• Alcohol Dependence Treatment;
  • Pharmacogenetics;
  • Opioid Receptor Genes;
  • OPRM1;
  • OPRK1;
  • OPRD1;
  • Naltrexone


  1. Top of page
  2. Abstract
  6. Appendix

Background: Pharmacotherapy of alcohol dependence (AD) is at an early stage of development; currently available medications have limited efficacy. It would be clinically valuable to identify, before initiation of a course of treatment, those patients who, based on genetic markers, are most likely to respond to a specific pharmacotherapy. A previous report suggested that a functional variant at the genetic locus encoding the μ opioid receptor (Asn40Asp) is such a marker, in short-term (3-month) treatment with the opioid-blocking drug naltrexone (NTX).

Methods: We studied polymorphic variants at each of the 3 opioid receptor genes—OPRM1, OPRD1, and OPRK1, which encode the μ, δ, and κ opioid receptors, respectively—including the OPRM1 Asn40Asp variant—as predictors of response to NTX or placebo in 215 alcohol-dependent male subjects who participated in Veterans Affairs Cooperative Study 425, “Naltrexone in the Treatment of Alcohol Dependence.”

Results: At the 3-month time point, treatment condition, age, and the pretreatment number of drinks per drinking day were all significant (p<0.05) predictors of the rate of relapse and time to relapse. Although NTX had no significant effect on relapse to heavy drinking in the overall sample in CSP 425, it significantly reduced relapse in the subgroup that provided DNA for analysis (i.e., the present study sample). There were no significant interactions between any individual single nucleotide polymorphisms studied and NTX treatment response.

Conclusions: These results do not support association of the OPRM1 Asn40Asp polymorphism with NTX treatment response for AD.

THE PHARMACOLOGICAL TREATMENT of alcohol dependence (AD) is early in its development. Although there are now 3 drugs approved by the U.S. Food and Drug Administration (FDA) for treatment of the disorder—disulfiram, naltrexone (NTX; both an oral formulation and a long-acting one), and acamprosate—these drugs produce substantial benefit in only a minority of patients (Bouza et al., 2005; Fuller et al., 1986; Garbutt et al., 1999; Kranzler and Van Kirk, 2001). Because these medications appear to exert their therapeutic effects through distinct mechanisms, knowledge of specific genetic predictors of response could make it possible for treatment to be matched to patients so as to optimize therapeutic response and minimize adverse effects.

In 1994, the FDA approved NTX, an opioid receptor antagonist, for the treatment of AD. Although efficacious in most placebo-controlled clinical trials (Bouza et al., 2005; Srisurapanont and Jarusuraisin, 2005), the VA Cooperative Study, the largest study of oral NTX treatment of AD to that date, failed to show an advantage of the medication over placebo treatment (Krystal et al., 2001). Meta-analyses have shown that NTX exerts a modest therapeutic effect (Bouza et al., 2005; Kranzler and Van Kirk, 2001; Srisurapanont and Jarusuraisin, 2005; Streeton and Whelan, 2001). This may, in part, explain why the medication has not been widely prescribed by practitioners (Mark et al., 2003). For example, only 1.9% of veterans with alcohol use disorders received NTX treatment in 2000 to 2001 (Petrakis et al., 2003). The recent COMBINE study (Anton et al., 2006) also showed NTX treatment to be efficacious.

Although clinical predictors of NTX response have been identified (Monterosso et al., 2001; Rubio et al., 2005; Volpicelli et al., 1995), little is known about the mechanisms behind the observed interindividual variation in response. Pharmacogenetic approaches provide a method for identifying mechanisms underlying interindividual response variation and, in particular, identifying patients who have a higher probability of benefiting from a particular medication. NTX is particularly well suited to pharmacogenetic analyses: first, its neuroreceptor targets and the genes encoding them are well described; and second, the presence of a family history of AD is known to be associated with superior clinical response (Jaffe et al., 1996; Monterosso et al., 2001; Rubio et al., 2005)—i.e., based on clinical data, we already know that family-genetic factors are important in modulating NTX response. At the 50 mg dosage, NTX produces a saturating blockade of μ opioid receptors (Lee et al., 1988), 20 to 35% blockade of δ opioid receptors (McCaul et al., 2003), and presumably lower occupancy of κ opioid receptors. Thus, the genetic loci that code for these opioid receptors, OPRM1, OPRD1, and OPRK1, respectively, are potential targets for pharmacogenetic studies of NTX treatment effects.

One obvious candidate variant for such an effect is the A118G single nucleotide polymorphism (SNP) in exon 1 of OPRM1. This polymorphism encodes an Asn40Asp amino acid substitution and is reported to be functionally important, but while there are several reports showing that this variant is functional, they are not easily placed in a consistent framework regarding the nature of the functional effects (Befort et al., 2001; Beyer et al., 2004; Bond et al., 1998; Zhang et al., 2005).

Oslin et al. (2003), in a sample of 130 European American (EA) subjects treated in one of 3 placebo-controlled trials for AD, found that individuals with one or 2 copies of the Asp40 allele (i.e., Asp40 carriers) who were treated with NTX showed a significantly greater reduction in risk of relapse to heavy drinking, but not in risk of relapse to any drinking. Outcomes among individuals treated with placebo did not differ based on genotype. The authors concluded that their findings were consistent with the literature demonstrating that the opioid system is involved in the reinforcing properties of alcohol and that allelic variation at OPRM1 is associated with differential response to medications active at the μ-receptor.

In the present study, we hypothesized that genetic variation at any of the genetic loci encoding opioid receptors might modulate NTX response, or response to AD treatment regardless of assignment to the active medication or placebo group. The treatment sample is a subset of that from the VA Cooperative Study of Naltrexone Treatment (Krystal et al., 2001). For OPRM1, we studied the Asn40Asp SNP, a variant, −2044c/a, which has been associated with the phenotype of comorbid alcohol and opioid dependence (Luo et al., 2003), and an additional intron 3 variant, rs648893, which has been associated with drug and alcohol dependence (Zhang et al., 2006). We studied 3 OPRD1 markers—a functional exon 1 variant, F27C (Gelernter and Kranzler, 2000), and 2 informative variants that do not affect coding sequence: T921C and rs678849. An OPRK1 marker (rs963549) was selected because its population genetic characteristics (high FST, a measure of genetic differentiation) suggested that it had undergone selection, and therefore might be of functional importance (Akey et al., 2002). We evaluated whether each of these variants was associated with treatment outcome in either the NTX group or the placebo group. We limited the variants studied at each locus to those with some degree of prior support for affecting function or phenotype, to contain the problem of multiple statistical testing.


  1. Top of page
  2. Abstract
  6. Appendix


Subjects in this study were veterans who met DSM-IV criteria for AD on the basis of the Structured Clinical Interview for DSM-IV (First et al., 1996) and who participated in the Veterans Affairs Cooperative Study 425, “Naltrexone in the Treatment of Alcohol Dependence,” a 12-month, double-blind, placebo-controlled, multicenter NTX treatment trial (Krystal et al., 2001). There were 3 treatment groups in the medication study (Krystal et al., 2001); subjects received NTX treatment (50 mg/day) for 12 months, or for 3 months followed by 9 months of placebo, or they received matched placebo treatment for 12 months. Treatment compliance was monitored through the use of medication bottle caps that recorded when the bottles were opened. Subjects were classified as “compliant” if they opened the medication bottle on at least 50% of the days, as prescribed. Of the 627 study participants, 251 subjects gave informed consent to participate in a genetic substudy as well; both genotype and phenotype data were available for 240 of these individuals. To increase sample homogeneity, 3 females, and 17 subjects who were neither EA nor African American (AA) were excluded. Of the remaining 220 subject records, 213 provided relapse data for the first 13 weeks of the trial, and 159 had nonmissing relapse data and were medication compliant in the first 13 weeks. Differences between the entire clinical trial sample and the present sample are summarized in Table 1.

Table 1.   Comparison of Baseline Characteristics and Outcomes of Patients With or Without Genotype Data
 With genotype dataWithout genotype dataAll subjects
  • *

    Difference between groups is significant at the 0.05 level by a chi-square test.

  • **

    ** Difference between groups is significant at the 0.01 level by a chi-square test.

  • Note that the analyzed sample includes only European-American and African-American males.

Mean age (years)49.7 ± 10.248.8 ± 9.849.1 ± 10
Percent male100.097.398.1
 European American (%)73.657.963.3
 African American (%)26.431.329.5
 Hispanic (%)
 Other (%)
Marital status:
Single (%)17.717.317.4
Married or living with partner (%)38.632.334.5
Divorced, separated or widowed (%)43.650.448.0
Education (years)13.3 ± 2.313.2 ± 2.013.2 ± 2.1
Disability (not related to alcoholism)
 Military—psychiatric (%)
 Military—medical (%)16.815.515.9
 Non-military (%)15.016.3%15.5%
Psychoactive substance use disorder (lifetime prevalence)
 Cannabis (%)
 Cocaine (%)
DSM-IV diagnoses (lifetime)
Major depression (%)18.2*11.6*13.9
Social phobia (%)11.8**5.5**7.7
Generalized Anxiety Disorder (%)
Post-Traumatic Stress Disorder (%)14.613.413.7
Antisocial Personality Disorder (%)13.9**5.3**8.1
Current smoker (%)72.771.371.8
Age began getting intoxicated regularly (mean years)22.8 ± 9.522.9 ± 9.722.9 ± 9.2
Age first had difficulty stopping before intoxication (mean years)30.8 ± 11.830.0 ± 11.030.3 ± 11.3
History of alcoholism, 1st relative84.184.584.4%
Percent days drinking (past 90 days)65.3 ± 29.667.3 ± 29.266.6 ± 29
Drinks per drinking day (past 90 days)14.0 ± 7.813.1 ± 8.013.4 ± 8
Brief Symptom Inventory1.7 ± 0.61.7 ± 0.51.7 ± 0.6

Below, we describe the dataset used in these analyses. Tables 1 and 2 compare subjects with or without genotype data; these differences are also presented in the beginning of “Results” . The subject sample (n=220) was comprised of 73.6% EAs and 26.4% AAs. The mean (SD) age was 49.7 (10.2) years. The majority of subjects (115 or 52.3%) had more than a high school education, with 87 (39.6%) having completed high school only and 18 (8.2%) having less than a high school education. Thirty-nine subjects (17.7%) were never married, 85 (38.6%) were currently married, and 96 (43.6%) were separated or divorced.

Table 2.   Comparison of Samples With or Without Genotype Information (Where Summary Drinking Measures Were Available) on Outcome Variables and Compliance for the First 13 Weeks
13 WeeksNaltrexone (short and long term) in subsample with genotype information N=149Placebo in subsample with genotype information N=64Naltrexone in subsample without genotype information N=229Placebo in subsample without genotype information N=123
Mean ± SD or percentMean ± SD or percentMean ± SD or percentMean ± SD or percent
  1. For the comparison of compliance (continuous variable) in sample with genotype information: t (194)=−2.08, p=0.04.

  2. For the comparison of compliance (continuous variable) in sample without genotype information: t (331)=0.63, p=0.53.

Rate of Relapse (%)36.948.438.442.3
Days to Relapse (primary endpoint)67.5 ± 33.958.8 ± 36.562.6 ± 33.661.8 ± 34.0
Percent Days Drinking8.9 ± 17.115.7 ± 24.512.8 ± 23.413.1 ± 22.1
Drinks per Drinking Day8.7 ± 6.28.7 ± 6.39.2 ± 8.69.1 ± 6.6
Medication compliance (yes/no variable) (%)84.668.868.165.0
Medication compliance (continuous variable)80.3 ± 26.671.2 ± 31.367.1 ± 32.369.4 ± 30.3


The study included baseline and monthly assessments, with longer interviews at 6, 12, and 18 months. Outcome variables included time to relapse during the first 3 months (number of days from randomization until relapse, with relapse defined as the first day of heavy drinking [6 or more drinks for men]), percent drinking days over 3 months, and number of drinks per drinking day over 3 months.

Genes and Genotyping

DNA was extracted from whole blood using standard salting-out methods. The markers studied are summarized in Table 3. For markers genotyped by the restriction fragment length polymorphism method, at least 8% of genotypes were repeated for quality control, with complete concordance. For markers genotyped by the TaqMan technique, i.e., a fluorogenic 5′ nuclease assay method (Shi et al., 1999) using the ABI PRISM 7900 Sequence Detection System (ABI, Foster City, CA), all assays were completed in duplicate, and discordant genotypes were discarded.

Table 3.   Markers Studied
GeneMarkerBase or amino acid changeChromosomePositionGenotyping method
  1. RFLP, restriction fragment length polymorphism.

rs17180961−2044c/a 154450752RFLP
rs648893c/t 154530742Taqman
rs2234918T921C 29010213Taqman
rs678849c/t 28965804RFLP

Statistical Analysis

Consistency with Hardy–Weinberg Equilibrium Expectations (HWEE) was tested using an exact test implemented in PowerMarker software, version 3.25 (Liu and Muse, 2005).

Logistic regression (using SAS PROC LOGISTIC) was used to assess whether relapse rate differed by genotype (each gene was considered separately; the combination of OPRD1 (T921C) and OPRK1 was also examined). The outcome variable was relapse within 13 weeks (1=Yes/0=No). Predictors were genotype (to reduce the number of comparisons, the rarer homozygote was generally combined with the heterozygote class to yield 2 levels of this variable), treatment (NTX, placebo), and their interaction. Covariates in the analysis included race, age, marital status, years of education (<12, 12, >12), drinking history (number of drinking days and of drinks per drinking day in the 90 days before randomization), and total score on the Brief Symptom Inventory (BSI; which reflects overall level of psychological distress; Derogatis and Melisaratos, 1983) at baseline and its interaction with genotype. Nonsignificant covariates were dropped from the model using backwards elimination at an α level of 0.10. The interaction between genotype and treatment was dropped from the model when nonsignificant and when estimating odds ratios for main effects of genotype and/or treatment.

Odds ratios and 95% confidence intervals (CI) were used to explain significant interactions and/or main effects. For each SNP, the main analysis included all subjects with SNP information in the merged data set, and secondary analyses were restricted to the sample of medication-compliant subjects.

A Cox proportional hazards model was used to test whether time to relapse differed by genotype in each treatment arm, with each marker considered separately. PROC PHREG in SAS was used for this analysis. The outcome variable was time to relapse in the first 13 weeks. Predictors were genotype, treatment, and their interaction recoded as dummy variables. Covariates were the same as in the logistic regression analysis. Backwards elimination was used to drop nonsignificant (at α=0.10) covariates from the model.

The General Linear Modeling (GLM) procedure in SAS was used to test for significant effects of each genotype on the continuous measures number of drinks per drinking day and percent drinking days in the first 13 weeks of the trial. Treatment, genotype, and their interaction were used as fixed effects. The same covariates as were used in all other genotype analyses were also considered in this analysis. Both outcomes were log-transformed because of positive skewness and residual plots and normal probability plots were used to confirm that the assumptions of the models were not violated.

Finally, interaction covariate analyses for smoking and BSI were used to evaluate whether baseline psychological distress modified gene effects. These analyses were restricted to OPRM1 (Asn40Asp), OPRD1 (T921C), and OPRK1. Brief Symptom Inventory (total) scores at baseline were inversely transformed to approximate normality.


  1. Top of page
  2. Abstract
  6. Appendix

Genotype distributions for all markers were consistent with HWEE in AAs (each p>0.35). In EAs, 2 markers showed nominally significant Hardy–Weinberg disequilibrium: OPRM1 rs17180961 (p=0.014), and OPRD1 F27C (p=0.022).

The subsample that provided DNA differed in several important respects from the overall CSP 425 study sample. Although the groups were comparable on age, the genotyped subgroup had nonsignificantly more EAs (67.9 vs 60.5%) and fewer AAs (25.0 vs 32.3%), and a significantly higher rate of psychiatric comorbidity [e.g., major depression (18.3 vs 11.1%, χ2(1)=6.3, p=0.01), social phobia (11.6 vs 5.2%, χ2(1)=8.7, p=0.003), and antisocial personality disorder (13.3 vs 4.9%, χ2(1)=14.0, p=0.0002); Table 1]. There was no significant difference in the relapse rate for subjects who received NTX based on whether they provided DNA (35.2 vs 39.8% in the NTX-treated nongenotyped group), but there was a numerically higher (but not statistically significant; p=0.4) rate of relapse among the placebo-treated genotyped subjects (48.6 vs 41.9%; Table 2). The genotyped group was more medication-compliant (defined as the percentage of days on which the medication bottle was opened during the period) than the nongenotyped group [85.2 vs 70.1%; χ2(1)=11.8, p=0.001].

During the 90 days before randomization, subjects reported drinking on 53.3 (SD=24.7) days, the mean number of drinks per drinking day being 13.9 (SD=7.8).

Relapse rates per treatment group and genotype are reported in Table 4. Treatment condition, age, and the number of drinks per drinking day at baseline were all significant (p<0.05) predictors of the rate of relapse and time to relapse, and hence were retained in all models for genotype effects.

Table 4.   Comparison of Naltrexone and Placebo Groups by Genotype Group (Excluding OPRM1−2044c/a and OPRD1 F27C, Where Minor Allele Frequency Was Too Low for Meaningful Comparisons)
NN (% Relapsed)NN (% Relapsed)
OPRM1 Asn40Asp (A118G)
 “AA”5024 (48.0)9835 (35.7)
 “G carrier”95 (55.6)3312 (36.0)
OPRM1 2044C/A
 “A carrier”64 (67.0)42 (50.0)
 “CC”4320 (46.5)9937 (37.4)
OPRK1 rs963549
 “CC”3418 (53.0)9638 (39.6)
 “T carrier”2811 (39.3)4915 (31.4)
OPRD1 T921
 “GG”156 (40.0)3611 (30.6)
 “AG”3117 (54.8)6017 (28.3)
 “AA”93 (33.3)3419 (55.9)
 “G carrier”20 (0.0)10 (0.0)
 “TT”5427(50.0)12646 (36.5)
OPRD1 rs678849
 “CC”218 (38.1)5320 (37.7)
 “CT”2513 (52.0)6219 (30.6)
 “TT”179 (52.9)3014 (46.7)
OPRM1 rs648893
 “AA”4020 (50.0)11238 (33.9)
 “G carrier”198 (42.1)3114 (45.2)

The results from the analyses of relapse rates are presented in Table 5. There were no significant interactions between individual SNPs and NTX treatment response. Although in the intent-to-treat analysis there was no significant main effect of NTX treatment (Krystal et al., 2001), in the available subsample of patients with genotype information for OPRM1 Asn40Asp, OPRK1, or OPRD1 rs678849, NTX treatment significantly reduced the odds of relapse. Subjects in the placebo group were about twice as likely to relapse as subjects in the NTX group. Estimated odds ratios and associated 95% CI are presented in Table 5.

Table 5.   Results From Analyses of the Effects of SNP Variants on Rate of Relapse (Logistic Regression Analyses)
GenotypeTreatment main effectGenotype main effectTreatment by genotype interactionSignificant covariates
  1. OR, odds ratio; P, placebo; A, active; SNP, single nucleotide polymorphism.

OPRM1 Asn40Aspχ2(1)=4.74, p=0.03 OR (P vs A)=2.10 95%CI=(1.08, 4.11)NSNSAge, drinks/drinking day in previous 90 days, drinking days in previous 90 days
OPRK1 rs963549χ2(1)=4.72, p=0.03 OR (P vs A)=2.08 95%CI: (1.07, 4.04)χ2(1)=3.51, p=0.06 OR (“CC” vs “T”)=1.84 95%CI=(0.97, 3.49)NSAge, drinks/drinking day in previous 90 days, drinking days in previous 90 days
OPRD1 T921Cχ2(1)=4.09, p=0.04 OR (P vs A)=2.05 95%CI: (1.02, 4.12)NSNSAge, drinks/drinking day in previous 90 days, drinking days in previous 90 day
OPRD1 rs678849χ2(1)=4.75, p=0.03 OR (P vs A)=2.09 95%CI: (1.08, 4.06)NSNSAge, race, drinks/drinking day in previous 90 days, Drinking days in previous 90 day
OPRM1 rs648893χ2(1)=4.38, p=0.04 OR (P vs A)=2.05 95%CI: (1.05, 4.00)NSNSAge, race, drinks/drinking day in previous 90 days

As shown in Table 5, of all SNPs considered, only the main effect of OPRK1 approached statistical significance: χ2(1)=3.51, p=0.06, such that “CC” subjects were 1.84 times more likely to relapse than were “T” carriers (95% CI:0.97–3.49). The “CC” genotype was associated with a marginally significantly increased risk of relapse relative to “TT,” but the differences between the heterozygote and either of the homozygotes were not statistically significant. A linear effect of genotype was not statistically significant. Although the main effect of OPRK1 was close to being nominally significant, the interactive effect of this SNP with treatment was not significant. Similarly, the interaction of medication condition with the OPRD1 and OPRK1 genotypes yielded no significant effects.

The results from the analyses of time to relapse are presented in Table 6. In these analyses, there were no significant interactions between genotype and treatment, but for the subsamples of patients with genotype information, NTX treatment decreased the likelihood of relapse by about 50%. The hazard ratio estimates and associated 95% CI are presented in Table 6. For OPRK1, there was a nonsignificant trend for a main effect of genotype: χ2(1)=3.58, p=0.06 [hazard ratio (“T” vs “CC”)=0.64, 95%CI: 0.40–1.02]. A greater number of drinks per drinking day before study entrance significantly increased the chance of relapse [HR=1.76, 95% CI: (1.15, 2.69)]. Increasing age was associated with a decreasing hazard rate [HR=0.97, 95% CI: (0.95, 1.00)]. The effect of the interaction between OPRD1 and OPRK1 was not significant.

Table 6.   Results From Analyses of the Effects of SNPs on Time to Relapse (Cox Model)
GenotypeTreatment main effectGenotype main effectTreatment by genotype interactionSignificant covariates
  1. HR, hazard ratio; CI, confidence interval; P, placebo; A, active; SNP, single nucleotide polymorphism.

OPRM1 Asn40Aspχ2(1)=4.78, p=0.03 HR (A vs P)=0.59 95%CI=(0.36, 0.95)NSNSAge, drinks/drinking day in previous 90 days, drinking days in previous 90 days
OPRK1 rs963549χ2(1)=5.97, p=0.01 HR (A vs P)=0.57 95%CI: (0.36, 0.90)χ2(1)=3.58, p=0.06 HR (“T” vs “CC”)=0.64 95%CI: (0.40, 1.02)NSAge, drinks/drinking day in previous 90 days, drinking days in previous 90 days
OPRD1 T921Cχ2(1)=6.00, p=0.01 HR (A vs P)=0.57 95%CI: (0.36, 0.89)NSNSAge, drinks/drinking day in previous 90 days, drinking days in previous 90 day
OPRD1 rs678849χ2(1)=3.67, p=0.05 HR (A vs P)=0.64 95%CI: (0.41, 1.01)NSNSAge, drinks/drinking day in previous 90 days, drinking days in previous 90 day
OPRM1 rs648893χ2(1)=4.29, p=0.04 HR (A vs P)=0.62 95%CI: (0.40, 0.98)NSNSAge, drinks/drinking day in previous 90 days

The interaction covariate analyses for smoking and BSI demonstrated that smoking had no significant effect on relapse and did not change the relationship between the genotypes and outcome. There was a significant interaction between the OPRM1 Asn40Asp genotype and BSI (χ2(1)=7.23, p=0.007). Brief Symptom Inventory was significantly associated with higher odds for relapse for Asp40 carriers but not for Asn40 homozygotes. For OPRK1, there was a trend-level main effect of BSI (p=0.06); higher scores on the BSI at baseline were associated with a higher rate of relapse.


  1. Top of page
  2. Abstract
  6. Appendix

We did not find reliable evidence that allelic variation at any of the loci studied moderated the response to NTX treatment. This is inconsistent with one previous report (Oslin et al., 2003). There are several possible explanations for these findings.

Our results could be false negatives. If this is the case, it most likely stems from limited statistical power derived from a study sample that, while reasonable for many clinical designs, is modest for the detection of pharmacogenetic (i.e., interaction) effects. Second, the difference in findings between our study and that of Oslin et al. (2003) could be attributable to differences in the clinical makeup of the sample. Krystal et al. (2001) have commented previously on the high level of severity of AD among the subjects in the VA Cooperative Study of Naltrexone. Third, our finding could be correct, and the previous report (Oslin et al., 2003), a false positive. Naltrexone response, as a phenotype, lacks many of the characteristics of traits for which genetic associations might be expected to be identifiable. Although implicated through findings associating it with a family history of alcoholism (Monterosso et al., 2001; Rubio et al., 2005; Volpicelli et al., 1995), NTX treatment response has not been proven to be a heritable trait. Also, this trait has not been well validated. That is, the differences in response profiles between NTX and placebo, while clinically important, are small (Bouza et al., 2005; Srisurapanont and Jarusuraisin, 2005). Many responders and nonresponders could be considered “borderline,” i.e., assigned to that category equivocally. Some responders in the NTX group are in the “responder” category because of NTX, but we cannot identify those individuals. However, it is only that category of responders—subjects who responded to NTX but who would not have responded to placebo—who are informative for a pharmacogenetic study. Further, relapse is most commonly defined using self-report data, which, although more sensitive than objective measures of drinking behavior, may not be precise enough to define a trait for pharmacogenetic analysis, particularly when measured using recall methods (Kranzler and Tennen, 2004; Kranzler et al., 2004).

If the increment in response for NTX over placebo is assumed to be 10%, this implies that for a sample of 300 individuals of whom 150 are on NTX, 15 responded because of NTX and could possibly provide genetic treatment predictor information. Even with a very large effect size, it would be extremely difficult to detect a valid association in the context of so many noninformative individuals. This might be ameliorated somewhat if the genetic predictor were a predictor of response in general, not just of response to a specific treatment. Such an association could still provide useful response prediction information and could be modulated by the pharmacologic agent under study, but it would then not be accurate to attribute any predictive power to a specific interaction with a particular medication.

A limitation to this study is the lack of samples from all 627 participants in the NTX clinical trial. This is partly explained by the differences in sample collection and delays in IRB approval for the DNA substudy among the 15 sites. By the time approval was received at some sites, some patients were no longer available for testing. Also, subjects could participate in the main study without participating in the DNA substudy. Thus, many patients had not provided a DNA sample by the conclusion of the clinical trial. We were not able to determine whether there was a systematic bias other than by site. In the parent sample from which the present study sample was drawn, the absolute 13-week relapse rate for NTX was 37.8%, and the response rate for placebo was 44.4%. In that study, there was no significant treatment effect for NTX. However, in the subset of subjects participating in the present study, there was a significant treatment effect. As the genetics substudy was voluntary, we speculate that this difference in treatment effect may be attributable to genetics substudy subjects tending to be more cooperative and possibly more motivated to comply with treatment instructions, than those who declined to participate. This is consistent with the significantly higher rate of medication compliance observed in the genotyped vs. the nongenotyped subjects. Another possible explanation is that study staff members who presented and explained the genetics substudy persuasively were more highly motivated in a way that resulted in better compliance and treatment outcome for the study participants.

The duration of NTX treatment may influence the relationship between opioid receptor genotype and clinical outcome. The expectation that the polymorphic variants studied here would influence clinical outcomes is based on the function of the opioid receptor under conditions of acute exposure to agonists and antagonists. However, chronic exposure to opioid receptor antagonists produces substantial adaptations in opioid receptor function (Lesscher et al., 2003), resulting in increases in opioid receptors on the cell surface, increased coupling of opioid receptors to second messenger systems, increased opioid receptor function, and tolerance to some features of opioid antagonist response. It is thus possible that these SNPs influence acute, but not chronic, responses to NTX. If this is the case, then the timing of relapses with respect to the initiation of NTX could have a significant impact on the ability to detect effects of polymorphism in opioid receptor genes on clinical outcome, i.e., relapses early in the course of treatment might be expected to show greater genotype moderation of NTX response than relapses later in treatment. For clinical purposes, chronic response would be expected to be the more important measure.

Although we queried variants at each opioid receptor gene, our negative results do not exclude predictive relationships between other SNPs mapping to any of these loci, and NTX treatment response. We did query SNPs from each major OPRM1 haplotype block (Zhang et al., 2006). We focused on the SNPs that we thought were the best candidates for a relationship with treatment response (including the only variant where such a relationship has been reported previously, OPRM1 Asn40Asp), rather than genotyping a more extensive set of SNPs or attempting to conduct a haplotype analysis, to avoid a more serious multiple testing issue.

None of the markers studied showed deviation from HWEE in AAs, but 2 (OPRM1 rs17180961 and OPRD1 F27C) did show nominal Hardy–Weinberg disequilibrium (HWD) in EAs. There are many possible explanations for significant deviation from HWEE (as discussed in Luo et al., 2006), including occult admixture, genotyping error, and true association with phenotype. We previously reported that several other OPRM1 SNPs were in HWD in substance-dependent subjects and concluded that in that case, the HWD reflected an association with phenotype (Zhang et al., 2006). We followed careful quality control procedures in genotyping as described above, so it is unlikely that genotyping error explains the HWD in this sample, but cannot distinguish between the other possibilities.

We remark again on several limitations of this study. First, although the patient sample available to us was larger than that ascertained in a prior positive study (Oslin et al., 2003) (which considered 71 NTX-treated and 59 placebo-treated EA subjects), it is still possible, as noted above, that we failed to observe a true effect because of the limited statistical power. The odds ratio for the genotype effect (OPRM1 Asn40Asp or A118G) in the study by Oslin et al. (2003) was 3.5 in the NTX group and 2.4 for the NTX and placebo groups combined. Our sample size for the same genotype was 98 NTX-treated “AA” subjects, 33 NTX-treated “G” carriers, 50 placebo-treated “AA” subjects, and 9 placebo-treated “G” carriers. Based on these numbers, our power to detect similar effect sizes is about 80 and 60%, respectively, for the NTX and placebo groups.

Second, the patient sample we studied was drawn from a parent study where NTX did not show a positive treatment effect overall, and therefore this might be considered a suboptimal sample in which to demonstrate pharmacogenetic predictors of treatment. (Two of the 3 studies from which the sample for the Oslin et al. (2003) study was drawn showed an NTX treatment effect.) Interestingly, however, and as discussed above, an NTX treatment effect was observed in that part of the larger study sample for which genotype information is available (i.e., the part of the sample discussed here), and would seem to mitigate this problem.

There is a growing understanding of genes that increase risk for AD—the past few years have witnessed the identification and replication of several risk genes acting via GABAergic (Edenberg et al., 2004; Covault et al., 2004; Lappalainen et al., 2005), muscarinic (Luo et al., 2005; Wang et al., 2004), serotonergic (Feinn et al., 2005), and opioidergic (Zhang et al., 2006) mechanisms. At present, the weight of evidence supporting an effect of the OPRM1 Asn40Asp (A118G) variant on NTX response is not, in our view, persuasive. One may hope reasonably that gene variants that predispose to differences in AD course, and the response to pharmacological—or nonpharmacological—treatment, may be identified in the near future, based on better knowledge of the genetic factors that lead to AD itself. We also propose that further study of OPRK1 (SNP rs963549, or other markers at that locus) as a possible predictor of treatment response is warranted. Although this variant has not yet been shown to be functional, its observed high FST supports it as a target for selection (Akey et al., 2002), which would make sense only if variant alleles resulted in phenotypically different outcomes.


  1. Top of page
  2. Abstract
  6. Appendix

The Investigators for the Department of Veterans Affairs Cooperative Study 425 Group were as follows: K. Drexler, Atlanta, GA; J. Hermos, Boston, MA; L. Rugle, O. Kausch, Cleveland, OH; B. Adinoff, Dallas, TX; J. Grabowski, R. Wancha, Detroit, MI; L. Madlock, Memphis, TN; M. FeBornstein, J. Pena, New Orleans, LA; P. Casadonte, New York, NY; S. Nixon, C. Shaw, Oklahoma City, OK; L. Haynes-Tucker, L. Moffet, Menlo Park, CA; I. Maany, Philadelphia, PA; G. Kaplan, Providence, RI; C. Stock, Salt Lake City, UT; P. Banys, San Francisco, CA; A. Saxon, Seattle, WA; W. Krol, Perry Point, MD (Study Biostatistician); and M. Miller, Albuquerque, NM (Study Pharmacist).


  1. Top of page
  2. Abstract
  6. Appendix

COI Disclosure: Dr. Krystal is a consultant to Merz Pharmaceuticals and Takeda Medical Industries.


  1. Top of page
  2. Abstract
  6. Appendix
  • Akey JM, Zhang G, Zhang K, Jin L, Shriver MD (2002) Interrogating a high-density SNP map for signatures of natural selection. Genome Res 12:18051814.
  • Anton RF, O'Malley SS, Ciraulo DA, Cisler RA, Couper D, Donovan DM, Gastfriend DR, Hosking JD, Johnson BA, LoCastro JS, Longabaugh R, Mason BJ, Mattson ME, Miller WR, Pettinati HM, Randall CL, Swift R, Weiss RD, Williams LD, Zweben A COMBINE Study Research Group (2006) Combined pharmacotherapies and behavioral interventions for alcohol dependence: the COMBINE study: a randomized controlled trial. JAMA 295:20032017.
  • Befort K, Filliol D, Decaillot FM, Gaveriaux-Ruff C, Hoehe MR, Kieffer BL (2001) A single nucleotide polymorphic mutation in the human mu-opioid receptor severely impairs receptor signaling. J Biol Chem 276:31303137.
  • Beyer A, Koch T, Schroder H, Schulz S, Hollt V (2004) Effect of the A118G polymorphism on binding affinity, potency and agonist-mediated endocytosis, desensitization, and resensitization of the human mu-opioid receptor. J Neurochem 89:553560.
  • Bond C, LaForge KS, Tian M, Melia D, Zhang S, Borg L (1998) Single-nucleotide polymorphism in the human mu opioid receptor gene alters beta-endorphin binding and activity: possible implications for opiate addiction. Proc Natl Acad Sci U S A 95:96089613.
  • Bouza C, Angeles M, Munoz A, Amate JM (2005) Efficacy and safety of naltrexone and acamprosate in the treatment of alcohol dependence: a systematic review. Addiction 99:811828.
  • Covault C, Gelernter J, Hesselbrock V, Nellissery M, Kranzler HR (2004) Allelic and haplotypic association of GABRA2 with alcohol dependence. Am J Med Genet 129B:104109.
  • Derogatis LR, Melisaratos N (1983) The brief symptom inventory: an introductory report. Psychol Med 13:595605.
  • Edenberg HJ, Dick DM, Xuei X, Tian H, Almasy L, Bauer LO, Crowe RR, Goate A, Hesselbrock V, Jones K, Kwon J, Li TK, Nurnberger JI Jr., O'Connor SJ, Reich T, Rice J, Schuckit MA, Porjesz B, Foroud T, Begleiter H (2004) Variations in GABRA2, encoding the alpha 2 subunit of the GABA(A) receptor, are associated with alcohol dependence and with brain oscillations. Am J Hum Genet 74:705714.
  • Feinn R, Nellissery M, Kranzler HR (2005) Meta-analysis of the association of a functional serotonin transporter promoter polymorphism with alcohol dependence. Am J Med Genet B Neuropsychiatr Genet 133:7984.
  • First MB, Gibbon M, Spitzer RL, Williams JBW (1996) User's Guide for the Structured Clinical Interview for DSM-IV Axis I Disorders—Research Version. Biometrics Research, New York.
  • Fuller RK, Branchey L, Brightwell DR, Derman RM, Emrick CD, Iber FL, James KE, Lacoursiere RB, Lee KK, Lowenstam I, Maany I, Neiderhiser D, Nocks JJ, Shaw S (1986) Disulfiram treatment of alcoholism. A veterans administration cooperative study. JAMA 256:14491455.
  • Garbutt JC, West SL, Carey TS, Lohr KN, Crews FT (1999) Pharmacological treatment of alcohol dependence: a review of the evidence. JAMA 281:13181325.
  • Gelernter J, Kranzler HK (2000) Variant detection at the delta opioid receptor (OPRD1) locus and population genetics of a novel variant affecting protein sequence. Hum Genet 107:8688.
  • Jaffe AJ, Rounsaville B, Chang G, Schottenfeld RS, Meyer RE, O'Malley SS (1996) Naltrexone, relapse prevention, and supportive therapy with alcoholics: an analysis of patient treatment matching. J Consult Clin Psychol 64:10441053.
  • Kranzler HR, Abu-Hasaballah K, Tennen H, Feinn R, Young K (2004) Using daily interactive voice response technology to measure drinking and related behaviors in a pharmacotherapy study. Alcohol Clin Exp Res 28:10601064.
  • Kranzler HR, Tennen H (2004) How to measure relapse in humans, in Drugs for Relapse Prevention of Alcoholism (SpanagelR, MannK eds), pp 2339. Birkhauser Verlag, Basel.
  • Kranzler HR, Van Kirk J (2001) Efficacy of naltrexone and acamprosate for alcoholism treatment: a meta-analysis. Alcohol Clin Exp Res 25:13351341.
  • Krystal JH, Cramer JA, Krol WF, Kirk GF, Rosenheck RA, Veterans Affairs Naltrexone Cooperative Study 425 Group (2001) Naltrexone in the treatment of alcohol dependence. N Engl J Med 345:17341739.
  • Lappalainen J, Krupitsky E, Remizov M, Pchelina S, Taraskina A, Zvartau E, Somberg LK, Covault J, Kranzler HR, Krystal JH, Gelernter J (2005) Association between alcoholism and GABRA2 in a Russian population. Alcohol Clin Exp Res 29:493498.
  • Lee MC, Wagner HN Jr., Tanada S, Frost JJ, Bice AN, Dannals RF (1988) Duration of occupancy of opiate receptors by naltrexone. J Nuclear Med 29:12071211.
  • Lesscher HM, Bailey A, Burbach JP, Van Ree JM, Kitchen I, Gerrits MA (2003) Receptor-selective changes in mu-, delta- and kappa-opioid receptors after chronic naltrexone treatment in mice. Eur J Neurosci 17:10061012.
  • Liu K, Muse SV (2005) PowerMarker: integrated analysis environment for genetic marker data. Bioinformatics 21:21282129.
  • Luo X, Kranzler HR, Zhao H, Gelernter J (2003) Haplotypes at the OPRM1 locus are associated with susceptibility to substance dependence in European-Americans. Am J Med Genet (Neuropsychiatr Genet) 120B:97108.
  • Luo X, Kranzler HR, Zuo L, Lappalainen J, Yang BZ, Gelernter J (2006) ADH4 gene variation is associated with alcohol and drug dependence in European Americans: results from HWD tests and case-control association studies. Neuropsychopharmacology 31:10851095.
  • Luo X, Kranzler HR, Zuo L, Wang S, Blumberg HP, Gelernter J (2005) CHRM2 gene predisposes to alcohol dependence, drug dependence, and affective disorders: results from an extended case-control structured association study. Hum Mol Genet 14:24212434.
  • Mark TL, Kranzler HR, Song X, Bransberger P, Poole V, Crosse S (2003) Physicians' opinions about medications to treat alcoholism. Addiction 98:617626.
  • McCaul ME, Wand GS, Kim YK, Bencherif B, Dannals RF, Frost JJ (2003) Naltrexone effects on mu- and delta-opioid receptor availability in alcohol dependence. Alcohol Clin Exp Res 27 (5 suppl): 21A.
  • Monterosso JR, Flannery BA, Pettinati HM, Oslin DW, Rukstalis M, O'Brien CP, Volpicelli JR (2001) Predicting treatment response to naltrexone: the influence of craving and family history. Am J Addict 10:258268.
  • Oslin DW, Berrettini W, Kranzler HR, Pettinati H, Gelernter J, Volpicelli JR, O'Brien CP (2003) A functional polymorphism of the mu-opioid receptor gene is associated with naltrexone response in alcohol-dependent patients. Neuropsychopharmacology 28:15461552.
  • Petrakis IL, Leslie D, Rosenheck R (2003) Use of naltrexone in the treatment of alcoholism nationally in the department of veterans affairs. Alcohol Clin Exp Res 27:17801784.
  • Rubio G, Ponce G, Rodriguez-Jimenez R, Jimenez-Arriero MA, Hoenicka J, Palomo T (2005) Clinical predictors of response to naltrexone in alcoholic patients: who benefits most from treatment with naltrexone? Alcohol Alcohol 40:227233.
  • Shi MM, Myrand SP, Bleavins MR, De La Iglesia FA (1999) High throughput genotyping for the detection of a single nucleotide polymorphism in NAD(P)H quinone oxidoreductase (DT diaphorase) using TaqMan probes. Mol Pathol 52:295299.
  • Srisurapanont M, Jarusuraisin N (2005) Opioid antagonists for alcohol dependence. Cochrane Database Syst Rev. (1):CD001867.
  • Streeton C, Whelan G (2001) Naltrexone, a relapse prevention maintenance treatment of alcohol dependence: a meta-analysis of randomized controlled trials. [Journal Article. Meta-Analysis]. Alcohol Alcohol 36:544552.
  • Volpicelli JR, Clay KL, Watson NT, O'Brien CP (1995) Naltrexone in the treatment of alcoholism: predicting response to naltrexone. J Clin Psychiatry 56 (Suppl 7): 3944.
  • Wang JC, Hinrichs AL, Stock H, Budde J, Allen R, Bertelsen, Kwon JM, Wu W, Dick DM, Rice J, Jones K, Nurnberger JI Jr., Tischfield J, Porjesz B, Edenberg HJ, Hesselbrock V, Crowe R, Schuckit M, Begleiter H, Reich T, Goate AM, Bierut LJ (2004) Evidence of common and specific genetic effects: association of the muscarinic acetylcholine receptor M2 (CHRM2) gene with alcohol dependence and major depressive syndrome. Hum Mol Genet 13:19031911.
  • Zhang H, Luo X, Kranzler HR, Lappalainen J, Yang BZ, Krupitsky E, Zvartau E, Gelernter J (2006) Association between two μ opioid receptor gene (OPRM1) haplotype blocks and drug or alcohol dependence. Hum Mol Genet 15:807819.
  • Zhang Y, Wang D, Johnson AD, Papp AC, Sadée W (2005) Allele expression imbalance of human mu opioid receptor (OPRM1) caused by variant A118G. J Biol Chem 280:3261832624.