8) 16 (13.2)a 14 (12.0) 9 (7.4)
Constipation 23 (6.4) 16 (13.2)a,c 5 (4.3) 2 (1.6)
Infections and infestations
Nasopharyngitis 41 (11.4) 13 (10.7) 8 (6.8) 20 (16.4)a,b
Investigations
c-Glutamyltransferase increased 27 (7.5) 18 (14.9)a,c 5 (4.3) 4 (3.3)
Safety population, n = 360. See Table S4 for full list of adverse events.
Greater than 5% difference between treatment groups for:
a
CR-CBZ versus LEV,
b
LTG versus LEV,
c
CR-CBZ versus LTG.
d
Due to MedDRA coding, the terms “epilepsy” and “convulsion” were used synonymous.

Table 3. Reasons for discontinuation

Reason for discontinuation,
no. (%) [95% CI] Carbamazepine (n = 121) Lamotrigine (n = 118) Levetiracetam (n = 122)
Adverse event/serious adverse event/death 39 (32.2) [23.9–40.5] 31 (26.3) [18.4–34.2] 21 (17.2)* [10.5–23.9]
Lack of efficacy (seizure) 3 (2.5) 2 (1.7) 4 (3.3)
Other 24 (19.8) 20 (16.9) 22 (18.0)
Consent withdrawn 20 (16.5) 13 (11.0) 11 (9.0)
Protocol violation/noncompliance 4 (3.3) 6 (5.1) 10 (8.2)
Investigator’s decision 0 1 (0.8) 1 (0.8)
Enrolled population, n = 361. *p < 0.01 for CR-CBZ versus LEV (Fisher’s exact test).

All deaths were considered by the investigator to be not conducted. To our knowledge, the present trial is the first
related or unlikely related to the study drug. However, the prospective comparison of LEV with the optimal and most
sponsor considered that for two deaths, a possible relation- widely used formulation (controlled-release) of CBZ in this
ship to study drug could not be ruled out: one patient taking population. Our results indicate that in this patient popula-
LTG had intestinal ileus followed by aspiration of vomit tion, efficacy of LEV monotherapy was comparable to
and cardiac arrest, and one patient taking LEV committed CR-CBZ, whereas tolerability was superior, leading to a
suicide, likely to have resulted from domestic problems. higher retention to treatment at week 58 (the primary out-
Clinically relevant changes in hematology and clinical come).
chemistry parameters were observed in a number of Our findings suggest that, overall, the efficacy of the
patients. Of note, serum c-glutamyltransferase (GGT) levels three drugs assessed by seizure freedom and time to first
were elevated in 22 patients on CR-CBZ compared with 6 seizure was similar. Of note, average doses of the drugs
on LTG and 8 on LEV. Three patients on CR-CBZ were during the trial were all lower than the predefined target
hospitalized because of drug-induced hepatotoxicity. doses (most notably for CR-CBZ), and also lower than those
recommended in the individual product labels.26,28,29 This
supports the recommendation that initial dosing of epilepsy
Discussion medication for elderly patients should be lower, with a
In this trial, a double-blind comparison of treatment out- slower titration schedule than that used in younger adults, to
comes associated with three AEDs, CR-CBZ, LTG and reduce the risk of tolerability issues.30 One could argue that
LEV, in elderly patients with newly diagnosed epilepsy was 1,000 mg/day LEV is too high a target dose compared to

Epilepsia, **(*):1–10, 2015

doi: 10.1111/epi.12926
8
K. J. Werhahn et al.

400 mg/day for CR-CBZ. Target doses were chosen based immediate-release compared to CR-CBZ, and therefore
on the assessment of benefit–risk to use the lowest dose with more dose-related AEs can be expected.30
sufficient efficacy evidence,27,31 to avoid adverse dropouts CR-CBZ was compared with LTG over 40 weeks in one
with a too high starting dose, and to achieve the best possi- previous study in 186 elderly patients.18 Similar to the find-
ble dose equivalence by choosing the presumed lowest ings from the present study, no significant difference
effective dose for each drug in accordance with the respec- between retention rates was reported; patients on CR-CBZ
tive drug label.26,28,29 We also do not think that target doses had numerically higher seizure freedom rates, while LTG
biased results, since dose levels were not fixed and investi- was associated with better tolerability. The higher retention
gators were free to adjust doses already during titration rates in Saetre et al.18 (67% and 73%, for CR-CBZ and
within a wide range. LTG, respectively) compared to the current trial (46% and
Despite similar efficacy results, clear differences were 56%, respectively) may be due to a higher number of con-
evident in the tolerability profiles of these drugs during the comitant diseases and a greater number of concomitant
trial, suggesting that the differences in retention rates at drugs, with patients being more susceptible to AEs and
week 58 were driven primarily by the drugs’ tolerability dropout. Differences in study design (such as exit criteria)
profiles rather than their efficacy. Although the overall and duration of treatment (40 vs. 58 weeks) further limit a
incidence of AEs was similar across all treatment groups, direct comparison to our study. A more recent study
time to first drug-related AE was shorter for CR-CBZ com- comparing LTG and LEV in a population aged ≥55 years,
pared with LTG and LEV. In addition, AEs leading to dis- showed similar 12-month retention rates for the two drugs;
continuation, as well as drug-related adverse events, were however, the study was retrospective and uncontrolled.33
more frequent in the CR-CBZ group compared with the One limitation inherent to the design of the current trial is
other groups. Of note, adverse dropout rates may have the inclusion of the LTG arm for which the study was not
been confounded by the fact that a considerable proportion powered. There were a number of reasons that led to this
of patients withdrew their consent (Table 3, CR-CBZ decision. At the time the study was designed (autumn 2005)
16.5%, LTG 11%, and LEV 9%) possibly related to an comparative data between LTG and CR-CBZ was not yet
AE. This assumption again would point to CR-CBZ as available for power calculations.18 Assuming similar reten-
being most affected, since numbers were highest in this tion rates, a comparison between LTG and LEV would have
arm. Suggesting inferior tolerability of CR-CBZ might also required a sample size (or an adaption of the sample size
be that a greater number of patients chose to continue during the trial) that was deemed not feasible and a noninfe-
treatment with LEV or LTG than with CR-CBZ at the end riority design was regarded as less informative for clinical
of the trial. practice. In clinical practice, on the other hand, based on the
The type and nature of AEs reported were generally data by Rowan et al.,17 LTG was perceived as the main
consistent with findings of previous studies.16–18 CR-CBZ alternative to LEV at the time. Lamotrigine was superior to
and LTG were associated with a higher incidence of gastro- immediate-release CBZ in previous trials and therefore a
intestinal disorders than LEV. This is not an unexpected benchmark for LEV effectiveness.16,17 For these reasons it
finding, as discontinuation of CR-CBZ for gastrointestinal was decided to make the comparison of CR-CBZ versus
disturbances had been noted in previous studies of the LEV the primary focus of this trial.
elderly.16–18 Carbamazepine was also more often associated A novel feature of this trial was the inclusion of patients
with increased GGT levels. Although approximately 60% with only one previous seizure, in whom differences in
of users of CBZ have an elevation in GGT and 10–15% retention rates were most pronounced, yet retention rates
elevation in alkaline phosphatase (ALP), this is usually not under LEV treatment retention still being 5.9% and 11.2%
a reason to stop the drug. However, unusually high GGT higher compared to CR-CBZ in the other two subgroups
values and/or ALP values should be closely monitored to indicated that this is not only a subgroup effect. Although
detect severe drug induced hepatotoxicity.32 the study was not powered to detect differences in single
Despite the incidence of new-onset epilepsy being high- subgroups, the study population as defined by inclusion and
est in the elderly,3 there has been a paucity of trials evaluat- exclusion criteria should reflect patient subgroups in the
ing the most effective treatment options for these patients. target population. Thus, we think that a group effect
Although carbamazepine is still often the treatment of between LEV and CBZ can be derived from study results. In
choice,12 there is a general lack of supportive evidence. addition, inclusion of patients with a single seizure is consis-
Two previous comparative trials in the elderly evaluating tent with the new definition of epilepsy by the International
immediate release CBZ versus LTG, and immediate release League Against Epilepsy (ILEA),22 that is, the requirement
CBZ, gabapentin, and LTG demonstrated that immediate- of a significant risk of seizure recurrence based on EEG and
release CBZ was as efficacious as its comparators in imaging, factors which significantly increase the likelihood
controlling seizures, but was associated with poorer tolera- of seizure recurrence.34 This definition was adopted because
bility outcomes.16,17 However, time to maximum plasma it reflects current clinical practice, in particular in the elderly
concentration is shorter and peak concentration is higher for population with a frequent structural etiology of epilepsy.
Epilepsia, **(*):1–10, 2015
doi: 10.1111/epi.12926
 9
Comparison of Epilepsy Drugs in Elderly

This randomized-controlled trial (RCT) provides evi- 2. Pugh MJ, Knoefel JE, Mortensen EM, et al. New-onset epilepsy risk
factors in older veterans. J Am Geriatr Soc 2009;57:237–242.
dence supporting the use of LEV as first-line treatment for 3. Olafsson E, Ludvigsson P, Gudmundsson G, et al. Incidence of
elderly patients with new-onset focal epilepsy, and points to unprovoked seizures and epilepsy in Iceland and assessment of the
the value of LTG as alternative. epilepsy syndrome classification: a prospective study. Lancet Neurol
2005;4:627–634.
4. Faught E, Richman J, Martin R, et al. Incidence and prevalence of
epilepsy among older U.S. Medicare beneficiaries. Neurology
Disclosure 2012;78:448–453.
5. Gidal BE, French JA, Grossman P, et al. Assessment of potential drug
We confirm that we have read the Journal’s position on issues involved interactions in patients with epilepsy: impact of age and sex. Neurology
in ethical publication and affirm that this report is consistent with those 2009;72:419–425.
guidelines. 6. Brodie MJ, Elder AT, Kwan P. Epilepsy in later life. Lancet Neurol
2009;8:1019–1030.
7. Huying F, Klimpe S, Werhahn KJ. Antiepileptic drug use in nursing
Acknowledgments home residents: a cross-sectional, regional study. Seizure 2006;15:
194–197.
The trial was designed by the corresponding author and coordinated by 8. Johnell K, Fastbom J. Comparison of prescription drug use between
KJW, ET, GK, and AE. Data were collected by the investigators and ana- community-dwelling and institutionalized elderly in Sweden. Drugs
lyzed by IZKS, University Medical Centre of the Johannes Gutenberg Aging 2012;29:751–758.
University, Mainz, Germany. All authors contributed to the manuscript and 9. Galimberti CA, Magri F, Magnani B, et al. Antiepileptic drug
approved the final version to submit it for publication. The authors vouch use and epileptic seizures in elderly nursing home residents: a
for the accuracy of the data and for the fidelity of the study to the protocol. survey in the province of Pavia, Northern Italy. Epilepsy Res
The authors would like to acknowledge UCB Pharma for providing an 2006;68:1–8.
educational grant to support this study. Special thanks go to Mrs. C. Beierle 10. Huber DP, Griener R, Trinka E. Antiepileptic drug use in Austrian
(statistical programming), Mrs. H. Hofmann (study nurse), and Mrs. A. nursing home residents. Seizure 2013;22:24–27.
Laupert (clinical monitor) for their enduring efforts throughout the study. 11. Garrard J, Harms S, Hardie N, et al. Antiepileptic drug use in nursing
home admissions. Ann Neurol 2003;54:75–85.
12. Johnell K, Fastbom J. Antiepileptic drug use in community-dwelling
and institutionalized elderly: a nationwide study of over 1,300,000
Funding older people. Eur J Clin Pharmacol 2011;67:1069–1075.
The legal sponsor of this trial was the University Medical Centre of the 13. Pugh MJ, Van Cott AC, Cramer JA, et al. Trends in antiepileptic drug
Johannes Gutenberg University Mainz, Germany. The study was supported prescribing for older patients with new-onset epilepsy: 2000–2004.
by an unrestricted educational grant from UCB Pharma and by a grant from Neurology 2008;70:2171–2178.
the Interdisciplinary Centre for Clinical Trials (IZKS), University Medical 14. Brodie MJ, Mintzer S, Pack AM, et al. Enzyme induction with
Centre of the Johannes Gutenberg University, Mainz, Germany, funding antiepileptic drugs: cause for concern? Epilepsia 2013;54:11–27.
numbers FKN 01KN0703 and FKN 01KN1103, IZKS Mainz of the 15. Vecht CJ, Wagner GL, Wilms EB. Interactions between antiepileptic
German Federal Ministry of Education and Research. UCB Pharma had no and chemotherapeutic drugs. Lancet Neurol 2003;2:404–409.
role in the trial design, conduct, data analysis, and manuscript preparation. 16. Brodie MJ, Overstall PW, Giorgi L. Multicentre, double-blind,
randomised comparison between lamotrigine and carbamazepine in
elderly patients with newly diagnosed epilepsy. The UK Lamotrigine
Elderly Study Group. Epilepsy Res 1999;37:81–87.
Disclosure of Conflicts of 17. Rowan AJ, Ramsay RE, Collins JF, et al. New onset geriatric epilepsy:
a randomized study of gabapentin, lamotrigine, and carbamazepine.
Interest Neurology 2005;64:1868–1873.
18. Saetre E, Perucca E, Isojarvi J, et al. An international multicenter
KJW has received consultation/advisory board/speaking honoraria from
randomized double-blind controlled trial of lamotrigine and sustained-
UCB Pharma, Pfizer, Eisai, GlaxoSmithKline, Jansen-Cilag, and Desitin
release carbamazepine in the treatment of newly diagnosed epilepsy in
and has been employed by UCB Pharma since January 2012. ET has acted
the elderly. Epilepsia 2007;48:1292–1302.
as a paid consultant to Eisai, Ever Neuropharma, Biogen Idec, Medtronics,
19. Werhahn KJ, Klimpe S, Balkaya S, et al. The safety and efficacy of
Bial, and UCB and has received speakers’ honoraria from Bial, Eisai, GL
add-on levetiracetam in elderly patients with focal epilepsy: a one-year
Lannacher, GlaxoSmithKline, Boehringer, Viropharma, Actavis, and UCB
observational study. Seizure 2011;20:305–311.
Pharma. ET has received research funding from UCB Pharma, Biogen-
€ 20. Affairs UN-DoEaS. World Population Ageing 2013 ST/ESA/SER.A/
Idec, Red Bull, Merck, the European Union, FWF Osterreichischer Fond
348; 2013.
zur Wissenschaftsf€ orderung), and Bundesministerium f€ur Wissenschaft
21. Fisher RS, van Boas Emde W, Blume W, et al. Epileptic seizures and
und Forschung. JDIU has received honoraria from UCB Pharma, Gerot-
epilepsy: definitions proposed by the International League Against
Lanach, Eisai, and GlaxoSmithKline. PB has no potential conflict of inter-
Epilepsy (ILAE) and the International Bureau for Epilepsy (IBE).
est to disclose. IU has received advisory board/speaker honoraria from
Epilepsia 2005;46:470–472.
UCB Pharma, Eisai, GlaxoSmithKline, and GL Lannacher. MD-S has
22. Fisher RS, Acevedo C, Arzimanoglou A, et al. A practical clinical
received honoraria from GlaxoSmithKline, Eisai, Bial, and UCB Pharma.
definition of epilepsy. Epilepsia 2014;55:475–482.
TK no conflict of interest. BS has received speaker honoraria and funding
23. Alsaadi TM, Koopmans S, Apperson M, et al. Levetiracetam
for research projects from UCB Pharma. VB has received speaker honoraria
monotherapy for elderly patients with epilepsy. Seizure 2004;13:58–60.
from UCB Pharma. CR has no conflict of interest. AE has no conflict of
24. Cramer JA, Leppik IE, Rue KD, et al. Tolerability of levetiracetam in
interest. GK has has received advisory board and speaker honoraria from
elderly patients with CNS disorders. Epilepsy Res 2003;56:135–145.
Desitin, Eisai, GlaxoSmithKline, Lundbeck, Pfizer, UCB Pharma, and
25. Glauser T, Ben-Menachem E, Bourgeois B, et al. ILAE treatment
ViroPharma.
guidelines: evidence-based analysis of antiepileptic drug efficacy and
effectiveness as initial monotherapy for epileptic seizures and
syndromes. Epilepsia 2006;47:1094–1120.
References 26. SmPC C. Carbamazepine controlled-release summay of product
characteristics.
1. Hauser WA, Annegers JF, Kurland LT. Incidence of epilepsy and 27. Brodie MJ, Perucca E, Ryvlin P, et al. Comparison of levetiracetam
unprovoked seizures in Rochester, Minnesota: 1935–1984. Epilepsia and controlled-release carbamazepine in newly diagnosed epilepsy.
1993;34:453–468. Neurology 2007;68:402–408.

Epilepsia, **(*):1–10, 2015

doi: 10.1111/epi.12926
10
K. J. Werhahn et al.

28. SmPC L. Lamotrigine summay of product characteristics. of seizures at study entry (modified intent-to-treat popula-
29. SmPC L. Levetiracetam summay of product characteristics.
30. Ficker DM, Privitera M, Krauss G, et al. Improved tolerability and
tion, n = 359).
efficacy in epilepsy patients with extended-release carbamazepine. Table S3. Treatment emergent adverse events reported
Neurology 2005;65:593–595. by ≥5% of patients in any treatment group (safety popula-
31. Reunanen M, Dam M, Yuen AW. A randomised open multicentre
comparative trial of lamotrigine and carbamazepine as monotherapy in
tion, n = 360).
patients with newly diagnosed or recurrent epilepsy. Epilepsy Res Table S4. Adverse events reported by patients in any
1996;23:149–155. treatment group (safety population, n = 360).
32. Bjornsson E. Hepatotoxicity associated with antiepileptic drugs. Acta
Neurol Scand 2008;118:281–290.
Table S5. Adverse events most frequently leading to
33. Arif H, Buchsbaum R, Pierro J, et al. Comparative effectiveness of 10 discontinuation in ≥2% of patients in any treatment group
antiepileptic drugs in older adults with epilepsy. Arch Neurol (safety population).
2010;67:408–415.
34. Kim LG, Johnson TL, Marson AG, et al. Prediction of risk of seizure
Figure S1. Days to first seizure after randomisation (from
recurrence after a single seizure and early epilepsy: further results from V0, modified intent-to-treat population, n = 359).
the MESS trial. Lancet Neurol 2006;5:317–322.

Supporting Information
Additional Supporting Information may be found in the
online version of this article:
Table S1. Study schedule.
Table S2. Retention rate at week 58 stratified by number

Epilepsia, **(*):1–10, 2015

doi: 10.1111/epi.12926