Available online at www.sciencedirect.

com

Psychiatry Research 158 (2008) 278 286 www.elsevier.com/locate/psychres

MMPI measures as signs of predisposition to mental disorder among adoptees at high risk for schizophrenia
Virva Siira a,, Karl-Erik Wahlberg a , Jouko Miettunen a , Kristian Lksy b , Pekka Tienari a
a b

Department of Psychiatry, University of Oulu, Oulu, Finland University Hospital of Oulu, Clinic of Psychiatry, Oulu, Finland

Received 25 February 2005; received in revised form 24 February 2006; accepted 7 December 2006

Abstract The DSM-III-R diagnoses of a group of adoptees were predicted by the MMPI (Minnesota Multiphasic Personality Inventory) schizophrenia-related scales in the Finnish Adoptive Family Study. The sample consisted of 60 high-risk (HR) adopted-away offspring of biologic mothers with a diagnosis of broad schizophrenia spectrum and 76 low-risk (LR) control adoptees. They were assessed with the MMPI before the onset of any psychiatric disorder at a mean age of 24 years. High scores on the Psychopathic Deviate scale predicted psychiatric disorder at 11-year follow-up. Furthermore, LR adoptees', but not HR adoptees', mental disorders could be predicted with the MMPI scales Psychopathic Deviate and Golden-Meehl Indicators. These scales measure schizophrenia-related personality traits, including asocial behavior, anhedonia, ambivalence, interpersonal aversiveness, and formal thought disturbances. 2006 Elsevier Ireland Ltd. All rights reserved.
Keywords: Mental disorder; MMPI; Finnish Adoptive Family Study; Offspring of schizophrenic mothers

1. Introduction Genetic factors have been found to contribute to the liability to schizophrenia (Gottesman and Shields, 1967; Kendler et al., 2000). This liability is manifested as a personal vulnerability at the behavioral level (Green, 1998; Meehl, 1962; Nuechterlein, 1987; Rosenthal, 1970; Zubin and Spring, 1977), and it can be measured by using liability indicators, predictors, prodromal signs, or traits (Ellison et al., 1998; ErlenmeyerKimling, 2000). The individual's underlying biological predisposition, combined with environmental stress, is
Corresponding author. Department of Psychiatry, University of Oulu, BOX 5000, FIN-90014 University of Oulu, Finland. Tel.: +358 8 5200163; fax: +358 8 333167. E-mail address: virva.siira@oulu.fi (V. Siira).

thought to account for the onset of schizophrenia (Nuechterlein, 1987; Rosenthal, 1970; Zubin and Spring, 1977). Longitudinal prospective high-risk studies suggest that the offspring of schizophrenic parents show an increased rate of schizophrenia (Erlenmeyer-Kimling et al., 1997; Parnas et al., 1993; Tienari et al., 2003). The risk of schizophrenia in relatives of schizophrenic patients increases along with their degree of familial relationship (Gottesman, 1991). Although the risk of schizophrenia is higher among relatives of schizophrenic patients compared with the general population, schizophrenia tends not to be transmitted directly (Gottesman, 1991). Neither of the parents of about 89% of schizophrenics have schizophrenia (Gottesman and Erlenmeyer-Kimling, 2001). Recently, psychiatric morbidity in offspring born to

0165-1781/$ - see front matter 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.psychres.2006.12.005

V. Siira et al. / Psychiatry Research 158 (2008) 278286

279

mothers with schizophrenia has been found to exceed the incidence of mental disturbances (schizophrenia and other mental disorders) in offspring born to mentally healthy mothers and mothers with affective illness (Schubert and McNeil, 2003). Research on vulnerability and liability indicators has demonstrated that children at high genetic risk who later develop schizophrenia have more cognitive deficits, thought disorder, and motor incoordination than children without genetic risk (Erlenmeyer-Kimling and Cornblatt, 1992; Erlenmeyer-Kimling et al., 2000; Fish et al., 1992; Hans et al., 1999; Ott et al., 2001). Impaired attention, poor memory functions, and eye movement abnormalities are thought to be close to the underlying biological functions of schizophrenia (biobehavioral markers) because they manifest relatively early in development (Cornblatt and Obuchowski, 1997; Cornblatt et al., 1999; Erlenmeyer-Kimling and Cornblatt, 1978). In addition to the high-risk study design, an adoption study of Minnesota Multiphasic Personality Inventory indicators demonstrated that adopted-away offspring of schizophrenics have more schizophrenia spectrum disorders than their adoptive controls (Haier et al., 1978). High-risk studies have revealed genetically influenced premorbid traits that may act as phenotypic indicators or potentiators of schizophrenia (Erlenmeyer-Kimling, 2000; Meehl, 1962). These findings illustrate the behavioral, emotional, and social adjustment problems that precede schizophrenia and psychotic symptoms (Amminger et al., 2000; Kugelmass et al., 1995; Olin et al., 1995). For some time, deviant personality traits in the Minnesota Multiphasic Personality Inventory (MMPI; Dahlstrom et al., 1982), such as impaired attention/ concentration, formal thought disturbances, severe disorganization, overt psychotic symptoms, lack of insight, social withdrawal, asocial behavior, anhedonia, low energy, and restricted affectivity, have been known to be associated with schizophrenia in genetic high-risk studies (Moldin et al., 1990a,b,c). In addition, MMPImeasured schizophrenia-related personality traits of unaffected parents of schizophrenic probands have been found to be more deviant than those of parents of reference children, indicating a genetic liability to schizophrenia (Subotnik et al., 2005). The predictive validity for schizophrenia of different compositions of MMPI variables of Moldin et al. (1990a) variables has been reported in the most recent high-risk studies (Carter et al., 1999). Bolinskey et al. (2001) identified high-risk individuals who later became schizophrenic based on their psychometric index and a new MMPI-derived scale.

In another study by MacCrimmon et al. (1980), fostered offspring of schizophrenic mothers were distinguishable from non-foster controls, but not from a control group of foster individuals based on MMPI clinical scale scores. This may indicate that adoptive and fostered children have an elevated risk for mental disorder independent of their genetic liability. Studies have indicated that separation from the biological parents may increase mental health problems in the child (Veijola et al., 2004; Miller et al., 2000). In the absence of definitive genetic or any other individual premorbid markers, prediction of the future mental health status is only possible based on statistical probabilities. In this research, a prospective longitudinal study design was used to explore psychometric deviance in personality traits as a possible vulnerability indicator of predisposition to psychiatric disorder. We tested the hypothesis that the adoption of a child will manifest as early deviant scores on schizophrenia-related MMPI measures, which predict future mental disorder in the adoptee. Secondly, we hypothesized that the genetic interaction between liability (a high risk to schizophrenia spectrum disorder) and deviant MMPI variables would precede the onset of later mental disorder in the adoptee. 2. Methods 2.1. Subjects The present subsample was derived from the national Finnish Adoptive Family Study of Schizophrenia series (Tienari et al., 1981, 2000, 2003). The children were born to biological mothers hospitalized and diagnosed with schizophrenia or paranoid psychosis in the Finnish psychiatric hospitals including all the resident populations on 1 January, 1960 and consecutive admissions after 1 January, 1960 to 1979. Manic-depressive, depressive, reactive, or psychogenic psychosis, organic brain syndrome, severe mental retardation, or primary alcoholism before the onset of schizophrenia were used as exclusion criteria for these women in the primary hospital sampling phase. Adoptees were excluded if they had been adopted by a relative, adopted abroad, or adopted after the age of 4 years. The design, sampling, and diagnostic procedures of the adoption study as a whole have been described earlier (Tienari et al., 1987a,b, 2000). The final sample of the Finnish Adoptive Family Study of Schizophrenia consisted of 190 offspring at high genetic risk (HR) who have biological mothers with DSM-III-R diagnoses of schizophrenia, odd-cluster

280

V. Siira et al. / Psychiatry Research 158 (2008) 278286 Table 1 Demographic variables of the high- and low-risk adoptees in the sample Variable Adoptees High risk (HR) n = 60 Mean age SD at MMPI assessment a Mean age SD at placement a Gender M:F b *Social class of the adoptive family b 24.57 10.06 years 18.17 15.39 months 27:33 I 13.6% II 40.7% III 35.6% IV 10.2% Low risk (LR) n = 76 23.82 9.33 years 16.89 13.83 months 34:42 I 13.2% II 51.3% III 32.9% IV 2.6% Sig. 0.65 0.61 0.98 0.26

personality disorder (plus avoidant PD), nonschizophrenic nonaffective psychosis, or affective psychosis in the broad schizophrenia spectrum (Kendler et al., 1996; Tienari et al., 2000, 2003). The control sample consisted of 192 adoptees at low genetic risk (LR) with biological mothers who had a nonspectrum diagnosis (DSM-III-R diagnoses not included in the broad schizophrenia spectrum, n = 41) or no psychiatric disorder (n = 151). The schizophrenic and control biological mothers were assessed for research diagnoses through hospital record checks, diagnostic interviews with the Present State Examination (PSE) (Wing et al., 1974), and personal interviews when possible. The psychiatrists who performed the interviews of the biological mothers were unaware of the adoptees' HR or LR status. The DSM-III-R (American Psychiatric Association, 1987) diagnoses were made of the best-estimate, most severe psychiatric disorder (Tienari et al., 2000). In the whole sample, 56% of the adopted-away children's fathers have been interviewed, but in the rest of the cases the father is unknown or dead. We did not include the data of biological fathers in the present sample due to the small sample size. The evaluations of the adoptees' psychiatric status included personal interviews at follow-up (Tienari et al., 2003). In addition, whenever personal interviews were not possible, phone interviews of the parents and register information searches were done. The psychiatrists responsible for the follow-up evaluations were blind to the information from prior assessments. The subsample used in this study consisted of 136 adoptees (60 HR and 76 LR) who did not have any psychiatric disorder at the time of assessment, and whose MMPI measures were completed. Comparisons between the HR and LR adoptees showed no differences with regard to mean age at MMPI assessment, gender, mean age at placement, or socioeconomic class (Statistics Finland 17, 1983). There were no significant differences in the demographic variables or level of psychiatric diagnoses of the biological mothers (Kendler et al., 1996) between the MMPI-assessed adoptees included in this study and the adoptees in the rest of the sample, with the exception of the adoptees' age. The adoptees in this study were younger than those in the total sample, t(345) = 3.66, P b 0.001. Therefore, the effect of age was statistically accounted for in the multivariate model (Table 1). 2.2. Follow-up DSM-III-R diagnoses of the adoptees

Social class of the adoptive family was determined according to the four-level Finnish socioeconomic classification, a low number indicating a high social class (Statistic Finland, 17, 1983). a T-test analyses were used for mean age differences. b Chi-square analyses for gender and social class differences.

divided based on the follow-up diagnoses. The three groups were as follows: (1) adoptees mentally healthy at follow-up (n = 99), (2) adoptees with psychiatric disorder but not schizophrenia at follow-up (n = 31), and (3) adoptees with broad schizophrenia spectrum disorder at follow-up (n = 6). No MMPI scale score predicted the subsequent onset of broad schizophrenia spectrum disorder. Predictions based on variables of psychometric deviance were studied in two groups of adoptees in the present study. The first group consisted of adoptees (n = 99) who were mentally healthy at follow-up (MHF) and the second group of adoptees (n = 37) diagnosed with any psychiatric disorder at follow-up (APDF) 11 years after the MMPI assessment. We did this because earlier findings support the idea that children of schizophrenic mothers have increased risk for all kind of mental disorders. The follow-up diagnoses of the adoptees in the APDF group were: antisocial personality disorder (PD) (1 case), borderline PD (3), narcissistic PD (2), histrionic PD (5), obsessive-compulsive PD (1), PD NOS (2), alcohol use (3), dysthymic disorder (2), anxiety disorder (6), and mild mood disorder (6), as well as schizophrenia (1), paranoid PD (1), schizoid PD (1), bipolar psychosis (1), depressive psychosis (1), and avoidant PD (1). The last six cases had a diagnosis in the broad schizophrenia spectrum disorder group. 2.3. Measurement of predictive variables

When preparing this report, we first studied the predictive power of MMPI scale scores in three groups who were mentally healthy at the initial assessment and were

The following scales of the MMPI were included this study: L (Lie) (Dahlstrom et al., 1982; Friedman et al.,

V. Siira et al. / Psychiatry Research 158 (2008) 278286

281

1989), F (frequency), K (correction/defensiveness), 4 (psychopathic deviate), the sum of scales 2 (depression), 7 (psychasthenia), 8 (schizophrenia), and 0 (Social Introversion) (Golden-Meehl Indicators, 2 + 7 + 8 + 0) (Golden and Meehl, 1979), the 86 scale (Minneapolis Veterans Administration Hospital-MMPI Research Laboratory, 1975), Rosen's Pz scale (Paranoid Schizophrenia) (Rosen, 1962), HOS (Hostility), HYP (Hypomania), PHO (Phobias), PSY (Psychoticism), REL (Religious Fundamentalism), SOC (Social Maladjustment) (Wiggins, 1966), and the SzP scale (Schizophrenia Proneness) (Bolinskey et al., 2001). 2.3.1. Profile validity Profiles with 40 items not responded (Bolinskey et al., 2001) and 3 of the following rules met were excluded: F raw score 26, L raw score 10, TestRetest scale 6, Carelessness scale 6 (Greene, 1980; Bolinskey et al., 2001). Using these criteria, seven cases were excluded. 2.3.2. Estimation of MMPI scores Some participants were assessed by an abbreviated version of Form R of the MMPI, which includes 400 items. The omission of 166 items from the 400-item Form R did not affect scoring on the clinical and validity scales of the MMPI, but the special scales comprised items that were not included in the short form of the MMPI. The principle of augmenting the omitted items was completed
Table 2 Descriptive statistics of the adoptees' MMPI scores Variable Cutoff score a HR adoptees (n = 60) Adoptees mentally healthy at follow-up (n = 38) Mean L F K 4 2+7+8+0 86 SzP Pz HOS HYP PHO PSY REL SOC
a b

with correlations in this study. The scores of the fulllength scales were estimated from the scores of the abbreviated scales using information from those who replied to all items (n = 64). Pearson's correlations between the estimated and observed full-length scores were 0.99 on the PSY, 0.97 on the SOC, 0.97 on the Pz, 0.96 on the 86, 0.93 on the SzP, 0.92 on the HYP, 0.91 on the HOS, 0.91 on the REL, and 0.84 on the PHO scales, supporting the prorating. The Pz + 1K scores (Rosen, 1962) and the Wiggins adult T scores, but in lieu of the T score, the raw scores for the 86 scale and the SzP scale were used in the statistical analyses. The 0.4 K correction was used for Pd and the 1 K correction for 7 and 8 on the 2 + 7 + 8 + 0 scale with standard adult T scores. 2.4. Statistical methods All of the MMPI variables had absolute values of Skewness 2.0 or Kurtosis 7.0 (Curran et al., 1996). Therefore, parametric methods of statistical analysis were chosen for the group comparison of the means of the MMPI scales (Bland, 1995). Factorial analysis of variance (ANOVA) was performed to examine the main effects of the follow-up diagnoses of the adoptees and their genetic risk status and the interaction between these variables on each of the MMPI scales (Armitage and Berry, 1987). Post hoc tests with Bonferroni correction for multiple comparisons were used to determine the pairs of means with significant differences. Adjusted odds ratios were

LR adoptees (n = 76) Any psychiatric disorder at follow-up (n = 22) Mean 52.95 61.05 56.36 61.23 230.82 21.00 11.41 54.32 44.91 46.50 43.18 48.41 38.82 54.50 SD 10.16 8.56 8.58 9.87 35.25 4.75 3.96 6.79 7.87 10.12 7.35 8.81 11.05 12.22 Adoptees mentally healthy at follow-up (n = 61) Mean 49.18 62.44 55.25 57.28 227.03 22.26 10.87 53.59 51.34 52.25 47.02 52.21 38.64 51.95 SD 7.35 12.65 7.98 11.26 36.10 4.68 3.99 8.28 11.05 9.63 11.75 12.63 10.12 9.94 Any psychiatric disorder at follow-up (n = 15) Mean 50.87 69.80 54.87 68.53 250.53 24.73 13.33 59.40 54.27 51.53 50.33 56.13 38.67 55.07 SD 8.54 14.82 8.91 11.50 28.87 5.90 3.92 9.86 11.88 11.89 9.70 13.35 8.55 5.75

Fb

P-value

SD 7.34 10.72 8.48 9.26 27.59 3.97 3.85 7.33 8.51 11.00 9.21 10.04 9.93 9.72

56.00 68.00 61.75 67.00 252.75 26.00 13.00 59.75 41.00 44.00 51.00 57.00 46.00 61.00

51.71 60.87 57.92 59.82 232.53 23.18 10.26 54.74 47.05 49.16 44.82 50.95 39.08 54.26

1.53 2.26 0.94 4.73 2.04 2.23 2.29 2.14 4.10 1.93 1.81 1.44 0.02 0.76

0.21 0.09 0.43 b0.01 0.11 0.09 0.08 0.10 0.01 0.13 0.15 0.23 1.00 0.52

Cut-off score used to indicate the high and low scores on MMPI scales in later logistic regression analyses. Factorial ANOVA, statistical significance of the corrected model, df = 3.

282

V. Siira et al. / Psychiatry Research 158 (2008) 278286

Table 3 Proportions (n, %) of deviant MMPI scores at the initial assessment of the adoptees and adjusted odds ratios when predicting any psychiatric disorder of the adoptees at follow-up Variable Group Adoptees mentally healthy at follow-up (n = 99) n L F K 4 2+7+8+0 86 SzP Pz HOS HYP PHO PSY REL SOC 28 25 27 20 22 28 30 23 19 18 27 24 29 29 % 28.3 25.3 27.3 20.2 22.2 28.3 30.3 23.2 19.2 18.2 27.3 24.2 29.3 29.3 Any psychiatric disorder at follow-up (n = 37) n 9 12 7 15 12 11 13 11 10 10 13 14 11 11 % 24.3 32.4 18.9 40.5 32.4 29.7 35.1 29.7 27.0 27.0 35.1 37.8 29.7 29.7 0.66 (0.261.66) 1.61 (0.683.82) 0.49 (0.181.30) 2.79 (1.176.62) 1.77 (0.744.22) 1.13 (0.482.68) 1.50 (0.653.49) 1.47 (0.613.55) 1.31 (0.543.33) 1.32 (0.513.38) 1.77 (0.734.25) 2.13 (0.915.00) 0.97 (0.412.29) 0.92 (0.392.19) Adjusted odds ratio (95% confidence interval)

(F = 3.97, df = 1, P = 0.05), the PHO (Phobias) scale (F = 5.37, df = 1, P = 0.02), and the PSY (Psychoticism) scale (F = 3.95, df = 1, P = 0.05). Furthermore, a statistically significant main effect of the follow-up diagnosis was found on the SzP (Schizophrenia Proneness) scale (F = 5.43, df = 1, P = 0.02) and the 4 (Psychopathic Deviate) scale (F = 9.32, df= 1, P b 0.01). 3.2. MMPI psychometric deviance variables as predictors of later psychiatric disorder The adoptees with high scores on scale 4 (Psychopathic Deviate) ( P = 0.02) at the initial assessment showed significantly more of Any Psychiatric Disorder at Followup (APDF) compared with the adoptees who had low scores on this scale, when adjusted for the adoptees gender, age at MMPI assessment, age at placement in the adoptive family, and genetic risk of schizophrenia in logistic regression analyses (see Table 3). 3.3. Follow-up psychiatric disorders among HR and LR adoptees None of the deviant scores on the MMPI predicted future mental disorder in the group of HR adoptees, as
Table 4 Proportions (n, %) of deviant MMPI scores at the initial assessment of the high-risk adoptees and adjusted odds ratios when predicting any psychiatric disorder of the adoptees at follow-up Variable Group Adoptees mentally healthy at follow-up (n = 38) n L F K 4 2+7+8+0 86 SzP Pz HOS HYP PHO PSY REL SOC 14 10 15 11 10 12 10 10 9 9 8 9 11 12 % 36.8 26.3 39.5 28.9 26.3 31.6 26.3 26.3 23.7 23.7 21.1 23.7 28.9 31.6 Any psychiatric disorder at follow-up (n = 22) n 6 5 5 7 4 4 5 4 8 7 5 7 7 8 % 27.3 22.7 22.7 31.8 18.2 18.2 22.7 18.2 36.4 31.8 22.7 31.8 31.8 36.4 0.48 (0.141.67) 0.85 (0.243.00) 0.38 (0.111.33) 1.05 (0.323.43) 0.58 (0.152.22) 0.43 (0.111.66) 0.85 (0.243.00) 0.62 (0.162.37) 1.73 (0.535.70) 1.21 (0.364.12) 1.23 (0.334.54) 1.56 (0.475.25) 1.20 (0.373.89) 1.31 (0.414.17) Adjusted odds ratio (95% confidence interval)

Note. The adoptee's gender, age at MMPI assessment, age at placement in the adoptive family, and genetic risk of schizophrenia were used as confounding variables. * P b 0.05.

calculated in logistic regression analyses to statistically account for the effect of the following confounding variables: adoptee's gender, age at MMPI assessment, age at placement in the adoptive family, and genetic risk of schizophrenia (McCullagh and Nelder, 1989). The 75th quartile was used as the cut-off point for all the other MMPI scales except HOS and HYP, for which the 25th quartile was used to show the schizophrenic direction from the pathological end of the distribution (Moldin et al., 1990a; Bolinskey et al., 2001). 3. Results 3.1. Descriptive analysis of MMPI scales As shown in Table 2, the overall results of factorial ANOVA, on which the variation in the mean of each MMPI scale was modelled with HR/LR status, followup diagnosis (mentally healthy/any psychiatric disorder), and their interaction, showed statistically significant findings on the 4 (Psychopathic Deviate) and HOS (Hostility) scales. The univariate F-tests showed a significant effect of HR/LR status on the F scale (Frequency) (F = 4.93, df = 1, P = 0.03), the HOS scale (F = 11.96, df = 1, P b 0.01), the HYP (Hypomania) scale

Note. The adoptee's gender, age at MMPI assessment, age at placement in the adoptive family, and genetic risk of schizophrenia were used as confounding variables.

V. Siira et al. / Psychiatry Research 158 (2008) 278286 Table 5 Proportions (n, %) of deviant MMPI scores at the initial assessment of the low-risk adoptees and adjusted odds ratios when predicting any psychiatric disorder of the adoptees at follow-up Variable Group Adoptees mentally healthy at follow-up (n = 61) n L F K 4 2+7+8+0 86 SzP Pz HOS HYP PHO PSY REL SOC 14 15 12 9 12 16 20 13 10 9 19 15 18 17 % 23.0 24.6 19.7 14.8 19.7 26.2 32.8 21.3 16.4 14.8 31.1 24.6 29.5 27.9 Any psychiatric disorder at follow-up (n = 15) n 3 7 2 8 8 7 8 7 2 3 8 7 4 3 % 20.0 46.7 13.3 53.3 53.3 46.7 53.3 46.7 13.3 20.0 53.3 46.7 26.7 20.0 0.82 (0.203.40) 2.89 (0.869.72) 0.61 (0.123.13) 8.99 (2.2835.39) 6.44 (1.7224.11) 3.11 (0.8810.94) 2.56 (0.788.44) 3.21 (0.9410.92) 0.81 (0.154.39) 1.43 (0.316.51) 2.41 (0.708.27) 2.70 (0.779.40) 0.74 (0.962.84) 0.64 (0.152.70) Adjusted odds ratio (95% confidence interval)

283

Note. The adoptee's gender, age at MMPI assessment, age at placement in the adoptive family, and genetic risk of schizophrenia were used as confounding variables. P b 0.05.

can be seen in Table 4. The LR adoptees who had high scores on scale 4 (Psychopathic Deviate) (P b 0.01), high sum scores of scales 2 + 7 + 8 + 0 (Golden-Meehl Indicators) (P = 0.01) at the initial assessment had significantly more Any Psychiatric Disorder at Follow-up (APDF) than the LR adoptees whose corresponding scores were low when adjusted for possible confounders in logistic regression (see Table 5). 4. Discussion In the Finnish Adoptive Family Study, the adoptee's deviant scores on the Psychopathic Deviate (4) and Hostility (HOS) scales of the MMPI at the initial assessment, when all adoptees were mentally healthy, were found to associate with the adoptees later psychiatric disorder at the follow-up assessment. However, when we statistically accounted for the adoptee's gender, age at MMPI assessment, age at placement in the adoptive family, and genetic risk of schizophrenia as confounding variables, only high scores on the Psychopathic Deviate scale predicted future mental disturbances. Our follow-up study suggests that the Psychopathic Deviate scale may be useful as a sign of predisposition to

the later onset of psychiatric disorder. It measures traits associated with asocial behavior, egocentricity, impatience, and shallow affect (Dahlstrom et al., 1982; Moldin et al., 1990b). Other studies have also shown that behavioral, emotional, and social adjustment problems precede schizophrenia and psychotic symptoms in highrisk studies (Amminger et al., 2000; Kugelmass et al., 1995; Olin et al., 1995). Our first hypothesis was hence confirmed. The exceptional character of adoptive samples was earlier confirmed by MacCrimmon et al. (1980). They found that their foster-reared group at high risk for schizophrenia and their foster-reared control group scored significantly higher on the Psychopathic Deviate scale than their community control group. The two foster groups did not differ significantly from each other, indicating that foster children are not fully comparable with normal children. The children in the present adoption sample and in the foster samples of McCrimmon et al. were placed for extrafamilial rearing mainly because of parental inability to take care of the child. The foster and adoptive families of these children are comparable for official selection and for absence of influence of the disturbed biological parent, but there are no environmental variables for accurate comparison of these families. Fostered and adopted-away children may be traumatized by their loss of their biological parents and the placement into a new family. Therefore, the results of an adoptive sample may not represent the results of children who have been living with their biological parents. Additional analyses made in the present study demonstrated that LR adoptees scored significantly higher on the F scale (Frequency) and the PSY scale (Psychoticism) than HR adoptees, and that HR adoptees scored significantly lower on the scales HOS (Hostility), HYP (Hypomania), and PHO (Phobias) than LR adoptees. Our earlier finding was thus confirmed for deviant scores on the scales HOS and HYP (Siira et al., 2004). However, we did not find any predictive validity of the MMPI variables among the HR adoptees born to mothers who had schizophrenia spectrum disorder. Instead, LR adoptees who showed any psychiatric disorder at follow-up had higher scores on the Psychopathic Deviate and the Golden-Meehl Indicator (sum of the MMPI scales Depression (2), Psychasthenia (7), Schizophrenia (8), and Social Introversion (0)) than LR adoptees who were mentally healthy at follow-up. When we statistically accounted for the confounding variables, the results of LR adoptees on the Psychopathic Deviate scale and the Golden-Meehl Indicator predicted the presence of any psychiatric disorder at follow-up. The

284

V. Siira et al. / Psychiatry Research 158 (2008) 278286

Golden-Meehl Indicator obtained here for any follow-up psychiatric disorder of the LR adoptees is consistent with the earlier findings (Chapman et al., 1982; Golden and Meehl, 1979). We have speculated about two explanations for the unsuccessful prediction of the follow-up diagnoses among HR adoptees based on the Hostility and Hypomania scales in the present study. Psychometric deviance in an adoptee does not necessarily lead to signs and symptoms without the presence of disruptive environmental influences, and we did not use environmental variables here. In addition, although not all adoptees had follow-up diagnoses during the present follow-up period, their mental health status may change later. Thus, our second hypothesis was not confirmed. Our results are contradictory to the finding of the GoldenMeehl Indicator being specific only for later onset of schizophrenia among children at high risk for schizophrenia (Bolinskey et al., 2001; Carter et al., 1999). However, the earlier high-risk studies did not include HR adoptees, and the results are therefore not comparable with our results concerning HR adoptees. In the earlier studies, the genotype-environment interaction was not accounted for because the children had been reared by a schizophrenic parent. Thus, the influence of the protective significance of the rearing environment (the adoptive family) has not been assessed in the earlier studies, as it has been in the Finnish Adoptive Family Study of Schizophrenia. It is possible that the present results on the unsuccessful prediction of mental disorders in the HR adoptees reflect the gene-environment interaction in the development of schizophrenia. Recent studies have indicated that the onset of schizophrenia spectrum disorders in HR adoptees depends more on the interaction between genetic and environmental factors than the onset of corresponding disorders in control adoptees (Tienari et al., 2004; Wahlberg et al., 1997, 2000, 2004). A favorable rearing environment in the adoptive family may prevent the expression of vulnerability factors and indicators in HR adoptees. On the other hand, the LR adoptees were not genetically vulnerable, and they were therefore not sensitive to the negative or positive effects of the genotype-environment interaction. In an earlier report from the Finnish Adoptive Family Study, Metsnen et al. (2004) also found that the future mental disorder of the HR adoptees could not be predicted based on thought disorders measured 16 years earlier. Their result is in line with the findings obtained here. In future studies, variables measuring the interaction between genetic and environmental factors need to be considered, to explore our hypothesis that the genotype-

environment interaction has an effect on the prediction of mental disorders in HR adoptees. In addition, the predictive validity of the MMPI needs to be studied using an index of multiple indicators among the adoptive family sample. Acknowledgments This research was supported in part by grant MH39663 from the Public Health Service, by a grant from the Scottish Rite Schizophrenia Research Program, N.M.J., U.S.A., The Academy of Finland, The Alma and K. A. Snellman Foundation, Oulu, Finland, Finnish Cultural Foundation, Finland, Emil Aaltonen Foundation, Tampere, Finland, and K-EVO. We thank Ilpo Lahti, M.D., Juha Moring, M.D., Mikko Naarala, M.D., Markku Seitamaa, M.A., and Anneli Sorri, M.D., for contributing to the psychological testing of the subjects in this study. References
American Psychiatric Association, 1987. Diagnostic and Statistical Manual of Mental Disorders, 3rd ed., rev. APA, Washington, DC. Amminger, G.P., Pape, S., Rock, D., Roberts, S.A., Squires-Wheeler, E., Kestenbaum, C., Erlenmeyer-Kimling, L., 2000. The New York High-Risk Project: comorbidity for axis I disorders are preceded by childhood behavioral disturbance. Journal of Nervous and Mental Disease 188, 751756. Armitage, P., Berry, G., 1987. Statistical Methods in Medical Research, Second Edition. Blackwell Scientific Publications, Great Britain. Bland, M., 1995. An Introduction to Medical Statistics. Oxford University Press, Oxford. Bolinskey, P.K., Gottesman, I.I., Nichols, D.S., Shapiro, B.M., Roberts, S.A., Adamo, U.H., et al., 2001. A new MMPI-derived indicator of liability to develop schizophrenia: evidence from the New York high-risk project. Assessment 8, 127143. Carter, J.W., Parnas, J., Cannon, T.D., Schulsinger, F., Mednick, S.A., 1999. MMPI variables predictive of schizophrenia in the Copenhagen High-Risk Project: a 25-year follow up. Acta Psychiatrica Scandinavica 99, 432440. Chapman, L.J., Chapman, J.P., Miller, E.N., 1982. Reliabilities and intercorrelations of eight measures of proneness to psychosis. Journal of Consulting and Clinical Psychology 50, 187195. Cornblatt, B.A., Obuchowski, M., 1997. Update of high-risk research: 19871997. International Review of Psychiatry 9, 437447. Cornblatt, B.A., Obuchowski, M., Roberts, S., Pollack, S., Erlenmeyer-Kimling, L., 1999. Cognitive and behavioral precursors of schizophrenia. Development and Psychopathology 11, 487508. Curran, P.J., West, S.G., Finch, J.F., 1996. The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis. Psychological Methods 1, 1629. Dahlstrom, W.G., Welsh, G.S., Dahlstrom, L., 1982. An MMPI Handbook. Clinical Interpretation, vol. 1. University of Minnesota Press, Minneapolis. Ellison, Z., van Os, J., Murray, R., 1998. Special feature: childhood personality characteristics of schizophrenia: manifestations of, or risk factors for, the disorder ? Journal of Personality Disorders 12, 247261 (Review).

V. Siira et al. / Psychiatry Research 158 (2008) 278286 Erlenmeyer-Kimling, L., 2000. Neurobehavioral deficits in offspring of schizophrenic parents: liability indicators and predictors of illness. American Journal of Medical Genetics 97, 6571. Erlenmeyer-Kimling, L., Cornblatt, B.A., 1978. Attentional measures in a study of children at high-risk for schizophrenia. Journal of Psychiatric Research 1, 9398. Erlenmeyer-Kimling, L., Cornblatt, B.A., 1992. A summary of attentional findings in the New York High-Risk Project. Journal of Psychiatric Research 26, 405426. Erlenmeyer-Kimling, L., Adamo, U.H., Rock, D., Roberts, S.A., Bassett, A.S., Squires-Wheeler, E., et al., 1997. The New York high-risk project: prevalence and comorbidity of axis I disorders in offspring of schizophrenic parent at 25-year follow-up. Archives of General Psychiatry 54, 10961102. Erlenmeyer-Kimling, L., Rock, D., Roberts, S.A., Janal, M., Kestenbaum, C., Cornblatt, B., Adamo, U.H., Gottesman, I.I., 2000. Attention, memory, and motor skills as childhood predictors of schizophrenia-related psychoses: the New York High-Risk Project. American Journal of Psychiatry 157, 14161422. Fish, B., Marcus, J., Hans, S., Auerbach, J., Perdue, S., 1992. Infants at risk for schizophrenia: sequelae of a genetic neurointegrative defect. Archives of General Psychiatry 49, 221235. Friedman, A.F., Webb, J.T., Lewak, R., 1989. Psychological Assessment with the MMPI. Lawrence Erlbaum Associates, Hillsdale, NJ. Golden, R.R., Meehl, P.E., 1979. Detection of the schizoid taxon with MMPI indicators. Journal of Abnormal Psychology 88, 217233. Gottesman, I.I., 1991. Schizophrenia Genesis: The Origins of Madness. WH Freeman, New York. Gottesman, I.I., Erlenmeyer-Kimling, L., 2001. Family and twin strategies as a head start in defining prodromes and endophenotypes for hypothetical early-interventions in schizophrenia. Schizophrenia Research 51, 93102. Gottesman, I.I., Shields, J., 1967. A polygenetic theory of schizophrenia. Proceedings of the National Academy of Sciences of the United States of America 58, 199205. Green, M.F., 1998. Schizophrenia from a Neurocognitive Perspective. Probing in Impenetrable Darkness. Allyn and Bacon, Boston. Greene, R.L., 1980. The MMPI: An Interpretive Manual. Grune & Stratton, New York. Haier, R.J., Rosenthal, D., Wender, P.H., 1978. MMPI assessment of psychopathology in the adopted-away offspring of schizophrenics. Archives of General Psychiatry 35, 171175. Hans, S., Marcus, J., Nuechterlein, K., Asarnow, R., Styr, B., Auerbach, J., 1999. Neurobehavioral deficits at adolescence in children at risk for schizophrenia: The Jerusalem infant development study. Archieves of General Psychiatry 56, 741748. Kendler, K.S., O'Neill, F.A., Burke, J., Murphy, B., Duke, F., Straub, R.E., Shinkwin, R., Ni Nuallain, M., MacLean, C.J., Walsh, D., 1996. Irish study of high-density schizophrenia families: field methods and power to detect linkage. American Journal of Medical Genetics (Neuropsychiatric Genetics) 67, 179190. Kendler, K.S., Myers, J.M., O'Neill, F.A., Martin, R., Murphy, B., MacLean, C.J., Walsh, D., Straub, R.E., 2000. Clinical features of schizophrenia and linkage to chromosomes 5q, 6p, 8p, and 10p in the Irish Study of High Density Schizophrenia Families. American Journal of Psychiatry 157, 402408. Kugelmass, S., Faber, N., Ingraham, L.J., Frenkel, E., Nathan, M., Mirsky, A.F., Shakhar, G.B., 1995. Reanalysis of SCOR and anxiety measures in the Israeli High-Risk Study. Schizophrenia Bulletin 21, 205217. MacCrimmon, D.J., Cleghorn, J.M., Asarnow, R.F., Steffy, R.A., 1980. Children at risk for schizophrenia: clinical and attentional characteristics. Archives of General Psychiatry 37, 671674.

285

McCullagh, P., Nelder, J.A., 1989. Generalized Linear Models, 2nd ed. Chapman and Hall, New York. Meehl, P.E., 1962. Schizotaxia, schizotypy and schizophrenia. American Psychologist 17, 827838. Metsnen, M., Wahlberg, K.-E., Saarento, O., Tarvainen, T., Miettunen, J., Koistinen, P., Tienari, K., 2004. Early presence of thought disorder as a prospective sign of mental disorder. Psychiatry Research 125, 193203. Miller, B.C., Fan, X., Grotevant, H., Christensen, M., Coyl, D., Van Dulmen, M., 2000. Adopted adolescents' overrepresentation in mental health counseling: adoptees' problems or parents' lower threshold for referral? Child and Adolescent Psychiatry 39 (12), 15041511. Minnesota Veterans Administration Hospital\MMPI Research Laboratory, 1975. An Atlas for the P-Code System of MMPI Interpretation. Veterans Administration Hospital, Minneapolis, MN. Moldin, S.O., Gottesman, I.I., Erlenmeyer-Kimling, L., Cornblatt, B.A., 1990a. Psychometric deviance in offspring at risk for schizophrenia: I. Initial Delineation of a Distinct Subgroup. Psychiatry Research 32, 297310. Moldin, S.O., Rice, J.P., Gottesman, I.I., Erlenmeyer-Kimling, L., 1990b. Psychometric deviance in offspring at risk for schizophrenia: II. Resolving heterogeneity through admixture analysis. Psychiatry Research 32, 311322. Moldin, S.O., Rice, J.P., Gottesman, I.I., Erlenmeyer-Kimling, L., 1990c. Transmission of a psychometric indicator for liability to schizophrenia in normal families. Genetic Epidemiology 7, 163176. Nuechterlein, K.H., 1987. Vulnerability models for schizophrenia: State of the Art. In: Hfner, H., Gattaz, W.F., Janzarik, W. (Eds.), Search for the Causes of Schizophrenia. Springer-Verlag, Giessen, Germany, pp. 297316. Olin, S.S., John, R.S., Mednic, S.A., 1995. Assessing the predictive value of teacher reports in a high risk sample for schizophrenia: a ROC analysis. Schizophrenia Research 16, 5366. Ott, S., Allen, J., Erlenmeyer-Kimling, L., 2001. The New York HighRisk Project: observations on the rating of early manifestations of schizophrenia. American Journal of Medical Genetics 105, 2527. Parnas, J., Cannon, T., Jacobsen, B., Schulsinger, H., Schulsinger, F., Mednick, S., 1993. Lifetime DSM-III-R diagnostic outcomes in the offspring of schizophrenic mothers. Archives of General Psychiatry 50, 707714. Rosen, A., 1962. Development of MMPI scales based on a reference group of psychiatric patients. Psychological Monographs 76, 125. Rosenthal, D., 1970. Genetic Theory and Abnormal Behavior. McGraw-Hill, New York. Schubert, E.W., McNeil, T.F., 2003. Prospective study of adult mental disturbance in offspring of women with psychosis. Archives of General Psychiatry 60, 473480. Siira, V., Wahlberg, K.-E., Miettunen, J., Lksy, K., Tienari, P., 2004. Psychometric deviance measured by MMPI in adoptees at high risk for schizophrenia and their adoptive controls. Journal of Personality Assessment 83 (1), 1421. Statistics Finland, 17., 1983. Tilastokeskus, Helsinki, Finland. Subotnik, K.L., Asarnow, R.F., Nuechterlein, K.H., Fogelson, D.L., Thorpe, T.I., Payne, D.L., Giannini, C.A., Kuppinger, H.E., Torquato, R.D., Mintz, J., Hwang, S.S., Gottesman, I.I., 2005. MMPI vulnerability indicators for schizophrenia and attention deficit disorder: UCLA family study of biological parents of offspring with childhood onset schizophrenia or ADHD. Behavior Genetics 35 (2), 159175.

286

V. Siira et al. / Psychiatry Research 158 (2008) 278286 Wahlberg, K.-E., Wynne, L., Oja, H., Keskitalo, P., Pyklinen, L., Lahti, I., Moring, J., Naarala, M., Sorri, A., Seitamaa, M., Lksy, K., Kolossa, J., Tienari, P., 1997. Gene-environment interaction in vulnerability to schizophrenia: findings from the Finnish Adoptive Family Study of schizophrenia. American Journal of Psychiatry 154, 355362. Wahlberg, K.-E., Wynne, L.C., Oja, H., Keskitalo, P., Anias-Tanner, H., Koistinen, P., Tarvainen, T., Hakko, H., Lahti, I., Moring, J., Naarala, M., Sorri, A., Tienari, P., 2000. Thought Disorder Index of Finnish adoptees and communication deviance of their adoptive parents. Psychological Medicine 30, 127136. Wahlberg, K.-E., Wynne, L.C., Hakko, H., Lksy, K., Moring, J., Miettunen, J., Tienari, P., 2004. Interaction of genetic risk and adoptive parent communication deviance: longitudinal prediction of adoptee psychiatric disorders. Psychological Medicine 34, 15311542. Wiggins, J.S., 1966. Substantive dimensions of self-report in the MMPI item pool. Psychological Monographs 80, 142 (Whole No. 630). Wing, J.K., Cooper, J.E., Sartorius, N., 1974. Measurement and Classification of Psychiatric Symptoms: An Instruction Manual for the PSE and Catego Program. Cambridge University Press, London. Veijola, J., Maki, P., Joukamaa, M., Laara, E., Hakko, H., Isohanni, M., 2004. Parental separation at birth and depression in adulthood: a long-term follow-up of the Finnish Christmas Seal Home Children. Psychological Medicine 34 (2), 357362. Zubin, J., Spring, B., 1977. Vulnerability a new view of schizophrenia. Journal of Abnormal Psychology 86, 103126.

Tienari, P., Sorri, A., Naarala, M., Lahti, I., Bostrm, C., Wahlberg, K.-E., 1981. The Finnish adoptive family study, family-dynamic approach on psychosomatics. A preliminary report. Psychiatry and Social Science 1, 107115. Tienari, P., Lahti, I., Sorri, A., Naarala, M., Wahlberg, K.-E., Rnkk, T., et al., 1987a. The Finnish Adoptive Family Study of schizophrenia: possible joint effects of genetic vulnerability and family interaction. In: Halweg, K., Goldstein, M.J. (Eds.), Understanding Major Mental Disorder: The Contribution of Family Interaction Research. Family Process, New York, pp. 3354. Tienari, P., Sorri, A., Lahti, I., Naarala, M., Wahlberg, K.-E., Moring, J., Pohjola, J., Wynne, L.C., 1987b. Genetic and psychosocial factors in schizophrenia: the Finnish adoptive family study. Schizophrenia Bulletin 13, 477484. Tienari, P., Wynne, L.C., Moring, J., Lksy, K., Nieminen, P., Sorri, A., Lahti, I., Wahlberg, K.E., Naarala, M., Kurki-Suonio, K., Saarento, O., Koistinen, P., Tarvainen, T., Hakko, H., Miettunen, J., 2000. Finnish adoptive family study of schizophrenia: sample selection and adoptee DSM-III-R diagnoses. Acta Psychiatrica Scandinavica 101, 433443. Tienari, P., Wynne, L.C., Lksy, K., Moring, J., Nieminen, P., Sorri, A., Lahti, I., Wahlberg, K.E., 2003. Genetic boundaries of the schizophrenia spectrum: evidence from the Finnish adoptive family study of schizophrenia. American Journal of Psychiatry 160, 15871594. Tienari, P., Wynne, L.C., Sorri, A., Lahti, I., Lksy, K., Moring, J., Naarala, M., Nieminen, P., Wahlberg, K.-E., 2004. Genotypeenvironment interaction in schizophrenia-spectrum disorder. British Journal of Psychiatry 184, 216222.