OUP user menu

The effect of 12-week open-label memantine treatment on cognitive function improvement in patients with alcohol-related dementia

Younghoon Cheon, Jaewoo Park, Keun-Ho Joe, Dai-Jin Kim
DOI: http://dx.doi.org/10.1017/S1461145708008663 971-983 First published online: 1 November 2008


There is compelling evidence that alcohol-induced neurotoxicity is related to glutamate excitotoxicity. It was hypothesized that the low-affinity NMDA receptor antagonist memantine would improve the cognitive function of patients with alcoholic dementia. The aim of this study was to test this hypothesis and to evaluate the effect of memantine on the cognitive improvement of patients with alcohol-related dementia (ARD). The study was designed as a 12-wk open-label study investigating the efficacy of 20 mg memantine, a low-affinity NMDA receptor antagonist, as a treatment for cognitive and behavioural problems in 19 patients with probable ARD according to the criteria for ARD proposed by Oslin and colleagues. The CERAD-K (Consortium to Establish a Registry for Alzheimer's Disease – Korean version) and several clinical assessment scales were completed before and after the 12-wk memantine treatment period. Significant improvements in the mean scores from baseline to final assessment were observed in the Global Deterioration Scale (p<0.05), Brief Psychiatric Rating Scale (p<0.01), Geriatric Quality of Life – Dementia scale (p<0.01) and Neuropsychiatric Inventory (p<0.01) at the end of week 12. The CERAD-K subscales of word list recall (p<0.05), word list recognition (p<0.05), time orientation (p<0.01), drawing an interlocking pentagon (p<0.05), and the total MMSE-K (Mini Mental State Examination – Korean version) scores (p<0.01) of the patients all showed significant improvement following the memantine trial. In this open-label study, patients with ARD treated with 20 mg/d memantine for 12 wk showed improvement on global cognition, quality of life and behavioural symptoms. The result of this study suggests the possible usefulness of memantine for the treatment of ARD. As this was an open-label study, the possibility that participants improved cognitively on their own due to protracted abstinence from alcohol cannot be discounted.

Key words
  • Alcohol-related disorder
  • dementia
  • memantine
  • psychological test


Chronic alcohol use may have direct or indirect neurotoxic effects on the brain that can lead to cognitive impairment. However, the precise relationship between alcohol and dementia remains unclear. Heavy alcohol drinking can lead to brain damage and increase the risk of various types of dementia (Harper and Matsumoto, 2005; Jarvenpaa et al., 2005). Mild to severe neurocognitive impairment has been reported in 50–80% of patients with alcohol-use disorders (Bates et al., 2002).

The prevalence of alcohol-related dementia (ARD) differs according to the diagnostic criteria used and the nature of the population studied. One study found that chronic alcohol use was a contributing factor in 21–24% of dementia cases (Thomas and Rockwood, 2001). Another review suggested that 23% of elderly patients with alcoholism also had comorbid dementia (Oslin et al., 1998). In a study of 130 nursing-home residents with a diagnosis of dementia, 24% were also diagnosed with ARD.

The clinicopathological issues and criteria regarding so-called ‘alcoholic dementia’ remain under debate. Alcohol-induced persisting amnestic disorder, alcohol-induced persisting dementia, and Wernicke–Korsakoff syndrome (thiamine deficiency) may constitute distinct disease entities, but they may also share some common features. Based on this theory, Oslin and colleagues proposed the broader diagnostic scheme and criteria for ARD, which may include cases of Wernicke–Korsakoff syndrome and also other cases of dementia that appear to be alcohol-related (Oslin et al., 1998). The criteria were based on DSM-IV and modelled after the NINCDS/ADRDA criteria for Alzheimer's disease and the California Criteria (Chui et al., 1992; McKahnn et al., 1984). Thus, categories of defined, probable, possible, and mixed ARD were proposed. We have adopted the diagnostic criteria for probable ARD outlined by Oslin and Cary (2003). Table 1 presents the core elements of diagnostic criteria for probable ARD.

View this table:
Table 1

Patients with chronic alcohol dependence show signs of regional brain damage and cognitive dysfunction. Although there is consensus concerning the deleterious effects of chronic alcohol dependence on cognition, the question still remains as to the exact pathophysiology or mechanism of alcoholic dementia. Chronic exposure to ethanol results in the adaptive up-regulation of NMDA receptor sensitivity, which can result in an increased vulnerability to glutamate-induced excitotoxicity (Dodd et al., 2000). The increased glutamate concentration leads to a long-lasting neurotoxic influx of calcium and the subsequent loss of neurons. This excitotoxic insult is one of the most important factors in the mechanism underlying ethanol-induced brain damage (Harper and Matsumoto, 2005).

Several authors suggested that there might be a relationship between repeated episodes of Wernicke–Korsakoff syndrome and ARD. For instance, repeated psychotic episodes and recurrent depressive episodes may lead to cognitive impairment in schizophrenia and depression, respectively. Repeated episodes of Wernicke–Korsakoff syndrome may cause cognitive deterioration and ARD, especially in persons with increased susceptibility (Moriyama et al., 2006; Sher, 2004). Furthermore, poor nutrition, multiple traumas and the malfunctioning of the liver, pancreas, and kidneys may also contribute to the development of ARD.

Despite the clinical importance of ARD, few medical treatments for ARD have been proposed and studied. There are some case reports regarding the efficacy of donepezil in patients with alcoholic Korsakoff syndrome (Angunawela and Barker, 2001; Casadevall-Codina et al., 2002; Iga et al., 2001). Meanwhile, Gibson and Barnaby (2003) and Kim et al. (2004) reported a case of successful ARD treatment with rivastigmine and donepezil, respectively.

Memantine's mechanism of action is a voltage-dependent, low-moderate affinity, uncompetitive NMDA receptor antagonism with fast-blocking/unblocking kinetics (Danysz et al., 2000). The low-moderate affinity NMDA receptor antagonism of memantine modulates the glutamatergic system probably through the Ca2+ influx blockade of NMDA receptors (Kornhuber and Weller, 1997) and is important because other NMDA receptor antagonist, such as amantadine, are high-affinity compounds with neuropsychiatric side-effects. The fast on/off kinetics are also important because this means that memantine sits on the receptor just long enough to prevent pathological activation of the glutamate receptors and then quickly departs when physiological activation of the glutamate receptors are needed (Kumar, 2004).

Memantine treatment also shows positive effects on memory in the treatment of other psychogeriatric disorders with memory loss, such as Alzheimer's disease, vascular dementia, Wernicke–Korsakoff syndrome or Parkinson's disease (Muller et al., 1995). Disturbances in the glutamatergic neurotransmission via NMDA receptors also play an important role in the pathophysiology of alcoholism. There is increasing evidence that uncompetitive NMDA receptor antagonists alleviate some of the effects, including tolerance, withdrawal and sensitization phenomena, induced by ethanol. The effectiveness of acamprosate for alcohol dependence is thought to be related to its ability to modulate the inhibitory gamma-amino butyric acid (GABA) and decrease glutamate activity at NMDA receptors (Littelton and Zieglgansberger, 2003), but it is only a weak NMDA receptor antagonist (Rammes et al., 2001). Memantine is a more specific non-competitive NMDA receptor antagonis, and similar to acamprosate, it has been proposed as an anti-craving and anti-relapse compound (Parsons et al., 1999). Several animal studies have suggested the usefulness of memantine for the treatment of alcohol dependence and ARD due to its anti-craving and cognitive enhancing properties in rats (Holter et al., 1996; Lukoyanov and Paula-Barbosa, 2001; Maler et al., 2005; Nagy, 2004).

Regarding memantine, there was a report that mentioned its efficacy in 16 demented patients with Wernicke–Korsakoff syndrome (Rustembegovic et al., 2003), and a case report of ARD that showed evidence of improvement on the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) test and FDG-PET scores with memantine treatment (Preuss et al., 2001).

It is hypothesized that memantine could be used to improve the cognitive function of patients with ARD. To test this hypothesis, we measured the cognitive functions of patients with alcoholic dementia before and after 12 wk treatment with memantine.



A total of 25 patients with probable ARD according to the criteria for ARD established by Oslin et al. (1998) were enrolled in this study. All patients were without pathological brain lesions (including no evidence of previous stroke, brain atrophy or space-occupying mass on brain MRI) at admission. The samples were from patients in our hospital who were admitted for detoxification, evaluation and treatment for alcohol dependence. To exclude the other causes of dementia, patients who had certain clinical signs and symptoms that suggest vascular dementia or Alzheimer's disease were excluded from our study. The average period between the last alcoholic drink and the start of treatment was 78.9 d (range 62–98 d) and all patients satisfied the following inclusion and exclusion criteria.

Inclusion criteria

(1) Patients with cognitive dysfunction and a Mini Mental State Examination (MMSE) score ⩽23; (2) a past history of significant alcohol abuse (defined as a minimum average of 35 standard drinks per week for men and 28 for women for a period >5 yr); (3) the period of significant alcohol use must have occurred within 3 yr before the initial onset of dementia; (4) patients must have had an abstinence period of at least 60 d.

Exclusion criteria

(1) Any DSM-IV Axis I disorder not defined in the inclusion criteria except for alcohol-related disorders; (2) one or more seizures without a clear and resolved aetiology; (3) the presence of language impairment, especially dysnomia or anomia; (4) the presence of focal neurological signs or symptoms (except ataxia, or peripheral sensory polyneuropathy); (5) a score >4 on the modified Hachinski Ischemia Scale; (6) patients who are pregnant or lactating; (7) prior treatment with memantine; (8) history of severe allergies or multiple adverse drug reactions; (9) clinically judged to be at serious suicidal risk.

Six of the patients were discharged from hospital before completing the study (two patients were discharged against medical advice, two patients were transferred to general hospital due to uncontrollable diabetes mellitus and suspected myocardial infarction, and two were transferred to a sanitarium for long-term care) and therefore excluded from the study analysis because they were not able to control various confounding factors such as alcohol consumption. Nineteen patients remained in the hospital for the duration of the study period. Most of the patients had one or more medical diseases, such as alcohol-related hepatic diseases (42% of subjects confirmed by abdominal ultrasonography), diabetes mellitus (26%), hypertension (21%), respiratory diseases (15%) and others (13%). Clinical data were obtained and assessed on the remaining sample of 19 patients who gave written informed consent for their participation in this study.


The study was planned as a 12-wk, open-label study of the efficacy of 20 mg memantine, a low-affinity NMDA receptor antagonist, as a treatment for cognitive and neurobiological problems in patients with ARD.

After signing the informed consent form, all patients completed a clinical interview that included questions regarding demographic information, income, family living arrangements, medical and psychiatric history, current medications, subjective evidence of memory problems, and ability to carry out everyday activities. The patient's history of alcohol consumption and related problems was obtained by checking medical documentation. In order to obtain more reliable information about the patient's drinking history, we also gathered information from the patient's relatives or caregivers by face-to-face interview in all cases of our study. The patients also completed a physical and neurological examination, together with laboratory tests and various cognitive and behavioural assessments (see below for details) at baseline and at the end of study (after week 12).

The following blood tests were completed before and after the 12-wk memantine treatment period: complete blood count with differential count, alanine transaminase, aspartate transaminase, total bilirubin/direct bilirubin, total protein/albumin, total cholesterol, triglyceride, high-density lipoprotein (HDL) cholesterol, blood urea nitrogen (BUN)/creatinine, homocysteine, folate, and vitamin B12. The serum level of homocysteine, in particular, was measured as a risk factor for dementia. Folate and vitamin B12 were also measured as biological markers for nutritional deficiency.

After the baseline evaluation was complete, memantine treatment was administered at a dosage of 5 mg/d. The dosage was increased weekly, in 5 mg/d increments, to 20 mg/d in week 4. Thereafter, the 20 mg/d dosage remained fixed until week 12. Tolerability was assessed by records of side-effects, withdrawals due to side-effects, laboratory values, and serious adverse events. Memantine was distributed five times by a nurse at weeks 0, 1, 4, 8, and 12. Medication compliance was checked every day during treatment. Patients who were discharged from the hospital during the study period were excluded from the study because they could not be controlled from various confounding factors such as alcohol consumption. The patients should have remained abstinent during the study period, and no other psychiatric medications were permitted except for non-benzodiazepine sedatives (i.e. zolpidem) for sleep control.

Assessment scales

The general cognitive state was determined using a validated Korean version of the Mini Mental State Examination (MMSE-K) and the Clinical Dementia Rating Scale (CDR; Hughes et al., 1982), together with the Global Deterioration Scale (GDS; Reisberg et al., 1982) in order to evaluate the severity of dementia.

The behavioural and psychiatric symptoms commonly observed in patients with dementia were assessed by the Neuropsychiatric Inventory (NPI; Cummings, 1997) and Brief Psychiatric Rating Scale (BPRS).

The Seoul Activities of Daily Living (S-ADL) and Seoul Instrumental Activities of Daily Living (S-IADL) were used to assess daily function (Ku et al., 2004). These scales were developed for reflecting the language expression and culture in Korean elderly.

The S-IADL is comprised of 15 items (household/independent living tasks that include using the telephone, shopping, housekeeping, money management, etc.). The informants rated each item as follows: 0=normal, 1=with some assistance, 2=with much assistance, 3=unable to do. The S-ADL is comprised of 12 items that relate to basic personal activities which include bathing, eating, dressing, etc. Each item was rated as follows: 0=fully independent, 1=help required, 2=unable to perform task.

Finally, the quality of life for each patient was assessed by the Geriatric Quality of Life – Dementia (GQOL-D) scale developed by Lee et al. (2004). This scale contains 15 items modified from the Quality of Life in Alzheimer's Disease (QOL-AD; Logsdon et al., 1999) and the WHO Quality of Life (WHOQOL; WHOQOL Group, 1993).

The Korean version of the Consortium to Establish a Registry for Alzheimer's Disease Assessment Packet (CERAD-K)

The CERAD is the standardized clinical and neuropsychological assessment battery for the evaluation of patients with Alzheimer's disease (Morris et al., 1988). This instrument was designed to create uniformity in the enrolment criteria and assessment methods for Alzheimer's disease, promoting the gathering of reliable information on Alzheimer's disease across research centres in the USA. The CERAD-K is a standardized Korean version of the CERAD having equivalent reliability and validity (Lee et al., 2002).

As measure of verbal production, semantic memory, and language (Verbal Fluency Test), subjects were asked to name as many as animals as possible in 1 min. The subject's score it the total number of animals named. Confrontational naming was assessed by showing subjects 15 line drawings from the Boston Naming Test that represent objects with a range of naming difficulty and frequency of occurrence. Word List Memory measures the subject's ability to learn new verbal information and it was assessed over three learning trials with a 10-word list; in each trial, subjects read the list of common nouns aloud, then immediately remembered as many words from the list as possible. After a brief delay with a figure construction distraction task, subjects were again asked to recall the 10-item word list, which is the Word List Recall (Delay) score. Word List Recognition test counts the number of 10 words presented in the word list memory task correctly recognized (Rec–yes). These words are presented amidst 10 distractor words. The number of distractor words correctly identified (Rec–no) is also counted.

Between the immediate and delayed word list recall, subjects copied four line drawings as a measure of Constructional Praxis. Constructional Praxis Recall measures recent memory and visuospatial function. It requires the subject to copy the memorized drawings among the preceded drawings in the constructional praxis test. The Trail Making is test of visual conceptual and visuomotor tracking. Part A (Trails A) measures attention and entails connecting randomly ordered numbers by drawing a line in sequence and has a strong motor speed and agility component. Part B (Trails B) measures executive function and entails connecting numbers and letters in alternating order and adds a strong complex set-shifting component reflecting mental flexibility. The psychologist stopped each test after 5 min. The times to complete each of the tasks were used as the two measures for the analyses. The Mini Mental State Examination (MMSE) is also included in this battery.

Statistical analysis

All statistical analyses were carried out with the Statistical Package for the Social Sciences version 11.0 for Windows (SPSS Inc., Chicago, IL, USA). Changes on the components of the CERAD-K and other scales from baseline were tested using a paired t test. A result was considered to be significant when the statistical test yielded a two-tailed probability (p value) of ⩽0.05.


Demographics, alcohol use history, and laboratory findings

Nineteen of the original 25 patients completed the research assessments at the end of the clinical trial. All completing the study were completely abstinent from alcohol because they were in a closed ward of the hospital during the period of study. All patients tolerated memantine well and there were no marked adverse events.

The mean age was 57.37 yr, and the patients were predominantly men (94.7%). Approximately 63% (12/19 patients) of patients had a family history of alcohol dependence, but the percentage of family history of dementia was relatively low (16%, 3/19 patients). Most patients were unemployed (84.2%) and were of lower to middle socioeconomic status with moderate to severe alcohol dependence. The mean number of years of problem drinking was 11.04±10.2 yr and the number of previous alcohol treatments was 2.79±1.75. Before admission, the patients consumed an average 13.09 drinks per drinking day, with 5.16 drinking days per week. Basic demographic information and drinking histories of the patients are included in Table 2.

View this table:
Table 2

The mean values of biological markers of drinking at study entry were within the normal range except for the elevated level of serum homocysteine (Table 3). The mean level of homocysteine was 21.84 µmol/l (reference value 5–15 µmol/l). Hyperhomocysteinaemia was observed in 12/19 patients (63.1%). No evidence of folate or vitamin B12 deficiency was found. All of the patients were within the normal ranges in the levels of folate and vitamin B12. Nine out of 19 cases had an elevated mean corpuscular volume, but the mean value of this level was in the normal range. Low HDL-cholesterol levels were found in 13/19 patients (68.4%) and only one patient had elevated levels of total cholesterol. The mean values of liver function tests were in the normal range, although eight patients were diagnosed as having alcohol-related hepatic diseases. No significant changes in laboratory findings were observed at the end of the study.

View this table:
Table 3

Neuropsychological assessments

All of the scores on each of the assessment scales improved after the memantine treatment (Table 4). Significant improvements in the mean scores from baseline to final assessment were observed at the end of week 12 on the GDS (p<0.05), BPRS (p<0.01), GQOL-D (p<0.01) and NPI (p<0.01) scales. Among the NPI subscales, improvements were primarily seen in delusion, sleep disturbance and eating habits. The scores on the other NPI subscales also showed improvement, but it was not statistically significant. In general, memantine reduced cognitive behavioural problems and improved subjective feelings of wellbeing measured by GQOL-D.

View this table:
Table 4

Neuropsychological evaluation using the CERAD-K

The results of the descriptive analysis showed that either the total score or the individual scores on each of the subscales in the CERAD-K improved after 12 wk of treatment with memantine (Table 5). Statistically significant improvement was observed in word list recall (p<0.05), word list recognition (p<0.05), total MMSE-K score (p<0.01), orientation in time on the MMSE-K (p<0.05) and the interlocking pentagon (p<0.05) subscale of the MMSE-K. The mean scores of verbal fluency, Confrontational naming, word list memory, constructional praxis, constructional praxis recall and trail making A were improved, but these improvements were not statistically significant. The mean score on the MMSE-K was especially elevated from 17.21 to 19.58.

View this table:
Table 5

In our study, 12/19 patients were unable to complete Trail B. Thus we analysed only the result of Trail A and excluded the result of Trail B. Trail B has a more complex physical layout, increased task demands, and it is more difficult to process and complete than Trail A (Arnett and Labovitz, 1995; Gaudino et al., 1995). In a previous study of the trail making test of Korean elderly, large numbers of failures (nearly half of candidates) in completing Trail B were reported (Seo et al., 2006).


Several well-controlled studies have demonstrated the improvement of cognitive function after the cessation of alcohol drinking in non-demented alcoholics (Brandt et al., 1983; Grant et al., 1984; Rourke and Grant, 1999). Oslin and Cary (2003) attempted to validate the diagnostic criteria for ARD patients and reported that ARD patients showed a lesser decline in cognitive functions than did Alzheimer's disease patients. They also found that the level of cognitive functions in ARD patients remained stabilized for 2 yr. Improvement in cognitive function, or at least in the lack of a progressive cognitive deficit, is one of the major factors used to determine whether a patient has ARD or Alzheimer's disease (Smith and Atkinson, 1997).

The study's research group consists of patients with a prominent history of alcohol use that showed neither the signs and symptoms of Wernicke's encephalopathy nor brain dysfunctions on MRI tests. Furthermore, our laboratory findings regarding the levels of folate and vitamin B12 measured at the beginning and the end of our study were in the normal range, which suggests that the possibility for brain pathology caused by malnutrition can be excluded to a certain degree. All of the evidence observed in the present study indicates that the cognitive dysfunctions seen in the patients who participated in the study were primarily caused by alcohol use.

Memantine may cause several adverse effects (dizziness, constipation, headache, allergic reactions, etc.). No marked adverse effects that could be attributed to the memantine treatment have been reported among the 19 patients up to the end of the study period. However, this could be interpreted more as a case where patients with cognitive dysfunctions fail to sense and complain of uneasiness than demonstrating there are no adverse effects of memantine. Furthermore, no significant changes in the laboratory findings, such as in liver function, were observed after the 12-wk period of memantine treatment. These results suggest that treatment with memantine might be safe and tolerable for patients with ARD.

One of the most striking features of the laboratory findings was the incidence of hyperhomocysteinaemia (12/19 patients, 63.1%). With regard to the level of folate and vitamin B12, the patients who were included the study had no evidence of folate or vitamin B12 deficiency at admission and baseline of the study. Therefore we did not need to prescribe any medication for those deficiencies. In several studies, elevated homocysteine levels have been associated with an increased risk of atherosclerotic sequelae (including death from cardiovascular causes, coronary heart disease, carotid atherosclerosis, and clinical stroke) and dementia (Sachdev, 2005; Seshadri et al., 2002). Homocysteine acts as a partial or complete agonist at the glutamate and glycine binding sites within the NMDA receptor complex, and plays a role in a shared biochemical cascade involving overstimulation of NMDA receptors, increased oxidative stress, DNA damage, the triggering of apoptosis and excitotoxicity, all of which are important mechanisms in neurodegeneration (Bleich et al., 2004; Mattson and Shea, 2003). There is also growing evidence that chronic alcoholism is associated with hyperhomocysteinaemia (Bleich et al., 2000a; de la Vega et al., 2001). It has been suggested recently that alcoholism-induced hyperhomocysteinemia may lead to alcohol-withdrawal seizure, alcohol-related cerebral atrophy and cognitive impairment (Bleich et al., 2000b, 2004; Robinson et al., 2005; Wilhelm et al., 2006). In the metabolism, homocyteine is either trans-sulfurated into cystathionine or it is remethylated to methionine by methionine synthase. In the case of actively drinking alcoholics, hyperhomocysteinaemia may be due to reduced intracellular folate, with some contribution from reduced activity of the enzyme methionine synthase caused by acetaldehyde, a breakdown product of ethanol (Barak et al., 2001). Moreover, alcohol consumption leads to a disturbed permeability of the blood–brain barrier. Thus, in hyperhomocysteinaemic patients, disruption of the blood–brain barrier results in exposure of the brain to near the plasma level of homocysteine, leading to NMDA medicated excitotoxicity (Lipton et al., 1997). The findings of our study also suggest the role of homocysteine as a risk factor for ARD; however, further study is required to confirm these findings.

The results of the study also show that the memantine treatment significantly increased the total scores on the GDS, BPRS, GQOL-D and NPI, improving not only the cognitive functions of the patients, but also their subjective quality of life. On the NPI, significant improvements were observed in the total scores and subscores, such as sleep disturbance and eating habits (p<0.01). In Alzheimer's disease patients, effects of memantine on behavioural symptoms measured by the NPI were reported by two randomized, controlled studies (Gauthier et al., 2005; Peskind et al., 2006). In those studies, memantine treatment showed a significant beneficial effect in comparison to placebo treatment in the NPI total score and agitation/aggression domain. The result of our study can be attributed to either the effect of memantine itself or the stabilization of brain function caused by the abstinence period subsequent to relatively long-term admission.

Global cognition measured by the MMSE improved by 2 points (from 17.21 to 19.59) in the present study. This result might be a relatively small improvement when compared to the finding of Preuss et al. (2001) that reported a 4-point increase in the MMSE after 5-wk treatment with memantine for patients with ARD.

Other case reports on treatment with donepezil and rivastigmine showed improvements between 2 and 7 points in the MMSE within a few weeks or a period of 3 months (Angunawela and Barker, 2001; Casadevall-Codina et al., 2002; Cochrane et al., 2005; Iga et al., 2001; Kim et al., 2004). Although it can be assumed that an extended period of assessment may have increased the credibility of the evidence on the effect of memantine, the study's 12-wk period of assessment is still thought to be sufficient to measure the effectiveness of memantine in patients with ARD.

According to the results of the present study, the cognitive function of patients with ARD who were given memantine tended to increase on the CERAD-K test. Statistically significant changes were observed in MMSE scores of word list recall (p<0.05), word list recognition (p<0.05), time recognition (p<0.05), ability to draw an interlocking pentagon (p<0.05) and in the total MMSE score (p<0.01). This result was similar to that of a previous study reported by Preuss et al. (2001) who reported the improvement of MMSE scores together with verbal fluency, word list recall and drawing in the CERAD-K after a 5-wk period of memantine treatment. In the present study, CERAD-K verbal fluency was not significantly improved.

Due to the nature of the study structure in which identical tests were conducted both in the baseline measurement and in the final measurement, these results cannot exclude the possibility that the learning effect interfered with the test result. However, as shown in Table 2, while the word list memory scores did not show statistically significant improvements, the word list recall and word list recognition scores did show statistically significant improvements. This increases the possibility that the efficiency of the recall and recognition processes, not merely the learning effect, had significantly improved.

We also suggested that the results were not primarily due to the learning effect after repeated measurements since the majority of patients did not remember having taken the test before, and many of them could not recognize the psychologist that performed the assessments. Moreover, as shown in the change in the other clinical assessment scales for cognitive and behavioural functions, the general quality of life and clinical manifestations prominently improved with the improvement of CERAD-K and MMSE scores.

In order to control for confounding factors of any kind of alcohol consumption, the study patients should have been abstinent for a period of 60 d prior to inclusion, and they were admitted to a detoxification clinic. Due to the inclusion criteria, we might control the positive effects of stabilization from acute alcohol intoxication. However, the results of our study can not completely rule out the possibility that the favorable effects of memantine were derived from long-term abstinence of alcohol and not from the memantine itself.

The patients' range of age of onset, and number of drinks per occasion were quite wide, and the number of subjects were relatively small (19 patients, only one woman was included in this study). It thus appears that our patients do not fully represent the population of ARD sufferers.

Most of patients in our study had physical problems. We can not completely rule out the possibility that these somatic diseases caused by chronic alcohol ingestion may influence the outcome of our study. However, most of the diseases were controlled well and stabilized during the period of study. Those patients who had serious or progressive medical problems requiring intensive medical attention were excluded before entry in our study; therefore, we suggest that the effect of physical problems on this study is relatively low.

Although it is frequently reported that decline in cognitive function may be caused by dementia and amnestic disorder attributable to chronic alcohol ingestion, there is a great lack of well-controlled and large sample-based studies on the related field while a number of such studies has rigorously been made on Alzheimer's disease or vascular disease. Such trend in clinical studies of ARD can be attributed to the difficulties of diagnosing ARD and also controlling alcohol drinking as one of key variables during clinical experiment.

A large number of clinical trials have been performed with the aim of treating cognitive impairment cases that were caused by chronic alcohol consumption, these trials reported that abstinence and an ample supply of nutrition may prevent alcohol-induced cognitive impairment and also improve cognitive function to a certain degree (Bartels et al., 2007; Bergman et al., 1998; Martin et al., 1998).

With regard to treatment for alcohol-induced cognitive impairment, there have been a great volume of cases that report the use of memantine (Rustembegovic et al., 2003), donepezil (Angunawela and Barker, 2001; Cochrane et al., 2005; Iga et al., 2001) and fluvoxamine (Martin et al., 1995) for the treatment of alcoholic amnestic disorder and alcoholic Wernicke–Korsakoff syndrome. However, the study of pharmacological treatment for ARD is extremely rare. To our knowledge, there exist only three cases on treatment for ARD, each of which introduced improvements through prescription of memantine (Preuss et al., 2001), donepezil (Kim et al., 2004), and rivastigmine (Gibson and Barnaby, 2003).

Consequently, our study on 19 ARD patients possesses a great significance in that it is the first study conducted on a larger sample of patients. However, more large-scale studies to identify the one with the highest efficacy among the three major cognitive enhancers for ARD will still need to be conducted.

There are a number of studies that show memantine's efficacy for treating Alzheimer's disease, vascular dementia, Wernicke–Korsakoff syndrome (Kumar, 2004) and HIV-induced cognitive impairment (Schifitto et al., 2007). Moreover, it is worth noting the recent findings on memantine's efficacy for alcohol dependence. As the NMDA receptor is involved in controlling alcohol dependence pathophysiology, there have been a series of studies on NMDA receptor antagonists such as acamprosate and memantine (Gass and Olive, 2008; Lovinger, 2004).

Many pharmacological trials of memantine on animals report that memantine possesses an anti-craving effect of alcohol via negative modulator action (Holter et al., 1996; Rammes et al., 2001) and reduces ethanol-reinforced behaviour (Piasecki et al., 1998), ethanol-withdrawal symptom and cognitive deficit (Lukoyanov and Paula-Barbosa, 2001). While a series of recent studies on patients with alcohol dependence show somewhat contradicting conclusions (Arias et al., 2007; Evans et al., 2007; Krupitsky et al., 2007), a great deal of researchers' attention and interest still remains on memantine and other NMDA receptor antagonists' efficacy on alcohol dependence (Maler et al., 2005; Nagy, 2004).

In the cases of ARD patients where abstinence is believed to play a critical role in preventing chronic cognitive impairment, memantine, already proven to be effective for improving cognitive function and suggested to be effective for treating alcohol dependence, is also expected to improve treatment of ARD patients.

The findings of our study can not yet be generalized. The study has certain limitations in that it was conducted as an open-label trial and despite consistent efforts to employ structured Oslin's criteria and additional inclusion/exclusion criteria throughout our experiment, there is still a possibility that other cognitive disorders were not completely excluded due to the characteristics of ARD that make confirmed diagnosis difficult. Although the patients in our study were cognitively stabilized from the acute effect of alcohol and had an abstinent period of at least 60 d, we can not rule out the possibility that the patients simply improved over time. Despite the aforementioned limitations that restrict generalizations of our findings, our study intends to present opportunities for applying memantine's pharmacological characteristics to future ARD treatment and emphasizes subsequent needs for more structured studies to prove our current findings.


This study is an open-label clinical trial approved by KFDA (Korea Food & Drug Administration) and financially supported by Lundbeck.

Statement of Interest



View Abstract