or Major Depression

or Major Depression

Psychopathology in Children of Parents With Opiate Dependence and/or Major Depression EDWARD V. NUNES, M.D., MYRNA M. WEISSMAN, PH.D., RISE B. GOLDSTE...

944KB Sizes 6 Downloads 47 Views

Psychopathology in Children of Parents With Opiate Dependence and/or Major Depression EDWARD V. NUNES, M.D., MYRNA M. WEISSMAN, PH.D., RISE B. GOLDSTEIN, PH.D., GAIL McAVAY, M.S., ANGELA M. SERACINI, PH.D., HELENA VERDELI, M.Sc., M.A., AND PRIYA J. WICKRAMARATNE, PH.D.

ABSTRACT Objective: To evaluate psychiatric disorders and impairment in school-age and adolescent children of opiate-dependent patients. Method: One hundred fourteen children, aged 6 to 17 years, of 69 white methadone maintenance patients with

(n = 30) and without (n = 39) major depression were evaluated for DSM-WR diagnoses by the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Epidemiologic version and best estimate, and by measures of functioning (Children’s Global Assessment Scale, Social Adjustment Inventory for Children and Adolescents, WISC, Peabody Picture Vocabulary Test), and compared with children of historical controls without substance abuse history. Results: Sons of opiate addicts with major depression were at increased risk for conduct disorder and global, social, and intellectual impairment compared with sons of opiate addicts without major depression and/or sons of controls with neither drug dependence nor depression. Sons of opiate addicts without major depression differed little from controls. Daughters of opiate addicts did not differ from controls in rates of disorders but had poorer social adjustment and nonverbal intelligence. Conclusions: Children of opiate-dependent patients, particularly sons of addicts with depression, may be at risk for a developmental path toward antisocial personality and poor social and intellectual functioning. Treatment settings such as methadone maintenance might afford an opportunity for primary and secondary prevention, both through early detection of childhood disorders and treatment of parental drug dependence and psychopathology. J. Am. Acad. Child Adolesc. Psychiatry, 1998, 37(11):1142-1151. Key Words: psychiatric diagnosis, opiate dependence, major depression,

conduct disorder

Controlled studies of psychiatric diagnosis in the children of adults with affective disorders (Harnmen et al., 1990; Weissman et al., 1987) and with alcoholism (Hill and Muka, 1996; Merikangas et al., 1985; Reich et al., 1993) have documented increased rates of psychiatric Accepted June 8, 1998. From the Department of GeneticEpiakmiology and the Depression Evaluation Service, New York State Psychiatric Institute and Department of Psychiaty, Columbia University College of Physicians and Surgeons, New York. This work was supported by NIH grants DA07201 (M.M.W), MH36197 (M.M. W), MH28274 (M.M. W), K20 DA00154 (E. KN.), and KO2 DA00288 (E. YN.). The authors are grate@ to Mary Rojas, Ph.D., Angela Aihla, Ph.D., Jim Johnson, Ph.D. (deceased), Christina Sobin, Ph.D., Joyce Edwarh, Ronald Brady, M.D., and the staff of Bridge Plam Treatment and Rehabilitation Services;Jim Koger, C.S.W, Tom Coyne, CS.W, and thestaffof the Long Island Jewish/HillsideMethadone Clinic; Dr. Richard Koeppel, Stuart Steiner, and the staff of the Koeppel Methadone Maintenance Treatment Program; Paul Casadonte, M.D., and the staffof the Manhattan VA Hospital Methadone Clinic; and Stephen Donovan, M.D., f i r his comments on the manuscript. Reprint requests to Dr. Nunes, New York State Psychiatric Institute (Unit 35), 722 West I68 Street, New York, NY10032; e-mail:[email protected] 0890-8567/98/3711-1142/$03.00/00 1998 by the American Academy of Child and Adolescent Psychiatry.

1142

disorders and other adverse childhood outcomes. To date, similar controlled studies with structured diagnostic assessment in children of drug-dependent parents have not been reported. Existing controlled studies have drawn mainly mixed samples of substance-dependent parents and have used dimensional measures of their children’s psychopathology, usually by parental report, demonstrating increased “internalizing” and “externalizing” problem scores (Wilens and Biederman, 1993). Opiate dependence is a severe form of drug dependence characterized by early onset, chronic course, and high rates of psychiatric comorbidity (Brooner et al., 1997; Khantzian and Treece, 1985; Lowinson et al., 1992; Rounsaville et al., 1982b).Thus, children of opiate addicts would be expected to be at risk for a variety of psychiatric disorders for which early diagnosis and treatment would be important. Two small studies of children of methadone-maintained, opiate-dependent patients found increased internalizing and externalizing problem scores (deCubas and Field, 1993; Wilens et al., 1995). The first published study of psychiatric diagnosis by

J. A M . ACAD. C H I L D A D O L E S C . PSYCHIATRY, 37:11, N O V E M B E R 1998

CHILDREN OF OPIATE-DEPENDENT PATIENTS

direct, structured interview in children of a mixed sample of opiate and cocaine addicts (Luthar et al., 1998) lacked controls but found rates of disorders higher than those of children in a community sample (Cohen et al., 1993), and comparable with rates in published studies of children of alcoholic (Hill and Muka, 1996) and depressed (Hammen et al., 1990) parents. Depression is a common co-occurring psychiatric disorder in opiate addicts (Brooner et al., 1997; Khantzian and Treece, 1985; Rounsaville et al., 1982b), is prevalent in their adult relatives (Luthar et al., 1992, 1993; Rounsaville et al., 1991), and is associated with poor treatment outcome (Kosten et al., 1986; LaPorte et d., 1981; Rounsaville et al., 1982a). Depression in parents has been associated with a variety of adverse outcomes in their offspring, including mood, conduct, and substance use disorders, learning difficulties, and medical problems. Both genetic influences and the deleterious effects of depression on parenting have been implicated (Cummings and Davies, 1994; Hammen et al., 1990; Weissman et al., 1986, 1987). This raises the question of whether some of the risk to children of opiate addicts may stem from parental depression. In this article we report results from a direct interview study of psychiatric diagnosis in school-age children of two groups of methadone-maintained, opiatedependent patients, those with and those without a history of major depression. Prevalence of disorders and measures of social, intellectual, and global functioning are presented in comparison with two historical control groups, children of probands with major depression but no substance use disorders and children of probands with neither depression nor substance abuse. All historical control probands were white, and this report is therefore restricted to methadone-maintained white probands. A forthcoming article will compare rates of psychopathology in offspring of opiate addicts of different ethnic groups. The availability of historical control groups with and without parental major depression affords a unique opportunity to examine simultaneously the associations of two potentially important parental risk factors, opiate dependence and depression, with risk to offspring. METHOD

Subjects Subjects were 114 children, aged 6 to 17 years, of 69 opiatedependent, white probands, recruited from four methadone main-

tenance clinics in the New York area. Admission to methadone maintenance requires opiate dependence of more than a year's duration and at least one previous treatment failure. Probands were required to be on a stable dose of methadone and to have one or more school-age biological children (6-17 years old) available to be interviewed. Probands were excluded if they had schizophrenia or were too medically ill to participate.

Procedures Methadone maintenance patients were recruited at the participating clinics and transported with their families from home to the New York State Psychiatric Institute for assessment. Informed consent was obtained from the parent probands and assent was obtained from the children. Fees of $50 per interview for parents and $35 per interview for children were offered. Interviewers were master's-level clinicians with at least 2 years of clinical experience with children. Interviewers received 40 hours of training in our department with methods developed and applied over a series of ongoing genetic-epidemiological projects (Weissman et d., 1987, 1993). Parent probands were interviewed about themselves with the Schedule for Affective Disorders and Schizophrenia-Lifetimeversion (Endicott and Spitzer, 1978), modified to derive DSM-IZI-R diagnoses. Diagnostic information on the children was derived from the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Epidemiologic version (K-SADS-E) (Orvaschel et al., 1982), modified for DSM-III-R, with two versions, one for interviewing adult informants about children and one for interviewing children about themselves. Separate, blinded interviewers interviewed parents about themselves, parents about each school-age child, and school-age children about themselves. Several studies of high-risk offspring were going on simultaneously using the same methods, so that interviewers were assigned to cases blind to the study and thus to probands' diagnostic status. Other assessments of the children included the Children's Global Assessment Scale (C-GAS) (Shaffer et al., 1983); the Social Adjustment Inventory for Children and Adolescents (SAICA) (John et al., 1987), which measures school and social functioning; the Vocabulary and Block Design subscales of the WISC, which correlate highly with overall Verbal and Performance IQ, respectively (Wechsler, 1974); and the Peabody Picture Vocabulary Test (Dunn and Dunn, 1981). For the purposes of analysis, the C-GAS was dichotomized at a cutoff point of 61. Scores of less than 61 have been shown to reflect clinically significant global impairment in children of depressed parents (Weissman et al., 1990). For each subject, the interviewer wrote a narrative clinical history based on all elicited information. When children were not available for direct interview, diagnostic data were derived solely through interviews of parents. Eighty-nine percent of children (101/114) were directly interviewed. Final diagnoses for both parent probands and offspring were derived from a best-estimate procedure (Leckman et al., 1982) in which expert clinician-diagnosticians,blind to data on other relatives, reviewed the SADS interviews, other interview and self-report data, and the clinical narratives and arrived at DSM-ZII-R diagnoses. Reliability of the best-estimate procedure was examined in 173 cases, randomly selected from this and two other ongoing studies, which were blindly evaluated by a second reviewer. Kappa coefficients of agreement (standard errors) between reviewers were good to excellent for all major DSM-III-R diagnoses: major depression, 0.86 (0.03); any anxiety disorder, 0.79 (0.06); attention deficit disorder, 0.64 (0.12); conduct disorder, 0.91 (0.05); suicide attempt, 0.80 (0.07); alcohol dependence, 0.92 (0.03); drug abuse/dependence, 0.89 (0.03).

J. AM. ACAD. C H I L D A D O L E S C . PSYCHIATRY, 37:11, N O V E M B E R 1798

1143

NUNES ET AL.

Historical Control Groups

computed using the methods of Kaplan and Meier (1958), and overall and painvise differences in these survival curves were tested using the appropriate log-rank tests (Mantel, 1996). In addition, to adjust for the effects of potentially confounding variables on these estimates of group differences, Cox proportional hazards regression analyses (Cox, 1972) were performed. Because the disorder status of offspring from the same family may not be independent, the assumption of independence of the outcome variable implicit in the proportional hazards model may be violated. Therefore, the methods of Lin and Wey (1989),as extended by Binder (1992), were used with SUDAAN software (Shah et al., 1996).These methods estimate the covariance matrix of estimated parameters in situations in which the model is misspecified because of correlation among sampling units. For the categorical outcome measure of global impairment (CGAS < 61), a multivariable logistic regression model was fitted to examine group differences, again using SUDAAN software to account for the nonindependence of siblings. Continuous measures of social and cognitive functioning were analyzed using a random-effects linear model, as implemented in the SAS statistical software package PROC MIXED procedure (SAS Institute, 1996),with a random effect of family included to account for nonindependence of siblings. Covariates included in the multivariable models were parental educational level, age of offspring, and current living situation of offspring, because they differed significantly across proband groups (Table I), and proband gender, because previous literature suggests it is associated with risk to offspring (Ripple and Luthar, 1996).

Two historical control groups of nonaddict probands with children evaluated between ages 6 and 17 years were drawn from the Yale Family Study (Weissman et al., 1987): 50 children of 28 adults with a lifetime history of major depression by Research Diagnostic Criteria and no history of substance use disorders, all of whom had been treated at the Yale Universiry Depression Research Unit in New Haven, Connecticut; and 45 children of 22 adults with no history of either depression or substance use disorders. The methods of the Yale Family Study have been described previously (Weissman et al., 1987). This study was designed to mirror the previous study in its core elements, including best-estimated childhood diagnoses based on SADS interviews of children and parents. M.M.W. supervised the design and data collection in both studies. Data Analysis Differences between proband groups in demographic and clinical characteristics were tested with one-way analysis of variance for continuous measures and tests for categorical measures. Lifetime rates of probable or definite psychiatric disorders, substance use, and suicide attempts in offspringwere estimated and compared across the relevant groups using survival analysis techniques to account for the varying ages of the offspring in the different groups and the fact that not all of the offspring had passed through the age of risk. Specifically,survival curves and resulting cumulativerates were

x2

TABLE 1 Demographic Characteristics of White Probands and Offspring by Proband Group Proband Group

Probands Gender: No. (%) Female Age, yr: mean (SD) Marital status: No. (%) Single Married Separated or divorced Widowed SES: No. (%) I, I1 (upper) I11 IV & V (lower) Education: No. (Yo) Some college or higher High school graduate or GED
Opiate With MDD

Opiate Without MDD

n=30

n = 39

23 (77) 39.1 (4.2)

19 (49) 36.7 (4.3)

20 (71) 41.0 (7.8)

2 (7) 18 (60) 8 (27) 2 (7)

7 (18) 21 (54) 10 (26) 1 (3)

21 (75) 6 (21) 1 (4)

19 (86) 2 (9) 1 (5)

.08 1

2 (7) 7 (23) 21 (70)

5 (13) 7 (18) 27 (69)

10 (36) 13 (46)

6 (27) 6 (27) 10 (45)

.075

6 (20) 11 (37) 13 (43)

11 (28) 18 (46) 10 (26)

13 (46) 13 (46) 2 (7)

12 (55) 7 (32) 3 (14)

.013

n = 48

n

=

66

MDD No Drug n

=

0

28

(0)

5 (18)

n

=

50

No MDD & No Drug n

=

22

12 (55) 43.8 (5.9) 0

n = 45

33 (50) 10.5 (3.0)

24 (48) 12.7 (3.3)

27 (60) 13.5 (2.6)

22 (46) 14 (29) 3 (6) 9 (19)

35 (53) 22 (33) 4 (6) 5 (8)

36 (72) 12 (24) 2 (4)

38 (84) 3 (7) 3 (7) 1 (2)

(0)

.064 <.001

(0)

19 (40) 11.4 (3.7)

0

p Value

.27 <.001

<.001

Note: MDD = major depressive disorder; SES = socioeconomic status; GED = general equivalency diploma.

1144

J. A M . ACAD. C H I L D A D O L E S C . PSYCHIATRY, 37:11, N O V E M B E R 1998

CHILDREN OF OPIATE-DEPENDENT PATIENTS

RESULTS

more likely to be living either with a single parent or with neither parent.

Demographic Characteristics of Probands and Offspring

Parental Drug Dependence, Major Depression, and Offspring Diagnosis

Demographic characteristics of the historical control and opiate-dependent proband groups, divided by presence of major depression, and their offspring are presented in Table 1. In comparison with the historical controls, the opiate-dependent probands were younger and had lower educational attainment. Offspring of opiate-dependent probands were younger and were

Table 2 presents the age-adjusted cumulative lifetime rates, estimated from product-limit survival analyses, of psychiatric diagnoses, substance use, and suicide attempts in offspring of proband groups, and log-rank tests for differences between survival curves. In examining this

TABLE 2 Age-Adjusted Lifetime Rates/ 100 of Psychiatric Diagnoses, Suicide Attempts, and Substance Use in White Offspring by Proband Group and Offspring Gender Proband Group

No. of probands No. of offspring Boys Girls Diagnosis in offspring: rate MDD Boys Girls Any affective disorder Boys Girls Any phobia Boys Girls Any anxiety disorder Boys Girls ADHD Boys Girls Conduct disorder Boys Girls Any substance abuseldependence Overall Suicide attempts Overall Any alcohol use* Boys Girls Any cigarette use** Boys Girls

No MDD & No Drug

Opiate With MDD

Opiate Without MDD

MDD No Drug

30

39

28

22

29 19

33 33

26 24

18 27

12.0 (2)f 34.1 (3)

10.0 ( 1 ) g 73.5 (4)"

42.1 (6)" 36.0 (6)

6.7 ( 1 ) 19.0 (3)

.032 .542

43.8 (6)" 100 (4)

16.0 (3)' 77.9 (8)"

52.2 (12)' 56.0 ( 1 1 )

27.8 (3) 46.0 (7)

<.001 .304

13.6 (3) 26.9 (3)

14.7 (3) 12.7 (2)

11.5 (3) 11.6 (2)

17.5 (4)

.492 .592

26.1 (6) 31.8 (4)

20.2 (5) 46.0 (8)

30.8 (8) 47.9 (10)

16.7 (3) 28.7 (7)

,455 ,783

20.7 (6) 0 (0)

15.4 (5) 10.0 (3)

15.4 (4) 4.2 ( 1 )

5.6 (1) 3.7 ( 1 )

.591 .523

45.6 (S)',. 30.0 (2)

9.1 ( 1 ) 0 (0)

32.2 (4) 11.8 (2)

14.3 ( 1 ) 14.4 (2)

.022 .617

21.9 (4)',d

0

9.1 (2)

2.9 ( 1 )

.I10

5.6 (1)

7.1 ( 1 )

.215

6.5 (3)

(0)

20.6 (4)'4

0

(0)

p Value

80.0 (9) 75.4 (7)

58.6 (4) 100 (6)

-

-

,084 .379

54.1 (7) 100 (7)

28.2 (3) 64.8 (7)

-

-

.I20 .343

-

Note: Values in parentheses represent the number of offspring who developed the disorder. MDD = major depressive disorder; ADHD = attention-deficit hyperactivity disorder;p value from log-rank test for differences by proband group. "p I .lo; 'p I .05; 'p 2.01 for contrast to no MDD & no drug group. dp 5 .lo; 'p I .05 for contrast of opiate MDD to opiate no MDD group.fp I .lo; gp 2.05; "p I .01; ip 5.001 for contrast to MDD no drug group. * Analyses for any alcohol use based on total n of 54 boys and 44 girls because of missing data. ** Analyses for any cigarette use based on total n of 50 boys and 41 girls because of missing data.

J. AM. ACAD. C H I L D A D O L E S C . PSYCHIATRY, 37:11, N O V E M B E R 1998

1145

NUNES ET AL.

70

,

65

60 55

50

8 45

9 40

- 30 3 25 .d

-5 35 20 15 10 5

0 0

1

2

3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 Age at Onset, y

+ Opiate + MDD

+MDD no Drug

Opiate No MDD

+ No MDD 8 No DNg

~~

Fig. 1 Cumulative rates of conduct disorder in male offspring by parental diagnosis. MDD disorder.

table it is important to note that the estimated cumulative rates are often higher than raw proportions would be because of censoring of younger offspring. Affective and Anxiety Disorders. Among male offspring, the overall tests for group differences were significant for major depression and for any affective disorder. This reflects higher rates of affective disorders in sons of control probands with major depression and no substance abuse, and it resembles findings from the larger study from which the historical controls were drawn (Weissman et al., 1987). Sons of opiate-dependent probands and of controls with neither major depression nor substance use disorders did not differ significantly in rates of affective disorders. Rates of affective disorders were higher among female offspring but did not differ by proband group. There were no significant differences in rates of anxiety disorders in offspring by proband group. Disruptive Disorders. There were no significant differences across proband groups in rates of attentiondeficit hyperactivity disorder, although this may be underdiagnosed based on interviews of parents and children alone. Among female offspring there were few cases of conduct disorder, and tests for group differences 1146

=

major depressive

were not significant. Among male offspring, the overall test for group differences in rates of conduct disorder was significant. Figure 1 displays the cumulative rate of conduct disorder by age for male offspring of each proband group. Pairwise comparisons showed that the rates of conduct disorder were significantly greater among sons of opiate addicts with major depression compared both with sons of opiate addicts without major depression and with sons of control probands with neither depression nor substance abuse. The rate of conduct disorder in sons of control probands with major depression and no substance abuse fell in between the other groups and did not differ significantly from any of them. In proportional hazards regressions, the hazard ratio (HR) for the painvise comparison between sons of opiate addicts with major depression versus sons of opiate addicts without depression was significant after adjustment for the potential confounders offspring age, offspring living situation, and proband education level (HR = 8.51 [1.14-63.5],p < .05). However, the H R for the comparison between sons of opiate addicts with major depression versus sons of controls with neither depression nor substance abuse (HR = 8.31 [1.13-60.91, p < .05) was no longer significant after adjustment for the poten-

J . AM. ACAD. C H I LD ADOLESC. PSYCHIATRY, 37:11, NOVEMBER 1998

C H I L D REN 0F 0I' I ATE - D E P E N D EN T PAT I E N TS

tial confounders (HR = 7.90 [0.44-143.11, not significant [NS]), due largely to widening of the confidence limits. Proband Gender and Conduct Disorder in Offspring

Among the male offspring of opiate-dependent probands, female proband gender was associated with conduct disorder. All nine cases of conduct disorder had female proband-parents (9/36 sons of female opiate addicts [estimated cumulative rate = 45%] versus 0126 sons of male addicts,p = .O1 by log rank test). The association could in part be attributable to the small number of male probands with major depression. Since there were no cases of conduct disorder in offspring of male probands, proband gender could not be included in the proportional hazards regression model. Substance Use

For the diagnosis of any substance abuse or dependence, the rates were low and did not differ significantly across proband groups; this finding is not surprising given the young age of the children. Occurrence of any cigarette or alcohol use, examined as possible early markers of risk for substance use disorders, did not differ across the two groups of offspring of opiate-dependent probands. These data were not available on the historical controls. The high estimated cumulative rates reflect the fact that by the mid-teenage years most of the offspring of opiate addicts had tried cigarettes and alcohol. Suicide Attempts

Suicide attempts were uncommon in this young sample, and overall tests of group differences were not significant. However, seven of the nine reported suicide attempts occurred among children of opiate-dependent probands. Global Functioning

Similar to the findings for conduct disorder, the proportion of children meeting the global impairment criterion (C-GAS < 61) was greatest (48%) among sons of opiate addicts with major depression, differing significantly from the rates both in sons of addicts without depression (6%) (odds ratio [OR]-unadjusted = 13.93 [2.33-83.401, p < .01; OR-adjusted = 13.09 [1.56109.691, p < .05) and sons of controls with neither depression nor substance use disorders (11%) (ORunadjusted = 7.43 [1.16-47.491, p < .05). The latter O R was substantially reduced after adjustment for

covariates (OR-adjusted = 3.73 [0.32-43.651, NS), suggesting demographic factors may account for some of the difference between sons of opiate addicts with major depression versus sons of controls with neither depression nor substance use disorders. The rate of impairment in sons of controls with major depression (35%) falls in between and does not differ significantly from that of the other groups. Mirroring the findings for conduct disorder, the effect of the covariate proband gender was significant (OR = 19.7 [1.7-225.6],p < .05) in the direction that sons of female probands had worse scores. In daughters rates of impairment are low, ranging from 12% to 2 1Yo across groups, and none of the groups differ significantly. Social and Cognitive Functioning in Offspring

Unadjusted and adjusted means for measures of childhood social and cognitive functioning are presented in Table 3. SAICA global scores, reflecting overall social adjustment, range from 1 to 4, with higher scores indicate poorer functioning. O n measures of intellectual functioning, lower scores indicate poorer functioning. Again mirroring the pattern for conduct disorder and global functional impairment, for social adjustment (SAICA), sons of opiate addicts with major depression had the worst scores, while sons of addicts without depression scored significantly better and did not differ significantly from controls. Adjustment for covariates did not affect the difference between groups within the opiate sample, but it rendered nonsignificant the difference between sons of depressed opiate addicts and controls with neither depression nor substance use disorders. On cognitive measures, sons of addicts with major depression scored significantly worse than both control groups, while sons of addicts without depression also scored significantly worse than both control groups on WISC Block Design. Again, adjustment for covariates rendered nonsignificant the overall group effects and most of the painvise group contrasts. This suggests demographic factors may account for some of the observed differences between sons of opiate addicts and nonaddict controls. However, the point estimates of the means changed little after adjustment, and none of the covariates in these analyses was significant. O n measures of verbal intelligence, female offspring did not differ significantly by parental diagnosis. However, for social adjustment (SAICA) and WISC Block Design, both groups of daughters of opiate addicts scored

J. AM. ACAD. C H I L D ADOLESC. PSYCHIATRY, 37:11, NOVEMBER 1998

1147

NUNES E T AL.

significantly worse than daughters of nonaddict controls. These differences were not affected by adjustment for potential confounds. DISCUSSION

Summary of Findings

School-age sons of opiate addicts with major depression were at increased risk for conduct disorder and global and social impairment, compared both with sons of opiate addicts without major depression and sons of historical controls with neither major depression nor substance use disorders, but not compared with sons of depressed patients without substance use disorders. The sons of opiate addicts without major depression resembled the sons of historical controls with neither major

depression nor substance use disorders on these outcomes. Conduct disorder and global impairment were also more likely if the proband-parent was female. These findings resemble increased externalizing problem scores observed in offspring of various substance-abusing groups (Wilens and Biederman, 1993) and in children of methadone-maintained patients (deCubas and Field, 1993; Wilens et al., 1995) and increased rates of disruptive disorders observed in children of alcoholics (Merikangas et al., 1985; Reich et al., 1993). The association of affective disease in parents, particularly maternal depression, with externalizing symptoms and disorders in children has been observed in other samples (Cummings and Davies, 1994; Hammen et al., 1990; Weissman et al., 1987), including a sample of children of female opiate and cocaine addicts where

TABLE 3 Unadjusted and Adjusted Means for Social and Cognitive Functioning by Proband Group and Offspring Gender Proband Group I. Opiate With MDD Male offspring Overall social adjustment* Unadjusted Adjusted** Peabody Picture Vocabulary Unadjusted Adjusted** WISC Vocabulary scale Unadjusted Adjusted** WISC Block Design scale Unadjusted Adjusted** Female offspring Overall social adjustment* Unadjusted Adjusted** Peabody Picture Vocabulary Unadjusted WISC Vocabulary scale Unadjusted WISC Block Design scale Unadjusted Adjusted**

11. Opiate Without MDD

1.78 (0.07)"' 1.73 (0.08)"'

1.54 (0.06) 1.52 (0.07)

91.23 (3.38)c,d.g 92.67 (3.86)'

100.68 (3.31)" 101.08 (3.57)

8.62 (0.74)'ah 9.08 (0.77)f

111. MDD No Drug

1.63 (0.07)" 1.63 (0.09)

IV. No MDD & No Drug

p Value

1.41 (0.08) 1.50 (0.10)

.016 .187

102.52 (3.25) 100.30 (3.98)

109.01 (3.62) 106.56 (4.46)

.014 .I75

10.07 (0.70) 9.99 (0.68)

11.48 (0.75) 11.16 (0.81)

11.28 (0.85) 10.91 (0.98)

,055 .278

9.11 (0.68)"g 9.07 (0.79)"~f

9.14 (0.71)',g 9.04 (0.77pf

11.46 (0.70) 11.06 (0.90)

11.45 (0.82) 11.58 (1.11)

,037 ,162

1.78 (0.09)' 1.76 (0.10)'

1.64 (0.08)' 1.65 (0.08)"

1.60 (0.09) 1.57 (0.11)

1.40 (0.08) 1.40 (0.10)

.040 .071

100.67 (4.56)

101.24 (3.16)

104.80 (3.40)

101.69 (3.07)

,844

9.78 (0.77)

10.13 (0.49)

10.65 (0.59)

10.24 (0.52)

.828

11.30 (0.64) 11.73 (0.77)

12.16 (0.59) 12.55 (0.70)

.004 .005

8.44 (0.81)"h 8.62 (0.86)c,h

9.45 (0.55)"g 9.43 (0.59)',g

~

Note: Values represent mean (SE). MDD = major depressive disorder;p value for differences by proband group. N's for each analysis vary because of missing data. Males: group I: n = 21 to 25; group 11: n = 20 to 28; group 111: n = 21 to 22: group IV n = 15 to 18. Females: group I: n = 9 to 16; group 11: n = 23 to 27; group 111: n = 16 to 18; group IV: n = 23 to 27. " p I .lo; ' p I .05; ' p I .01 for contrast to no MDD 81 no drug group. d p I .lo; ' p I .05 for contrast of opiate MDD to opiate no MDD group. fp I .lo; g p I .05; ' p I .01 for contrast to MDD no drug group. * Higher scores reflect poorer functioning. ** Adjusted for age at interview, proband education, proband gender, and whether or not offspring lives with both parents.

1148

I . A M . ACAD. C H I L D A D O L E S C . PSYCHIATRY,

37:11, N O V E M B E R 1998

C H I L D R E N OF OPIATE-DEPENDENT PATIENTS

maternal psychiatric illness in general was associated with increased externalizing problem scores in offspring (Luthar et al., 1998). This could reflect a common genetic diathesis for substance abuse and depression (Winokur et al., 1974), or a greater heritable vulnerability to psychopathology in female addicts (Ripple and Luthar, 1996), or impairment of parenting by depression (Cummings and Davies, 1994). Rather than a specific effect of depression, the findings could reflect a general effect of comorbidity, i.e., that patients with multiple diagnoses may pass on more vulnerability genes or be more impaired as parents. Conduct disorder was rare in female offspring in these samples and did not differ between groups. This may, in part, reflect a bias in the diagnostic criteria toward typically male symptoms of conduct disorder (Cottler et al., 1995; Goldstein et al., 1996; Rutherford et al., 1995). Both groups of daughters of opiate addicts (with and without major depression) scored significantly worse than the daughters of historical controls with neither depression nor substance abuse on social fimctioning and Performance IQ. However, the daughters of opiate addicts had low rates of global impairment, and had less evidence of intellectual impairment than sons, since they did not differ from historical controls in measures of Verbal IQ. Taken together, these findings are consistent with other evidence that girls in high-risk environments may be more resilient (Rutter, 1983). Lower scores on measures of Verbal and Performance I Q were observed in sons of opiate addicts with depression, while scores on Block Design, an index of Performance IQ, were low in both sons and daughters of opiate addicts, both with and without depression. Lower I Q scores compared to controls were similarly observed in a sample of children of men with substance use disorders (Moss et al., 1995). Low I Q has been associated with behavioral deviance in children and may be a causal risk factor or a marker for a common underlying vulnerability (Goodman et al., 1995). It is also possible that parental IQ, which was not measured in this study, could explain the variance in offspring IQ, and parental I Q should be measured in future studies. Methodological Limitations

The most serious limitation of this analysis is the use of historical control groups of nonaddict probands, ascertained at a different time in a different city. O n the posi-

tive side, the interviewers and best estimators for the historical controls were trained by the same investigator (M.M.W.) and followed a similar protocol. It is also important to note that the contrasts between groups within the opiate-dependent sample (with versus without major depression) are not subject to the same concerns, since these were ascertained simultaneously, and differences observed here were generally not attenuated in the multivariable models. However, the historical controls, compared with the opiate-dependent sample, were better educated and their children were more likely to live with both parents, and covariation for these attenuated most differences between offspring of opiate addicts versus controls. The lack of minority individuals is another limitation of the historical controls, restricting the design to white probands and limiting the generalizability of the findings. Clearly, future studies should include a racially mixed sample and utilize matched, nonaddict controls from the same communities as addict probands. Generalizability is also restricted in that we studied only treated opiate addicts drawn from a limited pool of cooperative clinics and who volunteered to participate, and participation required a significant time commitment and a degree of organization on the part of subjects. This could produce an ascertainment bias, for example toward lower levels of psychopathology if the most impaired opiate-dependent patients or those with the sickest children were excluded. A broader sample of opiate-dependent probands should be sought for future studies. Although this is the largest diagnostic study to date of school-age children of opiate addicts, and such samples are difficult to recruit, the sample size is limited from a statistical standpoint, as reflected in the wide confidence limits around parameter estimates. The children in this sample were young, and many had yet to pass into the age of risk for syndromal psychiatric disorders, particularly mood and substance use disorders. Furthermore, this study is not able to estimate the effect of other psychiatric disorders in probands, particularly antisocial personality, which is prevalent in opiate-dependent patients (Brooner et al., 1997) and heritable (Lykken, 1995). There were no cases of antisocial personality among the historical controls. The number and percentage of cases of opiate-dependent probands with antisocial personality (8 [21%] among those without major depression and 3 [lo%] among those with major

J. AM. ACAD. C H I L D A D O L E S C . PSYCHIATRY, 37:11, N O V E M B E R 1998

1149

NUNES E T AL.

depression, a group that contains more women), while consistent with the prevalence rates from a recent large study of DSM-111-R diagnoses in opioid addicts (25% overall, 15% in women) (Brooner et al., 1997), is too small for meaningful analyses. Larger studies would be needed, including longitudinal follow-up into adulthood and sufficient probands with antisocial personality as well as depression, to more precisely estimate risks of disorders and predictors of risk. Clinical Implications

Taken together, these findings suggest that children of opiate-dependent patients, particularly sons of female addicts with histories of depression, are at risk for a developmental trajectory toward antisocial personality and poor social and intellectual functioning. Longitudinal follow-up would be needed to determine whether these patterns translate into a higher risk for substance use disorders as the children age. Clinicians and programs treating opiate addicts should inquire about the mental health and adjustment of their patients’ children at home, at school, and with peers. Routine psychiatric screening of such children should be considered with the goal of early intervention and treatment, and the long-term outcome of such programs should be studied. The relative lack of impairment in offspring of methadone patients without depression is intriguing. One hypothesis would be that the patients without depression have had their predominant psychopathology, drug dependence, successfully treated by methadone, accompanied by improvement in their overall functioning and parental effectiveness.This is consistent with a retrospective study, which showed cessation of fathers’ substance abuse before their sons’ sixth birthdays was associated with reduced risk of both externalizing and internalizing problems (Moss et al., 1997). In contrast, opiate addicts with major depression may still have impaired functioning due to active psychopathology. The increased risk for externalizing disorders in children of depressed parents has been related to ineffective parenting, disturbed parent-child attachment, and marital conflict (Cummings and Davies, 1994).This suggests the hypothesis that treatment of both opiate dependence and comorbid psychiatric disorders in parents could have primary preventive effects against psychopathology and impairment in their children. Opiate dependence (Lowinson et al., 1992) and comorbid depression in addicts (Nunes et al., 1998; Woody 1150

et al., 1984) are treatable. Furthermore, behavioral therapies have been developed which target dysfunctional parent-child interactions that may engender antisocial behavior in children (Dishion et al., 1995). Treatment settings such as methadone maintenance offer an excellent opportunity both to treat parental psychiatric disorders and to address dysfunctional parenting. Yet the majority of opiate addicts remain outside the treatment system (Institute of Medicine, 1996), and even within that system adjunctive psychiatric care is often limited. Prospective studies examining psychopathology in children of opiate addicts as a function of their parents’ treatment are needed. Establishment of a link would provide an impetus for expanding the availability and scope of treatment for opiate-dependent patients and their children. REFERENCES Binder DA (1992), Fitting Cox’s proportional hazards models from survey data. Biomehika 79: 139-147 Brooner RK, King VL, Kidorf M, Schmidt CW Jr, Bigelow GE (1997), Psychiatric and substance use comorbidity among treatment-seeking opioid abusers. Arch Gen Psychiatry 54:71-80 Cohen l? Cohen J, Kasen S et al. (1993), An epidemiologic study of disorders in late childhood and adolescence, I: age- and gender-specific prevalence. J Child Psycho1 Psychiatry 34351-867 Cottler LB, Price RK, Compton WM, Mager DE (1995), Subtypes of adult antisocial behavior among drug abusers. JNerv Ment Dis 183:154-161 Cox DR (1972), Regression models and life tables. J R Stat Soc B 34:187-220 Cummings EM, Davies PT (1994), Maternal depression and child development. J ChildPsychol Psychiatry 35:73-112 deCubas MM, Field T (1993), Children of methadone-dependent women: developmental outcomes. Am J Orthopsychiatry63:266-276 DishionTJ, French DC, Patterson GR (1995),The development and ecology of antisocial behavior. In: Developmentalpsycho pa tho lo^, Vol2, Cicchetti D, Cohen DJ, eds. New York: Wiley, pp 421-472 Dunn LM, Dunn LM (1981), Peabody Picture Vocabulary Test-Revised. Circle Pines, MN: American Guidance Service Endicott J, Spitzer F X (1978), A diagnostic interview: the Schedule for Affective Disorders and Schizophrenia.Arch Gen Psycbiany 35:773-782 Goldstein RB, Powers SI, McCusker J, Mundt KA, Lewis BF, Bigelow C (1996), Gender differences in manifestations of antisocial personality disorder among residential drug abuse treatment clients. Drug Alcohol DPpend 4135-45 Goodman R, Simonoff E, Stevenson J (1995), The impact of child IQ, parent IQand sibling I Q o n child behavioral deviance scores. JChildPsychol Psychiatry 36:409-425 Hammen C, Burge D, Burney E, Adrian C (1990), Longitudinal study of diagnoses in children of women with unipolar and bipolar affective disorder. Arch Gen Psychiatry 47:1112-1117 Hill SY, Muka D (1996), Childhood psychopathology in children from families of alcoholic female probands. J Am Acad Child Adolesc Psychiany 35:725-733 Institute of Medicine (1996), Pathways of Addiction: Opportunities in Drug Abuse Research. Washington, DC: National Academy Press, p 193 John K, Gammon GD, Prusoff BA, Warner V (1987), The Social Adjustment Inventory for Children and Adolescents (SAICA): testing of a new semistructured interview. ] A m Acad ChildAdolesr Psychiatry 26398-911 Kaplan EL, Meier P (1958), Nonparametric estimation from incomplete observations.J A m Stathsoc 53:457-481

J. AM. ACAD. C H I LD ADOLESC. PSYCHIATRY, 37:11, NOVEMBER 1998

C H I L D R E N OF O P I A T E - D E P E N D E N T PATIENTS

Khantzian EJ, Treece C (1985), DSM-III psychiatric diagnosis of narcotic addicts. Arch Gen Psychiany 42: 1067-1071 Kosten TR, Rounsaville BJ, Kleber H D (1986), A 2.5 year follow-up of depression, life crises, and treatment effects on abstinence among opioid addicts. Arch Gen Psychiatry 43:733-738 LaPorte DJ, McLellan AT, O’Brien CP, Marshall JR (1981), Treatment response in psychiatrically impaired drug abusers. Compr Psychiatry 22:411-419 Leckman JF, Sholomskas D, Thompson WD, Belanger A, Weissman MM (1982), Best estimate of lifetime psychiatric diagnosis: a methodologic study. Arch Gen Psychiatry 393379-883 Lin DY, Wey LJ (1989), The robust inference for the proportional hazards model. J A m StatAssoc 84: 1074-1078 Lowinson JH, Marion IJ, Joseph H, Dole VP (1992), Methadone maintenance. In: SubstanceAbuse:A Comprehensive Tatbook, 2nd ed, Lowinson JH, Millman R, Ruiz P, eds. Baltimore: Williams & Wilkins, pp 550-561 Luthar SS, Anton SF, Merikangas KR, Rounsaville BJ (1992), Vulnerability to substance abuse and psychopathology among siblings of opioid abusers. JNervMentDis 180:153-161 Luthar S S , Cushing G, Merikangas KR, Rounsaville BJ (1998), Multiple jeopardy: risk and protective factors among addicted mothers’ offspring. Deu Psychopathol 10:117-136 Luthar SS, Merikangas KR, Rounsaville BJ (1993), Parental psychopathology and disorders in offspring: a study of relatives of drug abusers. J Nerv MentDis 181:351-357 Lykken D T (1995), The Antisocial Personalities. Hillsdale, NJ: Erlbaum, pp 92-109 Mantel N (1996), Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rpp 50: 163-170 Merikangas KR, Weissman MM, Prusoff BA, Pads DL, Leckman JF (1985), Depressives with secondary alcoholism: psychiatric disorders in offspring. ]StudAleuhol46:199-204 Moss HB, Clark DB, Kirisci L (1997), Timing of paternal substance use disorder cessation and effects on problem behaviors in sons. Am J Addict 6:30-37 Moss HB, Vanyukov M, Majumder PP, Kirisci L, Tarter R (1995), Prepubertal sons of substance abusers: influences of parental and familial substance abuse on behavioral disposition, IQ, and school achievement. Addict Behav 20:345-358 Nunes EV, Quitkin FM, Donovan SJ et al. (1998), Imipramine treatment of opiate-dependent patients with depressive disorders: a placebo-controlled trial. Arch Gen Prychiany 55:153-160 Orvaschel H, Puig-Antich J, Chambers W, Tabrizi MA, Johnson R (1982), Retrospective assessment of prepubertal major depression with the Kiddie-SADS-E. ] A m Acad Child Psychiatry 21:392-397 Reich W, Earls F, Frankel 0,Shayka JJ (1993), Psychopathology in children of alcoholics./ A m Acad ChildAdolesc Psychiaty 32:995-1002 Ripple CH, Luthar SS (1996), Familial factors in illicit drug abuse: an interdisciplinary perspective.Am ]Drug Alcohol Abuse 22: 147-172

Rounsaville BJ, Kosten TR, Weissman MM et al. (1991), Psychiatric disorders in relatives of probands with opiate addiction. Arch Gen Psychiatry 48:33-42 Rounsaville BJ, Weissman MM, Crits-Christoph K, Wilber C, Kleber H (1982a), Diagnosis and symptoms of depression in opiate addicts: course and relationship to treatment outcome. Arch Gen Psychiaty 39:151-156 Rounsaville BJ, Weissman MM, Kleber H, Wilber C (1982b), Heterogeneity of psychiatric diagnosis in treated opiate addicts. Arch Gen Psychiatry 39:161-168 Rutherford MJ, Alterman AI, Cacciola JS, Snider EC (1995), Gender differences in diagnosing antisocialpersonality disorder in methadone patients. Am ]Psychiatry 152:1309-1316 Rutter M (1983), Stress, coping and development: some issues and questions. In: Stress, Coping and Development in Children, Garmezy N, Rutter M, eds. New York: McGraw-Hill SAS Institute (1996), SAS/STATSoflware: Changesand Enhancements Through Release 6.11. Cary, N C SAS Institute Inc Shaffer D, Gould MS, Brasic J et al. (1983), A children’s global assessment scale (CGAS). Arch Gen Psychiany 40:1228-1231 Shah BV, Barnwell BG, Bieler GS (1996), S U D A N User; Manual, Krsion 6, 2nd ed. Research Triangle Park, NC: Research Triangle Institute Wechsler D (1974), Wechsler Intelligence Scale for Children Revised. New York: Psychological Corporation Weissman MM, Gammon GD, John K et al. (1987), Children of depressed parents: increased psychopathology and early onset of major depression. Arch Gen Psychiany 44347-853 Weissman MM, John K, Merikangas KR et al. (1986), Depressed parents and their children: general health, social and psychiatric problems. Am J Dis Child 140301-805 Weissman MM, Warner V, Fendrich M (1990), Applying impairment criteria to children’s psychiatric diagnosis.J A m Acad Child Adolesc Psychiany 29:789-795 Weissman MM, Wickramaratne P, Adam P et al. (1993), The relationship between panic disorder and major depression: a new family study. Arch Gen Psychiaty 50:767-780 Wilens TE, Biederman J (1993), Psychopathology in preadolescent children at high risk for substance abuse: a review of the literature. Haw Rev PsyS i a t y 4207-218 Wilens TE, Biederman J, Kiely K, Bredin E, Spencer TJ (1995), Pilot study of behavioral and emotional disturbances in the high-risk children of parents with opioid dependence. J Am Acad Child Adolesc Psychiatry 34:779-785 Winokur GA, Cadoret R, Dorzab JA, Baker M (1974), The division of depressive illness into depression spectrum disease and pure depressive illness. Int Pharmacopsychiatry 9:5-13 Woody GE, McLellan AT, Luborsky L et al. (1984), Severity of psychiatric symptoms as a predictor of benefits from psychotherapy: the Veterans Administration-Penn Study. Am JPsychiatry 141:1172-1177

J . AM. A C A D . C H I L D A D O L E S C . PSYCHIATRY, 37:11, N O V E M B E R 1998

1151