Cognitive changes in topiramate-treated patients with alcoholism: A 12-week prospective study in patients recently detoxified


Manit Srisurapanont, MD, Department of Psychiatry, Faculty of Medicine, Chiang Mai University, Muang, Chiang Mai 50200, Thailand. Email:


Aims:  The aim of this study was to determine the 12-week cognitive changes in topiramate-treated patients recently detoxified from alcohol.

Methods:  Participants were inpatients with DSM-IV alcohol dependence. All of them were discharged within 14 days after the initiation of topiramate treatment. The topiramate dose range was 50–300 mg/day. The Montreal Cognitive Assessment (MoCA) was used on day 0, day 29, day 57, and day 85. Differences of the MoCA total and seven subtest scores among four time-points were compared.

Results:  Thirty-eight participants (36 men and two women) had a mean ± SD age of 43.1 ± 8.6 years old. At enrollment, they were abstinent for a mean ± SD of 11.5 ± 5.3 days. Five, one, and three patients dropped out of the study on day 29, day 57, and day 85, respectively. On day 85, the mean ± SD dose of topiramate was 253.1 ± 60.8 mg/day. Alcohol consumption decreased drastically during follow up. At each time-point, 75%–80% of the participants were continuous abstainers. The mean ± SD MoCA total, language subtest, and delayed recall subtest scores increased significantly from day 0 to day 85, from 22.0 ± 4.7 to 24.7 ± 3.4 (P < 0.01), from 1.1 ± 1.0 to 1.3 ± 1.0 (P = 0.03), and from 2.7 ± 1.7 to 4.1 ± 1.0 (P < 0.01), respectively.

Conclusion:  Topiramate-treated patients recently detoxified from alcohol usually have an improvement of their cognitive function, especially in the language and delayed recall domains. This phenomenon may be caused by the greater influence of cognitive recovery associated with decreased drinking as compared with topiramate-induced cognitive impairment.

TOPIRAMATE IS A sulphamate fructopyranose derivative that facilitates γ-amino-butyric acid function and antagonizes glutamate activity at α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid and kainate receptors.1,2 Some experts, therefore, postulate that topiramate has a potential to decrease mesocorticolimbic dopamine activity after alcohol intake and to antagonize chronic changes induced by alcohol at the α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid and kainate glutamate receptors.3

Several lines of evidence have shown that topiramate is effective and safe for the treatment of alcoholism. In comparison to placebo controls, topiramate-treated patients are likely to have fewer alcoholic drinks per day, fewer drinking days, and fewer heavy drinking days.3,4 No serious adverse event of topiramate was found in these studies. In a recent randomized-controlled trial, topiramate was effective in the respects of time to first relapse, abstinence duration, weeks of heavy drinking, and percentage of subjects abstinent at 4 weeks.5 Note that, in this later study, as an agent approved for the relapse prevention of alcoholism, naltrexone was not found to be effective as compared with placebo. In this later study, except paresthesia, the adverse effects of topiramate and naltrexone were comparable.5

Although the mechanisms are not yet known, cognitive dysfunction is a well-recognized side-effect of topiramate. In a case series of six epileptic patients treated with topiramate in which topiramate-induced cognitive impairment was improved by donepezil, the authors postulated that central cholinergic malfunction might play a role.6 In neurological patients, almost a third of them may have mental slowing and word-finding difficulties.7 This dose-dependent adverse event is prominent not only in neurological patients (e.g. migraine, epilepsy) but also in healthy volunteers.8–10 In a randomized, placebo-controlled trial of topiramate for alcoholism, concentration difficulty was more common in topiramate-treated patients (14.8%), as compared with the placebo-treated group (3.2%). As concentration is a part of cognitive function, it is possible that topiramate also causes cognitive impairment in alcoholic patients.

Cognitive deficits are prevalent in people with chronic excessive alcohol consumption. These can be detected during the detoxification from alcohol, as well as short- and long-term alcohol abstinence.11,12 In severe cases, gross deficits may be evident, including Wernicke's encephalopathy, Korsakoff's syndrome, and alcohol-related dementia. For mild cases, deficits may be seen only in some areas, e.g. problem-solving, verbal and non-verbal abstraction, learning and memory. Even in apparently clinically healthy recently abstinent individuals with alcoholism, subtle frontal lobe dysfunction (e.g. visuospatial scanning skills, divided attention) can be detected.13 For an abstinent alcoholic patient with mild cognitive dysfunction, it may take several years before he/she can have a full recovery of cognitive function.14 The cognitive deficits found in patients with alcoholism affect not only their psychosocial function but also their ability to recover shortly or achieve treatment success.15

While topiramate can cause cognitive impairment, rapid recovery from cognitive deficits is relatively common in individuals with alcohol abstinence.16 Taken together, it is not yet known whether the positive effect of cognitive recovery or the negative effect of topiramate-induced cognitive dysfunction will be more influential in alcoholic patients treated with topiramate. We, therefore, carried out this study to explore the cognitive changes in topiramate-treated patients recently detoxified from alcohol.



This 12-week prospective study was carried out in patients recently detoxified from alcohol. Treatment settings included two university hospitals (one in Chiang Mai and one in Bangkok) and a drug dependence treatment center in Chiang Mai. The study was carried out in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans. The study protocol was approved by the Ethics Committee for Human Research or the Institution Review Board responsible for each participating site. Written informed consent was obtained from each participant after the study details had been fully explained.


Potential participants were male or female inpatients aged between 18 and 60 years old with alcohol-related disorders. Inclusion criteria were: (i) DSM-IV alcohol dependence; (ii) >1 week of ≥35 standard drinks in a man or ≥28 standard drinks in a woman during the 4 weeks prior to the admission; (iii) an Alcohol Use Disorders Identification Test (AUDIT) score of 8 or more;17 (iv) mild or no alcohol withdrawal (Clinical Institute Withdrawal Assessment of Alcohol Scale, revised [CIWA-Ar] ≤ 10);18 (v) CIWA-Ar, delirium item ≤1; (vi) CIWA-Ar, psychotic symptoms ≤1; (vii) likely to be discharged within 14 days; (viii) a body mass index ≥18 kg/m2; and (ix) intention to decrease or stop drinking. Exclusion criteria were: (i) previous or current illness of cognitive disorders, schizophrenia and other psychotic disorders, bipolar disorder or antisocial personality disorder; (ii) other substance dependence, except nicotine and caffeine dependence, during 6 months prior to the enrollment; (iii) being treated with antipsychotics, mood stabilizers, anticonvulsants, opioid analgesics, systemic steroids, carbonic anhydrase inhibitors, hydrochlorothiazide, metformin, pioglitazone or disulfiram; (iv) low risk of suicide during 4 weeks prior to the enrollment; (iv) physical illnesses, including narrow angle glaucoma, renal impairment, urinary stone, and epilepsy; (v) unstable medical conditions; and (vi) pregnancy and breast-feeding.

Drug administration and concomitant treatment

Topiramate was titrated as the dose-escalation schedule used in a previous study.4 However, based on the best judgment of physicians and patients, the doses were more flexible with a dose range of 50–300 mg/day. No dose adjustment was allowed after day 36 of topiramate treatment. Inpatient treatment programs included group therapy, 12-step facilities, and individual counseling for alcohol and drug use. All participants were discharged within 14 days after the initiation of topiramate treatment. Outpatient treatment programs included 2–3 sessions of individual counseling for alcohol and drug use.

Outcomes and assessment instruments

For an eligible patient, topiramate was started on the next day (day 1) after the screening (day 0). Drinking and cognitive function were assessed on day 0, day 29, day 57, and day 85. Drinking outcomes were assessed using the timeline follow-back.19 Cognitive function was examined using the Montreal Cognitive Assessment (MoCA).20 The MoCA is a brief screening tool developed for the assessment of mild cognitive impairment. The MoCA total score (0–30 points) reflects the global cognitive function. In addition, this scale can be divided into seven subtests, including visuospatial/executive (0–5 points), naming (0–3 points), attention (0–2 points), language (0–3 points), abstraction (0–2 points), delayed recall (0–5 points), and orientation (0–6 points). In a recent study, the MoCA showed satisfactory sensitivity and specificity for the identification of cognitive impairment in patients with substance use disorders.21 For the MoCA total and subtest scores, a higher score indicates better performance. To be used with Thai patients, the Thai version of this rating scale obtained from its official website ( was applied in this study.

Statistical analysis

It has been criticized that an intention-to-treat analysis applied in a study of medication-induced cognitive impairment may be problematic.22 Such analysis would include neuropsychological test results from patients receiving only low doses of the medication or exposed to the medication for only short periods of time. Similar to a recent study aiming to assess the dose-dependent fashion of topiramate-induced cognitive impairment,23 this study also used a completer analysis (i.e. analyzing only the data available at each time-point). Friedman's tests were used to assess the differences of MoCA total and subtest scores obtained at four time-points (day 0, day 29, day 57, and day 85). For the significant variables, Wilcoxon signed rank tests were applied to determine the pairs of difference. A value of P < 0.05 (two-tailed) was used to determine the statistical significance. All analyses were performed using spss software, version 17 (spss, Chicago, IL, USA).



Of 40 patients participating in this study, two of them were excluded because they were not discharged within the 14 days of topiramate treatment. A total of 38 participants (36 men and two women) had means ± SD age, age at first drink, and age at first having alcohol use disorders of 43.1 ± 8.6, 18.2 ± 4.4, and 31.0 ± 9.1 years old, respectively. Their mean ± SD number of hospitalizations for alcohol treatment was 3.0 ± 3.8. Thirty-one participants (81.6%) preferred whisky (40% alcohol). Twenty-one patients (55.3%) were taking psychotropic medications (see Table 1).

Table 1.  Characteristics and sociodemographic data of 38 alcohol-dependent patients treated with topiramate
Characteristics and sociodemographic dataValues
Age (years)43.1 ± 8.6
Male: female36:2 (94.7%:5.3%)
Education (years)10.2 ± 4.2
Education <12 years17 (44.7%)
Marital status 
 Married25 (65.8%)
 Single6 (15.8%)
 Divorced4 (10.5%)
 Separated or widowed3 (7.9%)
Alcohol type preference 
 White whisky (40% alcohol)22 (57.9%)
 Colored whisky (40% alcohol)9 (23.7%)
 Beer (5% alcohol)4 (10.5%)
 Home-made alcohol3 (7.9%)
Psychiatric comorbidity 
 Major depressive disorder2 (5.3%)
 Dysthymia1 (2.6%)
Medical comorbidity 
 Hepatitis3 (7.9%)
 Dyslipidemia3 (7.9%)
 Diabetes mellitus2 (5.3%)
 Hypertension2 (5.3%)
Psychotropic medications 
 Trazodone13 (34.2%)
 Fluoxetine5 (13.2%)
 Diazepam2 (5.3%)
 Escitalopram1 (2.6%)
Body mass index (kg/m2)21.4 (4.7)


Five, one, and three patients dropped out of the study on day 29, day 57, and day 85, respectively. Five and two participants were lost to follow up and could not tolerate the side-effects, respectively. Two patients discontinued topiramate due to upper gastrointestinal bleeding and trauma not related to topiramate treatment. For the whole sample, the mean ± SD of study duration was 72.8 ± 23.7 days.

Topiramate and alcohol outcomes

At enrollment (on day 0), the participants had only mild symptoms of alcohol withdrawal. Their mean ± SD CIWA-Ar score was 2.3 ± 2.8. The mean ± SD abstinence duration was 11.5 ± 5.3 days. Table 2 shows the topiramate doses and drinking outcomes at four time-points. The mean ± SD doses of topiramate on day 57 and day 85 were 267.2 ± 50.2 mg/day and 253.1 ± 60.8 mg/day, respectively. At each time-point, approximately 75–80% of the participants were complete abstainers. All drinking outcomes, including % of drinking days, % of heavy drinking days (a day with five drinks or more for men/four drinks or more for women), and γ-glutamyl-transferase (GGT) were decreased drastically from Day 29 until the end of the study.

Table 2.  Topiramate doses, drinking outcomes and cognitive function at admission, day 0, day 29, day 57, and day 85
 At admissionDay 0 (n = 38)Day 29 (n = 33)Day 57 (n = 32)Day 85 (n = 29)Significant difference*
  • *

    Friedman's test for cognitive function only.

  • During 4 weeks prior to assessment.

  • Wilcoxon signed rank tests: Day85 > Day 29 (P = 0.01); Day 57 > Day 29 (P = 0.02).

  • §

    Wilcoxon signed rank tests: Day 85 > Day 0 (P < 0.01), Day 85 > Day 29 (P = 0.02); Day 57 > Day 0 (P < 0.01) and Day 29 > Day 0 (P = 0.01).

  • Wilcoxon signed rank tests: Day 85 > Day 0 (P < 0.01); Day 85 > Day 29 (P = 0.02); Day 57 > Day 0 (P < 0.01); Day 57 > Day 29 (P = 0.01) and Day 29 > Day 0 (P = 0.02).

  • MoCA, Montreal Cognitive Assessment; NA, not available.

Topiramate doses and drinking outcomes
 Topiramate doses (mg/day) 0148.5 ± 8.6267.2 ± 50.2253.1 ± 60.8 
 % of drinking days88.1 ± 23.8NA7.0 ± 25.45.1 ± 18.14.2 ± 18.6 
 % of heavy drinking days86.1 ± 25.8NA4.4 ± 18.13.6 ± 17.73.9 ± 18.7 
 γ-glutamyl-transferase 528.9 ± 851.5107.9 ± 101.565.8 ± 43.173.8 ± 109.7 
 Continuous abstainers 029 (76.3%)24 (75.0%)23 (79.3%) 
Cognitive function (MoCA tests)
 Visuospatial/executive 3.2 ± 1.53.7 ± 1.43.6 ± 1.33.8 ± 1.3χ2 = 6.64; d.f. = 3; P = 0.08
 Naming 3.0 ± 0.22.9 ± 0.43.0 ± 0.23.0 ± 0.0χ2 = 3.67; d.f. = 3; P = 0.30
 Attention 4.7 ± 1.54.9 ± 1.05.2 ± 0.95.1 ± 1.2χ2 = 1.10; d.f. = 3; P = 0.78
 Language 1.1 ± 1.00.9 ± 0.91.4 ± 1.01.3 ± 1.0χ2 = 9.39; d.f. = 3; P = 0.03
 Abstraction 1.1 ± 0.91.3 ± 0.91.2 ± 0.91.1 ± 0.8χ2 = 3.26; d.f. = 3; P = 0.35
 Delayed recall 2.7 ± 1.73.4 ± 1.23.7 ± 1.34.1 ± 1.0χ2 = 16.53; d.f. = 3; P < 0.01§
 Orientation 5.7 ± 0.65.8 ± 0.55.8 ± 0.55.9 ± 0.4χ2 = 1.29; d.f. = 3; P = 0.73
 Total score 22.0 ± 4.723.3 ± 3.924.3 ± 3.224.7 ± 3.4χ2 = 13.52; d.f. = 3; P < 0.01

Cognitive changes

The means ± SD of MoCA total scores increased significantly from 22.0 ± 4.7 on day 0 to 24.7 ± 3.4 on day 85 (χ2 = 13.52; d.f. = 3; P ≤ 0.01) (see Table 2). Post-hoc Wilcoxon signed ranks tests also showed that the scores obtained on day 57 and day 85 were significantly higher than those of day 0 and day 29 (P < 0.05). The means (SD) of MoCA subtest scores for language and delayed recall were significantly different among four time-points (χ2 = 9.39; d.f. = 3; P = 0.03 and χ2 = 16.53; d.f. = 3; P < 0.01, respectively). Post-hoc tests suggested that the scores of both subtests obtained on day 85 and/or day 57 were significantly higher than those of day 29 (P < 0.05). There was a trend that the visuospatial/executive function also improved over time (χ2 = 6.64; d.f. = 3; P = 0.08).


It is possible that both the inpatient detoxification treatment program and topiramate helped decrease alcohol consumption and led to cognitive recovery in this sample. Topiramate-treated patients recently detoxified from alcohol, therefore, usually have an improvement of their cognitive function, especially in the language and delayed recall domains. This phenomenon may be caused by the greater influence of cognitive recovery associated with decreased drinking as compared with topiramate-induced cognitive impairment.

In the first 2 or 4 weeks of alcohol abstinence, the direct effects and the withdrawal symptoms of alcohol may interfere with an alcoholic patient's test performance by reducing motivation and effort. Because these severe and pervasive impairments are highly related with detoxification, neuropsychological assessment (e.g. cognitive function) during this period may be of little value.14 For this study, we decided to start topiramate at the end of the detoxification because this point of time is crucial for the initiation of a relapse prevention program. It has been suggested that relapse prevention medications should always be considered after detoxification.24 In addition, deficits in abstraction, comprehension, and memory frequently found at this stage suggest that cognitive-behavioral treatments for patients recently detoxified may be less effective.25

For the above-mentioned reasons, using day 0 cognitive data as the baseline data might not be justified for the assessment of cognitive changes in patients recently detoxified from alcohol. However, the post-hoc analyses of this study also showed that the MoCA total, language subtest, and delayed recall subtest scores on day 57 and/or day 85 were still higher than those obtained on day 29, which was approximately 40 days after the last drink of most subjects. In a previous study, it was found that the cognitive function of those abstinent for 42 days was similar to their matched controls.16 Taken together, it is possible that in this study, topiramate-treated patients with decreased drinking still had a chance to gain more cognitive function after 40 days of abstinence or 29 days of topiramate treatment.

To our knowledge, this is the first study of topiramate-induced cognitive impairment in patients with alcoholism. In respect to cognitive recovery, these findings are similar to those of a previous study showing that rapid recovery of cognitive deficits can be observed in abstinent alcoholic patients.16 The improvement of language and delayed recall in this sample appears to be in concordance with previous findings. In an inpatient study, Manning and colleagues also found that 4 weeks after detoxification, alcoholic patients had significant increases in performance scores of working memory, verbal fluency, and verbal inhibition.26

Although the topiramate doses given in this study (a mean of 253.1 mg/day) were relatively lower than those administered in some studies, they are the doses commonly used for the treatment of alcoholism. Two previous studies of topiramate-induced cognitive impairment applied this agent at higher doses in adults with partial seizure (mean dose of 414 mg/day)22 and in healthy volunteers (330 mg/day). However, so far, 300 mg/day has been the maximum dose of all randomized, placebo-controlled trials of topiramate for alcoholism.3–5 Taken together with the findings that topiramate doses of 192 mg/day and 384 mg/day can cause cognitive impairment,23 topiramate given in the present study might also cause cognitive impairment in this sample. Nevertheless, the cognitive recovery associated with decreased drinking might be more influential than that of topiramate-induced cognitive impairment. The mixed effect of these two factors, therefore, turned out to be positive.

A major issue of concern for this study is the lack of a placebo-controlled group and we could not determine the magnitude of topiramate-induced cognitive impairment. However, improving the study design by having a placebo-controlled group may also cause another problem. As a placebo may not be as effective as topiramate in decreasing alcohol consumption, cognitive recovery associated with decreased drinking in a placebo-controlled group may be less than that in a topiramate-treated group. If the cognitive recovery rates of both groups are different, the medication-induced cognitive deficits in both groups are still not comparable. Despite the above-mentioned concern, a study comparing cognitive changes between topiramate- and placebo-treated patients with alcoholism is still warranted. Applying some special statistical techniques or subgroup analyses should be considered in balancing the different amounts of alcohol consumption between groups.

Other than the lack of a control group, this study also has other limitations. First, the sample size of this study was relatively small. Type II errors cannot be excluded for many cognitive domains, e.g. the non-significant improvement of visuospatial/executive function (P = 0.08). Second, as alcohol consumption decreased drastically in this sample, the present findings may not be generalizeable to alcoholic patients treated with topiramate who relapse into heavy drinking. Last, as a 12-week study, the mixed long-term effect of alcohol abstinence and topiramate on cognitive function cannot be determined.

Although acamprosate and naltrexone are effective for the treatment of alcoholism, they are not available in some countries, e.g. Thailand. Although topiramate is effective for alcoholism, cognitive impairment associated with this agent may discourage its use in alcohol-dependent patients, who are likely to have cognitive deficits due to chronic excessive alcohol consumption. The present findings should reassure both patients and physicians that the positive effect of cognitive recovery associated with decreased drinking is more robust than the negative effect of topiramate-induced cognitive dysfunction. The risk and benefit on cognitive function appear to be in favor of topiramate treatment.

Topiramate-treated patients recently detoxified from alcohol usually have an improvement of their cognitive function, especially in the language and delayed recall domains. This phenomenon may be caused by the greater influence of cognitive recovery associated with decreased drinking as compared with topiramate-induced cognitive impairment.


This study was supported by grant 52-LR-055 from the Integrated Management for Alcohol Intervention Program (I-MAP), Thai Health Promotion Foundation. The funding body had no role in the decision to prepare and submit this manuscript. We are grateful for the administrative assistance given by Ms Siripan Jaranai.

Drs. Likhitsathian, Uttawichai, Yingwiwattanapong, and Wittayanookulluk have no conflict of interest. Dr. Saengcharnchai has received honoraria and/or travel reimbursement from AstraZeneca, Invida, Janssens-Cilag, and Thai-Otsuka. Dr. Srisurapanont has received honoraria, consultancy fees, research grants, and/or travel reimbursement from AstraZeneca, GlaxoSmithKline, Pfizer, Janssen-Cilag, Johnson & Johnson, Lundbeck, Thai-Otsuka, Sanofi-Aventis, and Servier.