Alcohol Abuse And Family

AM. J. DRUG ALCOHOL ABUSE, 28(3), 557–584 (2002)

THE HERITABILITY OF ALCOHOL ABUSE AND DEPENDENCE: A META-ANALYSIS OF BEHAVIOR GENETIC RESEARCH Glenn D. Walters* Psychology Services, Federal Correctional Institution-Schuylkill, P.O. Box 700, Minersville, PA 17954-0700

ABSTRACT A meta-analysis was performed on 50 family, twin, and adoption studies in which problem drinking and alcohol dependence served as the primary criterion measure. The results showed that far from being an established “fact,” the genetic foundations of alcohol misuse are modest and heterogeneous. A weighted mean f effect size of 0.12 (95% Confidence Interval ¼ 0.11 –0.12) was obtained for the total sample of 72 effect sizes. Four potential moderator variables (proband gender, sample nationality, pattern severity, year of publication) were examined with outcomes confirming that the heritability of alcohol misuse is stronger in males and in studies employing more severe definitions of abuse (alcoholism, alcohol dependence). When the effect size measure was restricted to studies using male subjects with more severe diagnoses of alcohol misuse, the unweighted mean f effect size was only 0.18 (95% Confidence Interval ¼ 0.15 –0.21), with an even smaller

*E-mail: [email protected] 557 Copyright q 2002 by Marcel Dekker, Inc.

www.dekker.com

558

WALTERS

weighted mean f effect size of 0.15 (95% Confidence Interval ¼ 0.12 – 0.18); results which indicate an upper limit of 30 –36% for the heritability of alcohol misuse. Key Words: analysis

Alcohol; Alcoholism; Genetics; Heredity; Meta-

In June 1990 the Journal of the American Medical Association (JAMA) published a paper by a group of researchers from the University of Texas Health Science Center in San Antonio claiming that a genetic locus for alcoholism on the Taq1 A1 allele of the dopamine D2 receptor (DRD2) gene had been identified.[1] The news made headlines in several newspapers, was written up in at least one national magazine, and was broadcast to millions over the evening news. The normally reserved New York Times called it a major breakthrough in the fight against alcoholism.[2] However, when JAMA published another paper 8 months later refuting Blum et al.’s original findings[3] the news met with none of the fanfare that had greeted the original Blum investigation. If the Bolos study was mentioned at all it, was relegated to the back pages of a handful of newspapers and trade magazines. Hence, much of the general public still believes that a gene for alcoholism exists,[4] even though the bulk of recently published studies on the DRD2 gene shed serious doubt on its status as a marker for alcohol misuse.[5 – 9] It is understandable that people would want to believe news that a gene responsible for a pattern that causes serious physical, social, and economic problems in three out of ten American families[10] had been discovered. After all, knowledge promotes control which, in turn, reduces fear and instills hope. However, if the knowledge is false or inaccurate then disappointment, frustration, and mistrust will surely ensue. The human genome is comprised of nearly 100,000 genes, 20,000 – 30,000 of which involve brain function.[11] That any one of these genes accounts for more than a modicum of variance in a behavioral pattern as intricate and multifaceted as alcohol misuse seems highly unlikely. The DRD2 gene may well contribute to alcohol misuse but only as part of a much larger polygenetic pattern. Schork and Schork[12] note that multiple genes can influence complex patterns like alcohol abuse and dependence in one of five ways: (1) general additive effect (the presence of several genes from a larger gene pool); (2) threshold effect (the proper number of genes); (3) epistasis or gene interaction (the proper combination of genes); (4) locus heterogeneity (more than one gene can activate the pattern); and (5) any combination of the four previously mentioned possibilities. The complexity of both genetics and alcohol abuse makes it that much more difficult to believe that any one gene is responsible for a major portion of the variance in alcohol misuse.

HERITABILITY OF ALCOHOL ABUSE

559

Before accepting the view that alcohol misuse is genetically transmitted and that heritability for alcohol abuse is upwards of 40– 60%,[13] we must first consider the growing body of behavior genetic research on this issue. Behavior geneticists seek to identify the relative inheritance of complex patterns like alcohol misuse with the aid of three primary methodologies: family studies, twin studies, and adoption studies. The rationale behind family studies is that if a trait or pattern is genetic then it should be more routinely observed in people with a family history of problem drinking. In other words, the trait or pattern should run in families. Outcomes obtained in early family studies on alcohol abuse and dependence revealed that people with a family history of alcohol misuse were three to four times more likely to personally misuse alcohol than persons without a family history of alcohol misuse.[14] In so much as most families share experiences as well as genes, the family method confounds biology (genetic inheritance) and environment (learning). This has led to the development of twin and adoption methodologies in behavior genetic research. The twin method is grounded in the fact that identical or monozygotic (MZ) twins are genetically identical, whereas fraternal or dizygotic (DZ) twins share half their genes in common. A genetic influence is therefore implied, at least in theory, when MZ twins display greater similarity or concordance for a trait, behavior, or pattern than DZ twins. However, twin research rests on a tenuous foundation of controversial assumptions. The first assumption made by supporters of the twin method is that parental mating is random or nonassortative. Studies addressing this issue in alcohol abusing populations have met with conflicting results, with some studies identifying moderate levels of assortative mating in the parents of alcohol abusers[15] and other studies showing no evidence of assortative mating in the mothers and fathers of alcohol abusers.[16] The equal environments assumption, a second supposition advanced by proponents of the twin method, maintains that MZ twins are raised in environments that are no more similar than the environments in which DZ twins are raised. Despite assertions by twin researchers that the equal environments assumption holds up under empirical scrutiny,[17] there is evidence that MZ twins develop significantly closer relationships[18] and have more contact with one another than DZ twins and that such closeness is associated with increased similarity in drinking patterns.[19] The adoption method compares individuals adopted in early childhood with their biological and adoptive parents on patterns like alcohol misuse. Although the adoption method is considered superior to the family and twin methods, it, like the other two methods, is not without limitations. Problems with the adoption method include failure to account for time spent with the biological mother prior to adoption, prenatal influences, and selective placement;[20] more specific concerns have been raised about the three major adoption cohorts in which alcohol abuse patterns have been studied. The Goodwin et al.[21,22] Danish adoption studies, for instance, have been criticized for employing confusing and

560

WALTERS

arbitrary classification criteria and a sample comprised largely of young adults who had not yet passed through the age of greatest risk for alcohol misuse.[23] The Swedish adoption studies,[24,25] on the other hand, have been reproached for relying on temperance board data because these data may confound alcohol and criminality.[26] Finally, research on Iowan adoptees[27,28] has been criticized for basing diagnoses of biological parent alcohol misuse on second-hand and potentially unreliable information.[29] There are several variables that potentially moderate the relationship between heredity and alcohol misuse. One such variable is gender. It has long been speculated that addictive liability for alcohol is higher in males than females. As Hill and Smith[30] point out, however, this does not rule out a genetic effect for female alcohol abuse and dependence. A second potential moderator of the gene – alcohol misuse relationship is the degree to which the subject sample derives from an ethnically homogeneous (Scandinavian countries) vs. ethnically heterogeneous (United States) population since more homogeneous samples have lower within group variance and a higher probability of achieving statistical significance than more heterogeneous samples.[20] The severity of alcohol abuse is a third potential moderating variable, for some researchers have found that heredity is more intimately involved in alcohol dependence than alcohol abuse or problem drinking.[31] The year a study was published is a fourth potential moderator variable, in part because it correlates significantly with design quality.[32] To reflect advances taking place in behavior genetic research on alcohol abuse/ dependence in the mid-1980s (e.g., routine use of blood tests for zygosity; development of criterion definitions and methods that were more structured and reliable than earlier definitions and methods), 1985 served as the cutting point for a study’s year of publication. Data for the present meta-analysis were gathered from behavior genetic (family, twin, adoption) studies on problem drinking and alcohol dependence in humans. Accordingly, molecular analyses, animal research, studies on substances other than alcohol, and investigations into alcohol use rather than misuse were excluded from this meta-analysis. In addition to explaining the overall gene – alcohol misuse relationship and its breakdown by moderator variable—gender (male, female), nationality (United States, foreign), pattern severity (more severe, less severe), and year of publication (before 1985, since 1985)—the present meta-analysis also produced effect sizes for family, twin, and adoption studies. The null hypothesis held that there would be no relationship between heredity, as measured by the three behavior genetic methodologies, and alcohol abuse/dependence. In the event the null hypothesis could be rejected it was reasoned that the gene – alcohol misuse relationship would be stronger for males, studies conducted outside the United States, samples displaying more severe patterns of alcohol misuse, and investigations published prior to 1985.

HERITABILITY OF ALCOHOL ABUSE

561

METHOD Selection of studies for this meta-analysis began with a computerized search of the PsycINFO and MEDLINE databases for studies published between 1970 and 2000 in which the following key words served as identifiers: alcohol, alcoholism, drinking, gene, heredity, family history, pedigree, twin, and adoption. Additional studies were gleaned from the reference sections of articles identified by the electronic search. This procedure yielded 50 behavior genetic studies on alcohol abuse/dependence (22 family studies, 18 twin studies, 10 adoption studies) and 72 total effect sizes. Several studies employed over-lapping samples and some investigators reported multiple outcomes for the same group of individuals. Multiple outcomes are listed only in studies where they provide novel information (e.g., separate diagnoses of alcohol abuse and dependence) and then only as an averaged composite before being combined with other studies.[21,33] A case-to-case statistical model in the form of a 2 £ 2 table of outcomes, whereby a phi coefficient was calculated by contrasting abuse status (absent vs. present) with genetic status (FH þ /MZ/biological parent þ vs. FH 2 /DZ/biological parent – ), was employed in this meta-analysis of behavior genetic research on alcohol misuse. The case-to-case model was preferred over the case-to-base rate model proposed by Gottesman and Carey,[34] in which a tetrachoric coefficient is calculated, because the case-to-case model allows for direct comparisons of subjects from the same sample, whether that entails contrasting people with and without a family history of alcohol abuse/dependence, MZ and DZ twins concordant and discordant for alcohol abuse/dependence, or adoptees with and without a history of biological parent alcohol abuse/dependence. It was reasoned that the case-to-case model more clearly captures the spirit of gene –alcohol misuse research. Accordingly, only family studies possessing a control or nonalcohol misusing comparison group are included in this meta-analysis. The phi coefficients obtained in this study were transformed into Fisher’s z for the purpose pooling the results of different studies and then backtransformed into phi coefficients. The minimum, maximum, median, weighted (by sample size) mean, and unweighted mean effect sizes (f ) were calculated, along with the 95% confidence interval for the weighted mean effect size. The statistical procedures used to combine studies, test for homogeneity, and compute a 95% confidence interval were based on the work of Hedges and Olkin.[35] Sampling error was calculated using the Schmidt – Hunter method —s2e ¼ ð1 2 r 2 Þ 2* kÞ=N—where r 2 represents the average weighted mean of the effect size, k the number of studies, and N the total sample size.[36] Fleiss[37] recommends use of the odds ratio instead of the phi coefficient to summarize categorical effect size data. This issue is taken up further in the “Discussion” section of this paper. Additional analyses were calculated using the four moderator variables of proband gender (male, female), sample nationality (United States, foreign),

562

WALTERS

pattern severity (more severe, less severe), and year of publication (before 1985, since 1985). The breakdowns for gender, nationality, and year of publication are self-evident, while pattern severity was coded less when the alcohol use pattern was described as abuse or problem drinking and more when the pattern was diagnosed as alcoholism or alcohol dependence. Samples were considered mixed if they contained subjects classified as both high and low severity. Gender was dummy coded using a three-category system (male ¼ 1, both ¼ 2, female ¼ 3), nationality using a two-category system (United States ¼ 1, foreign or nonUnited States ¼ 2), pattern severity using a three-category system (less severe ¼ 1 mixed ¼ 2, more severe ¼ 3), and year of publication using a two-category system (before 1985 ¼ 1; since 1985 ¼ 2).

RESULTS The 50 studies included in this meta-analysis of behavior genetic research on alcohol misuse are listed in Table 1 (family studies), Table 2 (twin studies), and Table 3 (adoption studies). Mean phi values of 0.12 (weighted) and 0.15 (unweighted) were found for the entire population of effect sizes ðk ¼ 72; total N ¼ 58; 887Þ: These values, along with the coefficients attained by family, twin, and adoption studies, are listed in Table 4. It should be noted, however, that heterogeneity was high for three groups of pooled effect sizes—all studies, Q ð71Þ ¼ 253:51; p , 0:001; family studies, Q ð30Þ ¼ 175:27; p , 0:001; and adoption studies, Q ð10Þ ¼ 38:49; p , 0:001—and low for twin studies— Q ð29Þ ¼ 40:07; p ¼ 0:08: Consequently, the analyses were broken down further by moderator variable—i.e., gender (male – female), location (US-foreign), pattern severity (more –less), and publication date (before 1985, since 1985)— but with only a few exceptions (i.e., US-adoption studies), the heterogeneity of the effect size estimates for the overall estimate as well as for family and adoption studies remained high. A multiple regression analysis of the four moderator variables on outcome (phi coefficients for all effect sizes included in the overall estimate) produced a multiple correlation of 0.38 and R 2 of 0.14. Beta weights for the four moderator variables that were regressed onto outcome were as follows: gender (2 0.25), nationality (2 0.14), severity (0.24), and year of publication (0.15). Only the gender and severity betas achieved statistical significance ðp , 0:05Þ: A statistical breakdown of effect sizes for all studies, family studies, twin studies, and adoption studies subdivided by moderator variable can be found in Table 5. In an attempt to identify the upper limit of heritability for alcohol misuse the 13 (11 twin, 2 adoption) studies with moderating conditions most favorable to a genetic interpretation of problem drinking (namely, male samples with more severe diagnoses of alcohol misuse) were analyzed as a group. The outcome

United States

Canada

United States

United States

United States

203

Sweden

Description

Alcoholics

Alcoholics

M Inpt and alcoholics F 56 M Inpt and alcoholics F 62 M Inpt and alcoholics F

F

24

500

M

183

M

Sex

Inpt and outpt alcoholics 50 M Inpt and alcoholics F from US

N

Location

Proband Subjects

608

944

230

662

441

300

85

Surgical pts

Description

Father

Father

Parents

Excessive alcohol consumption Hospitalization; neg. consequences from drinking

Alcoholism

Alcoholism

Father

Father

Alcoholism

Father

Diagnostic Criteria Alcohol abuse w/social consequences Alcoholism

Father

Relation

Family Members

10.0

6.2

16.1

0.9

1.2

0.0

9.7

3.4

23.2

27.6

4.2

4.9

22.0

26.2

Proband Control

Outcome (%)a

Family Studies on Alcohol Misuse

M Swiss and surgical F and psych pts M Mod. drinkers and F and nondrinkers F Mod. drinkers and nondrinkers M Hospital and employee F controls M Inpt with nonand alcohol F abuse dx M Pts w/medical and or affective F disorder dx

M

N Sex

Control Subjects

Table 1.

f

0.11

0.10

0.32

0.05

0.13

0.14

0.26

B

B

B

F

M

B

M

Sev

YP

[60]

[59]

[58]

[57]

[56]

[55]

Source

(continued )

US More Pre

For More Pre

US More Pre

US More Pre

US More Pre

US More Pre

For More Pre

Sex Nat

Moderating Variablesb

HERITABILITY OF ALCOHOL ABUSE 563

56

32

United States

United States

United States

United States

United States

United States

40

United Kingdom

F

M

Sex

Inpt alcoholics

Alcoholics

Description

M Alcoholic and halfsiblings F 51 M Inpt alcoholics 32 F Inpt alcoholics 39 M Dependent prob drinking 21 F Dependent prob drinking 72 M Alcohol and abusing F adol 166 M DSM-III alcohol abuse/dep

N

Location

Proband Subjects

M Nonalcohol and abusing F adol 455 M Nonalcohol abusers

1277

M

Description

Pts with physical illness unrelated to alcohol 56 F Nonalcoholic drug abusers 132 M Nonalcoholic and halfF siblings 7 M Inpt psych controls 32 F Inpt psych controls 723 M Nonalcohol abusers and F 989 F Nonalcohol abusers

40

N Sex

Control Subjects

Continued

Alcoholism

Parents

Alcoholism/problem drinking

1st and Alcoholism 2nd degree Parents Alcohol abuse or dependence

Parents

1st Alcoholic degree drinking style 1st Alcoholic degree drinking style Parents Alcoholism/problem drinking

Heavy drinking

Alcohol addiction; hospitalization; withdrawal symptoms

Father

Father

Diagnostic Criteria

Family Members Relation

Table 1.

10.6

30.8

36.1

40.3

20.4

22.5

17.9

50.0

59.4

23.8

28.6

20.0

12.5

2.5

56.8

62.0

46.4

2.5

Proband Control

Outcome (%)a

0.16

0.09

0.02

0.14

0.09

0.18

0.38

0.37

0.00

f

M

B

F

M

F

M

B

F

M

Sev

YP

US Mix

US Less

Post

Post

US More Post

US More Post

US More Pre

US More Pre

US More Pre

US More Pre

For More Pre

Sex Nat

Moderating Variablesb

[66]

[65]

[64]

[15]

[63]

[62]

[61]

Source

564 WALTERS

United States

Germany

M

F

38

38

M

66

49

Denmark

M

2806

United States

39

338

United States

89

United Kingdom

F

M and F M and F M and F M and F M

60

United States

67

DSM-III 971 F Nonalcohol alcohol abusers abuse/dep DSM-III 159 M Nonalcohol alcohol and abusers abuse/dep F Outpt 83 M Nonproblem and drinkers problem F drinkers 2135 M Nonalcohol DSM-III abusers and alcohol abuse/dep F 20,346 M Nonalcohol DSM-III-R abusers and alcohol F dep DSM-III-R 182 M Nonalcohol alcohol abusers abuse DSM-III-R 182 M Nonalcohol alcohol abusers dep Inpatients 64 M Inpatients w/alcoholw/depresism dx sion dx Inpatients 96 F Inpatients w/alcoholw/depresism dx sion dx DSM-III-R 247 M Nonalcohol alcohol abusers abuse Treatment of alcoholism

Treatment for alcoholism

42.1

71.0

1st and Father þ another 2nd 1st or 2nd degree degree alcohol abuse/dep

81.2

76.9

73.5

1st Alcoholism degree

1st Alcoholism degree

Parents

Parents

32.9

44.5

24.0

32.8

62.7

62.7

21.8

17.0

31.6

1st Alcoholism or degree problem drinking

16.9

55.1

1st and Problem 2nd drinking degree 1st Alcoholism degree

24.0

57.9

50.7

65.0

Alcohol abuse or dependence Alcoholism

Parents

Parents

0.18

0.18

0.50

0.12

0.10

0.09

0.13

0.40

0.06

0.15

M

F

M

M

M

B

B

B

B

F

Post

Post

Post

Post

Post

US Less

[71]

[43]

[70]

[69]

[68]

[67]

[48]

(continued )

Post

For More Post

For More Post

For More Post

For Less

US More Post

US Mix

For Less

US Mix

US Mix

HERITABILITY OF ALCOHOL ABUSE 565

M

F

166

F

296

640

M

M

73

916

Sex

N

DSM-III-R alcohol dep DSM-III R alcohol dep DSM-III-R alcohol dep DSM-III-R alcohol dep

DSM-III-R alcohol dep

Description

F

M

343

245

F

M

M

106

111

247

N Sex

Nonalcohol abusers

Nonalcohol abusers

Nonalcohol abusers

Nonalcohol abusers

Nonalcohol abusers

Description

Control Subjects

1st Alcohol abuse degree or dependence

1st Alcohol abuse degree or dependence

1st Alcohol degree dependence

1st and Father þ another 2nd 1st or 2nd degree degree alcohol abuse/dep 1st Alcohol degree dependence

Relation

Diagnostic Criteria

Family Members

52.4

45.6

23.8

49.7

75.3

20.4

19.8

6.0

19.8

44.5

Proband Control

Outcome (%)a

0.33

0.26

0.12

0.17

0.26

f

F

M

F

M

M

Sev

YP

US More Post

US More Post

US More Post

US More Post

US More Post

Sex Nat

Moderating Variablesb

[53]

[72]

Source

Note: adol ¼ adolescent; dep ¼ dependence; dis ¼ disorder; dx ¼ diagnosis; inpt ¼ inpatient; neg ¼ negative; outpt ¼ outpatient; prob ¼ problem; psych ¼ psychiatric; pts ¼ patients. a Outcome is the percentage of proband (alcohol abusing) and control (nonalcohol abusing) subjects with a family history of alcohol abuse, family being defined under relation and alcohol abuse being defined under diagnostic criteria. b Moderating variables: proband sex or gender (M ¼ male, F ¼ female, B ¼ both); proband nationality (US ¼ United States, For ¼ foreign or outside the United States), Sev ¼ severity of alcohol abuse in proband subjects (More ¼ more severe, Less ¼ less severe, Mix ¼ mixture of high and low severity), YP ¼ year of publication (Pre ¼ before 1985, Post ¼ since 1985).

United States

United States

Location

Proband Subjects

Table 1. Continued

566 WALTERS

Minnesota Twin Registry United DSM-III Alcohol States Abuse DSM-III Alcohol Dep DSM-III Alcohol Abuse DSM-III Alcohol Dep

Broad definition of alcohol abuse Swedish Twin Registry I Sweden Alcoholism

Finland

Finish Twin Cohort Finland Alcoholism

United States

Sweden United Kingdom Alcoholism

Alcohol Abuse Chronic Alcoholism Alcoholism Heavy Alcohol Use Alcoholism Alcoholism

Sweden

Finland

Diagnostic Criteria

Location

M M F F

39 30 24

M

95 50

M F M

F M

13 271

69 7 64

M M M M M and F M

Sex

58 27 172 198 (750)d 15

Nb

MZ Twins

25.0

26.7

59.0

74.0

12.6

13.0 0.0 10.9

8.0 26.3

54.0 71.0 26.0 75.0 22.0 33.0

Concordc

20

22

47

64

187

175 20 186

8 444

20

138 60 557 641

Nb

F

F

M

M

M

M F M

F M

M M M M M and F M

Sex

DZ Twins

5.0

27.3

36.2

57.8

9.1

5.7 0.0 6.4

13.0 11.9

28.0 32.0 12.0 63.0 16.0 30.0

Concordc

Table 2. Twin Studies on Alcohol Misuse

F

F

2 0.01 0.27

M

M

M

0.23

0.17

0.07

M F M

F M

2 0.08 0.18

0.12 0.00 0.07

M M M M B M

Sex

0.24 0.36 0.15 0.02 0.08 0.04

f

US

US

US

US

For

For For For

For US

For For For For For For

Nat

More

Less

More

Less

More

More More Less

More More

Less More More Less More More

Sev

[44]

[80]

[79]

[78]

[77]

[75] [76]

[74]

[73]

Source

(continued )

Post

Post

Post

Post

Post

Pre Pre Post

Pre Pre

Pre Pre Pre Pre Pre Pre

YP

Moderating Variablesa

HERITABILITY OF ALCOHOL ABUSE 567

M

M F

378

42 63 396 932

Australian Twin Registry Australia Alcohol Dependence Alcohol Dependence M F

F M

73 505

M F

54

Cloninger Type II

M

Sex

203

54

Nb

Cloninger Type I

Diagnostic Criteria

Virginia Twin Registry United Alcohol Dependence States Problem Drinking United DSM-IV Alcohol States Abuse DSM-IV Alcohol Dep Volunteer Twin Sample United Problem Drinking States Problem Drinking

United States

Location

MZ Twins

38.9 20.9

11.1

28.6

31.7

46.9 40.0

26.2

57.4

48.1

Concordc

231 534

24

12

436

55 316

154

65

M F

F

M

M

F M

F

M

M

Sex

DZ Twins

Continued

Nb 65

Table 2.

19.9 9.2

8.3

8.3

19.3

31.5 29.8

11.9

32.3

32.3

Concordc

0.20 0.11

0.04

0.20

0.14

0.16 0.11

0.17

0.25

0.16

f

M F

F

M

M

F M

F

M

M

Sex

For For

US

US

US

US US

US

US

US

Nat

More More

Less

Less

More

Less Less

More

More

Less

Sev

Post Post

Post

Post

Post

Post Post

Post

Post

Post

YP

Moderating Variablesa

[41]

[45]

[81]

[16]

[33]

Source

568 WALTERS

M

M F

131 215

M

709

710

M

753

74.0 67.0

53.2

28.6

31.3

76 175

588

842

1209

M F

M

M

M

52.6 60.0

43.2

21.1

21.6

0.22 0.07

0.10

0.09

0.11

M F

M

M

M

For For

US

US

For

Less Less

More

Less

Less

Post Post

Post

Post

Post

[54]

[84]

[83]

[82]

Moderating variables: proband sex or gender (M ¼ male, F ¼ female, B ¼ both); proband nationality (US ¼ United States, For ¼ foreign or outside the United States), Sev ¼ severity of alcohol abuse in proband subjects (More ¼ more severe, Less ¼ less severe, Mix ¼ mixture of high and low severity), YP ¼ year of publication (Pre ¼ before 1985, Post ¼ since 1985). b Number of twin pairs. c Pair-wise concordance for alcohol abuse. d The authors of this study failed to specify the number of MZ and DZ twins completing questionnaires on alcohol use and abuse patterns but did indicate that approximately 1500 pairs of twins were sent questionnaires. Given a 50% rate of return of mailed questionnaires it is estimated that approximately 750 twin pairs participated in this study. The phi was calculated from the percentages given by the authors and an N of 750 was used to pool these data with other twin and genetic studies in the derivation of a weighted mean f. SR ¼ self report.

a

Vietnam Era Twin Registry United Alcoholism States Canadian Twin Registry Canada Alcohol Misuse

Swedish Twin Registry II Sweden Temperance Board Registration World War II Twin Registry United Heavy alcohol States Consumption

HERITABILITY OF ALCOHOL ABUSE 569

13.0 48.7 70.6

39.4 7.0 7.8 24.1 0.9

18.2 9.1 33.3

3.7

Outcomeb

69 404 34

892 741 555 469 546

78 78 90

22

N

M M and F M and F

M F M M F

M M F

M and F

Sex

1.4 13.9 55.1

13.1 2.6 4.9 12.8 1.3

5.1 14.1 52.2

4.5

Outcomeb

Control Subjects

0.26 0.25 0.16

0.21 0.10 0.06 0.12 2 0.01

M B B

M F M M F

M M F

B

2 0.02 0.21 2 0.08 2 0.09

Sex

f

US US US

For For For For For

For For For

US

Nat

More Less Less

Less Less Less Less Less

More Less More

Less

Sev

Pre Post Post

Pre Pre Pre Post Post

Pre Pre Pre

Pre

YP

Moderating Variablesa

[27] [28] [88]

[86] [24] [87] [25]

[22]

[21]

[85]

Source

Moderating variables: proband sex or gender (M ¼ male, F ¼ female, B ¼ both); proband nationality (US ¼ United States, For ¼ foreign or outside the United States), Sev ¼ severity of alcohol abuse in proband subjects (More ¼ more severe, Less ¼ less severe, Mix ¼ mixture of high and low severity), YP ¼ year of publication (Pre ¼ before 1985, Post ¼ since 1985). b Percent of proband (alcohol abusing) and control (nonalcohol abusing) adoptees with at least one alcohol abusing biological parent.

a

M M and F M and F

M F M M F

89 172 307 108 114 23 39 49

M M F

55 55 6

Iowa Adoption Cohort United States Alcoholism United States Alcohol abuse United States Alcohol abuse

M and F

27

United States Alcohol abuse Danish Adoption Cohort Denmark Alcoholism Problem drinking Denmark Alcoholism Stockholm Adoption Study Sweden Alcohol abuse Sweden Alcohol abuse Sweden Severe alcohol abuse Sweden Alcohol abuse

Sex

N

Diagnostic Criteria

Location

Proband Subjects

Table 3. Adoption Studies on Alcohol Misuse

570 WALTERS

HERITABILITY OF ALCOHOL ABUSE Table 4.

Number of f estimates Maximum f Median f Minimum f Unweighted mean (f ) Weighted mean (f ) 95% Confidence intervala a

571

Effect Sizes for Studies on Alcohol Misuse Overall Effect

Family Studies

Twin Studies

Adoption Studies

72 0.50 0.13 2 0.09 0.15 0.12 0.11 – 0.12

31 0.50 0.14 0.00 0.18 0.12 0.11 –0.13

30 0.36 0.12 20.08 0.13 0.12 0.10– 0.13

11 0.26 0.11 2 0.09 0.10 0.12 0.09 – 0.14

Calculated from the weighted mean f and standard error of the weighted mean.

of this restricted analysis revealed a weighted mean f of 0.15 (95% confidence interval ¼ 0.12 –0.18) and unweighted mean f of 0.18 (95% confidence interval ¼ 0.15 –0.21) in a sample of homogeneous effect sizes, Q ð12Þ ¼ 15:67; p . 0:10: Heritability is calculated by doubling the mean effect size estimates of a correlational measure like the phi coefficient. This results in heritability estimates of 30% (weighted) and 36% (unweighted) for males with severe alcohol dependence, which is somewhat higher than the 20 – 26% heritability suggested by the full sample of twin and adoption studies.

DISCUSSION The outcome of this meta-analysis paints a somewhat different picture than the one drawn by the more enthusiastic proponents of the genetic view of alcohol misuse (Cloninger in Refs. [13,38,39]). The heritability of alcohol misuse, which can be estimated by doubling the effect sizes attained in the present meta-analysis by all twin and adoption studies, appears to range between 20 and 26%. Even when the analyses are restricted to studies most favorable to the genetic hypothesis—i.e., males diagnosed with severe forms of alcohol dependence— heritability does not appear to exceed 30– 36%, which is somewhat lower than the 40 – 60% rate normally cited in the literature. Therefore, Wilson and Crowe’s[40] quandary over whether we should identify people at risk for alcohol misuse for preventative purposes is deemed moot to the extent that the gene – alcohol misuse relationship is probably too weak and variable to permit reliable identification. Our limited time, energy, and financial resources might therefore be better spent clarifying the boundaries and parameters of this relationship than metaphorically spinning our wheels searching for phantom “alcoholism genes” independent of their environmental context.

Overall effect Number of f estimates Maximum f Median f Minimum f Unweighted mean (f ) Weighted mean (f ) 95% Confidence intervala Family Studies Number of f estimates Maximum f Median f Minimum f Unweighted mean (f ) Weighted mean (f ) 95% Confidence intervala 20 0.37 0.10 20.09 0.10 0.10 0.08–0.12

8 0.37 0.14 0.02 0.17 0.12 0.09–0.15

13 0.50 0.17 0.00 0.19 0.19 0.16–0.21

Female

38 0.50 0.15 0.00 0.16 0.14 0.13–0.16

Male

23 0.38 0.14 0.02 0.17 0.11 0.10–0.12

41 0.38 0.16 20.02 0.16 0.12 0.11–0.13

US

8 0.50 0.15 0.00 0.21 0.17 0.13–0.21

31 0.50 0.09 20.09 0.13 0.12 0.10–0.13

Foreign

Nationality

23 0.50 0.14 0.00 0.19 0.11 0.10–0.12

43 0.50 0.14 20.09 0.17 0.12 0.11–0.13

More

4 0.40 0.14 0.09 0.20 0.13 0.09–0.18

27 0.40 0.11 20.08 0.12 0.11 0.09–0.13

Less

Pattern Severity

Effects of Moderator Variables on the Genes – Alcohol Misuse Relationship

Proband Gender

Table 5.

12 0.38 0.14 0.00 0.18 0.16 0.13–0.19

29 0.38 0.11 20.09 0.13 0.14 0.12–0.15

Pre-85

19 0.50 0.15 0.02 0.18 0.11 0.10–12

43 0.50 0.14 20.01 0.16 0.11 0.10–0.12

Post-85

Year of Publication

572 WALTERS

a

9 0.27 0.07 20.08 0.08 0.11 0.07–0.18

3 0.10 20.01 20.09 20.00 0.04 20.00–0.09

20 0.36 0.14 0.02 0.15 0.12 0.10–0.14

5 0.26 0.12 0.06 0.14 0.14 0.10 2 0.17

4 0.26 0.20 20.02 0.16 0.22 0.15–0.29

14 0.27 0.16 20.01 0.15 0.12 0.10–0.15

7 0.21 0.06 20.09 0.06 0.10 0.07–0.13

16 0.36 0.11 20.08 0.11 0.11 0.09–0.13

Calculated from the weighted mean f and standard error of the weighted mean.

Twin studies Number of f estimates Maximum f Median f Minimum f Unweighted mean (f ) Weighted mean (f ) 95% Confidence intervala Adoption studies Number of f estimates Maximum f Median f Minimum f Unweighted mean (f ) Weighted mean (f ) 95% Confidence intervala 3 0.26 0.21 20.09 0.13 0.14 0.03–0.24

17 0.36 0.14 20.08 0.14 0.13 0.11–0.16

9 0.25 0.10 20.08 0.09 0.11 0.08–0.14

13 0.24 0.11 20.01 0.12 0.10 0.08–0.12

7 0.26 0.06 20.09 0.08 0.12 0.08–0.15

10 0.36 0.10 20.08 0.11 0.12 0.08–0.15

4 0.25 0.14 20.01 0.13 0.11 0.06–0.15

20 0.27 0.14 20.01 0.14 0.12 0.10–0.14

HERITABILITY OF ALCOHOL ABUSE 573

574

WALTERS

Four potential moderator variables were examined in this study: proband gender, sample nationality, pattern severity, and year of publication. Congruent with a number of individual studies in which male and female probands have been compared,[25,41 – 45] the heritability of alcohol misuse was stronger in males than females. While this may reflect a genuine male – female difference in genetic liability for alcohol misuse as represented by Cloninger’[46] Type II or malelimited alcoholism pattern, many of the comparisons involving females suffered from low power due to small sample sizes and decreased rates of alcohol misuse compared to males. When analyses were restricted to female studies with sample sizes larger than 100 ðk ¼ 13Þ the weighted effect size rose slightly (0.10 – 0.11) and the unweighted effect size showed moderate improvement (from 0.10 to 0.14), both figures of which approach the effect sizes attained for males in this metaanalysis. These findings lend support to Heath’s[39] assertion that the gene – alcohol misuse association may be as strong and consistent in women as it is in men. There is evidence from the results of this meta-analysis that the severity of alcohol abuse may moderate the gene – alcohol misuse relationship. One might be tempted to conclude from this that there are two types of alcohol misuse, one which is more severe and genetically influenced and the other which is less severe and mediated principally by environmental factors. These two patterns conform in a general way to Cloninger’s[46] Type II (male-limited) and Type I (milieu-limited) categories of alcohol misuse, respectively. However, the present findings are also compatible with a continuum view of alcohol misuse in which the continuum extends from mild to severe alcohol misuse and where genetic contributions vary both quantitatively and qualitatively at different points along the continuum. In contrasting the dichotomy and continuum views on the gene – alcohol misuse relationship we would be well advised to keep in mind that while the difference in effect sizes for studies using more and less severe definitions of alcohol misuse was modest to moderate, the confidence intervals for more and less severe definitions of alcohol misuse overlap significantly (Table 5). The third and fourth moderating variables examined in this meta-analysis (sample nationality, date of publication) had little appreciable effect on the results obtained by this study. This is good news for supporters of the genetic perspective on alcohol misuse for two reasons. Firstly, it confirms that the gene – alcohol misuse relationship is not restricted to ethnically homogeneous populations for it was just as likely to surface in a culturally diverse culture like the United States as it was in more ethnically homogeneous cultures such as those found in the Scandinavian countries. Secondly, a significant genetic effect is just as likely to occur in more recently published and presumably, more methodologically sound investigations as it is to appear in earlier and less methodologically rigorous studies. This second finding certifies that the gene – alcohol misuse correlation is not simply an artifact of poor quality research designs.

HERITABILITY OF ALCOHOL ABUSE

575

Overall, the four moderator variables accounted for only 14% of the variance in the effect sizes procured from this meta-analysis. This indicates that much of the heterogeneity in the gene – alcohol misuse relationship remains unexplained. Measurement error and interactions leading to nonshared environmental effects may account for a substantial portion of the heterogeneity in effect sizes. What variance remains once measurement error and nonshared environmental experience are extracted from the equation is probably attributable to variables that have not been routinely investigated in behavior genetic research on alcohol misuse. Either way, a great deal more research is required before we are in a position to offer firm conclusions as to the relationship between heredity and alcohol misuse. If the relative modesty of the effect sizes revealed in this meta-analysis is not enough to discourage simplistic genetic interpretations of alcohol misuse then the heterogeneity witnessed between the different studies should send a clear message of caution to even the most ardent of genetic reductionists. Comparing the present findings with outcomes registered in an earlier meta-analysis of crime studies[32] reveals that the gene – alcohol misuse relationship is no stronger, and is actually weaker for five of the six family, twin, and adoption contrasts (the weighted effect size for adoption studies being the one exception), than the gene – crime relationship (see Fig. 1). These findings insinuate that crime may be as strongly genetic as alcohol abuse. Other results denote that a portion of the variance traditionally ascribed to genetic differences in alcohol abuse may actually be a function of genetic differences in crime. Besides Cloninger’s[46] observation that the fathers of Type II alcoholics own more extensive records of prior criminality than the fathers of persons exhibiting the less genetically influenced Type I pattern, Cadoret and Gath[47] ascertained that childhood conduct disorder predicted later alcohol misuse in adoptees and correlated, albeit nonsignificantly, with biological parent alcohol misuse. Furthermore, Stabenau[48] determined that a diagnosis of antisocial personality disorder achieved a three-fold increase in liability for alcohol abuse, doubling the contributions of male gender and a family history of alcohol misuse in predicting personal misuse of alcohol. In interpreting the results of the present meta-analysis it should be kept in mind that the case-to-case method upon which the current analyses were based possesses less statistical power than the multivariate and model testing procedures used in many of the individual studies. However, multivariate and model testing procedures also tend to capitalize on large sample sizes, whereas the case-to-case approach provides a more equitable and conservative estimate of the gene – behavior relationship.[49] This may explain why the heritability estimates from this meta-analysis were lower than anticipated and may even slightly underestimate the gene – alcohol misuse relationship. Nevertheless, the case-to-case method permits inclusion of many more studies than could be accommodated

576

WALTERS

Figure 1.

A Comparison of Effect Sizes for Crime and Alcohol Misuse.

by multivariate or model testing procedures. By providing a statistic common to all three behavior genetic methodologies, the case-to-case approach furnishes a procedure by which family, twin, and adoption studies can be combined and compared. Despite its inclusiveness, the case-to-case method cannot encompass every relevant study. As such, individual studies not covered in this meta-analysis and excluded analyses from some of the studies included in the meta-analysis should be considered alongside the present meta-analytic results as part of a comprehensive evaluation of the proposed gene – alcohol misuse relationship. One such study was an investigation by Vernon et al.[50] in which the social learning concept of alcohol expectancies was found to be influenced by genetic factors, suggesting that heredity may be involved in alcohol misuse in a number of subtle and intricate ways. Advocates of the twin method will probably take issue with the fact that pair-wise rather than proband-wise concordance was used to calculate the effect

HERITABILITY OF ALCOHOL ABUSE

577

sizes for the twin study portion of this meta-analysis. However, the pair-wise method (concordant twin pairs/concordant twins pairs þ discordant twin pairs) seems more consistent with the way in which family and adoption study data were analyzed in this meta-analysis than the proband-wise method (2 £ concordant twin pairs/[2 £ concordant twin pairs] þ discordant twin pairs). Furthermore, there was very little difference in effect sizes between the pair-wise and proband-wise estimates. Substituting the proband-wise calculations for the pair-wise calculations increased the unweighted mean f effect size estimate by 0.01 (from 0.13 to 0.14), the weighted mean f effect size estimate by 0.01 (from 0.12 to 0.13), and elevated the 95% confidence interval from 0.10– 0.13 to 0.12– 0.15. Odds ratios were also calculated for all proband-wise twin comparisons, the outcome of which yielded an unweighted mean effect size of 2.15, a weighted mean effect size of 1.92, and a 95% confidence interval of 1.33 –2.61. These findings imply that neither the proband-wise method of calculating twin concordance nor the odds ratio approach advocated by Fleiss[37] significantly alter the pattern of results obtained from phi coefficients of pair-wise twin data. The outcomes achieved when behavior genetic research on alcohol abuse and dependence is subjected to meta-analysis show that these patterns are heritable to some extent, with the degree and type of influence still requiring further clarification. At this point in time the effect of genes on alcohol misuse cannot be denied, but neither can the effect of the environment. There has been a discernable shift within the alcohol abuse field over the past several years which has seen biological factors displace environmental and learning factors in explanations of problem drinking, a shift that threatens to accelerate as genetic mapping becomes a reality.[51,52] Whether shared environmental experience is as pivotal to alcohol misuse as it is to crime remains to be seen, although the interactive nature of nonshared environmental influence apparently plays a crucial role in the initiation and maintenance of alcohol misuse. In closing it is important to reiterate that the intent of this paper has not been to disparage or discourage genetic research on alcohol abuse and dependence, but rather to call for greater balance in our views on the subject. Rapprochement between the biological and learning camps is within our grasp but only if we are willing to reject genetic and environmental reductionism as a means of achieving a more comprehensive and interactive perspective on problem drinking.

ACKNOWLEDGMENTS The author would like to thank Scott Stoltenberg and Kerry Jang for their assistance in providing additional data from the Curran et al. and Jang et al.[53,54] studies, respectively. The assertions and opinions contained herein are the

578

WALTERS

private views of the author and should not be construed as official or as reflecting the views of the Federal Bureau of Prisons or the United States Department of Justice.

REFERENCES 1.

2. 3.

4. 5.

6.

7.

8.

9.

10.

Blum, K.; Nobel, E.P.; Sheridan, P.J.; Montgomery, A.; Ritchie, T.; Jagadeeswaran, P.; Nogami, H.; Briggs, A.H.; Cohn, J.B. Allelic Association of Human Dopamine D2 Receptor Gene in Alcoholism. J. Am. Med. Assoc. 1990, 263, 2055 –2060. Conrad, P. Public Eyes and Private Genes: Historical Frames, News Constructions, and Social Problems. Soc. Probl. 1997, 44, 139 –154. Bolos, A.M.; Dean, M.; Lucas-Derse, S.; Ramsburg, M.; Brown, G.L.; Goldman, D. Population and Pedigree Studies Reveal a Lack of Association Between the Dopamine D2 Receptor Gene and Alcoholism. J. Am. Med. Assoc. 1990, 264, 3156 –3160. White, M. 20/20. Drinking: Are You in Control? ABC News: New York, 2000, June 7. Cook, C.C.H.; Palsson, G.; Turner, A.; Holmes, D.; Brett, P.; Curtis, D.; Petursson, H.; Gurling, H.M.D. A Genetic Linkage Study of the D2 Dopamine Receptor Locus in Heavy Drinking and Alcoholism. Br. J. Psychiatry 1996, 169, 243 –248. Finckh, U.; von Widdern, O.; Giraldo-Velasquez, M.; Podschus, J.; Dufeu, P.; Sander, T.; Harms, H.; Schmidt, L.G.; Rommelspacher, H.; Rolfs, A. No Association of the Structural D2 Receptor (DRD2) Variation 311Cys with Alcoholism. Alcohol. Clin. Exp. Res. 1996, 20, 528– 532. Gelernter, J.; Kranzler, J. D2 Dopamine Receptor Gene (DRD2) Allele and Haplotype Frequencies in Alcohol Dependent and Control Subjects: No Association with Phenotype Or Severity of Phenotype. Neuropsychopharmacology 1999, 20, 640– 649. Goldman, D.; Urbanek, M.; Guenther, D.; Robin, R.; Long, J.C. A Functionally Deficient DRD2 Variant [Ser311Cys] is Not Linked to Alcoholism and Substance Abuse. Alcohol 1998, 16, 47 –52. Lee, J.-F.; Lu, R.-B.; Ko, H.C.; Chang, F.-M.; Shih-Juin, Y.; Pakstis, A.J.; Kidd, K.K. No Association Between DRD2 Locus and Alcoholism After Controlling for ADH and ALDH Genotypes in Chinese Han Population. Alcohol. Clin. Exp. Res. 1999, 23, 592 – 599. McAneny, L. Drinking a Cause of Family Problems for Three out of Ten Americans. Gallup Poll Release; Gallup News Service: Princeton, NJ, 1997, June 6.

HERITABILITY OF ALCOHOL ABUSE

11.

579

Sutcliffe, J.G. mRNA in the Mammalian Central Nervous System. Annu. Rev. Neurosci. 1988, 11, 157– 198. 12. Schork, N.J.; Schork, C.M. Issues and Strategies in the Genetic Analysis of Alcoholism and Related Addictive Behaviors. Alcohol 1998, 16, 71 – 83. 13. McGue, M. The Biological Genetics of Alcoholism. Curr. Dir. Psychol. Sci. 1999, 8, 109– 115. 14. Cotton, N.S. The Familial Incidence of Alcoholism. J. Stud. Alcohol 1979, 40, 89– 116. 15. Stabenau, J.R.; Hesselbrock, V.M. Family Pedigree of Alcoholic and Control Patients. Int. J. Addict. 1983, 18, 351 – 363. 16. Kendler, K.S.; Neale, M.C.; Heath, A.C.; Kessler, R.C.; Eaves, L.J. A TwinFamily Study of Alcoholism in Women. Am. J. Psychiatry 1994, 151, 707– 715. 17. Kendler, K.S.; Neale, M.C.; Kessler, R.C.; Heath, A.C.; Eaves, L.J. A Test of the Equal-Environment Assumption in Twin Studies of Psychiatric Illness. Behav. Genet. 1993, 23, 21 –27. 18. LaBuda, M.C.; Svikis, D.S.; Pickens, R.W. Twin Closeness and Co-Twin Risk for Substance Use Disorders: Assessing the Impact of the Equal Environment Assumption. Psychiatry Res. 1997, 70, 155 –164. 19. Rose, R.J.; Kaprio, J.; Williams, C.J.; Viken, R.; Obremski, K. Social Contact and Sibling Similarity: Facts, Issues, and Red Herrings. Behav. Genet. 1990, 20, 763 –778. 20. Walters, G.D.; White, T.W. Heredity and Crime: Bad Genes Or Bad Research? Criminology 1989, 27, 455– 485. 21. Goodwin, D.W.; Schulsinger, F.; Hermansen, L.; Guze, S.B.; Winokur, G. Alcohol Problems in Adoptees Raised Apart From Alcoholic Biological Parents. Arch. Gen. Psychiatry 1973, 28, 238– 243. 22. Goodwin, D.W.; Schulsinger, F.; Knop, J.; Mednick, S.; Guze, S.B. Alcoholism and Depression in Adopted-Out Daughters of Alcoholics. Arch. Gen. Psychiatry 1977, 34, 751 –755. 23. Murray, R.M.; Clifford, C.A.; Gurling, H.M.D. Twin and Adoption Studies: How Good is the Evidence for a Genetic Risk? In Recent Developments in Alcoholism; Galanter, M., Ed.; Plenum: New York, 1983; Vol. 1, 25 – 48. 24. Bohman, M.; Sigvardsson, S.; Cloninger, C.R. Maternal Inheritance of Alcohol Abuse: Cross-Fostering Analysis of Adopted Women. Arch. Gen. Psychiatry 1981, 38, 965 –969. 25. Sigvardsson, S.; Bohman, M.; Cloninger, C.R. Replication of the Stockholm Adoption Study of Alcoholism: Confirmatory Cross-Fostering Analysis. Arch. Gen. Psychiatry 1996, 53, 681 – 687.

580

26.

WALTERS

Lester, D. The Heritability of Alcoholism: Science and Social Policy. Current Issues in Alcohol/Drug Studies; Haworth Press: Binghamton, NY, 1989; 29– 68. 27. Cadoret, R.J.; Cain, C.A.; Grove, W.M. Development of Alcoholism in Adoptees Raised Apart From Alcoholic Biologic Relatives. Arch. Gen. Psychiatry 1980, 37, 561 – 563. 28. Cadoret, R.J.; Troughton, E.; O’Gorman, T.W.; Heywood, E. An Adoption Study of Genetic and Environmental Factors in Drug Abuse. Arch. Gen. Psychiatry 1986, 43, 1131 – 1136. 29. Searles, J.S. The Genetics of Alcoholism: Impact on Family and Sociological Models of Addiction. Fam. Dyn. Addict. 1991, 1, 8 – 21. 30. Hill, S.Y.; Smith, T.R. Evidence for Genetic Mediation of Alcoholism in Women. J. Subst. Abuse 1991, 3, 159 –174. 31. Pickens, P.W.; Svikis, D.S. Genetic Contributions to Alcohol Diagnosis. Alcohol Health Res. World 1991, 15, 272 –277. 32. Walters, G.D. A Meta-analysis of the Gene – Crime Relationship. Criminology 1992, 30, 595 –613. 33. van den Bree, M.B.M.; Johnson, E.O.; Neale, M.C.; Svikis, D.S.; McGue, M.; Pickens, R.W. Genetic Analysis of Diagnostic Systems of Alcoholism in Males. Biol. Psychiatry 1998, 43, 139– 145. 34. Gottesman, I.I.; Carey, G. Extracting Meaning and Direction From Twin Data. Psychiatr. Dev. 1983, 1, 35 –50. 35. Hedges, L.V.; Olkin, I. Statistical Methods of Meta-analysis; Academic Press: New York, 1985. 36. Hunter, J.E.; Schmidt, F.L.; Jackson, G.B. Meta-analysis: Cumulating Research Findings Across Studies; Sage: Beverly Hills, CA, 1982. 37. Fleiss, J.L. Measures of Effect Size for Categorical Data. In The Handbook of Research Synthesis; Cooper, H., Hedges, L.V., Eds.; Russell Sage: New York, 1994; 245– 260. 38. National Institute on Alcohol Abuse and Alcoholism, (1985). Alcoholism: An inherited disease. DHHS Publication No. ADM 85-1426. 39. Heath, A.C. Genetic Influences on Alcoholism Risk: A Review of Adoption and Twin Studies. Alcohol Health Res. World 1995, 19, 166 – 171. 40. Wilson, J.R.; Crowe, L. Genetics of Alcoholism: Can and Should Youth at Risk Be Identified? Alcohol Health Res. World 1991, 15, 11 –17. 41. Heath, A.C.; Bucholz, K.K.; Madden, A.F.; Dinwiddie, S.H.; Slutske, W.S.; Bierut, L.J.; Statham, D.J.; Dunne, M.P.; Whitfield, J.B.; Martin, N.G. Genetic and Environmental Contributions to Alcohol Dependence Risk in a National Twin Sample: Consistency of Findings in Women and Men. Psychol. Med. 1997, 27, 1381– 1396.

HERITABILITY OF ALCOHOL ABUSE

42.

43.

44.

45.

46. 47. 48. 49.

50.

51. 52.

53.

54.

55.

581

Jang, K.L.; Livesley, W.J.; Vernon, P.A. Gender-Specific Etiological Differences in Alcohol and Drug Problems: A Behavioural Genetic Analysis. Addiction 1997, 92, 1265 – 1276. Maier, W.; Lichtermann; Minges, J. The Relationship Between Alcoholism and Unipolar Depression—A Controlled Family Study. J. Psychiatr. Res. 1994, 28, 303– 317. Pickens, P.W.; Svikis, D.S.; McGue, M.; Lykken, D.T.; Heston, L.L.; Clayton, P.J. Heterogeneity in the Inheritance of Alcoholism: A Study of Male and Female Twins. Arch. Gen. Psychiatry 1991, 48, 19– 28. Prescott, C.A.; Hewitt, J.K.; Truett, K.R.; Heath, A.C.; Neale, M.C.; Eaves, L.J. Genetic and Environmental Influences on Lifetime AlcoholRelated Problems in a Volunteer Sample of Older Twins. J. Stud. Alcohol 1994, 55, 184 –202. Cloninger, C.R. Neurogenetic Adaptive Mechanisms in Alcoholism. Science 1987, 236, 410– 416. Cadoret, R.J.; Gath, A. Inheritance of Alcoholism in Adoptees. Br. J. Psychiatry 1978, 132, 252– 258. Stabenau, J.R. Additive Independent Factors That Predict Risk for Alcoholism. J. Stud. Alcohol 1990, 51, 164– 174. Walters, G.D. Behavior Genetic Research on Gambling and Problem Gambling: A Preliminary Meta-analysis of Available Data. J. Gambling Stud. 2001, 17, 255 –271. Vernon, P.A.; Lee, D.; Harris, J.A.; Jang, K.L. Genetic and Environmental Contributions to Individual Differences in Alcohol Expectancies. Pers. Individ. Differ. 1996, 21, 183 –187. Weiss, R.; Gillis, J. Clinton Scientists Celebrate “Working Draft” of Human Genetic Blueprint. The Washington Post 2000, June 7, A1. Crabb, J.C. Use of Genetic Analyses to Refine Phenotypes Related to Alcohol Tolerance and Dependence. Alcohol Clin. Exp. Res. 2001, 25, 288– 292. Curran, G.M.; Stotlenberg, S.F.; Hill, E.M.; Mudd, S.A.; Blow, F.C.; Zucker, R.A. Gender Differences in the Relationships Among SES, Family History of Alcohol Disorders and Alcohol Dependence. J. Stud. Alcohol 1999, 60, 825– 832. Jang, K.L.; Vernon, P.A.; Livesley, W.J. Personality Disorder Traits, Family Environment, and Alcohol Misuse: A Multivariate Behavioural Genetic Analysis. Addiction 2000, 95, 873 – 888. Amark, C. A Study in Alcoholism; Clinical, Social-Psychiatric and Genetic Investigations (D. Burton, Trans.). Acta Psychiatr. Neurol. Scand. 1951, Suppl. No. 70.

582

56.

WALTERS

Bleuler, M. Familial and Personal Background of Chronic Alcoholics. In Etiology of Chronic Alcoholism; Diethelm, O., Ed.; Thomas: Springfield, IL, 1955; 110– 166. 57. Jackson, J.K.; Connor, R. Attitudes of the Parents of Alcoholics, Moderate Drinkers and Nondrinkers Toward Drinking. Q. J. Stud. Alcohol 1953, 14, 596 – 613. 58. Oltman, J.E.; Friedman, S. A Consideration of Parental Deprivation and Other Factors in Alcohol Addicts. Q. J. Stud. Alcohol 1953, 14, 49 – 57. 59. Gregory, I. An Analysis of Family Data on 1000 Patients Admitted to a Canadian Mental Hospital. Acta Genet. Stat. Med. 1959, 9, 54– 96. 60. Pitts, F.N., Jr.; Winokur, G. Affective Disorders. VII. Alcoholism and Affective Disorder. J. Psychiatr. Res. 1966, 4, 37 –50. 61. Hassall, C. A Controlled Study of the Characteristics of Young Male Alcoholics. Br. J. Addict. 1968, 63, 193 –201. 62. Parker, F.B. Sex-Role Adjustment in Women Alcoholics. Q. J. Stud. Alcohol 1972, 33, 647 –657. 63. Schuckit, M.A.; Goodwin, D.A.; Winokur, G. A Study of Alcoholism in Half Siblings. Am. J. Psychiatry 1972, 128, 1132– 1137. 64. Parker, D.A.; Harford, T.C. Alcohol-Related Problems, Marital Disruption and Depressive Symptoms Among Adult Children of Alcohol Abusers in the United States. J. Stud. Alcohol 1988, 49, 306 – 313. 65. Pandina, R.J.; Johnson, V. Serious Alcohol and Drug Problems Among Adolescents with a Family History of Alcoholism. J. Stud. Alcohol 1990, 51, 278 –282. 66. Russell, M.; Cooper, M.L.; Frone, M.R. The Influence of Sociodemographic Characteristics on Familial Alcohol Problems: Data From a Community Sample. Alcohol. Clin. Exp. Res. 1990, 14, 221 –226. 67. Fletcher, K.D.; Price, D.K.; Cook, C.C.H. Problem Drinking and Family History. Br. J. Addict. 1991, 86, 1335– 1341. 68. Lewis, C.E.; Bucholz, K.K. Alcoholism, Antisocial Behavior and Family History. Br. J. Addict. 1991, 86, 177– 194. 69. Dawson, D.A.; Harford, T.C.; Grant, B.F. Family History as a Predictor of Alcohol Dependence. Alcohol. Clin. Exp. Res. 1992, 16, 572 – 575. 70. Goodwin, D.W.; Knop, J.; Jensen, P.; Gabrielli, W.F., Jr.; Schulsinger, F.; Penick, E.C. Thirty-Year Follow-up of Men at High Risk for Alcoholism. In Annals of the New York Academy of Science. Types of Alcoholics: Evidence from Clinical, Experimental, and Genetic Research; Babor, T.F., Hesselbrock, V., Meyer, R.E., Shoemaker, W., Eds.; New York Academy of Science: New York, 1994; Vol. 708, 97 –101. 71. Schuckit, M.A.; Smith, T.L. An 8-year Follow-up of 450 Sons of Alcoholic and Control Subjects. Arch. Gen. Psychiatry 1996, 53, 202– 210.

HERITABILITY OF ALCOHOL ABUSE

72.

73. 74.

75.

76.

77.

78.

79.

80.

81.

82.

83. 84.

583

Bierut, L.J.; Dinwiddie, S.H.; Begleiter, H.; Crowe, R.R.; Hesselbrock, V.; Nurnberger, J.I.; Porjesz, B.; Schuckit, M.A.; Reich, T. Familial Transmission of Substance Dependence: Alcohol, Marijuana, Cocaine, and Habitual Smoking. Arch. Gen. Psychiatry 1998, 55, 982 – 988. Kaij, L. Studies on the Etiology and Sequels of Abuse of Alcohol; University of Lund: Lund, Sweden, 1960. Partanen, J.; Brunn, K.; Markkanen, T. Drinking Behavior: A Study on Intelligence, Personality, and Use of Alcohol of Adult Twins; Finnish Foundation for Alcohol Studies: Helsinki, Finland, 1966. Jonsson, E.; Nilsson, T. Alkoholkonsumtion hos Monozygota och Dizygota Tvillingpar [Alcoholism in Monozygotic and Dizygotic Twins]. Nord. Hyg. Tidskr. 1968, 49, 21– 25. Gurling, H.M.D.; Murray, R.M.; Clifford, C.A. Investigations into the Genetics of Alcohol Dependence and into its Effects on Brain Function. Twin research 3: Epidemiology and Clinical Studies; Liss: New York, 1981; 87 –97. Hrubec, Z.; Omenn, G.S. Evidence of Genetic Predisposition to Alcoholic Cirrhosis and Psychosis: Twin Concordances for Alcoholism and Its Biological End Points By Zygosity Among Male Veterans. Alcohol. Clin. Exp. Res. 1981, 5, 207 – 215. Koskenvuo, M.; Langinvainio, H.; Kaprio, J.; Lo¨nnqvist, J.; Tienari, P. Psychiatric Hospitalization in Twins. Acta Genet. Med. Gemellol. 1984, 33, 321– 332. Romanov, K.; Kaprio, J.; Rose, R.J.; Koskenvuo, M. Genetics of Alcoholism: Effects of Migration on Concordance Rates Among Male Twins. Alcohol Alcohol. Suppl. 1991, 1, 137 – 140. Allgulander, C.; Nowak, J.; Rice, J.P. Psychopathology and Treatment of 30,344 Twins in Sweden II. Heritability Estimates of Psychiatric Diagnoses and Treatment in 12,884 Pairs. Acta Psychiatr. Scand. 1991, 83, 12– 15. Prescott, C.A.; Kendler, K.S. Genetic and Environmental Contributions to Alcohol Abuse and Dependence in a Population-based Sample of Male Twins. Am. J. Psychiatry 1999, 156, 34 –40. Kendler, K.S.; Prescott, C.A.; Neale, M.C.; Pedersen, N.L. Temperance Board Registration for Alcohol Abuse in a National Sample of Swedish Male Twins, Born 1902 –1949. Arch. Gen. Psychiatry 1997, 54, 178 –184. Swan, G.E.; Carmelli, D.; Cardon, L.R. Heavy Consumption of Cigarettes, Alcohol and Coffee in Male Twins. J. Stud. Alcohol 1997, 58, 182 –190. True, W.R.; Heath, A.C.; Bucholz, K.; Slutske, W.; Romeis, J.C.; Scherrer, J.F.; Lin, N.; Eisen, S.A.; Goldberg, J.; Lyons, M.J.; Tsuang, M.T. Models of Treatment Seeking for Alcoholism: The Role of Genes and Environment. Alcohol. Clin. Exp. Res. 1996, 20, 1577 –1581.

584

WALTERS

85.

Roe, A.; Burks, B. Adult Adjustment of Foster Children of Alcoholic and Psychotic Parentage and the Influence of the Foster Home. Memoirs of the Section on Alcohol Studies, Yale University, New Haven, CT. Q. J. Stud. Alcohol 1945, No. 3. Bohman, M. Some Genetic Aspects of Alcoholism and Criminality: A Population of Adoptees. Arch. Gen. Psychiatry 1978, 35, 269 –276. Cloninger, C.R.; Bohman, M.; Sigvardsson, S. Inheritance of Alcohol Abuse: Cross-Fostering Analysis of Adopted Men. Arch. Gen. Psychiatry 1981, 38, 861 –868. Cadoret, R.; Troughton, E.; Woodworth, G. Evidence of Heterogeneity of Genetic Effect in Iowa Adoption Studies. In Annals of the New York Academy of Science. Types of Alcoholics: Evidence from Clinical, Experimental, and Genetic Research; Babor, T.F., Hesselbrock, V., Meyer, R.E., Shoemaker, W., Eds.; New York Academy of Science: New York, 1994; Vol. 708, 59 – 71.

86. 87.

88.