• Lamotrigine;
  • Valproate;
  • Cognitive effects;
  • Mood effects;
  • Antiepileptic drugs


  1. Top of page
  2. Abstract

Summary:  Purpose: This study aimed at investigating the cognitive and mood effects of lamotrigine (LTG) versus valproate (VPA) and placebo (PBO).

Methods: By studying the effects in healthy volunteers, it is possible to separate the genuine effects of LTG from the cognitive improvements, caused by better seizure control. The study used a pretest–posttest comparison of 50 mg LTG, 900 mg VPA, or PBO in a double-blind single-dummy parallel-group design with 30 healthy volunteers. Study duration was 12 days (with a last control on day 13). Outcome measures included cognitive tests (FePsy neuropsychological test battery), mood scales (ASL; mood-rating scale), and a scale for subjective complaints (ABNAS Neurotoxicity scale). Total sleep time was controlled with actigraphic recordings. The results were analyzed by comparing the change over time (pretest with posttest) for the three treatments with Student's t tests.

Results:Cognitive tests: significant differences between the treatments were found for measurements of cognitive activation (i.e., three of the four simple reaction-time measurements showed statistically significant differences in change between PBO and LTG in favor of LTG (p = 0.03; 0.03; 0.04); two of four tests showed statistically significant differences in change between LTG and VPA, both in favor of LTG (p = 0.03; 0.05). Subjective complaints: the ABNAS-neurotoxicity scale reveals a significant reduction of drug-related cognitive complaints for the subjects taking LTG, relative to VPA (p = 0.02). Mood rating: significant changes were found on the scale assessing “tiredness,” showing increased tiredness/sedation for VPA relative to PBO (p = 0.02) and on the “timid scale” for LTG reporting “being more at ease” compared with both PBO and VPA (p = 0.02; 0.02). The general direction of change for the mood scales was toward “activation” for LTG (five of six scales improved), whereas for VPA, the reverse effect was found (four of six scales showed a change in the direction of “tiredness/sedation”).

Conclusions: Short-term treatment in normal volunteers with a low dose of LTG resulted in improved cognitive activation on simple reaction-time measurements, a more positive subjective report about the impact of drug treatment relative to VPA, and mood changes concurring with the activating effect demonstrated by the cognitive tests.

Effects of lamotrigine (LTG) on cognitive function have been studied in number of designs. No negative effects have been found in controlled volunteer studies after short-term (1-day) dosing of ≤240 mg of LTG (1), or after short-term treatment (4 weeks, titrated to a maximum of 7.1 mg/kg) (2). Moreover, the short-term effects were compared with those of carbamazepine (CBZ) also in normal volunteers, demonstrating differences in favor of LTG (3,4).

The effects in patients with epilepsy are more difficult to interpret, as no randomized controlled studies are available. Clinical studies in newly diagnosed patients (5) or in refractory patients (6,7) did not demonstrate clinically relevant cognitive side effects of LTG, concurring with the reported positive effects on “quality of life”(8).

As yet, no controlled comparisons are available with valproate (VPA). This comparison is clinically relevant, as LTG and VPA can be used in the same types of patients. Previous studies in patients with epilepsy showed, however, that it is difficult to isolate the genuine effects of LTG from the cognitive changes caused by changes in seizure frequency. Even in patients with low seizure frequency, it appeared to be difficult to interpret the cognitive effects of LTG because LTG may suppress epileptiform EEG discharge (9), which also may result in change of cognitive function (10). We therefore compared LTG with VPA in a controlled volunteer study, taking into account the disadvantage that only low doses can be used.


  1. Top of page
  2. Abstract

Trial design

The study used a pretest–posttest comparison for three treatments: LTG, VPA, and placebo in a double-blind single-dummy parallel-group design with 30 healthy volunteers. A parallel-group design was used to prevent extended admission (with the risk of subsequent selective drop-out), that would be needed with a crossover design.

Patients were sequentially allocated to a randomization schedule by the hospital pharmacy (Table 1).

Table 1.  Assessment flow chart and time schedule
 Day 0Day 1Day 4Day 8Day 10Day 12Day 13
  • a

     50 mg LTG as 25 mg b.i.d.

  • b

     600 mg VPA as 300 mg b.i.d.

  • c

     900 mg VPA as 300 mg t.i.d.

  • LTG, lamotrigine; VPA, valproate; PBO, placebo.

LTGBaseline25 mg 50 mg* StopControl
VPABaseline300 mg600 mg**900 mg** StopControl
Placebo LTGBaselinePBO PBO (b.i.d.) StopControl
Placebo VPABaselinePBOPBOPBO (t.i.d.) StopControl
Cognitive testsPretest    Posttest 
Blood samplesb1b2 b3b4b5b6
Sleep assessment    s1s2 


Thirty healthy male volunteers were recruited by advertisement in a local university newspaper by using the following inclusion and exclusion criteria:

Inclusion criteria

  • 1
    Gender: only male subjects were included in the trial
  • 2
    Age: between 18 and 35 years
  • 3
    Weight: within 20% of ideal body weight (Geigy scientific tables, Vol 3: page 324).

Exclusion criteria

  • 1
    Evidence of significant clinical abnormalities detected by full medical history, physical examination, ECG, and routine blood and urine analysis
  • 2
    Use of drugs known to affect central nervous system (CNS) performance
  • 3
    Drug abuse or evident abuse of alcohol (>4 units/day)
  • 4
    Participation in a clinical trial within 6 months before this study
  • 5
    Blood donation within the preceding 3 months
  • 6
    Previous use of antiepileptic drugs (AEDs) for>3 weeks in the preceding 4 years
  • 7
    Inability to give informed consent


The study was conducted in accordance with the ICH GCP guideline (1997) and approved by an independent ethics committee. Subjects gave written informed consent before entering the study.

Trial medication


The selected dose of LTG in this study was 50 mg, a dose that can be reached within a limited time period that is allowed for healthy volunteers in our country. Faster dose escalation within the same period would increase the risk of those side effects that are related to fast titration, especially rash (11). In the recent study by Martin et al. (2), a short-term dose of 50 mg was given to six volunteers and was well tolerated. Although this is clearly a low dose compared with the clinical dose that is used in patients with epilepsy, the study by Calabrese et al. (12) showed a clear effect on behavior of 50 mg/day in patients with bipolar I disorder.

During the trial, 25 mg was used as starting dose, and 50 mg was reached on day 8 (given as 25 mg, b.i.d.; see flow chart).


We decided to use the same dose of VPA that was applied in the only volunteer study (13) on VPA (i.e., 900 mg for 2 weeks) to be able to compare our results with this previous study. This study revealed mild to moderate effects exclusively on measures of motor speed.

During the trial, a starting dose of 300 mg was used; on day 4, the dose was increased to 600 mg (given as 300 mg, b.i.d.) and 900 mg on day 8 (given as 300 mg, t.i.d.; see flow chart).

The volunteers thus received either a dose of 50 mg LTG or 900 mg VPA or placebo tablets in a single-dummy design (half the placebo group received LTG–placebo; half, VPA–placebo). End-point outcome measures were assessed at day 12 (i.e., 4 days after reaching the maximal dose). Controls for stable plasma levels were performed on day 10. Similar titration schedules were used in the placebo arm, using single-dummy titration procedures.

A separate comment is needed about the comparability of the doses, as 50 mg LTG and 900 mg VPA are not entirely comparable. The dose of 50 mg LTG is subtherapeutic, whereas 900 mg of VPA is at the lower end of the therapeutic range. A higher dose of LTG was, however, not possible within the limitations set for a volunteer study, and although the cognitive and other behavioral effects of AEDs are often seen in volunteers at lower dose than needed for AED therapy in patients with epilepsy (14), this limits the results of this explorative study to effects seen at lower dose. Moreover, by using the dose of a previous volunteer study with VPA, we are able to compare our results with those of this previous study, but at the same time, limit the comparability with LTG.

Screening took place within 2 weeks before the study, consisting of clinical assessment, including medical history, physical examination, blood chemistry, and dipstick urine analysis. Demographic data, including age, level of education, medical diagnoses, use of drugs, smoking habits, and use of alcohol, were recorded. After medical screening, eligible subjects were made familiar with the cognitive test battery. All subjects were assessed during baseline. On day 1, the study medication was given with proper instructions, and blood samples were taken. On day 8, additional medication was given, together with appropriate instructions. Blood samples were taken. The same procedures were used on day 10. On day 12, reassessment (same outcome measures as assessed at baseline) was carried out. After these assessments, the medication was stopped without taper. On day 13, the subjects underwent a posttrial screening including blood sampling.

Outcome measures

The following outcome measures were assessed at baseline (pretest) and at day 12 (posttest). Posttests were performed 4 days after reaching the 50-mg dose of LTG and 900 mg VPA.

Cognitive tests

We used tests that have demonstrated sensitivity for detecting cognitive side effects of antiepileptic drugs. This selection was based on consensus meetings, e.g. the workshop on measurement of drug-induced cognitive impairment during the 21st International Epilepsy Congress in Sydney (15). Most tests are part of the FePsy neuropsychological test battery and are amply described elsewhere (16–19). The following areas/tests were included.

  • 1
    Psychomotor fluency The Finger Tapping Task, measuring motor speed and motor fluency in five consecutive trials for the index finger of the dominant and the nondominant hands separately.
  • 2
    Cognitive activation/alertness Simple reaction-time measurement on either auditory (800-Hz tones) or visual (a white square on the screen) stimuli that are presented at random intervals by the computer.
  • 3
    Information-processing speed Binary Choice Reaction Test, in which a decision component is introduced. The subject has to react differently to a red square, presented on the left side of the screen, and to a green square, presented on the right side. Reaction time here reflects not only motor speed but also the decision-making process. The Computerized Visual Searching Task (CVST), an adaptation of Goldstein's Visual Searching Task. A centered grid pattern has to be compared with 24 surrounding patterns, one of which is identical to the target pattern. The test consists of 24 trials and gives an indication of the speed of information processing and perceptual mental strategies.
  • 4
    Memory function Recognition of words and figures. The test stimuli are presented simultaneously or serially during a learning phase. In the simultaneous form, six words and four figures are presented with a presentation time of 1s per item. After a delay of 2 s, the screen shows one of these words/figures between distracters. The target item has to be recognized. In the serial presentation, recall of the order of the stimuli is required. Verbal Paired Associates I (immediate recall) from the Wechsler Memory Scale revised (20) in which verbal pairs have to be learned by association and later recalled.

To control for retesting effects, the study used a parallel group design in which this is only a relative validity threat, as this factor is equally active in each parallel group. Moreover, we used parallel tests or randomly presented items.

Cognitive setting

Subjects were instructed not to use alcoholic beverages from the evening before the testing days. In addition, use of coffee, tea, or chocolate was not allowed on the testing day. All tests were administered in the morning between 9 a.m. and 12 noon to prevent bias due to circadian rhythms. Each retesting was carried out in exactly the same hour as the baseline test. All tests were presented in a fixed order, as presented previously. Serum samples were taken directly after or before the testing session. Total sleep time was assessed by an actigraph on two separate days, day 10 and 12, to control for any marked changes in circadian patterns around the final test day.

Subjective rating of improvement

Subjects were given the ABNAS-neurotoxicity scale, assessing subjectively experienced side effects. This scale is described elsewhere (21–23).

Mood effects

The subjects were given a Dutch mood rating scale: “The Amsterdamse stemmingsschaal” (ASL; 24) to control for possible mood-modulating or psychotropic effects of LTG. In the ASL, the subjects are asked to evaluate how well the feelings of the last few days are described by a particular adjective (e.g., “sad,”“nervous,” or “unhappy,” on a 5-point scale ranging from “absolutely not” to “very well.” The rationale for using the ASL is that the scale was developed for psychopharmacologic experiments.

Blood sampling and laboratory tests

Blood samples of 15 ml for determining LTG and VPA serum levels were taken at baseline and on days 1, 8, 10, 12, and 13. Samples were taken in the morning on all occasions. The following laboratory tests were performed: biochemical profile [alanine aminotransferase (ALAT), aspartate aminotransferase (ASAT), γ-glutamyl transferase (γ-GT), alkaline phosphatase, creatine phosphokinase, bilirubin, total protein, albumin, Na+, K+, Ca2+, creatinine, urea]; routine hematology (hemoglobin, erythrocyte count, mean corpuscular volume (MCV), hematocrit, white cell count + differential count, platelet count); urine testing including pH, specific gravity, proteins, glucose, bilirubin, leukocytes, nitrite, urobilinogen, erythrocytes, hemoglobin, and nitrite.

Statistical analysis

A parallel-group study with 10 subjects per group is sufficient to detect medium-sized behavioral effects (i.e., changes of >0.7 standard deviation, following Cohen's conventions) (25,26) with a significance level of 5% and assuming a β or type-2 error risk of 20% (most tests have a discriminative power ∼80% or higher). Medium-sized effects are generally seen as behavioral effects of AEDs (19,27). As this was an explorative study, the results were analyzed by using Student's t tests to compare the changes over time for the three groups. Corrections for multiple testing were not applied to avoid that significance levels exceed the chance of finding the previously mentioned medium-sized effects. Statistical analysis was performed by using SPSS/PC V9.1.


  1. Top of page
  2. Abstract

All subjects were of male gender and were students or employed. Table 2 summarizes the main characteristics.

Table 2.  Age, gender, and hand dominance
Age (yr) (SD)22.1 (4.1)22.6 (3.0)21.8 (3.3)22.2 (3.4)
Gender10 male10 male10 male30 male
Dominant hand (%  right handed)80%100%80%86.7%

Table 3 shows the results of the serum level analysis of the AEDs and sleep recordings.

Table 3.  Serum levels and actigraph recordings
  1. Repeated measurement analysis of sleeping recordings: F value, 0.640 (df, 2) p = 0.53; SD in parentheses.

Serum level (mg/L)   
 Baseline0 (0)0 (0)0 (0)
 End point0.96 (0.26)55.2 (11.5)0 (0)
Actigraph (total sleep time   in hours)   
 1st assessment8.17 (2.38)8.18 (1.23)6.42 (1.14)
 2nd assessment7.47 (1.06)7.34 (1.34)7.05 (1.20)

The serum levels are in the lower range for both VPA and LTG. The actigraph recordings do not show significant differences in change over time for the three groups and an acceptable total sleep time. Table 4 shows the differences in change between baseline and end point for the cognitive tests.

Table 4.  Statistical evaluation of pre- and posttest scores for the cognitive tests
Cognitive areaLamotrigineValproatePlacebop Valuesb
  • LTG > VPA indicates a more favorable profile of LTG compared with VPA; LTG < VPA indicates a less favorable cognitive profile of LTG compared with VPA; LTG-VPA, no significant difference between LTG and VPA. Standard deviations in parentheses.

  • a

     1st is pretest during pretreatment baseline; 2nd is posttest after 12 days of treatment.

  • b

     results of analysis for change over time (baseline–end point) between the three treatment conditions based on Student's t-tests; statistically significant differences are underlined and in bold.

Psychomotor fluency       
 Finger-tap dom (number of taps)61.2 (8,1)59.8 (8.9)61.6 (8.8)64.6 (8.3)59.0 (4.4)62.0 (3.7)LTG-VPA p = 0.41
       LTG-PBO p = 0.98
       VPA-PBO p = 0.41
 Finger-tap ndom (number of taps)53.6 (6.7)52.9 (6.4)55.1 (8.9)56.4 (7.2)53.6 (3.8)55.8 (5.8)LTG-VPA p = 0.39
       LTG-PBO p = 0.61
       VPA-PBO p = 0.72
Cognitive activation and alertness       
 Auditory rt dom (milliseconds)209 (21)200 (13)226 (21)220 (28)226 (16)220 (17)LTG > VPA p = 0.03;
       LTG > PBO p = 0.03;
       VPA-PBO p = 0.99
 Auditory rt ndom (milliseconds)209 (15)205 (21)226 (20)224 (30)226 (16)229 (25)LTG > VPA p = 0.05;
       LTG > PBO p = 0.03;
       VPA-PBO p = 0.77
 Visual rt dom (milliseconds)245 (20)247 (20)254 (28)261 (19)276 (23)274 (30)LTG-VPA p = 0.28
       LTG < PBO p = 0.003
       VPA < PBO p = 0.04
 Visual rt ndom (milliseconds)251 (26)256 (21)261 (19)264 (32)274 (30)283 (34)LTG-VPA p = 0.44
       LTG > PBO p = 0.04;
       VPA-PBO p = 0.18
Information-processing speed       
 Binary choice rt (milliseconds)307 (43)284 (49)316 (52)292 (70)338 (46)323 (38)LTG-VPA p = 0.68
       LTG-PBO p = 0.11
       VPA-PBO p = 0.22
 Binary choice err. (number of errors)3.5 (5.3)4.5 (5.1)3.8 (2.8)4.9 (6.6)2.5 (2.6)2.2 (1.6)LTG-VPA p = 0.85
       LTG-PBO p = 0.37
       VPA-PBO p = 0.28
 CVST rt (seconds)9.0 (3.0)7.7 (2.3)7.9 (1.4)7.0 (0.8)8.3 (1.5)7.3 (1.4)LTG-VPA p = 0.23
       LTG-PBO p = 0.48
       VPA-PBO p = 0.61
 CVST errors (number of errors)1.6 (1.2)0.9 (1.5)2.2 (0.8)0.7 (0.8)1.6 (1.4)1.6 (1.3)LTG-VPA p = 0.63
       LTG-PBO p = 0.40
       VPA-PBO p = 0.72
Memory function       
 recogn words sim (number correct)21.7 (2.0)20.7 (3.2)21.3 (2.8)21.2 (1.9)21.0 (2.1)21.0 (1.8)LTG-VPA p = 0.96 LTG-PBO p = 0.78
       VPA-PBO p = 0.74
 recogn fig sim (number correct)15.6 (3.2)17.2 (4.0)15.4 (3.3)18.5 (4.0)15.1 (2.4)17.1 (3.3)LTG-VPA p = 0.67
       LTG-PBO p = 0.82
       VPA-PBO p = 0.52
 recogn words ser (number correct)17.3 (5.0)18.1 (4.2)19.5 (3.2)20.1 (2.7)19.5 (2.3)19.9 (3.2)LTG-VPA p = 0.12
       LTG-PBO p = 0.14
       VPA-PBO p = 0.94
 recogn fig ser (number correct)18.8 (2.7)19.0 (3.2)19.9 (3.1)20.7 (2.0)18.8 (3.7)20.1 (2.9)LTG-VPA p = 0.26
       LTG-PBO p = 0.67
       VPA-PBO p = 0.49
 word pairs 1 (number correct)19.8 (2.6)23.6 (0.5)19.0 (2.3)23.4 (1.3)20.2 (2.2)23.2 (1.3)LTG-VPA p = 0.44
       LTG-PBO p = 0.98
       VPA-PBO p = 0.44
 word pairs 2 (number correct)7.5 (1.0)8.0 (0.0)7.7 (0.5)7.9 (0.3)7.5 (0.5)7.9 (0.3)LTG-VPA p = 0.78
       LTG-PBO p = 0.78
       VPA-PBO p = 0.58

Significant differences in change between the groups were found for one area of cognitive function: the four tests, measuring cognitive activation and alertness (simple reaction-time measurement). Two of the four tests (auditory reaction-time measurement) show a statistically significant difference in change between LTG and VPA in favor of LTG (p = 0.03; 0.05). Three of the four tests show statistically significant differences in change between LTG and PBO in favor of LTG (p = 0.03; 0.03; 0.04). One test shows a significant difference in change between both VPA and LTG with PBO, both in favor of PBO (p = 0.003; 0.04).

Table 5 shows the results of the subjective rating scale and the mood scale:

Table 5.  Statistical evaluation of pre–posttest scores of the mood rating scales
AreaLamotrigineValproatePlacebop Valuesb
  • a

     1st is pretest during pretreatment baseline; 2nd is posttest after 12 weeks of treatment.

  • b

     Results of analysis for change over time (baseline–end point) between the three treatment conditions based on Student's t tests; statistically significant differences are underlined and in bold. LTG > VPA indicates a more favorable profile of LTG compared with VPA; LTG < VPA indicates a less favorable cognitive profile of LTG compared with VPA; LTG-VPA, no significant difference between LTG and VPA.

  • c

     Scores are stanine scores; higher = negative (i.e., more depression); standard deviation in parentheses.

Cognitive Complaints       
 ABNAS6.2 (3.8)4.1 (3.5)13.0 (11)9.8 (8.6)7.5 (8.2)4.8 (3.7)LTG > VPA p = 0.02
 Neurotoxicity scale (number of drug-related   cognitive complaints)      LTG-PBO p = 0.71
       VPA-PBO p = 0.06
 Depression3.7 (1.3)3.2 (1.3)3.9 (1.5)4.1 (1.5)4.0 (1.2)2.6 (1.0)LTG-VPA p = 0.29
       LTG-PBO p = 0.77
       VPA-PBO p = 0.18
 Timid6.4 (2.9)4.7 (1.6)3.7 (2.2)3.7 (1.3)4.0 (2.5)3.4 (1.6)LTG > VPA p = 0.02
       LTG > PBO p = 0.02
       VPA-PBO p = 1.0
 Bad temper5.0 (1.4)5.0 (0.8)5.1 (1.5)5.1 (2.1)4.4 (1.5)2.9 (1.2)LTG-VPA p = 0.86
       LTG < PBO p = 0.03
       VPA < PBO p = 0.02
 Angry4.3 (1.4)3.8 (1.4)3.7 (1.7)4.1 (1.7)4.4 (1.5)3.3 (1.2)LTG-VPA p = 0.81
       LTG-PBO p = 0.75
       VPA-PBO p = 0.94
 Tired3.8 (0.6)3.6 (0.5)4.1 (1.6)5.0 (1.6)3.5 (1.6)3.0 (1.1)LTG-VPA p = 0.09
       LTG-PBO p = 0.36
       VPA < PBO p = 0.02
 Anxious4.4 (1.4)3.7 (1.3)4.1 (1.3)4.3 (1.4)3.7 (0.8)3.3 (0.8)LTG-VPA p = 0.73
       LTG-PBO p = 0.21
       VPA-PBO p = 0.11

For the subjective cognitive complaints on the ABNAS-neurotoxicity scale, the difference in change between LTG and VPA is statistically significant in favor of LTG (p = 0.02). This indicates that subjects taking LTG complained less about the impact of drug treatment than did subjects taking VPA.

For the mood functions, several changes are statistically significant. For the “timid scale” the differences in change between LTG and VPA and between LTG and PBO are statistically significant, both in favor of LTG (p = 0.02; 0.02): the subjects taking LTG reported “being more at ease.” For the “bad temper scale,” the significant differences in change are caused by the positive change in PBO relative to both the subjects taking LTG and VPA (who do not show change during the trial: p = 0.03; 0.02). For the “tired scale,” there is a statistically significant worsening of scores for VPA (relative to PBO: p = 0.02), whereas the scores for LTG improve (although this does not reach statistical significance; p = 0.09).

In addition, the results presented in Table 5 show an overall trend of change.

Table 6 illustrates the reverse trend of change on the mood scales for LTG and VPA. Five of six scales improve in LTG, whereas the trend in VPA is the opposite, and none of the scales show improvement, whereas four of six scales show worsening. The direction of the scores on the scale show that “improvement” reflects a change toward more feelings of “activation,” whereas “worsening” reflects a change in the direction of “sedation” or “tiredness.”

Table 6.  Direction of change (improvement/no change/worsening) for the mood scales combined; lamotrigine versus valproate
  • Improvement reflects change in the direction of activation. Worsening reflects change in the direction of sedation/tiredness.

  • a

     Number of scales.



  1. Top of page
  2. Abstract

Most of the AEDs have negative effects on behavior, specifically on cognition, although these are mild for most of the commonly used drugs (i.e., CBZ and VPA), in contrast to the effects of some of the older AEDs such as phenobarbitone (27). For LTG, clinical information also suggests a relatively favorable cognitive profile (5,7,28). Experimental cognitive studies in volunteers confirmed these clinical data for the comparison between LTG and CBZ (4). Our study was aimed at finding confirmation for the comparison of LTG with VPA, while also comparing LTG with placebo.

The results of cognitive testing show almost no difference when LTG and VPA are compared with placebo. This reconfirms the mild cognitive profile of VPA as demonstrated by the Thompson and Trimble study (13). It also demonstrates a similar mild cognitive profile for LTG, at least at low dose. In addition, there is a modest difference between LTG and VPA on measures for cognitive activation and alertness. This effect seems selective, as it is consistently shown only on all activation tests, whereas none of the other investigated areas show changes in the same magnitude.

An aspect to consider is whether the results on the auditory and visual reaction-time tests reflect changes in cognitive activation or are merely an effect of improved motor speed. These tasks have, however, been extensively evaluated in experimental cognitive psychology, showing that the motor component is only marginally important for reaction-time speed, which reflects the “early stages” of information processing, generally labeled “activation” (“general receptivity to process information”) (29). Moreover, the task primarily assessing motor speed (i.e., finger-tapping task) does not show the same effect. We can therefore conclude that the effect that is found in this study reflects cognitive activation.

This activation effect is not caused by the difference in dosing for LTG and VPA, as most measures showing a significant difference between LTG and VPA also show a difference between LTG and PBO. Nonetheless, the effects must be interpreted cautiously, as the absence of significant change in other areas (positive and negative) also may be due to the short treatment period and relatively low dose. A longer treatment period and/or a higher dose may thus well result in other effects, or effects in other areas than those demonstrated in the present study.

The results on the measures for cognitive activation may be interpreted as suggestive for a specific mechanism of LTG (i.e., a general psychostimulant effect on the brainstem arousal and vigilance systems). This is in line with a previous study that combined cognitive testing and EEG recording. In this study (30), LTG treatment resulted in better cognitive scores on a neuropsychological battery, accompanied with a selective increase in EEG alpha reactivity during the attentive tasks. This was considered to be a sign of improved cortical activation. The observed improvement of cognitive activation also concurs with the assumed mechanism of action of LTG and the effect of such mechanisms on cognition, specifically on the ability to reduce calcium ion influx and the inhibition of glutamate release is considered to explain the in vivo observed ability of LTG to reduce cognitive impairment arising from episodes of ischemia (31,32). No effect was found, however, on the mechanism of long-term potentiation (LTP) in gerbils tested in the Water Maze (33). LTP is considered to represent the mechanisms of “learning,” and both results combined again suggest a selective effect of LTG specifically on activation and arousal and not on higher-order cognitive mechanisms (i.e., “learning or memory”), which concurs with the results of the present study.

A possible cognitive-activating effect also may be the result of a number of interfering factors, changed sleeping patterns being the most important. Our comparison of sleep time during the trial, using actigraph recordings, did, however, not show statistically significant differences.

Remarkably, the results of the mood assessment also are in the direction of “activation” for LTG. In general, the effects for LTG are in a different direction, compared with VPA: five of six of the mood scales show changes in the direction of “activation” for LTG, whereas four of six scales show changes in the direction of “sedation” and “tiredness” for VPA. This is in line with the cognitive results of our study, with previous studies in patients with, for example, bipolar disease (12), and with the assumed differential mechanisms of VPA and LTG on mood factors. Ketter et al. (34) classified the mood profiles of AEDs in two separate classes: the first class has sedating effects and is used in psychiatry for its anxiolytic and antimanic effects (e.g., for the treatment of agitation and aggression). These effects are assumed to be related to the potentiation of γ-aminobutyric acid (GABA)–mediated inhibitory neurotransmission that is the predominant mechanisms of action in this class. VPA belongs to this class of drugs. The major CNS-related side effects are fatigue and cognitive slowing. The second class of drugs has opposite (i.e., activating) effects, associated with the predominance of glutamate excitatory neurotransmission as the dominant mechanism of action. This class of drugs is used for its anergic profiles in psychiatry (e.g., for the treatment of depression, apathy, hypersomnia, and fatigue). Its assumed effect is cognitive activation. LTG belongs to this class of drugs. The major side effects are hyperactivity, hyperirritability, and possibly insomnia. The results on the mood scales in this study are in line with this classification and show a cognitive “activating” effect of LTG and a “sedation” effect for VPA.

We may conclude that this study suggests a relatively favorable cognitive profile for LTG, similar to VPA, at least at this low dose. In addition there may be a selective general psychostimulant effect of LTG that is demonstrated both in improved cognitive activation and in psychotropic effects (“activating” mood changes). If this effect can be confirmed in patients with epilepsy, it can be helpful, especially in treatment of chronic epilepsy, given the dominance of depression, apathy, and fatigue as comorbid symptoms (35).

Acknowledgment: This study was financially supported by a grant of GlaxoSmithKline. We thank Arno Schlosser for his advice during the period of designing the study.


  1. Top of page
  2. Abstract
  • 1
    Cohen AF, Ashby L, Crowley D, et al. Lamotrigine (BW430C), a potential anticonvulsant: effects on the central nervous system in comparison with phenytoin and diazepam. Br J Clin Pharmacol 1985;20:61929.
  • 2
    Martin R, Kuzniecky R, Ho S, et al. Cognitive effects of topiramate, gabapentin and lamotrigine in healthy young adults. Neurology 1999;15:3217.
  • 3
    Hamilton MJ, Cohen AF, Yuen AW, et al. Carbamazepine and lamotrigine in healthy volunteers: relevance to early tolerance and clinical trial dosage. Epilepsia 1993;34:16673.
  • 4
    Meador KJ, Loring DW, Ray PG, et al. Differential cognitive and behavioral effects of carbamazepine and lamotrigine. Neurology 2001;56:117782.
  • 5
    Brodie MJ, Read CL, Gillham R, et al. Lack of neuropsychological effects of lamotrigine compared to carbamazepine as monotherapy. Epilepsia 1999;40(suppl 2):94.
  • 6
    Banks GK, Beran RG. Neuropsychological assessment in lamotrigine treated epileptic patients. Clin Exp Neurol 1991;28:2307.
  • 7
    Aldenkamp AP, Mulder OG, Overweg J. Cognitive effects of lamotrigine as first line add-on in patients with localized related (partial) epilepsy. J Epilepsy 1997;10:11721.
  • 8
    Meador KJ, Baker GA. Behavioral and cognitive effects of lamotrigine. J Child Neurol 1997;12:S447.
  • 9
    Binnie CD. Cognitive effects of lamotrigine. Epilepsia 1995;36:31.
  • 10
    Aldenkamp AP, Overweg J, Gutter TH, et al. Effect of epilepsy, seizures and epileptiform EEG discharges on cognitive function. Acta Neurol Scand 1996;93:2539.
  • 11
    Richens A. Safety of lamotrigine. Epilepsia 1994;35:3740.
  • 12
    Calabrese JR, Bowden CL, Sachs GS, et al. A double-blind placebo-controlled study of lamotrigine monotherapy in outpatients with bipolar 1 depression. J Clin Psychiatry 1999;60:7988.
  • 13
    Thompson PJ, Trimble MR. Sodium valproate and cognitive functioning in normal volunteers. Br J Clin Pharmacol 1981;12:81924.
  • 14
    Trimble MR. Anticonvulsant drugs and cognitive function: a review of the literature. Epilepsia 1987;28:3745.
  • 15
    Aldenkamp AP, Timble MR. Cognitive side-effects of antiepileptic drugs: fact or fiction? Epilepsia 1996;37:82.
  • 16
    Alpherts WCJ. Computers as a technique for neurochological assessment in epilepsy. In: AldenkampAP, AlphertsWCJ, MeinardiH, et al., eds. Education and epilepsy. Lisse/Berwyn: Swets & Zeitlinger, 1987:1019.
  • 17
    Alpherts WCJ, Aldenkamp AP. Computerized neuropsychological assessment in children with epilepsy. Epilepsia 1990;31:3540.
  • 18
    Aldenkamp AP, Vermeulen J, Alpherts WCJ. Validity of computerized testing: patient dysfunction and complaints versus measured changes. In: DodsonWE, KinsbourneM, eds. Assessment of cognitive function. New York: Demos, 1992:5168.
  • 19
    Aldenkamp AP, Alpherts WCJ, Blennow G, et al. Withdrawal of antiepileptic medication: effects on cognitive function in children: the results of the multicentre “Holmfrid” study. Neurology 1993;43:4151.
  • 20
    Wechsler D. WMS-R, Wechsler Memory Scale–Revised. New York: The Psychological Corporation, Harcourt Brace Jovanovich, 1987.
  • 21
    Aldenkamp AP, Baker G, Pieters MSM, et al. The Neurotoxicity Scale; the validity of a patient-based scale, assessing neurotoxicity. Epilepsy Res 1995;20:22939.
  • 22
    Aldenkamp AP, Baker G. The Neurotoxicity Scale II: results of a patient-based scale assessing neurotoxicity in patients with epilepsy. Epilepsy Res 1997;27:16573.
  • 23
    Brooks J, Baker GA, Aldenkamp AP. The A-B Neuropsychological Assessment Schedule (ABNAS): the further refinement of a patient-based scale of patient-perceived cognitive functioning. Epilepsy Res 2001;43:22737.
  • 24
    De Sonneville LMJ, Schaap TH, et al. ASL: Amsterdamse Stemmingslijst. Lisse: Handleiding, Swets & Zeitlinger, 1984.
  • 25
    Cohen J. Statistical power analysis for the behavioral sciences. New York: Academic Press, 1977.
  • 26
    Cook D, Campbell DT. Quasi-experimentation: design & analysis issues for field settings. Boston: Houghton Mifflin, 1979.
  • 27
    Vermeulen J, Aldenkamp AP. Cognitive side-effects of chronic antiepileptic drug treatment: a review of 25 years of research. Epilepsy Res 1995;22:6595.
  • 28
    Aldenkamp AP, Baker G. A systematic review of the effects of lamotrigine on cognitive function and quality of life. Epilepsy Behav 2001;2:8591.DOI: 10.1006/ebeh.2001.0168
  • 29
    Davies DR, Parasuraman R. The psychology of vigilance. New York: Academic Press, 1982.
  • 30
    Marciani MG, Stanzione P, Mattia D, et al. Lamotrigine add-on therapy in focal epilepsy: electroencephalographic and neuropsychological evaluation. Clin Neuropharmacol 1998;21:417.
  • 31
    Leach MJ, Marden CM, Miller AA. Pharmacological studies on lamotrigine, a novel potential antiepileptic drug, II: neurochemical studies on the mechanism of action. Epilepsia 1986;27:4907.
  • 32
    Smith SE, Meldrum BS. Cerebroprotective effect of lamotrigine after focal ischemia in rats. Stroke 1995;26:11721.
  • 33
    Wiard RP, Dickerson MC, Beek O, et al. Neuroprotective properties of the novel antiepileptic lamotrigine in a gerbil model of global cerebral ischemia. Stroke 1995;26:46672.
  • 34
    Ketter TA, Post RM, Theodore WH. Positive and negative effects of antiepileptic drugs in patients with seizure disorders. Neurology 1999;53(5 suppl 2):S5367.
  • 35
    Hermann BP, Whitman S. Behavioral and personality correlates of epilepsy: a review, methodological critique, and conceptual model. Psychol Bull 1984;95:45197.