JLE

Epileptic Disorders

MENU

Topiramate: a prospective study on the relationship between concentration, dosage and adverse events in epileptic patients on combination therapy Volume 7, issue 3, September 2005

Figures

See all figures

Auteur(s) : Walter Fröscher1, Katja R Schier1, Markus Hoffmann2, Andreas Meyer1, Theodor W May3, Bernhard Rambeck3, Johannes Rösche1

1Department of Neurology and Epileptology, the Epilepsy Center Bodensee, Die Weissenau Germany
2Department of Psychiatry I, University of Ulm, Department Versorgungsforschung, Ravensburg
3Department of Biochemistry, Epilepsy Research Foundation, Bielefeld, Germany

Article reçu le 19 Août 2004, accepté le 14 Février 2005

Topiramate (TPM) was officially approved in Germany for combination therapy in July 1999 and for monotherapy in July 2002. The relationship between TPM serum concentration and efficacy and between serum concentration and toxicity is still under discussion. Reife et al. (1995a) reported as early as 1995 that in patients on TPM combination therapy, the mean TPM serum concentrations were generally higher for patients experiencing a given adverse event than for patients who did not experience that event. Significantly higher mean serum concentrations occurred in subjects experiencing anorexia or weight loss, symptoms related to cognition (e.g. impaired concentration, confusion), and symptoms related to emotional state (e.g. anxiety, depression). According to the data of Kockelmann et al. (2002) however, cognitive impairment appeared to be more an idiosyncratic feature of TPM than a dose-dependent event. In 2004, this result was modified by the authors in so far as they concluded that the severity of the cognitive side effects of TPM might, to a certain extent, be related to dosing (Kockelmann et al. 2004). Perucca (2000) stated: “Although at present there are not sufficient data to justify the measurement of serum TPM levels on a routine basis, this drug shows promise for further evaluation as a TDM (Therapeutic Drug Monitoring) candidate”.According to the experience with “old” anticonvulsants such as phenobarbital, phenytoin or carbamazepine, “therapeutic drug level ranges” are useful for optimizing dosages and identifying possible adverse events (Fröscher et al. 1999). In the case of TPM, the results of the correlation of adverse events with TPM serum concentration, and adverse events with TPM dosage are conflicting: in some studies with TPM as add-on therapy (Sachdeo 1998, Contin et al. 2002, Kockelmann et al. 2002, Ferrari et al. 2003), the TPM serum concentration allowed no reliable prediction of the occurrence of adverse events. In other studies of TPM add-on therapy, there were significant correlations between the TPM serum concentration and some adverse events e.g. anorexia, weight loss, cognitive and emotional symptoms (Reife et al. 1995a, Nagel et al. 2001). We therefore investigated the relationship between serum concentration and adverse events in the case of the new drug TPM. A significant correlation would allow the specification of an approximate upper limit for a “therapeutic range” of TPM serum concentrations. Furthermore, we examined whether adverse events of TPM correlate more closely with the serum concentrations than with the dosage.It should also be mentioned that the results of studies on the correlation of TPM serum concentration and seizure reduction are contradictory. Reife et al. (1995a) did not find a significant correlation between seizure reduction and TPM serum concentration. In other studies, seizure-free patients, as well as patients with a seizure reduction of at least 50%, had a TPM serum concentration of 2-25 μg/mL (Reife et al. 1995a, Patsalos 1999, Stephen et al. 2000, Burcet-Dardé 2002, Ferrari et al. 2003), thus the range of an effective serum concentration is very broad. According to Christensen et al. (2003) however, an optimal treatment response is most likely to be found at serum concentrations higher than 2 μg/mL, but no further increase in efficacy seems to occur at concentrations above 10.5 μg/mL. As a consequence of these conflicting results, the recommended target range of TPM given in the literature varies between 2-5 μg/mL (Fröscher et al. 1999) and 2-25 μg/mL (Glauser and Pippenger 2000).

Patients and methods

We included prospectively all 42 outpatients and inpatients with epilepsy, treated with TPM in our department of neurology and epileptology between December 1999 and September 2001. We included patients who had started with TPM before the beginning of our study as well as patients who received TPM for the first time during the study. Probable or certain adverse events of TPM were documented and blood samples were taken to determine the serum concentration of TPM at the same time. Medical history and clinical findings were documented on special forms. Examinations during this period and special investigations such as EEG were performed only if clinically necessary. In case of combination therapy, the serum concentrations of the following anticonvulsants were also measured: carbamazepine, lamotrigine, phenobarbital, phenytoin, valproic acid. The serum concentrations of other anticonvulsants were not measured because an appropriate method was not available in our hospital.

The record of adverse events was based on questioning (medical interview) and examination of the patients. The patients were asked if they had noticed any adverse events, with special emphasis on dizziness, abnormal thinking (mental slowing, can’t think as clearly, dull thinking), somnolence, ataxia, fatigue, confusion, impaired concentration and paresthesia. These 8 adverse events occurred in 5, double-blind, placebo-controlled, parallel group studies more often (≥ 10%) in patients on TPM treatment than in patients on placebo (Reife et al. 1995b, Sachdeo 1998). The 42 patients were examined up to 9 times (median 2) during the study period; 15 of the 42 patients were examined only once, the observation period of the other 27 patients was 6 to 554 days (median 74 days). The demographic and clinical characteristics of the patients are given in table 1( Table 1 ). At the first examination, 41 patients were treated with TPM in combination with up to 4 additional anticonvulsants (median 2), one patient was on TPM monotherapy. The study was performed before TPM was approved for monotherapy in Germany. Seventeen patients were treated with TPM in combination with one additional anticonvulsant, 14 patients in combination with 2 additional anticonvulsants, eight patients had 3 and 2 patients 4 additional anticonvulsants (table 2)( Table 2 ).

At the time of the first clinical examination, the duration of TPM treatment was 0.3 – 50 months (median 6 months, mean 10.6 ± 12.3 months).

The mean TPM dosage for 42 patients at the 109 medical examinations performed was 271 ± 155 mg (± SD). The median was 250 mg (range: 25 to 800). The corresponding body weight-related TPM dosage was 3.8 ± 2.3 mg/kg (mean ± SD) and 3.3 mg/kg (median; range: 0.2 to 10.9).

As a rule, changes of dosage were performed slowly (initial dose: 25 mg daily, increase of dosage was undertaken as 25 mg increments weekly). For the inpatients, TPM was determined in morning samples, before drug intake. For outpatients, blood samples had to be taken after the morning dose, because many of the patients attending our specialized outpatient clinic live quite a distance away from the hospital, and so blood examination could not be done before late morning. The mean interval between last drug intake and blood sampling in the whole group was 8.3 ± 5.5 h (± SD), the median was 6 h (range 1.3 to 24 h).

Separate evaluation of outpatients and inpatients was not performed, because the total number of patients was too small. Furthermore, some patients were treated as outpatients as well as inpatients during the study period. The study design does not allow for investigation of the relationship between TPM serum levels (or TPM dosages) and the reduction of seizure frequency caused by TPM, or the withdrawal of TPM due to adverse events.

The efficacy of TPM was not investigated for two reasons: the anticonvulsant efficacy of TPM is well known and, above all, we did not want to exclude those outpatients who did not reliably record seizure frequency.

TPM serum concentrations were determined by fluorescence polarization immunoassay/FPIA) developed by OXIS International Inc. (Portland, USA). The INNOFLOURTM TPM Assay is for use on TDx or TDxFLx analyzers. The Innofluor assay is a robust procedure for the measurement of TPM in clinical samples (Berry and Patsalos 2002, May et al. 2002). The serum concentrations of carbamazepine, phenobarbital, phenytoin and valproic acid were also determined by FPIA; lamotrigine was determined by HPLC.

Statistical analysis: statistical methods included descriptive statistics for specification of the total sample, and the Wilcoxon rank sum test for differences between patients without, respectively with adverse events relating to TPM serum concentrations and TPM dosages. To compare the non-independent TPM serum concentrations at dates with and without adverse events, we used the Wilcoxon signed rank test. For each test we required a minimum of 10 patients.

To examine the influence of the co-medication on the occurrence of adverse events, we performed a logistic regression analysis using the occurrence of the most frequent adverse event “abnormal thinking” as the dependent variable (0 = no/1 = yes). The dose (mg/kg) and the serum concentration of TPM were used as independent variables. In addition, the total number of antiepileptic drugs (model 1) and the number of antiepileptic drugs per substance group (see below, model 2) were included, but not the respective dosages (group 1 – marked sedation: clobazam, clonazepam, phenobarbital, primidone. Group 2 – moderate sedation: carbamazepine, oxcarbazepine, phenytoin. Group 3 – mild sedation: lamotrigine, valproic acid. Group 4 – other anticonvulsants. The last group comprises the anticonvulsants rarely used in this study: acetazolamide, gabapentin, levetiracetam, sulthiame, tiagabine, vigabatrin). The predictor variables were entered simultaneously into the models. All statistical analyses were performed with the SAS System (Version 8.2).
Table 1 Demographic and clinical characteristics of patients (n = 42).

Age, median (y)

36.5

- range

16-70

Gender (M/F)

21/21 patients

Epileptic syndrome

- focal

n = 36

- generalized

n = 4

- undetermined

n = 2

Etiology of epilepsy

- symptomatic/cryptogenic

n = 29

- idiopathic

n = 4

- undetermined

n = 9

Seizure (sz) history

No. of patients

- focal sz evolving to secondary generalized sz

22

- tonic-clonic sz, undetermined if with or without focal initiation

2

- simple focal sz

2

- complex focal sz

9

- tonic-clonic sz and absences

2

- absences

2

- tonic-clonic sz, complex focal sz, absences

1

- tonic-clonic sz, complex focal sz, absences, drop attacks, head nodding

1

- complex focal sz, absences, drop attacks

1

Compliance during the study (n = 42 patients, n = 109 investigations)

No (%) of investigations

- yes

98 (90%)

- no

3 (2.8%)

- uncertain

8 (7.3%)

Age (y) at onset of epilepsy, median (range)

6.5 (0.8 - 35)

Duration (y) of epilepsy at the beginning of the study, median (range)

30 (2 - 57)

Duration (y) of anticonvulsant treatment at the beginning of the study, median (range)

28 (2 - 56)


Table 2 Type of concomitant antiepileptic drugs (AED)Baseline: n = 41 patients, one patient was on TPM monotherapy.

n

dose (mg, median, range)

  • serum concentration
  • (μg/mL, median, range)


Phenobarbital (n = 11)

17

PB: 120 (15-325)

PB: 25.3 (5.5-52.8)

and primidonea (n = 6)

PRM: 844 (500-1000)

Carbamazepine

14

1300 (600-2850)

8.4 (4.7-15)

Valproic acid

13

1500 (150-2700)

70 (5-110)

Lamotrigine

10

375 (100-700)

3.8 (1.4-10)

Phenytoin

10

325 (200-450)

16.2 (8.4.-20.8)

  • Benzodiazepines
  • - clobazam


  • (6)
  • - 3


CLB: 10 (10-20)

not determined in most cases

- clonazepam

- 3

CZP: 5.5 (1-8)

Other AEDs (acetazolamide, gabapentin, levetiracetam, oxcarbazepine, sulthiame, tiagabine, vigabatrin)

9

not determined in most cases

aIn patients who were treated with primidone, only the concentration of phenobarbital was determined. In one patient the concentration of phenobarbital was not determined by error.

Results

At the first examination, 29 out of 42 patients (69%) experienced an adverse event, 33 out of 42 patients (79%) had, in at least one out of the 109 examinations, an adverse event (table 3)( Table 3 ). Adverse events that were observed in at least five (out of 42) patients, were evaluated further. The TPM serum concentration and the TPM dosage at the time of the first registration of these adverse events were compared with the corresponding values of patients who had no adverse event at the first examination (n = 13; tables 4 and 5( Table 4 )( Table 5 )). The differences in TPM serum concentrations and TPM dosages (mg/kg) for patients without an adverse event and patients with a given adverse event were statistically significant for abnormal thinking, impaired concentration, weight loss, dizziness, speech problems, somnolence, ataxia, increased seizure frequency and paresthesia. Only in the case of “fatigue” was the difference not significant.

The comparison of TPM serum concentration and dosage at the first examination of patients with the occurrence of any adverse events resulted in statistically significant differences for the serum concentration and for dosage (table 6)( Table 6 ).

( Figure 1 ) shows the 109 serum concentrations of all examinations of the 42 patients. Serum concentrations accompanied by an adverse event, and serum concentrations not accompanied by an adverse event are markedly different; however, a vast overlap between these two categories can be seen. This vast overlap can also be demonstrated by comparing the lowest TPM serum concentration of patients with a distinct adverse event, with the highest TPM serum concentration at which the patient did not have this adverse event. The statistical evaluation (Wilcoxon signed-rank test) was done only in those adverse events which were reported by at least ten patients. Only those patients could be included who had the distinct adverse event in at least one but not all examinations. Therefore, the number of patients is small for this comparison of the adverse events “abnormal thinking, impaired concentration and weight loss”. There was no statistical difference between TPM serum concentration or dosage and patients with and without these adverse events.

( Figure 2 ) indicates the proportion of patients with and without adverse events at the different examination times. This proportion did not change essentially during the whole study period.

( Figure 3 ) shows that the occurrence of adverse events at the first examination was dependent on TPM serum concentrations. Patients were experiencing adverse events at serum concentrations as low as 0.1 – 4 μg/mL. Above a TPM serum concentration of 4 μg/mL, the predominance of patients with adverse events is even more distinct. figures 4 and 5 show the occurrence of adverse events to be dependent on absolute dosage or the dosage related to body weight. Only with a TPM dosage of ≤ 100 mg or ≤ 1.5 mg/kg, respectively, do patients appear to be without any adverse events. The 13 patients without any adverse event at the first examination had a mean TPM serum concentration of 3.2 ± 3.6 μg/mL (mean ± SD).

In the logistic regression model 1 (with the total number of antiepileptic drugs) as well as in model 2 (with the number of antiepileptic drugs per substance group), only the TPM dose and not the serum concentration allowed the prediction of the appearance of “abnormal thinking” (table 7)( Table 7 ). The increase in the dose of 1 mg/kg raises the risk of “abnormal thinking” by a factor of 1.5 or 1.6, respectively. The total number of additional anticonvulsants and the number of anticonvulsants per group were not significant.
Table 3 Adverse events in patients treated with TPM at the first examination (53 adverse events in 29 patients), and during the whole study (123 adverse events in 42 patients).

Adverse events

1st examination

All examinations

n

% of 42 pat.

n

% of 42 pat.

Abnormal thinking

10

24

18

43

Impaired concentration

9

21

13

31

Weight loss

5

12

12

29

Dizziness

4

10

8

19

Speech problems

4

10

8

19

Ataxia

4

10

5

12

Fatigue

3

7

6

14

Exhaustion

2

5

4

10

Paresthesia

2

5

5

12

Loss of appetite

2

5

2

5

Somnolence

1

2

8

19

Dysarrythmia

1

2

3

7

Confusion

1

2

2

5

Nausea, vomiting

1

2

2

5

Increased seizure frequency

1

2

5

12

Impaired memory

1

2

3

7

Emotional lability

1

2

2

5

Leucopenia

1

2

2

5

Irritability, aggressiveness

0

0

3

7

Apathia

0

0

2

5

Skin lesion

0

0

2

5

Increased alertness

0

0

1

2

Elevated mood

0

0

1

2

Dyspnoea on exertion

0

0

1

2

Edema of the malleolar region

0

0

1

2

Incontinence of urine

0

0

1

2

Diarrhoea

0

0

1

2

Nervousness

0

0

1

2

Tremor

0

0

1

2


Table 4 TPM serum concentration at the first observation of an adverse event. Comparison with a TPM serum concentration from 13 patients without adverse events at the first examination.

Adverse event

  • Serum concentration of the patients with a given adverse event
  • (μg/mL)


  • Serum concentration of the 13 patients without adverse events (n = 13)
  • (μg/mL)


Wilcoxon rank sum test

Abnormal thinking

  • n = 18
  • 7.2 ± 4.9


3.2 ± 3.6

  • W = 138.
  • p = 0.003


Impaired concentration

  • n = 13
  • 6.7 ± 4


3.2 ± 3.6

  • W = 228.
  • p = 0.004


Weight loss

  • n = 12
  • 7.3 ± 5.1


3.2 ± 3.6

  • W = 202.
  • p = 0.0067


Dizziness

  • n = 8
  • 5.5 ± 4.7


3.2 ± 3.6

  • W = 115.
  • p = 0.03


Speech problems

  • n = 8
  • 6.4 ± 4.9


3.2 ± 3.6

  • W = 0 113.
  • p = 0.04


Somnolence

  • n = 8
  • 6.9 ± 5.


3.2 ± 3.6

  • W = 113.
  • p = 0.04


Fatigue

  • n = 6
  • 3.8 ± 3.5


3.2 ± 3.6

  • W = 73.
  • p = 0.14


Ataxia

  • n = 5
  • 7.1 ± 5.2


3.2 ± 3.6

  • W = 69.
  • p = 0.02


Increased seizure frequency

  • n = 5
  • 6.7 ± 4.1


3.2 ± 3.6

  • W = 68.
  • p = 0.02


Paresthesia

  • n = 5
  • 6.8 ± 3.8


3.2 ± 3.6

  • W = 68.
  • p = 0.02



Table 5 Dosage at the first observation of the adverse event. Comparison with TPM dose of 13 patients without adverse events at the first examination.

Adverse event

TPM dose of the patients with a given adverse event

TPM dose of the 13 patients without adverse events (mg/kg)

Wilcoxon rank sum test

Abnormal thinking

  • n = 18
  • 5.2 ± 2.8


2.1 ± 1.8

  • W = 134.5
  • p = 0.002


Impaired concentration

  • n = 13
  • 4.4 ± 2.2


2.1 ± 1.8

  • W = 225.
  • p = 0.006


Weight loss

  • n = 12
  • 5.1 ± 2.4


2.1 ± 1.8

  • W = 211.
  • p = 0.002


Dizziness

  • n = 8
  • 5.1 ± 2.5


2.1 ± 1.8

  • W = 121.5
  • p = 0.008


Speech problems

  • n = 8
  • 4.5 ± 3


2.1 ± 1.8

  • W = 117.
  • p = 0.02


Somnolence

  • n = 8
  • 5.3 ± 2.6


2.1 ± 1.8

  • W = 123.5
  • p = 0.006


Fatigue

  • n = 6
  • 2.9 ± 2.5


2.1 ± 1.8

  • W = 70.
  • p = 0.2


Ataxia

  • n = 5
  • 5.2 ± 1.7


2.1 ± 1.8

  • W = 72.5
  • p = 0.008


  • Increased
  • seizure frequency


  • n = 5
  • 4.4±1.7


2.1 ± 1.8

  • W = 69.
  • p = 0.02


Paresthesia

  • n = 5
  • 5.1 ± 1.4


2.1 ± 1.8

  • W = 71.
  • p = 0.01



Table 6 Comparison of TPM serum concentration and dose for patients with versus without adverse events at the first examination.

Without adverse events (n = 13 pat.)

  • With adverse events
  • (n = 29 pat.)


Wilcoxon rank sum test

TPM serum concentration (μg/mL)

3.2 ± 3.6

5.2 ± 4.2

  • W = 199.
  • p = 0.02


TPM dose (mg/kg)

2.1 ± 1.8

4.2 ± 2.8

  • W = 185.
  • p = 0.006



Table 7 Results of the logistic regression analysis to examine the influence of the co-medication on the emergence of “abnormal thinking” (the anticonvulsive co-medication, the TPM serum concentration and the TPM dose were used as effects).

Odds ratio

95%-CI-Limits

Model 1

Total number of anticonvulsants

0.8

0.2 – 2.2

TPM serum concentration

1.0

0.8 – 1.3

TPM dose (mg/kg)

1.5

1.03 – 2.2

Model 2

Number of markedly sedating anticonvulsants

2.2

0.4 – 10.6

Number of moderately sedating anticonvulsants

0.8

0.2 – 3.5

Number of mildly sedating anticonvulsants

0.6

0.1 – 3.1

Number of remaining anticonvulsants

0.2

0.01 – 2.5

TPM dose (mg/kg)

1.6

1.03 – 2.6

TPM serum concentration

1.0

0.79 – 1.4

Discussion

The proportion of patients with adverse events likely to be related to TPM was high in our study. At the first clinical examination, 29 out of 42 patients (69%) showed at least one adverse event; 33 out of 42 patients (79%) had an adverse event at least once during the study period. According to the different reports in the literature regarding children and adult patients with epilepsy, the frequency of at least one adverse event with TPM as add-on-therapy ranges from 17% (Schreiner 2003b) to 73% of patients (Rosenfeld et al. 1997). Cognitive or behavioural side effects were reported in almost 50% of patients in the study of Kellett et al. (1999), and most of these patients were receiving combination therapy. Kellett et al. (1999) supposed that their figures could be an underestimation of the true incidence of side effects as they did not specifically enquire about adverse events in all patients. The observation of Diaz-Obregon et al. (Diaz-Obregon et al. 2002) that neuropsychological function is not influenced by treatment with TPM is in contradiction with other results in the literature (Kockelmann et al 2004, Ortinski and Meador 2004). The frequency of adverse events with TPM monotherapy is lower than with combination therapy (Reife et al. 2000, Ben-Menachem et al. 2003, Gilliam et al. 2003, Privitera et al. 2003, Schreiner 2003a, Schreiner 2003b); however, the spectrum of adverse events seems to be identical (2003). In the TPM monotherapy studies of Gilliam et al. (2003), the number of adverse events occurring in more than 10% of the patients was considerable however, (paresthesia in 13% of patients with a low TPM dosage, and in 35% of patients with a higher TPM dose).

Some studies report a relationship between the frequency of adverse events caused by TPM and the duration of the TPM treatment; the adverse events occurring at the beginning of the treatment are assumed to disappear, at least partially with time (Reife et al. 1995b, et al. 1997, Gilliam et al. 2003). We did not observe this effect in our patients; at the first clinical examination they had been treated with TPM for six months (median, range 0.3 to 50 months), the percentage of patients with any adverse event did not decrease during the subsequent study period ( (figure 2) ).

In our study, the most frequent adverse events, i.e. adverse events occurring in at least 10% of the patients, were, at the first examination (in alphabetical order): ataxia, impaired concentration, dizziness, speech problems, abnormal thinking, and weight loss. During the whole study, the most frequent adverse events were (in alphabetical order): ataxia, impaired concentration, dizziness, exhaustion, increased seizure frequency, speech problems, paresthesia, somnolence, abnormal thinking, fatigue, weight loss. Obviously, during the course of the study, complaints of a sedating effect of TPM became more frequent. The most frequent adverse events at the first examination – abnormal thinking, impaired concentration, weight loss – remained the most frequent adverse events during the whole observation period.

The spectrum of adverse events associated with TPM in our study, corresponds, to a high degree, to the results of other studies in adults and children with TPM both as add-on-therapy and as monotherapy (Kellett et al. 1999, Sziklas et al. 1999, Thompson et al. 1999, Abou-Khalill et al. 2000, Arroyo et al. 2000, Biton et al. 2001, Contin et al. 2002, Guberman et al. 2002, Jette et al. 2002, Christensen et al. 2003, Dlugos et al. 2003, Dodson et al. 2003, Gilliam et al. 2003, Krakow et al. 2003, Mikaeloff et al. 2003, Reith et al. 2003); this is especially true for the comparison with the results of Reife et al. (1995b, 2000) in 5, double-blind, placebo-controlled parallel group studies. Their results were the basis of our list of questions regarding adverse events. In the study of Reife et al. (Reife et al. 1995b), adverse events related to the CNS were also the most frequent adverse events. In particular, “abnormal thinking” was also the most frequent adverse event. In the study of Krakow and Schreiner (2003) in patients on polytherapy however, adverse events related to the CNS were much less frequent.

The fact that the results of Reife et al. (Reife et al. 1995b) are similar to ours is probably caused by our choice of questions concerning the adverse events; the frequent adverse events shown in their study were also investigated systematically in our own study, whereas other possible adverse events (e.g. irritability, elevated mood, nausea or upper respiratory infection), which had a relatively high frequency in other studies (Arroyo et al. 2000, Montouris et al. 2000, Ritter et al. 2000, Arroyo et al. 2002, Gilliam et al. 2003) were not investigated systematically, and therefore were probably underestimated. The percentage of reported adverse events depends, to a high degree, on the specificity of questions (Grudzinski et al. 1998, Cramer et al. 1999, Kellett et al. 1999, Donati et al. 2000, Gilliam et al. 2004). The dosage schedule in our study was different however, from those in the studies reviewed by Reife et al. (1995b); the initial dosage of TPM and the dosage increments were lower in our study. This is important because it was demonstrated that the frequency of adverse events, and especially of adverse events related to the CNS, depends, to a greater degree, on the starting dose and the escalation regimen rather than the maintenance dosage (Biton et al. 2001, Mula et al. 2003). While Reife et al. (1995b) reported on daily dosages of 200 to 1,000 mg, the median daily dosage in our patients was 250 mg (range 25 to 800 mg), and the increase of dosage was not performed in 200 mg/d increments at weekly intervals, but in 25 (- 50) mg increments at weekly intervals. It should be underlined that in some other studies (Reith et al. 2003, Trimble et al. 2003) there was no clear correlation between adverse events and starting dose or titration schedule.

The determination of the serum concentration of TPM was aimed at finding a better correlation between TPM serum concentration and adverse events than could be expected between such adverse events and TPM dosage.

In our study, the adverse events “abnormal thinking, impaired concentration, weight loss, dizziness, speech problems, somnolence, ataxia, increased seizure frequency and paresthesia” showed a significant positive correlation with the serum concentration as well as with the dosage of TPM. Only the correlation with fatigue was not significant; the possible reason for this result might be the small number of patients. The positive correlation was at least as strong in the case of the dosage as in the serum concentration of TPM. Therefore, the predictability of adverse events is probably not much improved by the availability of the serum concentration of TPM. This assumption is supported by the broad overlapping of TPM serum concentrations accompanied by adverse events and serum concentrations not accompanied by adverse events. This was demonstrated graphically ( (figure 1) ), and by comparing the lowest TPM serum concentration (all dosage) of patients with the adverse events “abnormal thinking, impaired concentration, weight loss” with the highest TPM serum concentration at which the patients did not have these adverse events.

In our study with a TPM serum concentration of up to 4 μg/mL, 15 out of 25 patients (60%) complained of adverse events, with a serum concentration above 4 μg/mL 14 out of 17 patients (82%) complained of adverse events. Therefore, we recommend starting treatment with TPM, with the aim of obtaining a serum concentration below 4 μg/mL, while checking the efficacy of this relatively low serum concentration.

Some authors report on the positive correlation of dosages of TPM and adverse events, e.g. a significant dependence of weight loss on the TPM dosage (monotherapy and/or combination therapy) was demonstrated (Nagel et al. 2001, Widmann et al. 2002, Gilliam et al. 2003); an analogous dependence was demonstrated in case of paresthesia, diarrhea and upper respiratory infections (Gilliam et al. 2003). A dependence on the TPM dosage was also demonstrated with the adverse events “difficulty with concentration/attention, speech problems, confusion, nausea, abdominal pain” (Privitera et al. 2003); the degree of statistical significance was not reported.

The dependence of “abnormal thinking” and “somnolence” on the TPM dosage has to be underlined, as it is in contrast to the statement that cognitive impairment (particular verbal fluency, working memory and speed of information processing) appears to be more an idiosyncratic feature of TPM than a dose-dependent one (Kockelmann et al. 2002).

Up to a dosage of 100 mg or 1.5 mg/kg, respectively, the majority of patients did not complain of adverse events (6 out of 8, or 8 out of 12 patients respectively, without adverse events), above this dosage the majority of patients did complain of adverse events (figures 4, 5). With regard to the adverse events, patients should be treated with low dosage TPM. This well-tolerated, low dosage is below the dosage that is usually recommended in the literature. The dosage for seizure prevention is typically given as between 25 and 900 mg (in TPM monotherapy, higher dosages are recommended than when given in combination with enzyme inducing drugs (Mula et al. 2003, Chadwick et al. 1999, Topamax®-Fachinformation/Product monograph 2003, Sander et al. 2003). In the case of TPM as an add-on drug, 200 to 400 mg is the usual recommended daily maintenance dosage (Stefan 2003).

Our recommendation of a target dosage of 100 mg TPM or less (1.5 mg/kg or less) with respect to possible adverse events is not unrealistic. A daily dose of 50 to 100 mg TPM as monotherapy, or as combination therapy has proved to be effective in different studies (Arroyo et al. 2000, Kelly et al. 2002, Privitera 2003, Privitera et al. 2003, Dodson et al. 2003, Prange 2004). Our recommendations are limited by the relatively small number of patients.

The problem of adverse events occurring already at low doses of TPM indicates a narrow therapeutic range. Occasionally, the treatment with TPM has to be stopped because of adverse events, although a high dosage might be very effective in seizure prevention. This situation is similar to the situation with phenobarbital, which seems to be as effective as TPM and whose sedative effect is comparable or equal to that of TPM (Lathers et al. 2003).

Our results indicate that the correlation of several adverse events with TPM dosage seems to be comparable to that with TPM serum concentration. However, this does not mean that monitoring of serum concentration of TPM is superfluous. Nevertheless, it reduces the relevance of the knowledge of the TPM serum concentration for dosage titration. Primarily dosage titration is based on the clinical situation (occurrence of seizures, occurrence of adverse events) Trimble et al. 2003.

Acknowledgements

We are especially thankful to Maria Friese, Dr. med. Kirsten Habedank, Dr. med. Jürgen Kellenbenz, Dr. med. Tim Petersdorff, Claudia Piepenbrink, Jan Raape, Dr. med. Johann Rothmeier (Department of Neurology and Epileptology, DIE WEISSENAU, Ravensburg) for the clinical examination of patients.

The authors also thank the following people for their technical support in this study: Dr. med. Martin Häfele and the staff of the clinical laboratory in Weissenau and Dr. Uwe Jürgens and Mrs. K. Witte-Bölt, Department of Biochemistry, Gesellschaft für Epilepsieforschung, Bielefeld.

Approval was obtained from the Ethics Committee of the University of Ulm (No 14/2000). The work was presented partially as poster at the 44th Annual Meeting of the German Section of the ILAE, Freiburg, May 20 to 22, 2004 (Fröscher et al. 2004).