Low-dose add-on memantine treatment may improve cognitive performance and self-reported health conditions in opioid-dependent patients undergoing methadone-maintenance-therapy

An important interaction between opioid and dopamine systems has been indicated, and using opioids may negatively affect cognitive functioning. Memantine, a medication for Alzheimer's disease, increasingly is being used for several disorders and maybe important for cognitive improvement. Opioid-dependent patients undergoing methadone-maintenance-therapy (MMT) and healthy controls (HCs) were recruited. Patients randomly assigned to the experimental (5 mg/day memantine (MMT+M) or placebo (MMT+P) group: 57 in MMT+M, 77 in MMT+P. Those completed the cognitive tasks at the baseline and after the 12-week treatment were analyzed. Thirty-seven age- and gender-matched HCs, and 42 MMT+P and 39 MMT+M patients were compared. The dropout rates were 49.4% in the MMT+P and 26.3% in the MMT+M. Both patient groups' cognitive performances were significantly worse than that of the HCs. After the treatment, both patient groups showed improved cognitive performance. We also found an interaction between the patient groups and time which indicated that the MMT+M group's post-treatment improvement was better than that of the MMT+P group. Memantine, previously reported as neuroprotective may attenuate chronic opioid-dependence-induced cognitive decline. Using such low dose of memantine as adjuvant treatment for improving cognitive performance in opioid dependents; the dose of memantine might be a worthy topic in future studies.

rather than buprenorphine 8 , and may ameliorate some of its cognitive deficits [9][10][11][12][13] . However, contradictory results have also been reported, viz., that patients on MMT have more neuropsychological impairments than do currently abstinent former opioid abusers 14,15 . Bracken et al. 16 found that opioid-dependent patients on MMT exhibited poor performance in psychomotor speed, selective attention, and impulsivity, which implies a cognitive impairment caused by MMT. In addition, after MMT is discontinued, patients may relapse to opioid use 17 . MMT alone may not be sufficient for treating opioid dependency.
Memantine (M), a non-competitive N-methyl-D-aspartate (NMDA) receptor antagonist, has been used for more than 15 years in Europe to treat a variety of neurological diseases. Relatively recently, it has been widely prescribed (usually $ 20 mg/day) to treat moderate-to-severe Alzheimer's disease (AD) because of its reported benefits for people with AD 18 . Memantine, a widely recognized as NMDA antagonist 19,20 , but its mechanism of action is not fully understood.
Memantine at higher doses (7.5-20 mg/kg; s.c.) attenuates morphine-induced tolerance, physical dependence, and drug-seeking effects in animals 21,22 , its mechanism was believed to be its ability to antagonize NMDA receptors. Moreover, memantine has also benefits on people who display cognitive decline, including brain tumor 23 , Parkinson's disease 24 , alcoholism 25 , and psychiatric disorders 26 .
Memantine has been suggested to suppress the development, expression and maintenance of opioid dependence and to reduce morphine self-administration in laboratory animals 8,21,27 . Afterwards, Bisaga et al. 28 had a pilot study to demonstrate the effect of memantine attenuating the expression of opioid withdrawal symptoms and lowering the severity of precipitated withdraw in large dosage (60 mg, PO qid) and very small sample (five opioid-addicts with cross-over design). Although, afterwards, Krupitsky et al. 29 demonstrated that memantine significantly reduces the withdrawal symptoms in detoxified opiod-dependents and provided a rationale of using the NMDA antagonist memantine to treat opioid addicts. In their study, they used the initial dose of 10 mg/day was gradually increased to the final dose of 30 mg/day over a period of 1 week. Although, significant effective of memantine in previous studies to reduce withdrawal symptoms and severity of opioid dependence, dose-related adverse effects of memantine have been noticed, including dizziness, restlessness, headache, hallucinations, vomiting, hypertonia, and a feeling of pressure within the head 30 . Using such large dose of memantine in the previous study may be considered another issue in the future.
More recently, a relative lower-dose (0.2 mg/kg) of memantine, abolished morphine-induced conditioned-place-preference behavior in rats because of its N-methyl-D-aspartate (NMDA) receptor antagonist was reported 31 for the neuroprotective effects. The dose condition could be a very special and highly original finding which is never reported ever for memantine. The extent of memantine dose in treating opioid-addicts remains unspecific and inconsistent, using such large dosage of memantine similar to Bisaga et al. (2001) and Krupitsky et al. (2002) may not only increase side effects but also confuse the detailed mechanism of medication. Using ultra low dosage might be better to demonstrate the specific effect of the memantine in studying its neuroprotective effect.
Since using large dose may increase side effects and might bias the mechanism of therapeutic effect of using memantine. Previously, an addict-animal model using low dose of memantine was reported the ultra-low dose effect, that addictive behavior was changed (Chen et al., unpublished data). Calculating the dosage in the animal to human, the ultra-dose of 5 mg/day memantine was applied in this study. In this study, we proposed to investigate whether a low dose of memantine would attenuate chronic opioid-induced dependent behavior and have beneficial on cognitive improvement. A doubleblind, randomly stratified clinical trial with add-on low dose of memantine (5 mg/day, oral) or placebo in opioid-dependent users undergoing MMT for 12 weeks was conducted. We hypothesized that (a) chronic opioid users had worse cognitive performance compared to the health controls; (b) methadone-maintenance-therapy may delay cognitive decline; (c) methadone-maintenance-therapy add-on lowdose of memantine may improve cognitive performance compared to MMT only. To verify the effects and mechanism of memantine, a  double-blind, randomly stratified clinical trial was conducted to investigate the effects of low-dose memantine on chronically opioid-dependent patients' cognitive performance and self-reported health condition undergoing MMT.

Results
Demographic data. At the end of this 12-week follow-up study, 81 opioid-dependent participants (MMT1P: n 5 42; MMT1M: n 5 39) completed tasks and their data were analyzed ( Figure 1). Dropout rates of two experimental groups were 26.3% and 49.4%, respectively. The required methadone dose at baseline was not significantly different between the two MMT groups. However, a urine test revealed significantly more amphetamine users in the MMT1P group than in the MMT1M group (Table 1).
Cognitive performance. The HCs performed significantly better than did the MMT1P and MMT1M participants on all cognitive tasks was found in the HCs ( Table 2). After 12-weeks of add-on memantine treatment, a repeated measurement of variance analysis with two factors, Group with two levels (MMT1P and MMT1M) 3 Time with two levels (baseline and 12 weeks), was done to examine whether there was an interaction between Group and Time for each cognitive index (Table 3).
There     Opiate performance index and urinary drug test. After the 12-week MMT treatment, there was no significant difference in the number between the two MMT groups (Table 4). Subsequently, McNemar's test confirmed this result determined using the OTI. The results of the urinary drug test were consistent with the OTI results, that no significantly increase of heroin or other opiate users in the MMT1M group during this 12-week treatment was found. In addition, for the criminality and health indices of the OTI, a trend of improvement was noticed, although not significant (Table 5).

Discussion
Cognitive performance. We found that opioid-dependent patients on MMT were cognitively impaired compared with the HCs on tests of executive function and attention 4,32 related to frontal networks 3 . However, methadone is also an active agonist drug that leads to opioid dependence and opioid abuse. The MMT may cause cognitive impairment, in which previous studies have reported that patients on MMT have more neuropsychological impairments than do currently abstinent former opioid abusers 14,15 .
We found a larger improvement of cognitive and executive function in the MMT1M group than in the MMT1P group. Moreover, on some of the measures the improvements in the MMT1M group have been taken to the levels of HCs, indicating the effect of memantine. This finding implied an association between memantine and cognitive improvement. For the attention performance tested using CPT, however, only the HRT by Block Change index improved in the MMT1M group. This finding may indicate that memantine protects opioid-dependent patients from slower psychomotor speed and increased impulsivity caused by MMT 16 .
Although Saab et al. 33 reported that memantine affected the cognitive flexibility in mice memory. Wesierska et al. 34 , who compared the effects of different doses of memantine on working memory training in adult male rats, reported a dose-dependent and therefore suggested that a mild NMDAR blockade using low-dose memantine would improve spatial working memory in drug-naïve rats undergoing a highly challenging task. For these inconsistent results in animal models, the causality between the therapeutic effect of memantine and cognitive domains should be investigated. Moreover, in our study, adding a low dose of memantine (5 mg/day) to MMT improved cognitive performance; therefore, a large dose (60 mg/ day by Bisaga et al. (1997) and 10-30 mg/day by Krupitsky et al., 2002) used in previous studies to investigate its effectiveness in improving cognitive performance or opioid tolerance/withdrawal symptoms might not be necessary. Bisaga et al. (1997) reports treating opioid-dependent patients undergoing opiate withdrawal with memantine (37 mg/day) but do not suggest that a dose of 60 mg/kg or less is effective for treating addictive behavior. Different doses of memantine might be another topic when investigating the therapeutic effect of MMT1M on cognitive functions.
MMT and the urinary drug test. About half of the MMT1P group (49.4%) and about one-fifth (ca. 26.3%) of the MMT1M group did not complete the tasks, which indicates a positive effect of MMT1M. A higher dropout rate in the MMT1P group than in the MMT1M group was consistent with a meta-analysis 35 on patients with vascular dementia, which implied that memantine had a positive effect on remaining in the treatment. Because methadone is released slowly in unmetabolized metabolized form from liver and I N-demethylated in the chrome P450 enzyme, including CYP2D6 which has been reported tobe varied across different ethnics [36][37][38]    This study has some limitations. Although used to wean opioiddependent patients from heroin and morphine, methadone itself is an addictive synthetic opioid. Whether methadone can be replaced by a non-addictive substance requires additional studies. Our total sample size (n581) had power. 0.8, which detected a medium effect (effect size50.5) for a repeated ANOVA. The dropout rate was higher in the MMT1P group than in the MMT1M group, longer follow-up duration might be needed to confirm the effect of add-on memantine in a longer-term treatment, e.g., 24 or 48 weeks, with a larger sample. The higher dropout rate was found in the MMT1 group which might influence the generalization of our results. Using per protocol analysis may bias the results, for example, participants who experienced sever adverse effect, poor financial support and family support were excluded might make the analysis reach significant level 40 . Moreover, lacking of craving data and prescription of benzodiazepines, the clinical relevance needs further study. Furthermore, because of relatively lower average methadone dose and shorter clinical experience of methadone use in Taiwan compared to other countries 6,41 , our finding might not be applicable to other ethnics.

Methods
Study design. This study, a double-blind, randomly clinical trial. The methods were carried out in "accordance" with the approved guidelines. The all experimental protocols were approved by the Institutional Review Board of National Cheng Kung University. The randomization strategy for treatment was simple randomization using excel's random number generator. Signed informed consent was obtained from all participants before they enrolled in this study. Participants who were opioiddependent or opioid abusers were randomly assigned to one of two groups: methadone-maintenance therapy (MMT) plus add-on placebo (P) (MMT1P: MMT plus one daily placebo capsule) and methadone-maintenance therapy plus add-on memantine (M) (MMT1M: MMT plus one daily 5 mg M sustained-release capsule). Methadone dosage was increased or decreased by 5 mg when necessary in response to each participant's clinical situation. All participants were evaluated before and after the 12-week treatment.
Participants. Opioid-dependent participants who came to the Department of Psychiatry at National Cheng Kung University Hospital were recruited. Each participant, including the healthy controls (HCs), was asked to take the Chinese version of the Mini International Neuropsychiatric Interview (MINI) to screen their psychiatric conditions and confirm that all diagnoses of opioid dependence met the DSM-IV-TR (American Psychiatric Association 2000) diagnostic criteria. The Chinese version of the MINI has been reported as reliable and has been validated 42 . All potential participants with other major and minor mental illnesses, including alcohol abuse disorder, alcohol dependence disorder, and illicit substance-use (other than heroin and morphine) disorders were excluded. All healthy controls were free of any major and minor mental illness.
The pre-MINI exclusion criteria were: being pregnant or nursing an infant; having taken any anti-inflammatory medications within 1 week before the study; having a major Axis-I DSM-IV diagnosis other than morphine dependence; and having a history of uncontrolled major physical increases in aspartate aminotransferase (AST), alanine aminotransferase (ALT), blood urea nitrogen (BUN), and creatinine levels.
Each opioid-dependent participant was given a urine drug test at each visit to determine whether they had used other substances, e.g., amphetamines or morphine, and was asked to perform neuropsychological tasks and complete the Opiate Treatment Index (OTI) interview at baseline and at the end of the 12-week follow-up. In addition, each opioid-dependent participant was required to do cognitive tasks before and after the end of this study. The HCs were required to do cognitive tasks only once-at baseline-as a control comparison. All the HCs were interviewed using the Modified Chinese Version of the Modified Schedule of Affective Disorders and Schizophrenia-Lifetime (SADS-L) 43 to confirm that they had no mental illness or history of substance dependence or substance abuse. We recruited 134 opioiddependent Han Chinese outpatients from the methadone therapy clinic in the Department of Psychiatry at National Cheng Kung University Hospital, and 37 ageand gender-matched HCs.
Wisconsin Card Sorting Test. The Wisconsin Card Sorting Test (WCST) measures the ability to perform certain types of executive functions: categorization, abstraction reasoning, maintaining sets, set switching, strategic planning, and modulating impulsive responding 44 . The test requires that certain cognitive functions be intact, for example, attention, working memory, and visual processing. Participants are required to try out different rules to find a correct method for sorting the cards. The original test used cards that had to be sorted into piles in front of four stimulus cards. The participants were instructed to infer the matching principle from the feedback provided: ''correct'' or ''incorrect'', depending upon whether they correctly guessed the rule. The cards could be matched by number (1, 2, 3, or 4), color (yellow, green, blue, or red), or shape of the symbols (star, triangle, circle, or cross). The rule was applied for a run of trials and then changed without warning (Mei, 1998). The interrater liability is 0.88-0.93, within-rater reliability is 0.91-0.96, and test-retest reliability is 0.57. Performance on the WCST was scored in terms of the total number of errors (TNE), perseverative errors (PE), conceptual level responses (CLRs), number of categories completed (NCC), and trials to complete the first category (TCC).
Continuous Performance Test. The Conners' Continuous Performance Test (CPT) 45 lasts for several minutes to assess the maintenance of focused attention. Optimal performance requires an adequate level of arousal, combined with an element of executive control to resist distraction and inhibit responses to stimuli resembling targets. Respondents are required to press the space bar on a computer keyboard when any letter other than ''X'' appears. The inter-stimulus intervals are 1, 2, and 4 seconds, with a display time of 250 ms.
The CPT produces a standard set of performance measures that include the number of errors of omission and errors of commission.  Opiate Treatment Index. The OTI is an evaluation tool for treating opiate users 46 . The original OTI has a multidimensional structure that measures six independent domains: drug use, HIV-risk-taking behavior, social functioning, criminality, health, and psychological adjustment. It has good inter-rater reliability 47 . The Chinese version of OTI contains drug use, criminality and health domains modified by Zhao et al. 48 .
Statistical analyses. Data are means 6 standard deviation. A repeated measurement with a mixed design, Group (MMT1P; MMT1M) 3 Time at two points during the study (pre-treatment 5 baseline; post-treatment 5 after 12 weeks of treatment) was done to determine whether there was an interaction between these sets of two factors.
McNemar's test was conducted to explore changes in urine drug test values between baseline and the end of this study in each patient group. Significance was set at p ,0.05.SPSS 18.0 was used for all statistical computations.