, 201:611

A mu opioid receptor gene polymorphism (A118G) and naltrexone treatment response in adherent Korean alcohol-dependent patients


    • Department of PsychiatrySchool of Medicine, Pusan National University
  • Cheol-Min Kim
    • Haein Hospital
  • Sam-Wook Choi
    • Department of PsychiatryCollege of Medicine, University of Ulsan
  • Young-Myo Jae
    • Department of PsychiatryBongseng Memorial Hospital
  • Hae-Gook Lee
    • Department of PsychiatryCollege of Medicine, The Catholic University
  • Bong-Ki Son
    • Department of PsychiatryHallym University Medical Center, Chuncheon Sacred Heart Hospital
  • Jeong-Gee Kim
    • Department of PsychiatryMaryknoll General Hospital
  • Young-Sung Choi
    • St. Andrew’s Neuropsychiatric Hospital
  • Han-Oh Kim
    • Keyo Medical FoundationKeyo Hospital
  • Seong-Yeon Kim
    • Division of Management Information ScienceDong-A University
  • David W. Oslin
    • Department of Psychiatry, University of PennsylvaniaPhiladelphia VA Medical Center and VISN 4 MIRECC
Original Investigation

DOI: 10.1007/s00213-008-1330-5

Cite this article as:
Kim, S., Kim, C., Choi, S. et al. Psychopharmacology (2009) 201: 611. doi:10.1007/s00213-008-1330-5



Previous studies have demonstrated an association between genetic polymorphisms of the μ opioid receptor gene (OPRM1) and response to naltrexone treatment. The Asp40 variant genotype previously shown to be associated with naltrexone treatment response is known to be relatively common among Koreans.


This study was conducted to prospectively investigate the relationship between genotype and response to open-label naltrexone treatment in Korean alcohol-dependent subjects.

Materials and methods

Sixty-three alcohol-dependent subjects were prescribed naltrexone for 12 weeks in combination with cognitive behavioral therapy. Thirty-two subjects were adherent, taking the medication at least 80% of the treatment days [16 Asn40 (A/A) patients and 16 Asp40 variant (A/G or G/G) patients].


Subjects adherent to naltrexone treatment with one or two copies of the Asp40 allele took a significantly longer time than the Asn40 group to relapse (p = 0.014). Although not significant, the Asn40 group treated with naltrexone had a 10.6 times greater relapse rate than the Asp40 variant group. There was no significant difference between the Asn40 group and the Asp40 variant group treated with naltrexone in rates of abstinence.


These results demonstrating a higher therapeutic effect of naltrexone in Korean alcohol-dependent individuals with the Asp40 variant genotype than the Asn40 genotype are consistent with previous study results in individuals of European descent. This is the first study to examine the pharmacogenetics treatment response to naltrexone in non-European subjects.


NaltrexoneMu opioid receptorPolymorphismKorean alcohol dependent


Alcohol dependence is one of the most disabling health problems worldwide (Murray and Lopez 1996). Among various methods of treating alcohol dependence, pharmacological treatment is increasingly being adopted. Naltrexone, an opioid receptor antagonist, was first demonstrated to be effective in reducing relapse in alcohol-dependent patients by Vopicelli et al. (1990, 1992) and has been shown to be efficacious in a number of studies (O’Malley et al. 1992; Oslin et al. 1997; Chick et al. 2000; Anton et al. 2001; Monterosso et al. 2001; Monti et al. 2001; Morris et al. 2001). The use of naltrexone is based on animal studies that demonstrate the involvement of the endogenous opioid system in alcohol-induced reward (Patel and Pohorecky 1989; De Witte 1996). There are three subtypes of endogenous opioid receptors: mu, delta, and kappa (Wood 1982). Among them, the μ opioid receptor is thought to play an important role in the fluctuation of drinking behaviors in alcohol-dependent patients. Alcohol activates the mesolimbic dopamine reward circuit through μ opioid receptor, promoting excessive alcohol intake behavior (Widdowson and Holman 1992). Naltrexone has been reported to block euphoria or the reward induced by alcohol consumption (Volpicelli et al. 1995; King et al. 1997). In other words, naltrexone helps prevent relapse in alcohol-dependent patients probably by blocking reinforcement of alcohol drinking, which is related to an increase of endogenous opioids.

In most studies investigating therapeutic effects in alcohol dependence, naltrexone has been shown to have meaningful effects compared to placebo (O’Malley et al. 1992; Volpicelli et al. 1992; Oslin et al. 1997; Chick et al. 2000; Anton et al. 2001; Monti et al. 2001; Morris et al. 2001), although a few study results failed to show a significant difference from placebo (Kranzler et al. 2000; Krystal et al. 2001). However, naltrexone does not show therapeutic effects in all alcohol-dependent patients, and it is clear that not everyone feels “high” from an increase of endogenous opioids after alcohol consumption. The fact that naltrexone does not have therapeutic effects in some alcohol-dependent patients has raised the issue of which patients might benefit from its effect. Human laboratory studies have shown that alcohol increases endogenous opioids more in patients with a family history of alcohol dependence (Gianoulakis 1996). In addition, Monterosso et al. (2001) and Rohsenow et al. (2007) reported that therapeutic effects of naltrexone in alcohol-dependent patients are greater in those with a family history of alcohol dependence than in those without such a family history. Such study results suggest that there is a difference in the efficiency of the endogenous opioid system in those with family history of alcohol dependence and those without it, and this ultimately suggests a possible genetic relationship to treatment response.

Sequence changes in the codon of the μ-opioid receptor gene (OPRM1) can cause a significant change in its function (Bond et al. 1998; Zhang et al. 2005). Two genetic polymorphisms, A118G (Asn40Asp) and C17T (Ala6Val), in exon 1 of OPRM1 are drawing the most attention (Oslin et al. 2003). According to Bond et al. (1998), the reactivity of the μ opioid receptor to β-endorphin triples in the presence of the G allele carrier genotype in OPRM1 compared to the A/A genotype, and recently, Zhang et al. (2005) suggested that G allele carrier genotype might be associated with lower OPRM1 messenger RNA (mRNA) and protein expression. In addition, van den Wildenberg et al. (2007) compared the difference of cue-induced craving in 84 A/A genotype alcohol-dependent patients and 24 alcohol-dependent patients with the G carrier genotype and stated that G carrier alcohol-dependent patients had a significantly higher cue-induced craving than A/A alcohol-dependent patients. Kim et al. (2004) demonstrated that Koreans with the G allele had significantly more drinking days than those with the A/A genotype.

Recent study results suggest the possibility of different treatment responses due to opioid receptor blockade based on the OPRM1 genetic polymorphism. Oslin et al. (2003) demonstrated that alcohol-dependent patients with one or two copies of the Asp40 variant who were adherent to medication had significantly better treatment response (73.9%), as measured by the absence of any heavy drinking, than patients with the Asn40 variant (49.0% response). In patients receiving placebo, there was no association between response and genotype, with rates of response similar to those homozygous for the Asn40 allele treated with naltrexone. More recently, results from the NIAAA supported COMBINE study found that 87% of those patients assigned to medical management who completed 16 weeks of treatment and who had one or two copies of the Asp40 allele had a positive outcome to treatment compared to 49% response among those who were homozygous for the Asn40 allele and treated with naltrexone (Anton et al. 2008). However, Gelerntner et al. (2007) found no association between OPRM1 and treatment outcomes among participants in the VA Cooperative study of naltrexone. It is clear from the VA Cooperative study that substantial post-randomization selection bias was present. There was a significant difference in medication adherence between those participating and those not participating in the genetics subcomponent.

Based on the preceding studies, this study aims to prospectively investigate the impact of the A118G genetic polymorphism of OPRM1 in Korean alcohol-dependent patients who were adherent to naltrexone treatment. In contrast to the low allele frequency of the Asp40 allele found in those of European descent (Bergen et al. 1997; Bond et al. 1998; Sander et al. 1998), Koreans have high allele frequency such that half of Koreans will have as least one copy of the Asp40 allele (Kim et al. 2004). Therefore, if the G carrier genotype turns out to be related to treatment response, it is expected to be more helpful in treating Korean alcohol-dependent patients.

Materials and methods

Study subjects

Eligible subjects were recruited from the in- and outpatient programs of one of 15 participating hospitals. Inclusion criteria included having a diagnosis of alcohol dependence according to the Diagnostic and Statistical Manual of Mental Disorder, Fourth Edition (DSM-IV; American Psychiatric Association 1994) as determined by a psychiatric specialist and being Korean and between 18 and 65 years of age and, according to the Timeline Follow-Back (TLFB), having drunk at least 15 standard drinks (SD) per week for 28 days and at least five SD (three SD for women) for more than 1 day a week, immediately before the last admission or the first outpatient visit. Subjects were excluded if any of the following applied: a diagnosis of substance dependence other than alcohol, nicotine, and caffeine based on DSM-IV; use of any opioid in the last 30 days prior to randomization; use of psychotropic medications over the last month; presence of a serious physical illness such as AIDS, acute hepatitis, or severe hepatocyte injury that shows an increase in bilirubin levels; presence of psychosis, bipolar disorder, dementia, suicide ideation, or manic symptoms; or active engagement in other addiction treatment.

Based on the inclusion and exclusion criteria, 72 subjects (including five women) were consented for participation. Nine subjects signed consent but did not start treatment, leaving 63 subjects who initiated naltrexone treatment. Of the 63 subjects, 31 subjects (including five women) did not fulfill the a priori requirement of taking naltrexone at least 80% of the treatment days and were excluded from the final analyses. All in-person-screened individuals signed an informed consent form approved by each site’s institutional review board.

Study method

At the initial screening visit, several study-specific procedures were performed, including the informed consent procedure; demographic assessment of the patients, including gender, age, marital status, academic achievement, and occupation; a medical history; a physical examination; and renal and liver function tests. A drinking history, including the age at which the subject first started drinking, the age at which drinking caused an alcohol-related problem for the first time, average drinking amount per drinking day over the past year, monthly average number of drinking days over the past year, existence of withdrawal symptoms, and family drinking history were assessed. The drinking amount per drinking day during the most recent 4 weeks while not hospitalized was assessed using the TLFB.

On the first study day (baseline), eligible subjects were started on naltrexone 25 mg once a day for the first 3 days and 50 mg for the remaining days of a 12 week trial. Subjects returned 4, 8, and 12 weeks after initiating medication. At each session, drinking was assessed using the TLFB. Adherence to medication was assessed by self report. At baseline and at each study visit, the subject also met with a psychiatrist for a 45-min cognitive behavioral therapy session tailored from the Project MATCH study, which was standardized as four sessions.

DNA analysis

Approximately 10 ml of ethylenediaminetetraacetic acid-treated venous blood was obtained for DNA extraction from each subject. Genomic DNA was extracted from blood samples by standard methods (e.g., Lahiri and Schnabel 1993). A118G polymorphism was genotyped by using the polymerase chain reaction/restriction fragment length polymorphism method of Gelernter et al. (1999). Genotyping was conducted in batches according to each of the studies. For each batch of genotyping, all 72 DNA samples were genotyped in duplicate, and no discordant genotypes were generated.

Statistical analysis

Descriptive statistics for the subjects were summarized. Association of the genotype with adherence was evaluated using Fisher’s exact test. The baseline demographics and clinical characteristics are compared across the genotype using t test and χ2 test. It was hypothesized that a gene by treatment interaction would not be expected in patients non-adherent to treatment. For that reason, the effects of genotypes on rate of abstinence and relapse and on the time to first relapse were analyzed with only adherent subjects. Logistic regression analysis was used to evaluate the effects of ASP40 allele type on relapse and abstinence from drinking, and χ2 test was used to assess the difference of relapse rate and abstinence rate based on ASP40 allele type, and Cox regression analysis was used to investigate the difference of time to relapse. Backward elimination method was used at sls = 0.10 level when selecting variables expected to be significant. All p values were two-sided, and data were analyzed using SAS 9.1. All variables with t statistics or χ2 values having p value of 0.05 or less were considered possibly significant. Relapse was defined as drinking more than five SD per drinking day for men even for once during the trial and over three SD for women.


OPRM1 genotypes and adherence

Of the 63 subjects who initiated treatment (including five women), 25 subjects (including one woman) had the A/A genotype (A/A group hereafter), and 38 subjects (including four women) had the G carrier genotype (G carrier group hereafter). The number of subjects who did not meet the a priori definition of adherence for the 12 weeks of the study period (non-completers, NC hereafter) was 31 (including five women), with nine (including one woman) from the A/A group and 22 from the G carrier group (including four women). Adherence to treatment was not significantly associated with genotype group (df = 1, χ2 = 2.9, p = 0.09). The subjects who were adherent to naltrexone for 12 weeks included 32 men, of which 16 were in the A/A group and 16 were in the G carrier group (Fig. 1).
Fig. 1

Flow chart of subjects. A/A A/A genotype, G carrier A/G or G/G genotype

Demographic characteristics and drinking history

As noted in Table 1, there were no differences between excluded subjects and adherent subjects in average age and academic achievement (years). There were similarly no differences between the excluded group and the adherent group on baseline clinical characteristics. Comparing genotype groups, there were no differences in baseline demographics or baseline clinical characteristics.
Table 1

Baseline demographics and clinical characteristics of the subjects


Excluded subjects (E), (n = 31)

Adherent subjects (n = 32)

Differences in genotype group

Differences between excluded and adherent

A/A (n = 9)

G carrier (n = 22)

A/A (n = 16)

G carrier (n = 16)


Age (year)

44.8 ± 7.0

42.4 ± 9.2

46.3 ± 11.8

49.2 ± 8.4



Education (year)

12.6 ± 3.7

11.8 ± 3.5

11.5 ± 3.9

11.8 ± 4.9



Married (%)

8 (100)

13 (59)

13 (81)

12 (75)



Employed (%)

5 (63)

11 (55)

12 (75)

10 (67)



Clinical characteristics

Age drinking stated (year)

22.3 ± 5.5

20.1 ± 8.5

17.9 ± 2.5

22.0 ± 5.9



Onset age of ARP (year)

35.6 ± 10.3

30.3 ± 9.4

32.8 ± 12.9

39.6 ± 11.3



Drinking day/montha

22.3 ± 8.4

23.5 ± 7.0

17.9 ± 9.2

22.8 ± 7.6



Drinks/drinking dayb

9.1 ± 4.2

10.9 ± 5.6

14.3 ± 7.5

11.2 ± 5.3



Total heavy drinking day 4 weeks before screening

20.0 ± 6.0

18.8 ± 7.8

15.6 ± 6.7

21.5 ± 6.1



History of several alcohol withdrawal (%)

5 (63)

15 (68)

11 (69)

12 (75)



Family history of ARP (%)

4 (50)

15 (71)

11 (69)

8 (50)




28.4 ± 15.1

28.4 ± 15.3

37.6 ± 27.2

32.8 ± 26.7




37.3 ± 22.3

31.2 ± 21.6

37.4 ± 17.4

47.8 ± 66.8




147.8 ± 227.0

166.7 ± 292.9

81.4 ± 81.4

246.1 ± 369.7



Values represent mean ± standard deviations for continuous measures or number (percentage) for categorical measures.

Adherent Those who completed 12-week trial and took 80% or more medication prescribed, Excluded 63 subjects − Adherent, G carrier A/G or G/G genotype, A/A A/A genotype, ARP alcohol-related problem, ALT alanine transaminase, AST aspartate aminotransferase, GGT gamma glutamyl transferase, ns not significant

aAverage drinking day per month for 1 year prior to present admission or first visit

bAverage drinking amount on drinking day for 1 year prior to present admission or first visit

Effects of genotype on the relapse rate, time to relapse, and abstinence rate

The relapse rates of the A/A group and the G carrier group were 37.50% and 18.75%, respectively (Table 2). Logistic regression analysis showed a tendency to affect the relapse rate [OR (A/A vs G carrier genotype) = 10.608, p = 0.072]. Included in the regression analysis were variables that were expected to have significant effects on the occurrence of relapse, including marital status, employment status, and drinks/drinking day (SD) over the past year. The average time to relapse was 59.9 days for the A/A group and 73.3 days for the G carrier group. Cox regression analysis demonstrates that the time to relapse was significantly longer in the G carrier group than in the A/A group [HR (A/A vs G carrier genotype) = 13.623, p value = 0.014; Table 2, Fig. 2].
Fig. 2

Survival analysis for time to first relapse in subjects with G carrier genotype vs A/A genotype [HR (A/A vs G carrier) = 13.623, p = 0.014]. A/A A/A genotype, G carrier A/G or G/G genotype

Table 2

Results from analysis of the effects of genotypes on rate of abstinence and relapse and on the time to first relapse



Genotype main effect


Significant covariates


G carrier

Abstinent rate (%)



χ2(1) = 0.1979


Drinking day/montha

OR (A/A vs G carrier) = 0.657


(95% CI 0.103, 4.192)

Age drinking started

Relapse rate (%)



χ2(1) = 3.228


Drinks/drinking day (SD)b

OR (A/A vs G carrier) = 10.608


(95% CI 0.807, 139.459)


Time to first relapse (day)



χ2(1) = 5.995


Drinks/drinking day (SD)b

HR (A/A vs G carrier) = 13.623


(95% CI 1.684, 110.200)


OR Odds ratio, HR hazard ratio, A/A A/A genotype, G carrier A/G or G/G genotype, SD standard drink

aAverage drinking day per month for 1 year prior to present admission or first visit

bAverage drinking amount on drinking day for 1 year prior to present admission or first visit

The abstinence rates of the A/A group and the G carrier group during the study period were 50.00% and 43.75%, respectively (Table 2). There was no significant difference in the abstinence rate based on genotype [OR (A/A vs G carrier genotype) = 0.657, p = 0.656]. The logistic regression analysis included covariates for academic achievement (years), the age of first alcohol consumption, and monthly average days of drinking in the past year.


Results from this prospective study confirm prior research demonstrating an association between clinical response to naltrexone and having one or two copies of the Asp40 allele for the OPRM1 gene. This implies that naltrexone treatment is more effective in alcohol-dependent Koreans with the G allele genotype than in those with the A/A genotype. The results are significant in those patients who were adherent to treatment. A gene by treatment interaction would not be expected in patients non-adherent to treatment. However, non-adherence to treatment was a significant factor in both genetic groups and clearly impacts treatment response and the interpretation of these findings. These findings may help to explain some of the variability in response to naltrexone seen both clinically and in randomized clinical trials. Important limitations of this study include the lack of a placebo control group, the relatively small sample size of the trial, and the small representation of female participants.

Prior studies by Oslin et al. (2003) and Anton et al. (2008) demonstrated an association between having one or two copies of the Asp40 allele and main outcome measures such as time to relapse, relapse rate and percentages of days abstinent, heavy drinking days, and a good clinical outcome. Our results demonstrated an association with time to relapse but not relapse. While the odds ratio was similar to both the Oslin et al. (2003) and Anton et al. (2008) findings, the sample size of this study was not sufficient to demonstrate significance. Gelerntner et al. (2007) found no association between OPRM1 and treatment outcomes among participants in the VA Cooperative study of naltrexone. An important limitation of these prior studies highlighted by the VA Cooperative study is the substantial post-randomization selection bias present as a retrospective analysis. There was a very significant difference in medication adherence between those participating and those not participating in the genetics subcomponent. As noted by the authors, the sample size was also insufficient to exclude the possibility of an effect of genotype. As noted in this article, the discrepancies between the three studies highlight the dangers of relying solely on post hoc data analysis.

The clinical results from this study and others are consistent with a growing literature on the implications of the Asp40 allele on the function of the mu receptor. Perhaps most compelling is the work done on the translation to mRNA, which resulted in 1.5-fold lower mRNA levels and more than ten-fold lower OPRM1 protein levels in the presence of the Asp40 allele (Zhang et al. 2005). In addition, both Wand et al. (2002) and Hernandez-Avila et al. (2003) found that individuals with one or two copies of the Asp40 allele had greater hypothalamo–pituitary–adrenocortical axis activation induced by the opioid receptor antagonism. In clinical settings, the Asp40 allele was demonstrated to impart increased sensitivity to an agonist when measured by pupillary response (Lotsch et al. 2002). In addition, several studies now demonstrate an association between the Asp40 allele and the increased requirements for analgesia for pain (Klepstad et al. 2004; Chou et al. 2006; Lotsch et al. 2006; Oertel et al. 2006), opioid sensitivity in heroin addicts (Zhang et al. 2007), and greater hemodynamic response in mesolimbic area (Filbey et al. 2008). Moreover, in three different studies, this polymorphism has been linked to clinical response in nicotine dependence, which is another disorder involving the dopamine-opioid reward system (Lerman et al. 2003; Ray et al. 2006; Munafo et al. 2007). In accordance with the hypothesis that certain individuals demonstrate a heightened opioid response to alcohol, the Asp40 allele has also been demonstrated to affect both high and craving states. Ray and Hutchison (2004) demonstrated that the Asp40 polymorphism was associated with greater reports of high and “stimulation” during IV administration of alcohol to heavy or hazardous drinkers, further supporting the implications for alcohol-dependent individuals. This effect was then blocked by pretreatment with naltrexone (Ray and Hutchison 2007). Similarly, Ooteman et al. (2007) and van den Wildenberg et al. (2007) showed that Asp40 allele carriers had greater cue-induced craving when exposed to alcohol.

It is important to consider the role of treatment adherence in defining a phamacogenetic effect. In the present study, we a priori defined adherence as taking medication for 80% of the treatment days. We hypothesized a pharmacogenetic effect only in patients who were adherent to treatment. The rationale is clear: If there is a biological interaction between naltrexone and the mu receptor encoded with the Asp40 allele, one would expect this interaction only in the presence of naltrexone. However, much of the basis of contemporary clinical trial design rests on the concept of intent-to-treat hypotheses. While there were no apparent baseline differences among those that were adherent to medication and those that were not, we must assume that non-adherence is another factor associated with treatment response. The higher order interactions between adherence, response, and genotype are beyond the scope of the present study and will require substantially more patients than included in this report.

In summary, results from this prospective study confirm the possibility of targeting naltrexone therapy for specific genetic phenotypes. The results extend prior studies conducted almost exclusively in patients of European descent to Asians. As noted previously, Asian populations have a greater allele frequency of the Asp40 allele, making them an important population for study (Kim et al. 2004). However, it is also known that Asians have a genetic vulnerability to reduced alcohol metabolism, which leads to aversive consequences when drinking. Given this heterogeneity and likely competing genetic influences, it is important to establish the gene–gene interaction and the clinical relevance of each of these genotypes in various populations.


This work was supported by Je-il Pharmaceutical Co., Seoul, South Korea.

Copyright information

© Springer-Verlag 2008