K»

UMEÅ UNIVERSITY MEDICAL DISSERTATIONS New series No 101 - ISSN 0346 - 6612

From the Department of Child and Youth Psychiatry, University of Umeå, Umeå, Sweden

ADOPTION STUDIES ON PSYCHIATRIC ILLNESS Epidemiological, Environmental and

Genetic Aspects

by

Anne-Liis von Knorring

Umeå University Umeå 1983

ADOPTION STUDIES ON PSYCHIATRIC ILLNESS Epidemiological, Environmental and

Genetic Aspects

AKADEMISK AVHANDLING

som med vederbörligt tillstånd av Medicinska fakulteten vid Umeå universitet för avläggande av medicine doktorsexamen

kommer att offentligen försvaras i föreläsningssalen, Psykiatriska kliniken, torsdagen den 19 maj 1983 kl. 09.15.

av

ANNE-LIIS von KNORRING

med. lic.

Umeå 1983

Umeå University Medical Dissertations New Series No 101-ISSN 0346-6612

ABSTRACT

Adoption studies on psychiatric illness. Epidemiological, environmental, and genetic aspects. Anne-Liis von Knorring, Department of Child and Youth Psychiatry, University of Umeå, S-901 85 Umeå, Sweden.

The aim of this study is to evaluate the outcome of adoptions and to study the gene-environment influences on psychiatric illness as well as sick-leave patterns. The material consists of 2 966 adopted persons born between 1917 and 1949, their 5 932 adoptive parents and 5 438 identified biological parents.

Adopted persons had a higher incidence of personality disorders and substance abuse than non-adopted controls. Adopted men also had an increased incidence of neuroses. Adopted women had an increased sick-leave because of somatic complaints, especially upper respiratory tract infections and abdominal complaints of short duration. Somatiza­tion i.e. more than 2 sick-leaves/year because of somatic complaint together with nervous complaints was more frequent among adopted women. Women with somatization could be separated into 2 types according to the pattern of sick-leave. Type 1 ("high frequency") had frequent sick-leaves for psychiatric, abdominal and back com­plaints. They also had a high frequency of alcohol abuse. Type 2 ("diversiform") had more diverse complaints and had fewer sick-leaves because of nervous complaints.

High frequency somatizers had biological fathers with teenage onset of criminality and frequent registrations for alcohol abuse. Diversiform somatizers had the same genetic background as adopted men with petty criminality or male limited alcoholism.

No specific genetic influences on treated depression or substance abuse were found in this study. However, a non-specific vulnerability of the biological mother influenced on the risk of depression and substance abuse among adopted women.

There were some indications that placement in the adoptive home between 6 and 12 months of age was associated with reactive neurotic depression in adult life. Otherwise early negative experiences in term of unstable placements before adoption did not significantly influence on psychiatric illness in adulthood.

Affective disorders in the adoptive father were associated with treat­ment for depressions or substance abuse in the adoptee. Low social status in the part of the adoptive father increased the risk of somati­zation of both types in the adoptee.

Key words: Adoption - adoption outcome - psychiatric illness - gene-environment - early experience - somatoform disorders - affective disorders - substance abuse - sick-leave.

UMEÅ UNIVERSITY MEDICAL DISSERTATIONS New series No 101 - ISSN 0346 - 6612

From the Department of Child and Youth Psychiatry, University of Umeå, Umeå, Sweden

ADOPTION STUDIES ON PSYCHIATRIC ILLNESS Epidemiological, Environmental and

Genetic Aspects

by

Anne-Liis von Knorring

IJmeå University Umeå 1983

2

ABSTRACT

The aim of this study is to evaluate the outcome of adoptions and to study the gene-environment influences on psychiatric illness as well as sick-leave patterns. The material consists of 2 966 adopted persons born between 1917 and 1949, their 5 932 adoptive parents and 5 438 identified biological parents.

Adopted persons had a higher incidence of personality disorders and substance abuse than non-adopted controls. Adopted men also had an increased incidence of neuroses. Adopted women had an increased sick-leave because of somatic complaints, especially upper respiratory tract infections and abdominal complaints of short duration. Somatiza­tion i.e. more than 2 sick-leaves/year because of somatic complaint together with nervous complaints was more frequent among adopted women. Women with somatization could be separated into 2 types according to the pattern of sick-leave. Type 1 ("high frequency") had frequent sick-leaves for psychiatric, abdominal and back com­plaints. They also had a high frequency of alcohol abuse. Type 2 ("diversiform") had more diverse complaints and had fewer sick-leaves because of nervous complaints.

High frequency somatizers had biological fathers with teenage onset of criminality and frequent registrations for alcohol abuse. Diversiform somatizers had the same genetic background as adopted men with petty criminality or male limited alcoholism.

No specific genetic influences on treated depression or substance abuse were found in this study. However, a non-specific vulnerability of the biological mother influenced on the risk of depression and substance abuse among adopted women.

There were some indications that placement in the adoptive home between 6 and 12 months of age was associated with reactive neurotic depression in adult life. Otherwise early negative experiences in term of unstable placements before adoption did not significantly influence on psychiatric illness in adulthood.

Affective disorders in the adoptive father were associated with treat­ment for depressions or substance abuse in the adoptee. Low social status in the part of the adoptive father increased the risk of somati­zation of both types in the adoptee.

Key words: Adoption - adoption outcome - psychiatric illness - gene-environment - early experience - somatoform disorders - affective disorders - substance abuse - sick-leave.

3

The present thesis is based on the following papers:

I. Bohman, M. and von Knorring, A-L.: Psychiatric illness among adults adopted as infants. Acta Psychiat. Scand. 60, 106-112, 1979.

II. von Knorring, A-L.. Bohman, M. and Sigvardsson, S.: Early life experiences and psychiatric disorders: an adoptee study. Acta Psychiat. Scand. 65, 283-291, 1982.

III. Sigvardsson, S., von Knorring, A-L., Bohman, M., Cloninger, C. R. : An adoption study of somatoform disorders I. The relation­ship of somatization to psychiatric disability. Submitted Arch. Gen. Psychiat.

IV. Cloninger, CR., Sigvardsson, S., von Knorring, A-L., Bohman, M.: An adoption study of somatoform disorders II. Identification of two discrete somatoform disorders. Submitted Arch. Gen. Psychiat.

V. Bohman, M., Cloninger, C.R., von Knorring, A-L., Sigvards­son, S.: An adoption study of somatoform disorders III. Cross-fostering analysis and genetic relationship to alcoholism and criminality. Submitted Arch. Gen. Psychiat.

VI. von Knorring, A-L., Cloninger, C.R., Bohman, M., Sigvards­son, S.: An adoption study of depressive disorders and sub­stance abuse. Submitted Arch. Gen. Psychiat.

4

CONTENTS

GENERAL BACKGROUND 6

Adoption practices in Sweden 6 Adoption research 7

PURPOSE OF THE STUDY 10

REVIEW OF THE LITERATURE 11

Outcome studies of adoptions 11 Maternal deprivation and psychopathology 14 The psychotic-neurotic distinction of depression 16 Genetics of affective disorders 17 Affective disorders and alcoholism 20 The transmission of alcohol abuse and criminality 21 Somatoform disorders and illness behaviour 22

MATERIALS AND METHODS 27

The cohort of adoptees 27 Register data 28 Reliability of psychiatric diagnoses 30 Life events 32 Statistical analyses 32

RESULTS 34

Psychiatric illness among adoptees 34 Somatic illness among adoptees 34 The relation of somatic and psychiatric illness 35 Early childhood experiences and psychiatric disorders 37 Postnatal childhood experiences and somatization 38 Environmental influences on affective disorders and 38 substance abuse

5

Genetic inheritance of affective disorders and substance 39 abuse Genetic inheritance of somatization 40 Life events related to genetic and/or social predisposition 41 Gene-environment correlation and interaction of 41 somatization

SOURCES OF ERROR 43

DISCUSSION 46

Somatic complaints and psychopathology 46 Environmental influences and psychiatric disability 48 Genetic influences on psychopathology 51 Gene-environment interaction on psychopathology 53

SUMMARY OF THE MAIN RESULTS 57

APPENDIX I 59

ACKNOWLEDGEMENTS 60

REFERENCES 61

Paper I 81 Paper II 89 Paper III 99 Paper IV 133 Paper V 165 Paper VI 201

6

GENERAL BACKGROUND

The practice of adoption occurs in some form in most cultures. In ancient Rome the law only permitted adoption in the case of childless couples who needed an heir. The adopters had then to be past child-bearing age and the adoptee had to be an adult. The older European adoption laws were influenced by the Roman law. In North-America, however, the adoption laws had a different approach. Adoption was regarded as a way of ensuring a secure home and parentage for a homeless child.

In Sweden the first adoption law came into force 1917 and was influen­ced both by Roman law and North-American laws. Adoption according to the first Swedish law was only permitted if the couple was child­less. The adoption could be cancelled under certain conditions 1) if the child turned out to have unknown disablements or diseases 2) if the parents or the child commited serious offences against each other. From 1944 onwards couples with biological children were also allowed to adopt children. In 1958 the adoption law was changed in order to strengthen the child's legal position in the adoptive family. The adoptee's ties to biological relatives were totally severed and the adoptee was granted the same rights in the adoptive family as children bora of the adoptive parents. Since 1971 it has been impossible to cancel an adoption. As there have been very few changes in adoption practices 1917-1949, the cohort of adoptees in this study (born 1917-1949) show no great differences concerning placement procedure and selection.

Adoption practices in Sweden

It was estimated that about 1 % of the population was adopted by couples who were not relatives or step-parents in the 30's and 40's (Jonsson 1964). Private adoption agencies are prohibited and the majority of infant adoptions are arranged by the local municipalities of the child welfare committees. The organization of the adoption agen­cies varies throughout Sweden but the principal task for each agency

7

is to prepare and participate in the placement of the children. The work of such an agency also includes assessment of the ability of the prospective adoptive parents to bring up and take care of the child.

Adoption research

As a ready-made ex-post facto experiment in social and genetic re­search , the adoptive situation has been used to focus on two main approaches :

I. _Studies_of the effect of adoption

Research has for example been carried out to determine which factors in the child, the adoptive parents and society influence the adoptive process and to what extent this process can be forecast prior to placement. Such evaluation studies try to evaluate the extent to which adoption has been effective as a social measure.

In Sweden the adoptive parents are often older and tend to have a higher level of education than biological parents. This is a consequen­ce of the adoption law and of practice. Adoptive parents have to be at least 25 years old before they are allowed to adopt children. Con­sequently their marriages are of longer standing than the marriages which produced biological children (Bohman 1970). Thus it can be assumed that there is a restriction range in the environmental factors regarding adopted children. Personal qualities, however, constitute the most important factor for the development and adjustment of the adoptee - including the adopters' attitudes to adoption and to illegiti­mate birth and to infertility. "The age, income and social class of adopters are of far less importance" (Pringle 1969).

II. Genetic research

Genetic and environmental factors are difficult to separate, especially in behavioural sciences. Both social and genetic factors are inherited but one way to separate them is through adoption studies. Genetically

8

unrelated individuals live together and genetically related individuals live apart. The adoptive situation is in itself a ready-made experiment for testing hypotheses concerning the importance of heredity and childhood environment. These kinds of studies can determine the extent to which familial resemblances are due to genetic or environ­mental similarity. The environmental influences distinct from genetic variables, can be eavlutated - as well as the interaction of genetic and environmental factors.

The correlation between the adoptee and the biological or adoptive parents or other environmental variables concerning certain traits or diseases can be measured. Such studies can start from the adoptees (the adoptee's families method), or from the parents (the adoptees study method) (See Fig. 1).

The term "cross-fostering method" is applied to studies in which postnatal influences can be isolated e.g. by studying adoptees from an affected biological background reared in a non-affected adoptive home and vice versa. The prevalence of a trait or disease can thus be related to both genetic and environmental influences.

9

Adoptees' families' method

"V Index

Adoptees study method

O7O

Index Control

0-r-tS ©-—{E

Control

O Female

•

0

Male

Either sex

Affected

o

•

Unaffected

Individual whose status is under investigation

Biological parentage

Adoptive parentage

Fig. 1. Research designs of adoption studies (modified from Plomin, DeFries & McClean 1980).

10

PURPOSE OF THE STUDY

The two main approaches within adoption research (see pp. 6-7) are employed in this study. The following questions are considered:

Out-come studies

Do adopted persons as adults differ from the general population as regards :

1. Psychiatric illness (frequency, duration, pattern) 2. Somatic complaints (frequency, duration, pattern) 3. The relation between psychiatric and somatic illness?

Genetic-environmental studies

A. To what extent are psychiatric illness and/or special patterns of sick-leave influenced by:

1. Childhood experiences prior to placement in the adoptive home 2. Familial environment during childhood and adolescence?

B. Do genetic factors influence: 1. Affective disorders and substance abuse 2. Special patterns of sick-leave?

C. How is the interaction between genetic and environmental fac­tors?

11

REVIEW OF THE LITERATURE

Outcome studies of adoptions

As adoption laws and practices vary in time and from country to country it can be difficult to compare the outcome of adoptions in different studies.

USA

The earliest follow-up of the long-term out-come of adoptions showed that the children who had been legally adopted had a better outcome than foster children (van Theis 1924). This study followed the sub­jects up to the age of 18-40 years. The better outcome of the adopted children was attributed to the fact that these children were adopted and placed in their adoptive homes before the age of 5 years. Only a small proportion of the adoptees were found "incapable".

Adopted persons had a higher intellectual ability than was to be expected from the socio-economic status of their biological parents (Skodak 1949, Skeels 1965).

A 10-year follow-up study concerning private adoptions showed that 70 % of the families functioned satisfactorily (Witmer 1963). The girls showed a better adjustment than the boys. Compared with a group of biological children the adopted children seemed to be at a slight disadvantage regarding personal and social functioning, but not intellectually. Different results were found in another 10-year follow-up study (Lawder 1969, Hoopes 1969), which revealed no significant differences between adopted and biological children concerning beha­viours rated by parents and teachers.

France

Schiff (1978) studied children of lower class origin who were adopted early into high status homes and who also |iad a sibling or half-sibling

12

reared by the biological mother. The adopted children had higher IQ ratings on average than their biological siblings and the children reared by their biological mothers had far more educational problems.

England

In one study (Seglow 1972) adopted children were compared to the general population and to children of unmarried women. At the age of 7 the adopted children were better achievers in school than the illegitimate non-adopted children. The adopted children also turned out favourably compared to the general population.

Children who had spent their first two years in an institution and were later adopted, restored to their biological parents or fostered on a long-term basis were compared in a study by Tizard (1977). The adoptions were considered successful at an investigation carried out when the subjects were 8 years old. Compared to the other groups "adoption was clearly the best solution to the child's need".

A 20-year follow-up study concerning one group of children first fostered and later adopted and another group immediately adopted as infants was made by Raynor (1980). Nearly all the children came to their families during their first year of life and 85 % of the parents rated the experience of adoption as very satisfactory. There were no differences between those who had adopted their child directly and those who had fostered first.

Sweden

An earlier study of adopted persons followed to adulthood showed that most of the adoptions turned out satisfactorily (Thysell 1952, 1965). The opportunities offered the adopted children for social, psycholo­gical and educational development were better than if they had re­mained with their unmarried mothers. The adoptive parents reported "some trouble" in 20 % of the cases and "serious problems" in 5 % of the cases. Problems were more common among the girls.

13

A 10-year follow up study (Bohman 1970, 1971) of adopted children showed a slight deviation in a negative direction concerning beha­viour. The deviation was greater for boys. Compared to a group of foster children the adopted boys had a much better outcome. At the age of 15 years the adopted children did not differ from their class­mates. The children in foster-care (although many of them were adopted at an older age) and the children who stayed with their biological mothers (although unwanted during pregnancy) tended to show social maladjustment (Bohman 1972). At the age of 19 it was found that the boys in foster care and those reared by their biolo­gical mothers had been rejected for military service for emotional problems and in some cases intellectual deficiency reasons. The adopt­ed youngsters had the best outcome in this respect (Bohman 1978a).

During the last few decades the adoption situation has changed in Sweden in that practically no children born here are adopted away any more. The reason for this is that abortion is legal for any reason during the first trimester of the pregnancy. The adopted children nowadays are mostly from Asia or Latin-America. Follow-up studies concerning international adoptions have shown that there is a good outcome if the children are adopted early. If the children go to the adoptive home at a later age there are more initial problems but as a whole their adjustment is quite good (Gardell 1979, Cederblad 1979).

In series of psychiatric in- and out-paitents some studies have found that there is an överrepresentation of adoptees. The frequencies reported differ from 2.4-13.3 % (Goodman 1963, Schechter 1960). This is to be compared to the frequency of 1-2 % of adoptees in a normal population. Other authors have not found any överrepresentation of adopted adults among psychiatric patients, but on the other hand the representation of adoptees among Children's Service patients was twice what would be expected (Brinich 1982).

Conclusions to be drawn from the above mentioned studies: Adopted persons have a good outcome when compared with persons raised in fosterhomes or by biological parents from a socially disadvantaged

14

group. Compared to the general population the outcome for adopted persons is as good or slightly poorer.

Maternal deprivation and psychopathology

The short-term effects of parent-child separation have been described by many researchers (Spitz 1946, Bowlby 1958, Robertson 1971, Yarrow 1964), but not all (Davenport 1970). The reaction is most marked in children aged between six months and four years. Children differ in their response to separation, because of differences in their temperamental characterstics before the separation, and because children used to brief ordinary separations are less distressed by unhappy traumatic separations (Stacey 1970). According to psycho-dynamic theory attachment and early bonding to the mother is seen as the essential precursor to later social relationships (Bowlby 1969, 1973). Infants usually develop an attachment to a specific person at the age of 6-12 months (Ainsworth 1973, Rutter 1979). The concept of bonding implies selective attachment which persists over time (Cohen 1974). The difference between attachment and bonding is shown in a study concerning 4-year-old children reared in institutions who were more clinging and tended to follow people around more than family-reared children, but they were also less likely to show selective bonding or deep relationships (Tizard 1975).

Similar results are found in animal studies. Infant Rhesus monkeys who had experienced more rejection showed more distress after se­paration (Hinde 1970).

One relevant question in this connection is how childhood experiences influence personality development. In psychoanalytically oriented psychiatry it is thought that early childhood experiences bear some causal relation to psychiatric illness in adult life. The relationship between parental deprivation and later depression has been widely studied with parental loss as the experience most commonly assessed. Psychoanalytical literature suggests that depression is the most likely outcome of parental loss (Abraham 1927, Freud 1917, Bowlby 1961).

15

The effect of the death of a parent is often studied, because this event is easily defined. The incidence of parental loss through death before the age of 15 has been found to be higher among psychiatric patients in several studies. In studies concerning depressed patients the patients had experienced bereavement to a greater extent than controls (Beck 1963, Forrest 1965, Dennehy 1966, Brown 1961, 1966). Another study showed a correlation between bereavement before the age of 6 years and recurrent depressive disorders (d'Elia 1971). Other researchers, however, have not found these correlations at all (Gregory 1966, Munro 1966, Roy 1978). In most studies separately defined age ranges for parental loss have not been used and this may be one reason why the results diverge.

Parental loss has also been shown to be associated with the persona­lity tfaits of dependency and hypochondriasis (Birchnell 1975, 1978). This may influénce the utilization of services, but the results are contradictory concerning the direction. Tennant (1980a, b) found from a community survey that parent-child separation due to parental death did not bear any significant relation to psychiatric symptomato­logy in adult life but was associated with referral to psychiatric services. On the other hand Brown (1978) found that depressed persons who had experienced childhood loss were less likely to be referred to a psychiatrist.

Within a population of depressed patients, those who attempted suicide showed a higher rate of parental loss than those who did not (Hill 1969). The same association was found among psychoneurotic and sociopathic patients who attempted suicide (Greer 1964).

Parental loss is not only caused by the death of a parent, it can also arise through separation but this is more difficult to define. Separa­tion can take the form of divorce between the parents, but all divor­ces do not mean that the contact between the parent and child is cut off. In spite of these difficulties some authors have found a correla­tion between separation and psychopathology. Paternal loss due to separation or divorce occurred with much greater frequency among a

16

large in- and out-patient population than in 2 control groups (Brill 1966). This finding occurred in conjunction with a great variety of diagnoses without any specific category. In a more recent work (Ten-nant 1982) the relationship of childhood separation to adult depres­sion, anxiety and 'general psychological morbidity' was assessed. Childhood separation occurring up to five years of age bore no rela­tion to psychopathology. From five years of age separations caused by parental illness were related to morbidity.

Deprivation during childhood does not necessarily result only from death of a parent or separation. It can also take the form of depri­vation of loving care or stimulating education towards self-realization. Depressed patients were found to have experienced the lack, loss or absence of an emotionally sustaining relationship before adolescence to a greater extent than controls (Raskin 1971, Jacobson 1975). In the light of these findings H. Perris (1982a) hypothesized that persons who had experienced rejection during childhood ought to be more vulnerable to life events and suffer depressions more easily. Her results supported the hypotheses. Rejected and depressed patients, had consistently experienced fewer life events than patients reared by stimulating mothers.

The psychotic-neurotic distinction of depression

During the last two decades many different classifications for depres­sive disorders have been proposed. There are classifications which involve from one up to nine categories (Kendell 1976). There are also dimensional categories with varying numbers of dimensions. Lewis (1938) did not think that the existing classifications at that time were useful. He accepted however that the phenomenology and ethiology of depressive disorders were heterogeneous. The situation does not seem to have changed much since then.

Most controversies have been about the psychotic-neurotic (or endo­genous» reactive) distinction. Confusion has arisen from the change­able meaning of the term itself (Kendell 1976).

17

There is some association between melancholic symptom cluster and lack of environmental stress (Rosenthal 1966, Perris H. 1982a), but it is weak. There is however little relationship between the type and severity of the preceeding stresses and the symptomatology of the illness itself (Paykel 1974, Perris H. 1982a). Some studies have been made which suggest that endogenous and neurotic depressions are distinct illnesses and that there is a similar clear-cut boundary be­tween depressive illness as a whole and anxiety states (Carney 1965, Gurney 1972). Other researchers have tried to show a bimodal distri­bution of scores on either a bipolar psychotic/neurotic factor or a discriminant function of symptomatology, but the overall distribution of symptomatology has tended to be unimodal (Kendell 1970, Ni Bhrol-chain 1979).

Genetics of affective disorders

Kraepelin (1921) noted about manic-depressive insanity that a "here­dity taint" could be demonstrated in the majority of cases.

Two main approaches can be employed in investigating the importance of the contribution of genetic factors to psychopathology. 1) Clinical studies which consist of three different methods - family studies, twin studies and adoptee studies. 2) Biological studies which can be divi­ded into three categories - association studies, linkage studies and chromosome studies (Perris C. 1976).

This review is focused on clinical studies. Most genetic studies con­cerning affective disorders have studied "endogenous" or "primary" or "major" or "psychotic" affective disorders. In older studies calcu­lation of the morbidity risk for mano-depressive psychosis (MDP) among the relatives of patients with MDP disorders have shown very wide variations. For parents the morbidity risk in different studies has been 3.4 % - - 23.4 % (Slater 1938, Strömgren 1938, Sjögren 1948, Stenstedt 1952). For siblings the range was 2.7 % - - 22.7 % (S jögren 1948, Kallmann 1952). For children the range was 6.0 % — 24.1 % (Stenstedt 1952, Luxemburger 1942). Overall the means are above the

18

median for the general population (0.7 %) (Perris C. 1976).

Leonhard (1957-1969) divided the affective psychoses into bipolar and unipolar types. The diagnostic criteria based on this classification have varied in the different studies. Some have labelled patients with a depressive period and a period of hypomania as bipolar (Weissman 1978, Perris C. 1966) while others have required a major manic and depressive period before diagnosing bipolar affective disorder (Clo-ninger 1979). There have also been differences concerning the cri­teria for unipolar affective disorder. In some studies (Perris C. 1966) three separate periods were requried and in other studies only one episode of depression was required (Angst 1966). Both Angst and Perris found that bipolar subjects had first-degree relatives with bipolar affective disorder and uni-polar subjects had first-degree relatives with unipolar affective disorder. This was also confirmed in a later study (Perris C. 1982).

In more recent family studies the division into unipolar and bipolar is used also by others. The morbidity risk of affective illness amongst relatives of bipolar probands is higher than for relatives of unipolar probands. The risk for siblings of bipolar probands varies from 12.5 % — 42.0 % (Gershon 1977, Winokur 1970a). For parents of bipolar probands the risk varies between 34.2 % — 41.0 % (Winokur 1970a). Among siblings of unipolar probands the risk in different studies varies between 12.9 — 31.3 % (Smeraldi 1977, Mendlewicz 1975) and for parents 16.0 % — 22.9 % (Shopsin 1976, Winokur 1971a). Unipolar affective disorders (depression) is very common in the general popu­lation according to some recent epidemiologic research where more standardized diagnostic procedures have been used. Helgason (1979) found a 14.4 % risk of depression for females and an American study has found even higher life time prevalences women, 25.8 % and men 12.3 % (Weissman 1978). The risk of affective disorder amongst relati­ves of patients with unipolar illness thus seems not to be much higher than for the general population. However, the risk of affective illness amongst relatives of bipolar patients is higher than for the general population. Thus there is some support for a genetic transmission of

19

bipolar illness but not consistently for unipolar affective disorder.

Twin studies have consistently shown a higher concordance rate for manic-depressive psychosis or affective psychosis for monozygotic twins (0.50-0.92) (Harvald 1965, Kallmann 1952). The concordance for dizygotic twins varied from 0.03 to 0.24. In a later study (Bertelson 1977) it was found that concordance for bipolar illness in monozygotic twins was 0.79 and in dizygotics 0.19. For unipolar illness the con­cordance rates were lower: 0.54 in monozygotics and 0.24 in dizygo­tics.

The classification of depressive disorders has followed a different course in Europe than in the USA. In Europe unipolar depression has been kept separate from neurotic-reactive depression (Kielholz 1973, Angst 1981), while in the USA attention has been focused on primary and secondary depression (Robins 1972). In DSM-III (1980) dysthymic disorder is the same as depressive neurosis. The disorder is not expected to be severe or long enough to be classified as major de­pression but then it is stated that a duration in adults of 2 years is required.

A twin study of non-endogenous or reactive depression showed a low rate of concordance in monozygotic twins (2 of 16 pairs) and no con­cordance in dizygotic twins (0 of 14 pairs) (Shapiro 1970). In a family study of neurotic-reactive patients the morbidity risk for bipolar affective disorder was very low but the overall morbidity risk of affective disorders was surprisingly high (Perris C. 1982).

Adoption studies on affective disorders are few. Mendlewicz (1977) studied families of adult bipolar probands who had been adopted during infancy. The rate of affective disorder among biological pa­rents was much higher than among the adoptive parents or the bio­logical parents of normal adoptees or adoptees with poliomyelitis. Cadoret (1978) studied a small sample of parents and offspring se­parated early in life by adoption. The results of this study support the hypothesis of the involvement of a genetic factor in unipolar

20

depressive illness. The Danish adoption studies indicate that there is an excess of suicide among the biological relatives of adoptees with affective disorder or other psychiatric disorders leading to suicide (Kety 1979, Shulsinger 1979).

A stress diathesis model for affective disorders has been described (Pollitt 1972, Akiskal 1975) in which certain specific life events pre­cipitate affective disorder in vulnerable individuals. Genetic loading can be regarded as an important vulnerability factor for the occur­rence of depression. Pollit found in his study of depression that severe viral or bacterial infections or other physical stressors and, to a lesser extent, severe psychological stressors were associated with a lower morbidity risk in first-degree relatives. On the other hand H. Perris (1982b) could find no support for this hypothesis in her stu­dy. Patients with a positive family history of depressive disorders did not differ significantly from those with a negative family history as regards the occurrence of life-change events prior to onset of the illness.

Affective disorders and alcoholism

Alcoholism and affective disorders are reported to occur in the same individual and in different members of the family more often than is to be expected by chance (Woodruff 1973, Winokur 1971a, b). Patients with a family history of alcoholism were more likely to have an early onset of depression (before the age of 40). The term 'depressive spectrum disease1 was used for the group of depressed patients who had relatives who were alcoholics. Suicide is also reported more frequently among alcoholics, as among depressed patients (Beskow 1981, Dahlgren 1977, Medhus 1974), but there are some contradictory studies that do not show any higher rates of alcoholism among rela­tives of patients with affective disorders (Perris C. 1982, Stenstedt 1952). Angst (1966) found a higher morbidity risk of alcoholism among fathers of depressed patients and neurotic patients. An adoption study of the daughters of alcoholic biological fathers showed no more depressions than adopted controls (Goodwin 1977). Genetic marker

21

studies do not provide any consistent evidence of genetic overlap between alcoholism and depressive disorder or antisocial personality (Cloninger 1979b).

The conclusion to be drawn from the above-mentioned studies is that there is a genetic transmission of bipolar affective illness. The results concerning unipolar affective disorder are not consistent, but most studies support a genetic transmission. There is little support for thé hypothesis that alocholism and depression share a common genetic basis.

The transmission of alcohol abuse and criminality

It is commonly accepted that "alcoholism runs in families" (Jellinek 1945, Åmark 1951, Winokur 1970b). The discussion has centered on the problem of whether this inheritance is related more to inborn factors or is rather dependent on sociocultural and psychological factors. Both twin studies (Kaij 1960, Partanen 1966) and adoption studies (Bohman 1978b, Cadoret 1978, Goodwin 1973) have supported the hypothesis that genetic factors contribute to the manifestation of alcoholism.

In a study using the cross-fostering method (Cloninger 1981) two types of alcoholism were identified. Both had distinct genetic and environmental causes. Type 1 (milieu limited type) affected both men and women and those affected were susceptible to postnatal environ­mental influences - such as prolonged contact with the biological mother, late age at placement, several stays in hospital prior to placement and lower occupational status of adoptive parents. Type 2 alcoholism only affected males. The subjects in this group were less influenced by postnatal factors but the susceptibility was highly hereditary from biological fathers. Inheritance from biological mothers (Bohman 1981) was also important in the development for alcoholism in women.

22

The concept of alcohol abuse related to criminality has not always been taken into account when transmission of criminality is studied (Hutchings 1975). In one study (Bohman 1982) criminality was sepa­rated from alcohol abuse in the analysis of adopted males. Most vio­lent crimes were connected with alcohol abuse. Non-alcoholic criminals were characterized by the commission of a small number of petty offences. This group of petty criminals had an excess of petty cri­minality among their biological fathers. The number of unstable pre-adoptive placements was high in the petty criminality group. Low socio-economic status was associated with alcohol-related criminality. Either congenital predisposition or postnatal experiences were shown to be sufficient for the development of criminality, but specific com­binations of genetic and environmental antecedents modified the addi­tive effects of the individual risk factors. There was no genetic overlap between predisposition to petty criminality and alcoholism (Cloninger 1982). The congenital predisposition to criminality had to be more severe for a woman to be affected. Postnatal factors were more important for the development of criminality than for alcoholism in women (Sigvardsson 1982b).

Somatoform disorders and illness behaviour

The distinction between somatic and psychiatric disorders is sometimes hard to make. Pain experience for example is influenced by both an organic pain generator and by personality characteristics or environ­mental factors (Mohammed 1978). There is a group of patients for whom no sure evidence of an organic cause can be found and in whom all attempts to reveal the cause of the pain fall. Patients in this group tend to show evidence of primary affective or personality disorders, including hysteria, conversion reactions, anxiety and hypochondriasis (Pilowsky 1976). Patients with psychiatric disorders often have complaints that may suggest somatic diseases (Hagnell 1966).

23

Hysteria

The ancient Greeks considered hysteric phenomena a disorder affect­ing women attributable to abnormal movements of the uterus (hystera). The term and diagnosis of hysteria thus has an old tradition, but there has been much confusion and disagreement about it. Briquet (1859) in his work described women with "extreme sensitivity of the nervous system". As a result of this his patients manifested a variety of pains involving almost every area of the body and they also had attacks of acute anxiety and sensory-motor paralyses. Somewhat later Charcot began to study hysteria and tried to understand the causes of the syndrome. From this the whole psychodynamic psychiatry movement emerged. Pupils of Charcot, e.g. Janet (1931), contributed to the theoretical formulations concerning hysteria. Attention slowly shifted away from clinical diagnostic categories to the psychodynamic mechanisms that underlay symptoms and that provided at least a partial causal explanation. This influenced the terms used so that hysteria became conversion hysteria. Guze (1963) on the other hand defined "hysteria" as a syndrome which starts early in life, occurring mainly in females and which shows recurrent symptoms in many differ­ent organ systems. This syndrome exhibits a certain amount of stabili­ty and is valid at follow-up (Perley 1962, Martin 1979). The term "hysteria" thus has many different meanings and this has been con­fusing and has made it impossible to use the term in scientific or clinical contexts. According to DSM-III (1980) the syndrome defined by Perley & Guze (Briquet's syndrome) is called somatization dis­order. This is a chronic syndrome of recurrent symptoms in many different organ systems that begins before the age of 30 and is associated with psychiatric distress but no physical disorder. The disorder affects women but rare cases in men have been reported. The utility of a distinction between isolated conversion symptoms and Briquet's syndrome has been demonstrated by follow-up studies (Martin 1979).

Family studies show that Briquet's syndrome and sociopathy cluster in the same families instead of segregating as independent traits (Clo-

24

ninger 1975). Women with a definiteci Briquet's syndrome have an excess of relatives with multiple somatic complaints who often do not meet the full criteria for the diagnosis (Woerner 1968). There is also some support for this view provided in 2 adoption studies (Cadoret 1976, Crowe 1974). There were however too few cases in these stu­dies to permit any firm conclusions.

Hypochondriasis

Gillespie (1928) proposed a strict, clinical definition of this disorder, which is essentially the same as hypochondriasis according to DSM-III. Hypochondriasis requires preoccupation with an unrealistic fear or conviction of having a serious disease and impaired social or occu­pational function despite medical evaluation and reassurance. This has been rejected by many investigators, e.g. Kenyon (1976) who suggest­ed that the term should only be used as a descriptive adjective when there is a morbid preoccupation with health or body. "Symptom choice is affected by sociocultural and other variables". Depressive mood is a frequent accompaniment and he suggested that in many cases the condition is primarily an affective one.

Masked dep re s sion

Masked depression is a condition which manifests itself chiefly at a somatic level according to Kielholz (1972). Lopez Ibor (1968) pointed out that the following features are present: phasic course, diurnal fluctuation, mild inhibition, shift in mood, tiredness, sleep disturban­ces, loss of concentration and anxious sense of failure. This term should only be used in connection with the diagnosis of affective disorders. Even in ordinary depressive syndrome pain is a common symptom (von Knorring 1975).

Registered_sick-leave and its^ relation to psychiatric and somatic _ disability

In earlier days the most common measure of health or sickness was

25

the mortality figures (Nord-Larsen 1979). Interest has now become focused on morbidity as well as mortality. Health can be measured in different ways. It can be "freedom from illness" or defined as positive health - the best known definition of this kind is that given by WHO (1958). "Health means more than freedom from disease, freedom from pain, freedom from untimely death. It means optimum physical, mental and social efficiency and well-being". But indexes of negative health may still be useful and there is no need for measures to subdivide positive health into too many aspects (Brorson 1977). There is a choice of several available methods for defining the presence or absence of disease such as the use of interviews, rating scales and symptom-checklists. Health indexes, a measure which summarizes data from two or more components and which is supposed to reflect the health status of an individual or a defined group, can then be used.

In Sweden the sick-listing system can be regarded as a health-index. All inhabitants over 16 years of age have been covered by compulsory health insurance since 1955. Anyone who works in the home, is a student, or has a minimal earning level is financially compensated for periods of illness. Sick-listing is not an objective measurement. Sick-leave can be taken at a person's own discretion for 7 days but a doctor's note is required from the 8th day onwards. Many factors influence the taking of sick-leave, personality, social factors and work conditions are as important as the symptomatology (Nyström 1979b).

Sick-leave is reported to be higher among both male and female psy­chiatric patients (Horsman 1970). More frequent sick-leave is also reported among alcoholics (Pell 1970, Berglin 1974). In two studies of female alcoholics (Medhus 1974, Dahlgren 1979) the disability rate increased markedly in the years before registration by the Temperan­ce Board or in the years after the first treatment period in a psy­chiatric department for alcoholism. It has also been observed that alcoholics are treated more often for psychiatric complaints and mus-culo-skeletal complaints like low-back-pains (Westrin 1973, Larsson 1976). As Dahlgren (1981) has pointed out there have been secular

26

changes in female alcohol consumption. Women born during the 20's and 30!s had a late onset of diagnosed alcoholism, and the latency-time from the regular use of alcohol to the onset of alcoholism was 24 years. Women born during the 40's became alcoholics much earlier and their course was more fulminant being accompanied by psychiatric disorders, social problems and somatic complications. The latency times from regular use of alcohol to alcoholism was very short, about 5 years.

Criminals, who also have a higher rate of substance abuse, have a greater morbidity according to sick-insurance cards (Johansson 1981).

Repeated sick-leave (Ferguson 1972) has been associated with neuro­sis, alimentary, upper respiratory tract disorders and injury. Re­peated short-term sick-leave was associated with an increased fre­quency of psycho-social disturbances (Leijon 1982).

27

MATERIALS AND METHODS

The cohort of adoptees

The cohort consists of all persons born out of wedlock in Stockholm during the years 1917-1949, who were placed in non-related adoptive homes, i.e. 2 966 adoptees, 1 354 males and 1 612 females. 751 were born during the period 1917-1929 and 2 215 were born 1930-1949. Most of the adoptees (58 %) were placed in adoptive homes during their first year of life and 26 % were placed during their second year. A further 9 % were placed before the age of 3 years. Finally 7 % after 3 years of age (placement-age is not synonymous with age at which the adoption gained legal force). The mean age of placement for the adoptees born 1917-1929 was 1.5 years and the mean age of placement for those born 1930-1949 was 12 months. The adoptees placed after the age of 3, however, were excluded from the analyses. In Paper I the cohort consisted of 2 323 persons, but it was later discovered that some of the adoptees had been placed with relatives. 494 (16.7 %) biological fathers have not been identified in this cohort.

This was quite a wide variation in the type of care before placement in the adoptive home: 10.7 % were placed directly in the adoptive home from maternity hospital; 47.2 % were cared for by the biological mother only before placement; 5.9 % were in institutions only; 23.8 % were cared for both by the biological mother and in an institution and 12.3 % were cared for by the biological mother, in an institution and in a foster home. Sometimes the child was temporarily in care in different institutions or foster homes. 10.6 % had experienced 4 or more temporary placements - the highest number being 18.

Children born out of wedlock were chosen because at that time every child who was born of an unmarried mother was appointed a Child Welfare Officer and the case was recorded by the Child Welfare Of­fice. A record was also kept of cancellations of the appointment and the reason for it, e.g. if the mother married the father of the child or if the child was placed in an adoptive home and legally adopted.

28

Register data

Data about the adoptees and their biological and adoptive background were obtained from:

a) Official records such as the criminal register and the Temperance Board register. These contain information about criminal ver­dicts, registrations of offences against the Temperance Act, treatment in institutions for alcoholics and supervision because of alcohol abuse.

b) Records kept by the adoption agency and the Child Welfare Office. These contain information about preplacement history, age of the adoptee at placement, age, residence and occupational status of the adoptive parents at the time of placement.

c) Records kept by the National Health Insurance Offices - the Insurance-cards (Försäkringskort or so-called F-cards). These contain information about sick-leave periods, diagnoses or symp­toms. Information about hospital-care is also registered, but this kind of information is not completely accurate. Until the end of 1973 the chief complaint or diagnosis for each period of sick-leave was recorded manually. From 1974 these records were computerized and the diagnoses were no longer noted.

Firstly the F-cards for 2 118 adoptees born 1930-1949 were collected and were used for further studies of sick-leave pat­terns .

1 677 adoptees remained for the study of sick-leave patterns after exclusions on the grounds of late placement, uncertain paternity, emigration, early death or missing data. Each sick-ness-occasion was assigned one of 18 categories according to the chief complaint or the diagnosis if certified by a physician. The categories are: 1) Headache, 2) Pneumonia, 3) Astma, 4) Upper Respiratory Infections, 5) Cardiac Disorders, 6) Peptic Ulcer

29

Disease, 7) Liver and Gallbladder Disorders, 8) Dyspepsia, that is other trancient gastrointestinal complaints, 9) Uro-genital Disorders, 10) Backache, 11) Other Joint Pains, 12) Anemia, 13) Asthenia, that is non-specific complaints of fatigue or weakness, 14) Psychiatric Disability, 15) Pregnancy Complications, 16) Accidents occurring at Work, 17) Out-of-Work Accidents, 18) Other complaints or disorders.

The F-cards for 649 adoptees born 1917-1929 were also collected, but these were not used in the sick-leave study because these persons were quite old when the national insurance system started (27 to 39 years old). These F-cards were only analyzed manually to discover the adoptees with psychiatric disabilities.

An F-card was selected for a non-adopted control from the same local Insurance Office. The control was matched pair-wise for sex, age and residential area. Thus the non-adopted control may be regarded as a fairly representative sample of a Swedish adult population. There was good agreement between adopted probands and non-adopted controls concerning occupational status, accord­ing to a modified British classification (H.S.M.O., London 1951). As the chief complaints and diagnoses were not recorded after 1973 only the period 1955-1973 was considered for the sake of uniformity and completeness.

Records concerning adopted and non-adopted persons from psychiatric departments and mental hospitals all over Sweden. If the person had been on sick-leave for at least two weeks with a pscyhiatric diagnosis (no. 290-309 ICD-8) a record was sought from psychiatric hospitals and departments. The National Health Board has a register of in-patient treatment in psychiatric de­partments and psychiatric hospitals. This register was started in 1972 but it was used when searched for the adoptees and their parents involved in the depression-study. In this way some more records were found, mostly however persons already identified as ill. In all, 218 adoptees with a record were discovered.

30

Psychiatric diagnoses were made from these records using the following criteria: 1) Schizofrenie syndrome (Feighner criteria), 2) Other non-affective psychoses, 3) Non-affective neuroses (Feighner criteria), 4) Personality disorders (Feighner criteria 1972), 5) Organic cerebral syndrome verified by X-ray or EEG, 6) Abuse of alcohol or drugs defined as (i) at least one year of regular abuse, (ii) the substance abuse was the major problem for the patient or society, (iii) criteria of Feighner for probable abuse were fulfilled, 7) Unipolar affective disorder (C. Perris criteria 1966), 8) Bipolar affective disorder (C. Perris criteria 1966), 9) Cycloid psychosis (C. Perris criteria 1974), 10) De­pressive syndrome NUD (Major depression according to DSM-III), 11) Neurotic reactive depression (d'Elia criteria 1974), 12) Un­classified .

Reliability of psychiatric diagnoses

Initially 96 randomly selected records concerning both adopted and non-adopted persons were evaluated by 4 independent, trained psy­chiatrists (C. Perris, L. von Knorring, M. Bohman, A-L. von Knor-ring). The above-mentioned diagnostic categories were used. As the different diagnoses were not similarly weighted all the different affec­tive disorder diagnoses were put together. Schizofrenie syndrome and other non-affective psychosis were also put together. There were 7 diagnostic groups to consider for the reliability-test. Totally all four psychiatrists agreed about the diagnosis in 61.5 % of cases (from the classifiable records they agreed in 72 % of cases) and three of four agreed in 77 % (in 88 % for the classifiable records). The best agree­ment was for alcohol and drug abuse all four raters agreed about 81 % of these cases.

In the depression study firstly 62 records concerning adopted persons were selected. In this sample all the above-mentioned diagnostic categories were taken into account for reliability calculations. All four raters diagnosed the records independently, 56 persons remained in the affective disorder group and 59 persons were assigned to the

31

substance abuse group. For the affective cases the rates of agreement amongst the psychiatrists were as follow: all four in agreement 33.9 % (19 out of 56 cases); at least 3 out of 4 in agreement 62.5 %; at least 2 out of 4 in agreement 85.7 %. The final diagnoses were made after consensus.

The parents' records were blindly rated by two of the psychiatrists (A-L. von Knorring and L. von Knorring). Here there was an 80 % agreement about the diagnoses (k = 0.64, Z = 12.8 p < 0.001). For alcohol abuse there was 100 % agreement. Not all the adoptees with a diagnosis of alcohol and/or substance abuse in their psychiatric records were included in the Temperance Board register. Only 29 % of the females and 58 % of the males had one or more registrations.

No of registrations with the Temperance Board

Males Females

0 1-4 >4 0 1-4 >4

Psychiatric diagnosis 16 of alcohol abuse and/ or drug abuse (42%) males N = 38 females N = 21

Psychiatric diagnosis 3 of primary affective disorder and secondary (75%) alcohol abuse males N = 4 females N = 4

Psychiatric diagnosis 4 of affective disorder only (67%) males N = 6 females N = 42

5 17 15 6

(13%) (45%) (71%) (29%) -

1 - 3 1 -

(25%) - (75%) (25%) -

2 - 40 2 -

(33%) - (95%) (5%)

Table 1. Cross-tabulation of Temperance Board registration and psy­chiatric registration of alcohol abuse.

32

In this group of 59 adoptees with a psychiatric diagnosis of substance abuse 8-4 men and 4 women - had died of complications due to alcoholism (such as cirrhosis of the liver, brain injuries or acute intoxication). Only two of these men and one woman appeared in the Temperance Board records.

Life events

Life events precipitating a depressive episode were rated from the medical records. It was assumed that stressful events were reported in the records. Most of these records contain a satisfactory amount of information. The explicit criteria described by Pollitt (1972), H. Perris (1982) and Samuelsson (1982) were used. Only undesirable, external ("independent") events occurring within 3 months of onset of the first depressive symptoms in a treated episode were accepted. The specific events are both somatic (e.g. surgery) and psychosocial (e.g. parental death, divorce) (see Appendix I).

Statistical analyses

Chi-square tests with Yates' correction were used to evaluate differen­ces between frequency distributions. Pearson's was used to test hypotheses of a correlation (Colton 1974). In Study No VI the propor­tion of parents with psychiatric disorders was computed in terms of life-time risks. No age adjustment was required because tests were carried out comparing age-matched controls. Comparisons of the proportion of sick parents in the matched pairs were made using McNemar's chi-square (Cloninger 1979a).

Several series of discriminant function analyses were carried out with the help of SPSS programmes (Nie 1975). For more detailed informa­tion see the separate papers Nos II, III, IV. Discriminant function analyses were used to predict psychiatric disorders on the basis of the placement variables and to identify sick-leave variables that distinguish subgroups of adoptees from one another.

33

Admixture analyses was used to analyse the distribution of a discri­minant function derived to distinguish subgroups in a larger popula­tion e.g. as in this study to distinguish female somatizers from other women. If means differ widely even small subgroups can be detected in samples of a practical size. The admixture analysis was carried out using the Fortran program SKUMIX (MacLean 1976). For further information see Paper No IV.

To analyse the effects of the various combinations of categorial back­ground classifications (congenital and postnatal) on the observed personal status FUNCAT analysis using linear logistic models was carried out (Helwig 1979). For further information see Paper No V.

34

RESULTS

Psychiatric illness among adoptees (Papers I and III)

A higher frequency of psychiatric illness, defined as being on sick-leave for more than 2 weeks with a psychiatric diagnosis, was found among the adoptees than in matched non-adopted controls (see Paper I Table 2). The number was greater for the adopted females than for the adopted males. The adopted women also had more frequent and longer periods of psychiatric sicke-leave (see Paper III Table 1). For the adopted men, however, the frequency of sick-leave due to psy­chiatric complaints was no greater, but there was a tendency for it to last longer (see Paper III Table 3).

It was found from psychiatric records that the over-representation in the male group was due to substance abuse, personality disorders and neurotic disorders. For the female group there was also an over-representation of substance abuse and personality disorders (see Paper I Table 3). The number of adopted women who had been on sick-leave due to psychiatric complaints but had no hospital records from psychiatric facilities (see Paper I Table 4) was significantly larger than that for non-adopted controls.

Somatic illness among adoptees (Paper III)

The adopted women had an increased frequency of physical com­plaints, but they had no increased duration (see Paper III Table 1). The non-adopted controls had fewer sick-leave periods but these occasions tended to be longer in duration. In contrast the adopted women more often had brief but frequent sick-leaves each year. More female adoptees than controls had 2 or more sick-leave periods per year (16.8 % versus 11.5 %). The overall excess of sick-leave amongst female adoptees was mostly due to complaints for transient problems that did not require medical certification i.e. sick-leave lasting less than 7 days (see Paper III Table 2). Upper respiratory infections and abdominal complaints were more common reasons for sick-leave among

35

adoptees than non-adoptees. The adopted women had more than three times as many children as age-matched non-adopted controls, so they had more sick-leave for pregnancy complications per year, but not

per child.

The adopted men did not have any excess of sick-leave complaints nor were there any differences within individual categories of somatic complaints (see Paper III Table 3).

The relation of somatic and psychiatric illness (Papers III and IV)

Among the 859 adopted women 27.4 % and among the 818 adopted men 17.7 % had one or more periods of sick-leave, the chief complaint or diagnosis being psychiatric. Adoptees with and without psychiatric disability were compared as regards the pattern of somatic sick-leave. Possible predictors of psychiatric disability were frequency and dura­tion in each of the 16 specific somatic categories and the overall measures of somatic frequency, duration and diversity (the ratio of the number of categories to the total number of occasions). Another variable for somatic frequency was whether or not the subject was a somatizer (having 2 or more occasions of sick-leave due to somatic complaints per year). The significant discriminating somatic variables can be seen in Paper III Table 4 (for women) and Paper III Table 5 (for men). Limited diversity of complaints, long duration of abdominal and upper respiratory and urogenital complaints, high frequency of accidents outside work and complaints of fatigue and weakness were associated with psychiatric disability in both men and women. The 11 variables (shown in Paper III Table 4) accounted for 20.32 % of the variability in risks for psychiatric disability in women, and the 16 variables (in Paper III Table 5) accounted for 22.8 % of the variance in psychiatric disability in men. Being a somatizer accounted for more than half of the explained variance in women and about one-third of the explained variance in men. Overall 36.2 % of adopted women with psychiatric registrations were somatizers compared to 9.5 % of other adopted women. 141 (16.4 %) adopted women were either registered or predicted from the statistical analysis to have psychiatric disability.

36

The incidence of somatizers among women with registered psychiatric disability was 36 % while the incidence in women with predicted or registered psychiatric disability was 59.6 %. The incidence of somati­zers in women with no registered or predicted psychiatric disability was 0.4 % (see Paper III Table 6). In other words 98.6 % of the somatizers were either registered for psychiatric disability (59.0 %) or were predicted to have psychiatric disability (39.6 %).

In average somatizers were 1.67 years younger than the others. This difference accounted for less than 2 % of the variability in risk for somatizers although it was significant. The variables shown in Paper III Table 7 were used as predictors in a discriminant analysis to distinguish the adopted somatizers from the non-adopted women. These variables were then used as predictors in another discriminant analysis to distinguish the adopted somatizers from other adopted women (see Paper III Table 8).

The 859 adopted women were split into 2 samples by stratified random sampling of the 144 somatizers and 715 others. A discriminant function was derived to distinguish between somatizers and non-somatizers in one sample (the criterion sample), and was applied to the other sample (the replication sample). This function accounted for 68 % of the variability in the somatizer status. The criterion sample had 98.8 % accuracy of classification and the replication sample had 97.4 % accuracy of classification (see Paper III Table 9). In particular a high frequency of disability from complaints in the head, back and the abdomen were associated with somatizer status (see Paper III Fig. 1).

The distribution of the discriminant function scores for somatization in both the criterion sample and the replication sample suggests that both the samples were trimodal, with one antimode separating the somatizers and non-somatizers and a second antimode separating the somatizers into 2 groups (see Paper IV Fig. 2). Admixture analysis of the distribution scores in the replication sample showed that the two components accounted for the observed distribution much better than did one distribution, and furthermore three components described the

37

data better than two did (see Paper IV Table 1). Skewness was significant regardless of whether the number of distributions was one, two or three. The extreme group of somatizers (type 1) did not overlap at all with non-somatizers. The intermediate component (type 2) overlapped slightly with each extreme group (see Paper IV Fig. 1 and Table III). Type 1 somatizers had a high total frequency but limited diversity of complaints and their sick-leaves were consistently for abdominal and back problems. Type 2 somatizers took sick-leave less often but for more diverse complaints (see Paper IV Fig. 3). Type 1 somatizers had the highest occurrence of psychiatric com­plaints whereas type 2 somatizers had the lowest rate of psychiatric disability relative to total frequency of sick-leave (see Paper IV Table 4). Backache was an infrequent complaint for type 2 and common for type 1. The frequency and duration of abdominal problems were greater in type 1 somatizers (see Paper IV Table 2).

Divorce and low occupational status were more common in both types of somatizers and they had more pregnancy complications per child. They had fewer children than others, age into account (see Paper IV Tables 4 and 5). Type 2 somatizers were intermediate to the other groups for all the social and demographic characteristics measured. Somatization was also associated with alcohol abuse and criminality (Paper IV Table 6). 29.7 % of type 1 and 10.8 % of type 2 somatizers had been registered for alcohol abuse and/or criminality compared to 2.6 % of other adopted women.

Early childhood experiences and psychiatric disorders (Paper II)

Psychiatric illness amongst adoptees, defined as in- or out-patient treatment at a psychiatric department, had no specific association with any special kind of preplacement care before adoption. Nor was there any association between length of care and psychiatric illness. A higher divorce rate was found among the adoptees who had been in care in institutions only.

38

There was no linear relationship between age at placement in an adoptive home and the incidence of psychiatric disorder i.e. psy­chiatric disorders were no more frequent among those who were placed late. Concerning the hypothesis of critical periods the cohort was classified into four subgroups according to age at placement 1) 0-5 months 2) 6-11 months 3) 12-17 months 4) 18 months or later. There was only a slight increase in the number of psychiatric dis­orders among those placed at 6-17 months. Those adoptees with a diagnosis of reactive-neurotic depression were placed at the age of 6-11 months inc. significantly more frequently than expected (see Paper II Fig. 1).

There was no correlation between length of institutional care and psychiatric disorders, or number of placements before adoption. Series of multivariate analyses in order to predict psychiatric dis­orders from early life experiences showed that the inter-groups (psychiatric illness or not) differences were small compared to intra-group variance. Thus there was a considerable overlap between the groups.

Postnatal childhood experiences and somatization {Paper V)

Postnatal variables that distinguished both groups of somatizers and the non-somatizers were occupational status of the adoptive father, age at placement, contact with biological mother lasting more than 6 months and alcohol abuse by the adoptive father (see Paper V Table 4) but these 4 variables only weakly distinguished the 3 groups. Alcohol abuse by the adoptive father was relatively common among type 2 somatizers.

Environmental influences on affective disorder and substance abuse (Paper VI)

The risk of psychiatric illness was greater in the adoptive parents of the cases with affective disorders and substance abuse compared to controls, but the difference was not statistically significant. When

39

parents were divided according to sex it was found that the excess was due to a higher frequency of psychiatric illness among adoptive fathers. (8.7 % versus 1.7 % b/c = 9/1 p = 0.01) (see Paper VI Table 4). Both males and females with affective disorders or substance abuse had an excess of sick adoptive fathers. There was no diag­nostic concordance in the parent-child pairs (see Paper VI Table 6). The excess of sick adoptive fathers was almost entirely due to affec­tive disorders. The temporal sequence of illness was evaluated in the 7 families in which the adoptive father was treated for affective dis­order and the adoptee was treated for substance abuse or depression, 4 of the 7 adoptees were treated before their parents were treated, 1 was treated the same year and 2 were treated afterwards.

Genetic inheritance of affective disorders and substance abuse (Paper VI)

The lifetime risk of psychiatric illness was almost the same in the biological parents of the cases with affective disorders and substance abuse as in their matched controls. Female cases had sick biological mothers more often than did their matched controls, but if the pro­bability of this difference was conservatively adjusted for the number of tests the excess of the sick biological mothers was marginally significant or at least a trend (p < 4 x 0.015 = 0.06). Other differen­ces in the biological parents were clearly insignificant (see Paper VI Table 3). There was no evidence of familial aggregation of biological parents and adoptees with the same diagnosis (see Paper VI Table 5). The excess of illness in biological mothers of female cases was not due to over-representation of any specific diagnostic category. The largest contribution was that due to an excess of substance abuse amongst the biological mothers of adoptees with affective disorders compared to the mothers of adoptees with no disorders (7.1 % versus 0.9 % p < 0.025) but this was not individually significant after correction for multiple tests. The possibility of etiological heterogeneity among the different subtypes of affective disorders was evaluated (see Paper VI Table 7). There was no significant heterogeneity among the biological parents of psychotic and non-psychotic depressives. There were

40

sufficent cases for the statistical calculations of non-psychotic de­pression (N = 40) and psychotic depression (N = 11) but not for bipolar or cycloid psychoses (N = 5).

Genetic inheritance of somatization disorder (Paper V)

The genetic background was classified by a discriminant analysis of an independent sample of biological parents of adopted men, as in earlier reports. The genetic background was divided into the 5 cate­gories that were most characteristic of men with type 1 (mild or severe) alcoholism, type 2 (moderate) alcoholism, petty criminality only or neither alcohol abuse nor criminality (low risk). There was a significant excess of type 2 somatizers amongst women with a genetic predisposition the same as that characteristic of men with petty crimi­nality, compared to daughters at low risk (see Paper V Table 1). There was also an excess of type 2 daughters with a genetic pre­disposition to type 2 (male limited) alcoholism compared to low risk daughters and to daughters with a type 1 (milieau limited) alcoholic background. The daughters with the same background as male limited alcoholism had an excess of type 2 somatization (see Paper V Table 2). The daughters of petty criminals had an excess of both crime only and type 2 somatization. Daughters with a background charac­teristic of men with milieau limited alcohol abuse had a significant excess of alcohol abuse themselves but no other disorders.

In order to distinguish the 2 types of somatizers and the non-somati­zers variables concerning the biological parents were identified. The 10 variables selected were arranged in decreasing value on the second discriminant function, which distinguished the 2 types of somatizers (see Paper V Table 3). They accounted for 4.5 % of the variability among the 3 groups. Somatizers were distinguished from non-somati-zers by the combination of criminality, alcohol abuse and low occupa­tional status among their biological parents. Type 1 somatizers were distinguished from other somatizers by biological fathers with teenage onset of non-property crimes and many registrations for untreated alcohol abuse. Type 2 somatizers were distinguished from other soma-

41

tizers by biological fathers with property crimes and treatment for alcoholism.

Life events related to genetic and/or social predisposition (Paper VI)

A precipitating life event occurring within 3 months prior to the onset of depression was found in 14 of the 51 cases. A variable life event pattern was found in another 14 cases. To test the diathesis-stress hypothesis the adoptees were subdivided according to the presence or absence of precipitating events and compared with the percentage of adoptees in each group who had at least one parent ill (see Paper VI Table 8). There were no significant differences among groups as regards illness in either biological or adoptive parents.

Gene-environment correlation and interaction of female somatization (Paper V)

To evaluate the extent of selective placement relevant to somatization the extent of correlation between biological parent and post-natal variables was tested with a stepwise discriminant analysis of the 3 groups of women. Only 1 variable of the 14 included was deleted - the number of alcohol abuse registrations of the biological father. Gene-environment correlation accounted only for 4 % of the variability.

The risk of type 1 somatization was determined for each of the four combinations of developmental background generated by classification of congenital and postnatal background into high and low risk groups (see Paper V Fig. 1). Congenital and postnatal backgrounds were classified independently based on the biological parent and adoptive placement variables (see Paper V Tables 3 and 4). The pattern of interaction between the dichotomous congenital and postnatal factors was evaluated using a linear logistic model. Significant additive con­tributions were made by both congenital and postnatal backgrounds, but no non-additive gene-environment interaction was detected.

42

The interaction of congenital and postnatal factors contributing to somatizing type 2 was evaluated (see Paper V Fig. 2). Significant additive contributions were made by both congenital and postnatal backgrounds. There was no significant evidence of non-additive interaction.

A cross-fostering analysis with the three classifications (type 1, type 2 somatization and neither) was made to evaluate the full pattern of interaction. The 9 possible combinations of congenital and postnatal backgrounds can be seen in Paper V Table 5. Here it was possible to reject an additive logistic model with no gene-environment interaction. The significant interaction was mainly due to the high risk of type 2 somatizing when both congenital and postnatal backgrounds were most characteristic of the other type of somatizing and when a woman with type 2 congenital background was reared in an adoptive home that was most characteristic of the other type of somatizer.

There was no significant overlap between the genetic predisposition to type 1 and 2 somatization (see Paper V Table 6), but postnatal in­fluences resulted in highly variable risks of type 2 somatization regardless of genetic background. The observed risks varied about twice as much in daughters with a low risk congenital background (12.5 % to 22.4 %) or type 2 congenital background (21.6 % to 34.9 %) and nearly six-fold in daughters with a type 1 congenital background (5.0 % to 28.3 %) (see Paper V Table 5).

43

SOURCES OF ERROR

Adoption studies are a powerful tool for evaluating the environmental influences as distinct from genetic variables. However, this method also has its weaknesses. Most adoptees are told at an early age that they are adopted. According to one Swedish study (Bohman 1970) 81 % of a population of adoptees were told before the age of 7 years. This cohort of adoptees were younger (born 1955-57) than the adop­tees in this study. Earlier the adoptive parents were not recommended to tell their children that they were adopted. It may be assumed that the adoptees born 1917-1949 are not so often aware of the fact that they are adopted. There is very little information about the contact between the adoptees and their biological parents. Bohman (1973) found that 5 % of the adoptive parents had some personal contact with the biological parents. In these cases the contact mostly is between the adoptee and his family and the biological mother. Knowing about and being in contact with the biological parents prevents the distinct separation of the genetic and environmental backgrounds. This may influence the results in such a way as to make the genetic factors look more important.

However, there can also be a bias in the opposite direction as regards psychopathological disorders. The biological parents have more social and psychiatric problems (Bohman 1970) such as alcoholism and crimi­nality. In a study of married mothers who had their children adopted it was found that these mothers had significant deviation on the clinical scales of the MMPI (Horn 1975). There was a group of women within this group who deviated extremely. The two scales with the highest elevation were the psychopathia and the schizofrenia scales. This implies that there is a bias against a genetic hypothesis. There will be a higher risk of psychopathology in the offspring, which will reduce the differences between index and control groups especially in studies about sociopathy (criminality) and schizofrenia. This seems to be less important for depressions as there is no deviation on the depression scale for mother that had their child adopted. The effect of this bias would be particularly great when the traits of interest

44

involve the variables of age at onset, incomplete penetrance, and polygenes. On the other hand this kind of bias is reduced by using the adoptees' families method instead of the adoptees study method.

Adoption studies also have other limitations. The correlation of a trait between the biological mother and the adoptee is not purely genetic. Intrauterine factors also influence and may explain, at least to some extent, this kind of correlation. The correlation between the biological father and the adoptee is more probably genetic, but in this case there are other difficulties. Blood-typing disproves paternity in about 20 % of the fathers who deny paternity of illegitimate children (Social­styrelsen 1976). Thus the estimated number of false fathers in this material is 20 % at the highest, and most probably much lower as the majority of fathers in this study accepted paternity. As the biological mothers tend to name as the father a better adjusted male if several are possible, this kind of error will influence the results so that a genetic hypothesis cannot be sustained. On the other hand children of unidentified fathers do have the same adjustment as children of identified fathers (Bohman 1980). This implies that having an un­identified father does not bias the results.

The information in this study is based on various registers, records etc. Swedish registers are considered unusually complete. As Rydelius (1981) has pointed out the utilization of a register for information has one obvious advantage, it is insensitive to any bias on the part of the researcher. The interviewer or researcher is unable to influence the registration. On the other hand registers also have several limita­tions as sources of information. Temperance Board registrations reflect social class and personality. Individuals from the lower class run a greater risk of being registered (Bjurulf 1971). It is known from epidemiologic studies (Öjesjö 1980) that 70 % of alcoholics are registered either by the Temperance Board or a psychiatric depart­ment. Non-registered alcoholics were fairly representative of the total population of registered alcohol abusers (Kaij 1970). The information found in the Criminal Records is probably less sensitive to variations in social status than that in social and medical registers (Johansson

45

1981). On the other hand these data do not measure criminal beha­viour that is not known to society.

Hospital registers have been thought to be less sensitive to social factors. Smedby (1974) considered after comprehensive empirical research that the state of health of an individual determines whether he becomes a heavy user of medical care. E. Johansson (1981) has worked with the information from National Insurance Cards concerning criminals and has concluded that it is not possible to measure the "purely medical" but rather the social aspects of illness. To a varying degree this is true also in cases of hospitalization, apart from a few instances with absolute indications.

Personality also influences the illness behaviour. Bond & Pearson (1969) found that personality not only influenced the experience of pain but it also influenced the ability to communicate about pain.

Information from hospital records may have some disadvantages. The quality of information varies, but some basic information is available. According to a reliability study (Hostetter 1983) full medical records and interview schedules showed a high concordance of diagnosis. There are also some advantages in using information from records: the patient himself may have difficulty in remembering retrospectively and is more likely to provide false information. Data about life events may not be consistently reported in the records but negative and severe life events are supposed to be reported to a greater extent. In this case the bias is not in any specific direction.

Hospital records about parents were found less often than expected from epidemiological data. One explanation is that we only searched the records of psychiatric departments and hospitals and thus missed any information given when the parent consulted a general practi­tioner or a private doctor. However, the number of cases missed for this reason is likely to be the same for both proband and control group. The differences found thus seem reliable.

46

DISCUSSION

Somatic complaints and psychopathology

It was hypothesized that adopted persons are overall more vulnerable to psychiatric illness and somatic complaints than the general popula­tion. The results in this study support the hypothesis. This is in line with the results of earlier research which has found that the outcome of adoption is good on the whole, but that the risk of mal­adjustment is greater.

There was a higher incidence of alcohol abuse and personality dis­orders among both adopted men and women but the former also had higher incidence of neurosis which the latter did not. There was no överrepresentation of affective disorders in either sex. Female adop­tees with psychiatric disability saw a psychiatrist less often than both male adoptees and controls. It is conceivable that they consulted a general practitioner instead.

Adopted women had an increased frequency of somatic complaints, but adopted men did not. The adopted women had an excess of sick-leave mostly due to transient complaints lasting less than 7 days, mainly upper respiratory infections and abdominal complaints.

Further analyses were carried out in order to discover any special patterns of sick-leave. Somatic sick-leave on more than 2 occasions per year was correlated to psychiatric complaints. The group of persons who had this kind of sick-leave pattern were called soma-tizers, because of the correlation to psychiatric disability. There were fewer male somatizers than female. The observation time in this study lasted 8-18 years and therefore the somatizers had at least totally 16 sick-leaves. At this point it is not possible to define for how long time of observation is needed in clinical work for a classification of this syndrome from similar patterns of sick-leave. Female somatizers were younger than non-somatizers, which may reflect an association with alcohol abuse.

47

During the last few decades (Dahlgren 1981) women have started to drink at a much younger age than in earlier generations. Adopted women had an excess of somatization compared to non-adopted con­trols. The sick-leaves were mostly because of backache and abdominal distress. The symptoms were often self-limited and transient, but abdominal disorder and psychiatric disability were of longer duration. The somatizers could be divided into two separate types, which differ­ed both as regards the extent of their disability and the pattern of their psychatric and somatic complaints.

Type 1 (high frequency) somatizers had very frequent sick-leaves with psychiatric, abdominal and back complaints. They also had a high frequency of registrations for alcoholism (19 %) and crime (10 %), 10 times greater than that for non-somâtizers. It has been observed that alcoholics have an excess of medical treatment for psychiatric disability and musculoskeletal disabilities like low-back ailments (Dahl­gren 1979, Medhus 1974, Larsson 1976, Westrin 1973). It would be useful to know how often somatization precedes alcohol abuse and vice versa.

Type 2 (diversiform) somatizers were more common than type 1 and had more diverse somatic complaints in general, and a lower propor­tion of all their sick-leaves involved chiefly psychiatric complaints or backache. Alcohol abuse and criminality were not as common as among type 1 somatizers, although they were four times more frequent than among non-somatizers.

Somatization of both types was associated with a high divorce rate, low occupational status and youth. Similar observations have been made by others (Leijon 1982). One explanation for the youth of the somatizers may be the secular trend of female alcoholics in Sweden. Younger female alcoholics have a more fulminant course with more medical and social complications than older female alcoholics (Dahlgren 1981). On the other hand patients with psychiatric disorders have more somatic symptoms when younger (Guze 1972).

48

The somatizers identified in this study bear some similarities to pa­tients suffering from clinical psychiatric syndromes described earlier. Over 150 years ago Benjamin Rush (1812) described hypochondriasis as follows: "The symptoms of this form of dearangement as they appear in the body are dyspepsia, congestiveness or diarrhoea..., strong venerai desires or absence of venerai desires, pains in the limbs..., a disposition to faint, wakefulness, indisposition to rise out of bed...". According to DSM III patients with hypochondriasis are preoccupied with different somatic signs and are convinced they have a serious disease. However little is known about the illness beha­viour, i.e. whether or not they consult a physician or if they turn to friends with their problems. Hypocondriasis is considered to be a chronic state (Kenyon 1964). Briquet's syndrome or somatization disorder according to DSM-III is also similar to the symptoms ex­hibited by the group of somatizers in this study. It is characterized by abdominal distress, back-pain and psychiatric disability combined with a chronic course. Somatizing could also be a manifestation of masked depression which according to Lopez Ibor (1969) is phasic in its course. The courses of "endogenous" depression (Nyström 1979a) and neurotic depression (d'Elia 1974) are similar in some respects to the course found for somatizers - transient, self-limited and recurrent over several years.

It is not possible however to draw any definite conclusions about the relationship between clincially defined syndromes and the more common groups of somatizers, defined according to epidemiological sick-leave register data. Therefore it would be necessary to study the group of somatizers clinically in order to obtain a more detailed description of their somatic and psychiatric status.

Environmental influences and psychiatric disability

Early negative experiences such as bereavement are found to increase the risk of psychatric illness later in life (Brown 1966, d'Elia 1971). Our contention was that institutional care during infancy, or unstable placement procedures were such negative experiences. It has been

49

suggested that 6-12 months is a critical age both as concerns the effect of separation (Spitz 1945, Schaffer 1968) and the establishment of social relationships. Disturbances during this critical period are supposed to increase risk of depression and divorce.

Placement variables did not explain the manifestation of psychatric disorders in general or the divorce rate. This is in line with the findings of an earlier study of adoptees where placement variables did not explain maladjustment or underachievement while in school (Boh­man 1979).

However, reactive neurotic depression was significantly correlated with placement at the age of 6-12 months. This could be a a chance effect. However a closer analysis showed that most cases (7 out of 10) in the reactive neurotic group who were placed at the age 6-12 months had been cared for by their biological mother or a foster mother until placement in the adoptive home. Thus there is some support - but weak - for the hypothesis that separation at the age of 6-12 months will affect later development.

Early object loss plays a central role in psychoanalytic theories con­cerning the origin of depression. It has been assumed that the ori­ginal trauma, which is caused by object-loss in early childhood, can be reactivated later by new losses (Mendelsson 1974). Our findings provide some support for this theory. The adoptees with reactive neurotic depression had the highest rate of divorce (57 % of those ever married) and thus have experienced object-losses. The evidence however is not convincing because there is only a weak statistical significance.

Earlier research on the influence of psychiatric illness in parents focused very much on the mother's influence on the child. However, in recent years interest in the father has increased (Earles 1976). A few studies have been carried out which have found that the father's mental health influences on the attitudes and treatment-seeking be­haviour of the children, especially the sons (Rydelius 1981, Nettel-

50

bladt 1981). In line with these studies, we found that children whose adoptive fathers had affective disorders were more likely than other adoptees to seek psychiatric treatment themselves as adults. This non-genetic influence occurred regardless of the sex or diagnosis of the adoptee, but is strongest in adopted sons with substance abuse.

Our data also suggest that the children did not simply imitate the illness of their adoptive parents because the children were often treated earlier than were their parents. Onset may depend on other variables such as feelings of self-insufficiency and attitudes about seeking help from psychiatrists. It has been suggested that cognitive styles of self-assessment may be learned in childhood and influence later self-esteem, self-reliance, non-specific vulnerability to stress etc. (Beck 1967, Raps 1982, Valliant 1982, Coopersmith 1967). Since such characteristics seem to be acquired at an early age parents have been vulnerable longer than their children and would seek treatment before their children. Adoptive parents are, however, selected for their ability to provide a stable environment for their children. There­fore most potential parents with an early onset of psychiatric dis­ability were not permitted to adopt children. Our data support the idea that the home environment, and not only imitation, is important as a determinant for disability and treatment for both depressive disorders and substance abuse. This has also been found in other studies (Brown 1978, Shepherd 1966).

Low social status of the adoptive father was one environmental factor which increased the risk for both types of somatization. Type 2 somatizers also had had longer contact with their biological mothers before placement in adoptive homes and alcohol abuse was more common in their adoptive fathers. Early separation from the biological mother and early placement was more characteristic for the type 1 somatizers.

The lower classes are reported to have a higher frequency of joint and back-problems (Carlsson 1979), and low social status is also related to an increased frequency of psychiatric and somatic illness (Jonsson 1967). Absenteeism, especially for short-term illness, is

51

highest among manual workers with a low income (Eriksen 1980). Our finding that somatizers had a lower status agrees with these earlier studies. Among the general population 60 % have the same social status as their parents (Jonsson 1964), the figure is even higher for the lowest group. The correlation to the low social group of the adoptive fathers may be the same as the correlation found in the general population. It is not possible to draw any conclusions from this study about whether this association to low social status between the adoptee and the adoptive father is environmental. Cross-fostering studies would be required for elucidation of this point.

Genetic influences on psychopathology

The female cases with depression and substance abuse more often than controls had biological mothers who had a psychiatric diagnosis although the figure was not statistically significant after corrections for multiple tests. Similarily other studies have observed a higher concordance in mother-daughter pairs than in other parent-child relationships (Stenstedt 1952, Smeraldi 1979). However, although there was a correlation concerning psychiatric illness in general, we observed no increased concordance between the adoptees and their parents for the same diagnosis. Genetic factors therefore exert only weak and non-specific influences on vulnerability to depression and substance abuse. This is similar to the findings for non-psychotic depressions in intact Swedish families (Perris C. 1982) and Danish twins with non-endogenous depression (Shapiro 1970).

There was no genetic overlap between affective disorders and sub­stance abuse. This is in line with the study of daughters of alcoholic fathers where no excess of depression was found unless they were raised by the father (Goodwin 1977). On the other hand substance abuse and other non-affective disorders accounted for most of the excess of psychiatric illness in the mothers of affective disorder patients. A study of the relationship of untreated alcohol abuse in biological parents to depressive disorders in children would provide a more thorough means of testing of the hypothesis that alcoholism and

52

depression share a common genetic basis in some families. At present however our results, together with Goodwin's findings, suggest that cultural inheritance is the major determinant in any familial aggre­gation of depression with alcoholism. This casts doubt on Winokur's "depressive spectrum" hypothesis, that depression and alcohol abuse are simply sex-modified expressions of the same genotype in some families.

We did not find any evidence to support the theory that psychotic and non-psychotic depressions are genetically distinct. None of the biological parents of psychotic depressives and none of the biological parents of bipolar cases had an affective psychosis. This is consistent with the discussion of Lewis (1938) and Kendell (1976) that there is no 'point of rarity' between psychotic and neurotic depressions. Akiskal (1978) in his follow-up of neurotic depressions found that this group was heterogenous. In this study we have used a multicategory classification based mainly on ICD-9. Patients with reactive neurotic depression according to our definition consisted of a heterogeneous group some of whom had an environmental precipitant and others a personality trait of general nervousness without any clear, severe precipitant other than a "break-down".

Reactive depressions are more common in younger people while the "endogenous" depressions increase with age, being most common over 50 (Post 1968). As the majority of our adoptees are below 50 we can assume that if we repeat the same study in 10 years time we will find more "endogenous" or "psychotic" depressions in the material.

Our results contrast to the findings of Mendlewicz (1977) who found a genetic correlation between 29 bipolar adoptees and their biological relatives in Belgium. Our cases consisted mostly of non-psychotic depressives which may explain the differing results. Our data indicate that there may be a minimal, genetic contribution to the familial aggregation of individuals with a specific diagnosis of primary major depression. The familial aggregation observed in earlier studies of major depression was probably due largely to sociocultural effects.

53

Our earlier findings on the inheritance of alcohol abuse indicated that the genetic influence on some forms of alcoholism (milieau limited) was weaker than on other forms (male limited) of alcohol abuse (Cloninger 1982). This may also be the case for depressive disorders. Support for this is also found in genetic research about personality traits as they are differentially heritable (Zuckerman 1979, Loehlin 1982).

According to our hypothesis somatization as shown in this study is genetically correlated to alcohol abuse and criminality. Four familial patterns have been distinguished. In one group of families the men are often petty criminals, but they seldom abuse alcohol, and the women are type 2 somatizers and/or petty criminals. In the second group of families the women are type 2 somatizers but they themselves are neither alcoholics nor criminals. Their male relatives often have type 2 alcoholism (male limited type) which can be associated with criminality, but seldom have criminal behaviour in the absence of alcohol abuse. In the third group of families the women are type 1 somatizers and their male relatives have a teenage onset of criminality with frequent registrations by the Temperance Board, but infrequent convictions for property crimes. About 30 % of the type 1 somatizers were also registered for alcohol abuse or criminal behaviour. As female alcoholism is more difficult to detect the frequency of alco­holism may be much higher, which indicates that the majority of type 1 somatizers are alcohol abusers. In the fourth group of families there was no excess of criminality or somatization, but men and women are at a high risk level for type 1 alcohol abuse (milieau limited). These findings are comparable with those of some earlier studies of somato­form disorders (Arconac 1963, Cloninger 1975, 1978) and imply that at least some of the somatizers would fit into this from clinical materials defined Briquet's syndrome.

Gene-environment interaction on psychopathology

According to the diathesis-stress hypothesis a greater genetic and/or social predisposition is required for the onset of depression without a precipitating life event. To test this hypothesis we subdivided the

54

adoptees according to the presence of precipitating events within 3 months prior the onset of depression and compared the percentage of adoptees in each group who had at least one parent ill. The rating of the life events was made from the information in the records using the criteria of Pollitt, H. Perris and Samuelsson. There were no significant differences among the groups as regards illness in either biological or adoptive parents. This is quite contradictory to the results of Pollitt (1972), but is in line with the findings of H. Perris (1982b). Our findings imply that "stressful" life events in adult life do not contribute to the genetic and social liability to depressive disorders. This casts doubt on the relevance of comparing clinical depressive disorders with untreated dysphoria in the general popu­lation. Judgements about the severity of putative stressors made by non-patient populations may be irrelevant to their presumed effect as précipitants of pathological depressive episodes. The coincidental occurrence of stressfull events may act as the non-specific triggers in the onset of illness. Genetic factors and childhood environment play a permissive role that is not modified by adult events. Later stressors may possibly determine onset at a specific time.

Both genetic and environmental factors predispose a subject to soma­toform disorders. The same genetic factors that predispose both men and women to petty criminality are also expressed as type 2 somati­zation in women.. An unskilled occupational status on the part of the adoptive father is also associated with an increased risk of both petty criminality amongst men and type 2 somatization amongst women. Alcohol abuse in adoptive father is associated with an increased risk of type 2 but not type 1 somatization. Low social status had a differ­ent pattern of interaction with the 2 types of genetic predisposition. The environmental factors that were most characteristic of type 2 somatizers (contact with biological mother and adoptive father's alcohol abuse) did not increase the risk of type 1 somatization. The environ­mental factors that were most characteristic of type 1 somatizers (early separation from biologal mother and early placement in adoptive home) did not make a significant additive contribution to the other type of somatizers. Nevertheless there was a non-additive interaction

55

between genetic and environmental determinants of type 2 but not type 1 somatization. Early placement which was characteristic of type 1 somatizers, might also reflect the lower status of biological mothers. This also occurs in type 2 somatizers in the presence of genetic susceptibility and low social status in the adoptive home. Genetic and environmental factors make additive contributions with no interaction to each somatoform disorder when they are considered separately. When they are considered jointly the interactions are not additive.

The variables regarding the biological parents that distinguished the 2 types of somatizers and the non-somatizers only accounted for 4.5 % of the variability. The information about environmental conditions is restricted to rather crude data. Important environmental variables such as emotional atmosphere or physiological factors in the adoptive family are not included in this study.

Type 2 somatization accounted for 12 - 18 % of the population of adopted women. When a syndrome is as common as this, it is expected to be more than one combination of genetic and environemntal factors leading to the final common pathway to the phenotype. Two types of predisposition, towards petty criminality and male limited alcoholism, that are genetically independent were each associated with diversiform somatization. The same genetic factors that lead to petty criminality and male limited alcoholism in men are expressed as diversiform somati­zation in women. In some families type 2 somatization is the counter­part to mild petty criminality among men and the severe female cases are criminals. This is the same pattern as observed for Briquet's syndrome and antisocial personality.

Low MAO activity is one factor that has been shown to be associated with sensation seeking, monotony avoidance, impulsiveness, criminality and alcohol abuse. These could possibly be factors that are shared by somatization, petty criminality and type 2 alcoholism (von Knorring 1978, Buchsbaum 1976, Zuckerman 1979, Fowler 1980, Coursey 1982). It would be valuable in further research to conduct clinical studies comparing objectively self-report measures of diability with clinical

56

evaluations and neuro-physiological measures or test. Using this strategy it would be possible to compare the sensitivity and spe­cificity of alternative assessment techniques.

The results of this study are in line with earlier findings about psychiatric * illness. Genetic predisposition is a prerequisite, but to develop psychiatric disorders familial and sociocultural factors are also necessary. The intrinsic ways in which genetic predisposition is mediated through enzymatic, biochemical and neuroendocrinological patterns to personality traits are yet to be discovered.

57

SUMMARY OF THE MAIN RESULTS

Adopted women, although adopted at quite an early age, seem to be more vulnerable than women in the general population. They have more psychiatric illness, due to substance abuse and personality disorders. The adopted women also had an excess of sick-leave mostly of short duration, with complaints concerning upper respiratory infection and abdominal symptoms. They also had a higher frequency of somatization, i.e. more than 2 sick-leaves per year because of somatic complaints, associated with nervous complaints.

Sick-leaves were not more frequent among adopted men, but sick-leaves for psychiatric reasons were longer. The psychiatric disability was mainly due to substance abuse, personality disorders and neuro­sis.

Early negative experiences, such as unstable preadoptive placements did not influence subsequent psychiatric illness. There was some indication of the existence of a critical period for separation, i.e. between 6-12 months of age, because placement in the adoptive home at that age was associated with reactive neurotic depression in adult life.

Affective disorders in adoptive fathers were significantly associated with treated depression or substance abuse in adoptees.

There is no significant evidence of genetic influence on occupational disability and treatment for depression or substance abuse, but there was a trend towards a non-specific vulnerability in mother-daughter pairs. No genetic overlap was found between affective disorders and substance abuse. There was no evidence that psychotic and non-psychotic depressions are genetically distinct.

There was no support for the diathesis-stress hypothesis, i.e. that a greater genetic and/or social predisposition is required to bring about the onset of depression without a precipitant life event.

58

Somatization was correlated with psychiatric disability. 99 % of the somatizers had either registered or predicted psychiatric disability, but only 36 % of the women with psychiatric disability had somati­zation syndrome.

The somatizers could be separated into 2 distinct types. Type 1 (high frequency) somatizers had frequent psychiatric, abdominal and back-complaints. They also had a very high frequency of alcohol abuse and criminality. Type 2 (diversiform) somatizers had more diverse somatic complaints in general and a lower proportion of their sick-leaves involved psychiatric complaints. They also had a higher frequency of alcohol abuse than non-somatizers, but not as high as that among type 1 somatizers.

Somatization was genetically correlated to alcohol abuse and crimina­lity. High frequency somatizers had biological fathers with a teenage onset of criminality and frequent registrations with the Temperance Board. Diversiform somatizers had either the biological background of petty criminality or male-limited alcoholism.

Low social status on the part of adoptive fathers was associated with somatization in women. Diversiform somatizers had a longer contact with their biological mothers before adoption. Alcohol abuse was more common among the adoptive fathers of diversiform somatizers.

Both genetic and environmental factors predispose a person to somato­form disorders. There was a non-additive interaction between genetic and environmental determinants of diversiform but not high frequency somatization.

59

APPENDIX I. <3 months 3-6 months

A. SOMATIC STRESSORS

1. Severe infection (pneumonia, nephritis, syste­mic infection lasting at least one week or requiring hospitalisation).

2. Mild viral infection (influenza, mononucleosis etc, lasting 3 to 6 days).

3. Gross overwork (physical exhaustion due to excessive activity).

4. Weight loss due to purposal restrictive dieting (at least 5 kg in 3 months).

5. Use of drugs precipitating depression (reser-pine, corticosteroids and other drugs reported to precipitate depression but not alcohol or sedatives or other hormones like contracep­tives) .

6. Other severe physical illness (cancer, neuro­logical illness, cardiac infarction or physical damage).

SEVERE AND INDEPENDENT PSYCHOLOGICAL STRESSORS

1. Death of a household member, closest friend, first degree relative, grandchild, half-sib or spouse (including stillbirths and abortions after first trimester).

2. Serious physical illness in member of house­hold, first-degree relative, closest friend, grandchild, half-sib or spouse.

MODERATE AND INDEPENDENT PSYCHOLOGICAL STRESSORS

1. Unwanted change in financial status (self fired or unemployed more than one month, spouse fired or unemployed more than one month or other change leading to clearly lowered living standard).

2. Separation or divorce from spouse. 3. Involved with violence. 4. Met with a serious accident. 5. Major changes in home condition due to damage,

repair etc.

60

ACKNOWLEDGEMENTS

Many have participated in several ways in this work and I wish to express my sincere gratitude especially to:

My husband Lars for supporting me emotionally and for helping me with the classification of records.

Marika and Björn for being the best children in the world.

My tutor and supervisor professor Michael Bohman for introducing me into this field and always being most enthustiastic in discussions.

My co-author and instructor of psychiatric genetics professor Robert Cloninger for giving inspiration in this work.

My co-author Sören Sigvardsson Ph.D. for always supporting me and being very helpful with computer work.

Professor Carlo Perris for showing great interest and being very helpful in many ways during this work.

Professor Lars Jacobsson for his kind support.

Greta Lundgren for very skilful assisting and data-collection.

Gunilla Berglund for competent secreterial help and typing of manu­scripts .

Pat Shrimp ton B.A. for her revision of the English langugage.

The work has been supported in part by the Swedish Medical Re­search Council grant B 82-21X-03789-11, Research Grant from the Medical Faculty of the University in Umeå, and U.S. Public Health Service grants AA-03539, MH 31302.

61

REFERENCES

Abraham K. (1927): Notes on the psychoanalytic investigation and treatment of manic-depressive insanity and allied conditions. In: Selected papers on psychoanalyses. Hogarth Press. London.

Ainsworth M.D.S. (1973): The development of infant-mother attach­ment. In: Caldwell B.M. & Ricciuti H.N. (eds): Review of child development research. Vol. 3. University of Chicago Press. Chicago.

Akiskal H.S., McKinney W.T. (1975): Overview of recent research in depression: Integration of ten conceptual models into a comprehensive clinical frame. Arch. Gen. Psychiat. 32: 285-305.

Akiskal H. S., Bitar A. H., Putzantian V.R., Rosenthal T. L., Parks W., Walker W. (1978). The nosological status of neurotic depression. A prospective three- to four-year follow-up examination in light of the primary-secondary and unipolar-bipolar dichotomies. Arch. Gen. Psychiat. 35: 756-766.

American Psychiatric Association Committe on Nomenclature and Statis­tics: (1980): Diagnostic and statistic manual, ed. 3. American Psy­c h i a t r i c A s s o c i a t i o n . W a s h i n g t o n D C .

Angst J. (1966): Zur Ätiologie und Nosologie endogener depressiver Psychosen. Springer. Heidelberg-New York.

Angst J., Grigo H., Lanz M. (1981): Classification of depression. Acta Psychiat. Scand. Suppl. 290: 23-28.

Arconac O., Guze S.B. (1963): A family study of hysteria. New Engl. J. Med. 268: 239-242.

Beck A.T., Seith B.B., Tuthill R.W. (1963): Childhood bereavement and adult depression. Arch. Gen. Psychiat. 9:295-302.

62

Beck A.T. (1967): Depression: clinical, experimental and theoretical aspects. Harper & Row. New York.

Berglin C-G., Rosengren E. (1974): Arbetsprestationer och pensione­ring hos 868 alkoholister. Läkartidningen 71: 3520-3521.

Berteisen A., Harvald B., Hauge M. (1977): A Danish twin study of manic-depressive disorders. Brit. J. Psychiat. 130: 330-351.

Beskow J. (1979): Suicide and mental disorder in Swedish men. Acta Psychiat. Scand. Suppl. 277.

Birchnell J. (1975): The personality characteristics of early bereaved psychiatric patients. Social Psychiat. 10: 97-103.

Birchnell J. (1978); Early parental death and the clinical scales of MMPI. Brit. J. Psychiat. 132: 574-579.

Bjurulf P., Sternby N. H . , Wistedt B. (1971): Definitions of alcoho­lism, relevance of liver disease and temperance board registrations in Sweden. Q. J. Stud. Ale. 32: 393-405.

Bohman M. (1970): Adopted children and their families. A follow-up study of adopted children, their background, environment and adjust­ment . Proprius. Stockholm.

Bohman M. (1971): A comparative study of adopted children, foster children and children in their biological environment born after un-desired pregnancies. Acta Paediat. Scand. Suppl. 221.

Bohman M. (1972): Unwanted children - a prognostic study. Child Adoption. 2: 13-25.

Bohman M. (1973): Adoptivbarn och deras familjer. En undersökning av adoptivbarn, deras bakgrund, miljö och anpassning. Almqvist & Wiksell. Stockholm.

63

Bohman M. (1978a): An eighteen-y ear, prospective, longitudinal study of adopted boys: In Anthony E. J. et al. (eds): The child in his family: vulnerable children. John Wiley. New York.

Bohman M. (1978b): Some genetic aspects of alcoholism and crimina­lity: a population of adoptees. Arch. Gen. Psychiat. 35: 269-276.

Bohman M., Sigvardsson S. (1979): Long-term effects of early institu­tional care: a prospective longitudinal study. J. Child Psychol. 20: 111-117.

Bohman M., Sigvardsson S. (1980): A prospective, longitudinal study of children registered for adoption. A 15-year follow-up. Acta Psy­chiat. Scand. 61: 339-355.

Bohman M., Sigvardsson S., Cloninger C.R. (1981): Maternal inheri­tance of alcohol abuse. Cross-fostering analysis of adopted women. Arch. Gen. Psychiat. 38: 965-969.

Bohman M., Cloninger C.R., Sigvardsson S., von Knorring A-L. (1981): Predisposition to petty criminality in Swedish adoptees. I. Genetic and environmental heterogeneity. Arch. Gen. Psychiat. 39: 1233-1241.

Bond M.R., Pearson I.B. (1969): Psychological aspects of pain in women with advanced cancer of the cervix. J. of Psychosom. Res. 13: 13-19.

Bowlby J. (1958): The nature of the child's tie to his mother. Int» J. Psychoanal. 39: 350-375.

Bowlby J. (1961): Childhood mourning and its implications for psy­chiatry. Am. J. Psychiat. 118: 481-498.

Bowlby J. (1969): Attachment and loss. Vol. 1. Attachment. Hogarth. London.

64

Bowlby J. (1973): Attachment and loss. Vol. 2. Separation, anxiety and anger. Hogarth. London.

Brill N.O., Liston E.H. (1966): Parental loss in adults with emotional disorders. Arch. Gen. Psychiat. 14: 307-314.

Brinich P.M., Brinich E.B. (1982): Adoption and adaptation. J. Nervous Mental Dis. 170: 489-493.

Briquet P. (1859): Traité de l'Hysterie. J.B. Balliers et fils. Paris.

Brorsson B. (1977): Att mäta hälsa. SPRI rapport. Civiltryck AB. Stockholm.

Brown F. (1961): Depression and childhood bereavement. J. Ment. Sci. 107: 754-777.

Brown F. (1966): Childhood bereavement and subsequent psychiatric disorder. Brit. J. Psychiat. 112: 1035-1041.

Brown G.W., Harris T. (1978): Social origins of depression: a study of psychiatric disorder in women. Tavistock. London.

Buchsbaum M.S., Cousey R.D., Murphy D.L. (1976): The biochemical high-risk paradigm: behavioral and familial correlates of low platelet MAO activity. Science 194: 339-341.

Cadoret R.J., Cunningham L., Loftus R., Edwards J. (1976): Studies o f a d o p t e e s f r o m p s y c h i a t r i c a l l y d i s t u r b e d b i o l o g i c a l p a r e n t s : I I I . Medical symptoms and illnesses in childhood and adolescence. Am. J. Psychiat. 133: 1316-1318.

Cadoret R.J. (1978): Evidence for genetic inheritance of primary affective disorder in adoptees. Am. J. Psychiat. 135: 463-466.

65

Cadoret R.J., Gath A. (1978): Inheritance of alcoholism in adoptees. Brit. J. Psychiat. 132: 252-258.

Carlsson G., Arvidsson O., Bygren L-O., Werkö L. (1979): Liv och hälsa. En kartläggning av hälsoutvecklingen i Sverige. Liber. Stock­holm.

Carney M.W.P., Roth M., Garside R.F. (1965): The diagnosis of depressive syndromes and the prediction of ECT responce. Brit. J. Psychiat. Ill: 659-674.

Cederblad M. (1978): De äldre adoptivbarnens anpassning. Läkartid­ningen 75: 4677-4678.

Classification of occupations. (1951): Her Majetsy's Stationary Office. London.

Cloninger C.R., Reich T., Guze S.B. (1975): The multifactorial model of disease transmission: III. Familial relationship between sociopathy and hysteria (Briquet's syndrome). Brit. J. Psychiat. 127: 23-32.

Cloninger C.R. (1978): The link between hysteria and sociopathy: an integrative model of pathogenesis based on clinical, genetic and neuro-physiological observations. In Akiskal H.S., Webb W.L. (eds): Psy­chiatric diagnosis : Explorations of biological predictors. Spectrum Press. New York.

Cloninger C. R., Miller J. P., Wette R., Martin R. L., Guze S.B. (1979a): The evaluation of diagnostic concordance in follow-up stu­dies: I. A general model of causal analysis and a methodological critique. J. Psychiat. Research 15: 85-106.

Cloninger C.R., Reich T., Wetzel R. (1979b): Alcoholism and the affective disorders: familial association and genetic models. In Good­win D., Erickson C. (eds): Alcoholism and the affective disorders. Spectrum Press. New York.

66

Cloninger C.R., Bohman M., Sigvardsson S. (1981): Inheritance of alcohol abuse. Cross-fostering analysis of adopted men. Arch. Gen. Psychiat. 38: 861-868.

Cloninger C. R., Sigvardsson S., Bohman M., von Knorring A-L. (198 2 ) : P r e d i s p o s i t i o n t o p e t t y c r i m i n a l i t y i n S w e d i s h a d o p t e e s . I I . Cross-fostering analysis of gene-environment interaction. Arch. Gen. Psychiat. 39: 1242-1247.

Cohen J. (1960): A coefficient of agreement for nominal scales. Edu­cational Psychological Measurement 20: 37-46.

Cohen L.J., Campos J.J. (1974): Father, mother and stranger as elicitors of attachment behaviours in infancy. Developm. Psychol. 10: 146-154.

Colton T. (1974): Statistics in medicine. Little, Brown & Company. Boston.

Coopersmith S. (1967): The antecedents of self-esteem. W.H. Free­man. San Francisco.

Coursey R.D., Buchsbaum M.S., Murphy D.L. (1982): 2-year follow-up of subjects at risk for psychopathology on the basis of platelet MAO activities. Neuropsychobiology 8: 51-56.

Crowe R.R. (1974): An adoption study of antisocial behaviour. Arch. Gen. Psychiat. 31: 785-791.

Dahlgren L., Myrhed M. (1977): Alcoholic females. II. Causes of death with reference to sex difference. Acta Psychiat. Scand. 56: 81-91.

Dahlgren L., Ideström C-M. (1979): Female alcoholics. V. Morbidity. Acta Psychiat. Scand. 60: 199-213.

67

Dahlgren L. (1981): Kvinnor och alkohol - missbruket ökar. Läkar­tidningen 78: 786-788.

Davenport H.T., Werry J.S. (1970): The effect of general anesthesia surgery and hospitalization upon the behaviour of children. Amer. Orthopsychiat. 36: 595-608.

Dennehy C.M. (1966): Childhood bereavement and psychiatric illness. Brit. J. Psychiat. 112: 1049-1069.

Earles F. (1976): The fathers (not the mothers): their importance and influence with infants and young children. Psychiatry 39: 209-226.

d'Elia G., Perris C. (1971): Childhood environment and bipolar and unipolar recurrent depressive psychosis. Proc. 4th U.E.P. Congr., pp. 53-62. Almquist & Wiksell. Stockholm.

d'Elia G., von Knorring L., Perris C. (1974): Non-psychotic de­pressive disorders: a ten-year foUow-up. Acta Psychiat. Scand. Suppl. 255: 173-186.

Eriksen T.E. (1980): Den ökade sjukfrånvaron - en statistisk analys. Delegationen för Social Forskning. Rapport Departementens Offset Central. Stockholm.

Feighner J., Robins E., Guze S., Woodruff R., Winokur G., Munoz R. (1972): Diagnostic criteria for use in psychiatric research. Arch. Gen. Psychiat. 26: 57-63.

Ferguson D. (1972): Some characteristics of repeated sickness ab­sence. Brit. J. Industr. Med. 29: 420-431.

Forrest A.D., Fraser R.H., Priest R.G. (1965): Environmental factors in depressive illness. Brit. J. Psychiat. Ill: 243-253.

68

Fowler C.J., von Knorring L., Oreland L. (1980): Platelet MAO activity in sensation seekers. Psychiatry Research. 6: 273-279.

Freud S. (1917): Mourning and melancholia. In Strachey J. (ed): the standard edition. Vol. 14. Hogarth Press. London.

Gardell I. (1979): Internationella adoptioner. Rapport från Allmänna Barnhuset. Stockholm.

Gershon E.S. (1977): Genetic and biologic studies of affective illness. In Gershon E.S., Belmaker R.H., Kety S.S., Rosenthal M. (eds): The impact of biology on modern psychiatry. Plenum Press. New York & L ondon.

Gillespie R.D. (1928): Hypochondria: its definition, nosology and pshycopathology. Guys Hosp. Rep. 78: 408-460.

Goodman J.D., Silbers tein R., Man dell W. (1963): Adopted children brought to child psychiatric clinic. Arch. Gen. Psychiat. 9: 451-456.

Goodwin J.D., Schulsinger F., Hermansen L. et al (1973): Alcohol problems in adoptees raised apart from alcoholic biological parents. Arch. Gen. Psychiat. 28: 238-243.

Goodwin D.W., Schulsinger F., Knop J., Mednik S., Guze B. (1977): Alcoholism and depression in adopted-out daughters of alcoholics. Arch. Gen. Psychiat. 34: 751-755.

Greer S. (1964): The relationship between parental loss and attempted suicide. Brit. J. Psychiat. 110: 698-705.

Gregory I. (1966): Retrospective data concerning childhood loss of a parent. II. Category of parental loss by decade of birth, diagnosis and MMPI. Arch. Gen. Psychiat. 15: 362-368.

69

Gurney C., Roth M., Garside R.F., Kerr T.A., Shapira K. (1972): Studies in the classification of the affective disorders. The relation­ship between anxiety states and depressive illness. II. Brit. J. Psychiat. 121: 162-166.

Guze S.B., Perley M.J. (1963): Observations on the natural history of hysteria. Am. J. Psychiat. 119: 960-965.

Guze S.B., Woodruff R.A., Clayton P.J. (1962): Sex, age and the diagnosis of hysteria (Briquet's syndrome). Am. J. Psychiat. 129: 121-124.

Hagneil H. (1966): A prospective study of the incidence of mental disorder. Scandinavian University Books. Uniskol Lund.

Hagnell O., Rorsman B. (1978): Maskerad depression - en svår och viktig psykiatrisk diagnos i somatisk sjukvård. Läkartidningen 75: 779-780.

Harvald B., Hauge M. (1965): Hereditary factors elucidated by twin studies. In Neel J.V., Shaw M.W., Schyll W.J. (eds) Genetics and the epidemiology of chronic diseases. U.S. Dept of Health, Education & W e l f a r e W a s h i n g t o n D . C .

Helgason T. (1979): Epidemiological investigations concerning affective disorder. In Schou M., Strömgren E. (eds): Origin, prevention and treatment of affective disorders. 241-255.

Helwig J.T., Council K.A. (eds) (1979): Statistical analysis system u s e r ' s g u i d e . S A S I n s t i t u t e I n c . C a r e y N C .

Hill O.W. (1969): The association of childhood bereavement with suicidal attempt in depressive illness. Brit. J. Psychiat. 115: 301-304.

70

Hin de R.A., Spencer-Booth Y. (1970): Individual differences in the responces of rhesus monkeys to a period of separation from their mothers. J. Child Psychol. Psychiat. 11: 159-176.

Hoopes J. L., Sherman E. A., Law der E. A., Andrews R. G., Lower K.D. (1969): A follow-up study of adoptions (vol. II): post placement functioning of adopted children. Child Welfare League of America. New York.

Horn J.M., Green M., Carney R., Erickson M.T. (1975): Bias against genetic hypotheses in adoption studies. Arch. Gen. Psychiat. 32: 1365-1367.

Hostetter A.M., Egeland J.A., Endicott J. (1983): Amish study, II: consensus diagnoses and reliability results. Am. J. Psychiat. 140: 62-66.

Hutchings B., Mednik S.A. (1975): Registered criminality in the adoptive and biological parents of registered male criminal adoptees. In Fieve R.R., Rosenthal D., Brill H. (eds): Genetic research in psychiatry. John Hopkins University Press. Baltimore.

Jacobson S., Fasman J., DiMascio A. (1975): Deprivation in the childhood of depressed women. J. Nerv. Ment. Dis. 160: 5-14.

Janet P. (1931): L'etat mental des hystériques: études sur divers symptômes hystériques. Alcan. Paris.

Jellinek E.M. (1945): Heredity of the alcoholic. Alcohol, sicence and society. O. J. Stud. Ale. 5: 105-114.

Johansson E. (1981): Recidivistic criminals and their families: mor­bidity, mortality and abuse of alcohol. A longitudinal study of earlier youth prison inmates and a control group and their families in three generations. Scand J. Soc. Med. Suppl. 27.

71

Jonsson G., Kälvesten A-L. (1964): 222 Stockholmspojkar. En social-psykiatrisk undersökning av pojkar i skolåldern. Almqvist & W iksell.

Jonsson G. (1967): Delinquent boys, their parents and grandparents. Acta Psychiat. Scand. Suppl. 195.

Kaij L. (1960): Alcoholism in twins: studies on the etiology and sequels of abuse of alcohol. Almquist & Wiksell. Stockholm.

Kaij L. (1970): Biases in a Swedish social register of alcoholics. Soc. Psyciat. 5: 216-218.

Kallmann F.J. (1952): Genetic principles in manic-depressive psycho­ses: depression. Proc. 42d Ann. Meeting Am. Psychopthol. Assoc.

Ken dell R.E., Gourlay J. (1970): The clinical distinction between psychotic and neurotic depression. Brit. J. Psychiat. 117: 257-260.

Ken dell R.E. (1976): The classification of depressions: a review of contemporary confusion. Brit. J. Psychiat. 129: 15-28.

Kenyon F.E. (1976): Hypochondrical states. Brit. J. Psychiat. 129: 1-14.

Kety S. (1979) : Disorders of the human brain. Scientific American 8: 202-214.

Kielholz P. (1972): Masked depression. Hans Huber Publishers. Bern.

Kielholz P. (1973): Classification of depressions. In Angst J. (ed): Classification and prediction of outcome of depression. Schattauer. Stuttgart.

von Knorring L. (1975): The experience of pain in patients with depressive disorders. A clinical and experimental study. Umeå Univ. Med. Diss. No. 2.

72

von Knorring L., Oreland L. (1978): Visual averaged evoked re­sponses and platelet MAO activity as an aid to identify a risk group for alcohol abuse. Prog. Neuro-Psychopharm. 2: 385-392.

Kraepelin E. (1921): Manic depressive insanity and paranoia. S. Livingstone. Edinburgh.

Larsson S. (1976): Sick-leave pattern in chronic alcoholism. Social-med. Tidskrift. 4: 269-275.

Lawder E. A., Lower K. D., Andrews R. G., Sherman E. A., Hill J. G. (1969): A follow-up study of adoptions: post placement functioning of adoption families. Child Welfare League of America. New York.

Leijon M., Mikaelsson B. (1982): Upprepade korta sjukfall kan vara symtom på socialmedicinsk problematik. Soc. Förs. nr 11: 8-11.

Leonhard K. (1957-69): Aufteilung der endogenen psychosen. lst-4th editions. Akademieverlag. Berlin.

Lewis A.J. (1938): States of depression: their clinical and aetiological differentiation. Brit. Med. J. ii: 875-878.

Loehlin J.C. (1982): Are personality traits differentially heritable? Behaviour Genetics 12: 417-428.

Lopez Ibor J.J. (1969): Depressive Äquivalente. In Hippius H. & Sehlbach H. (eds) Das depressive Syndrom. Int. Symp. Berlin. Urban & Schwarzenberg. Munich/Berlin/Vienna.

Luxenburger H. (1942): Das Zirkuläre Irrsein. In A. Gutt (ed): Handbuch der Erbkrankheiten. Vol. 4. Georg Thieme Verlag. Leipzig.

MacLean C.J., Morton N.E., Elston R.C., Yee S. (1976): Skewness in commingeled distributions. Biometrics 32: 695-699.

73

Martin R.L., Cloninger C.R., Guze S.B. (1979): The evaluation of diagnostic concordance in follow-up studies. II. A blind prospective follow-up of female criminals. J. Psychiat. Research. 15: 107-125.

Medhus A. (1974): Morbidity among female alcoholics. Scand. J. Soc. Med. 2: 5-11.

Mendelson M. (1974): Psychoanalytic concepts of depression. Spectrum Pubi. New York.

Mendlewicz J., Klotz J. (1975): The offspring of manic-depressive patients. In Sankar S. (ed): Mental health in children. Vol. 1. PJD P u b i . L t d . W e s t b u r y N . N .

Mendlewicz J., Rainer J.D. (1977): Adoption study supporting genetic transmission in manic-depressive illness. Nature 268: 327-329.

Mohammed S.N., Weisz G.M., Waring E.M. (1978): The relationship of chronic pain to depression, marital adjustment and family dynamics. Pain. 5: 285-291.

Munro A. (1966): Parental deprivation in depressive patients. Brit. J. Psychiat. 112: 443-457.

Nettelbladt P. (1981): The man and his family. A psychiatric study of back-ground variables of relevance for the man's mental adaptation and relations to his wife and child. Doctoral Diss. Lund.

Ni Bhrolchain M., Brown G.W., Harris T.Q. (1979): Psychotic and neurotic depression: clinical characteristics. Brit. J. Psychiat. 134: 94-107.

Nie N.H., Hull C.H., Jenkins J.G. (1975): Statistical package for the social sciences, ed. 2. McGraw-Hill Book Co. New York.

74

Nord-Larsen M. (1979): At måle helbred. The Danish National Insti­tute of Social Research. Study 38. Teknisk Forlag. Köpenhamn.

Nyström S. (1979a): Depressions: factors related to 10-year prog­nosis. Acta Psychiat. Scand. 60: 225-238.

Nyström S. (1979b): Sjukskrivning vid sjukdom med social bakgrund. Läkartidningen 76: 918-919.

Partanen J., Bruun K., Markkanen T. (1966): Inheritance of drink­ing behaviour. The Finnish foundation for alcohol studies. Helsinki.

Pell S., D1 Alonzo C.A. (1970): Sickness absenteeism of alcoholics. J. Occup. Med. 12: 198-210.

Perley M.J., Guze S.B. (1962): Hysteria - the stability and useful­ness of clinical criteria. A quantitative study based on a follow-up period of 6-8 years in 39 patients. New Eng. J. Med. 266: 421-426.

Perris C. (1966): A study of bipolar (manic-depressive) and unipolar recurrent depressive psychoses. Acta Psychiat. Scand. Suppl. 194. Vol. 42.

Perris C. (1974): A study of cycloid psychoses. Acta Psychiat. Scand. Suppl. 253.

Perris C. (1976): Frequency and hereditary aspects of depression. In Gallant D.M., Simpson G.M. (eds): Depression: behavioral, bioche­mical, diagnostic and treatment concepts. John Wiley & Sons, Inc. New York.

Perris C., Perris H., Ericsson U., von Knorring L. (1982): The genetics of depression. A family study of unipolar and neurotic-react­ive depressed patients. Arch. Psychiat. Nervenkr. 232: 137-155.

75

Perris H. (1978a): A multifactorial study of life events in depressed patients. Umea University Medical Diss. No. 78.

Perris H., von Knorring L., Perris C. (1978b): Genetic vulnerability for depression and life events. Neuropsychobiology 8: 241-247.

Pilowsky I., Spence N.D. (1976). Illness behavior syndromes asso­ciated with intractable pain. Pain 2: 61-71.

Plomin I., DeFries J.C., McClearn G.E. (1980): Behavioral genetics. A primer. W.H. Freeman & company. San Francisco.

Pollitt J. (1972): The relationship between genetic and precipitating factors in depressive illness. Brit. J. Psychiat. 121: 67-70.

Post F. (1968): The factor of aging in affective illness. In Coppen A., Walk A. (eds): Recent developments in affective disorders. Brit. J. Psychiat. Spec. Pubi. No. 2: 105-116.

Pringle M.L., Kellmer (1967): Adoption-facts and fallacies. A review of research in USA, Canada and Great Britain 1948-1965. Studies in Child Development. Longmans. London.

Raps C.S., Peterson C.R., Reinhard K.E., Abramson L.Y., Seligman M.E.P. (1982): Attributional style among depressed patients. J. Abn. Psychol. 91: 102-108.

Raskin A., Boothe N.A., Reatig N.A., Schultenbrandt J.C. (1971): Factor analyses of normal and depressed patients1 memories of paren­tal behavior. Psychol. Rep. 29: 871-879.

Raynor L. (1980): The adopted child comes of age. George Allen & Unwin. London.

Robertson J., Robertson H. (1971): Young children in brief separa­tions: a fresh look. Psychoanal. Study of the Child. 26: 264-315.

76

Robins E., Munoz R.A., Martin S., Gentry K.A. (1372): Primary and secondary affective disorders. In Zubin J., Freyhan F.A. (eds): Disorders of mood. John Hopkins Press. Baltimore.

Rorsman B. (1970): Disability of psychiatric patients. Social Psychiat­ry 5: 219-223.

Rosenthal S.H., Klerman G.L. (1966): Content and consistency in the endogenous depressive pattern. Brit. J. Psychiat. 112: 471-484.

Roy A. (1978): Vulnerability factors and depression in women. Brit. J. Psychiat. 113: 106-110.

Rush B. (1812): Medical inquires and observations upon the disease of the mind. Philadelphia, Kimber & Richardson (Reprinted as Special Edition, Classics of Medicine Library. Birmingham Alabama 1979).

Rutter M. (1979): Attachment and the development of social relation­ships. In Rutter M. (ed): Scientific foundations of developmental psychiatry. Heinemann Medical. London.

Rydelius P-A. (1981): Children of alcoholic fathers. Their social adjustment and their health status over 20 years. Acta Paediatr. Scand. Suppl. 286.

Samuelsson S. (1982): Life events and mental disorder in an urban female population. Acta Psychiat. Scand. Suppl. 299.

Schaffer H.R. (1968): Social learning and identification. In Lunzer E.A., Morris J.F. (eds): Development in human learning. Staples Press. London.

Schechter M.D. (1960): Observations on adopted children. Arch. Gen. Psychiat. 3: 45-56.

77

Schiff M., Duyme M., Dumaret A., Stewart J., Tomkiewicz S., Fein­gold J. (1978): Intellectual status of working-class children adopted early into upper-middle-class families. Science 200: 1503-1504.

Schulsinger F., Kety S., Rosenthal D., Wender P.H. (1979): A family study of suicide. In Schou M. & Strömgren E. (eds): The origin, prevention and treatment of affective disorders. Academic Press. New York.

Shapiro R.W. (1970): A twin study of "non-endogenous" depression Acta Jutlandica 42: no 2.

Shopsin B., Mendlewicz J., Suslak L., Silbey E., Gershon S. (1976): Genetics of affective disorders. II. Morbidity risk and genetic trans­mission. Neuropsychobiology. 2: 26-36.

Seglow J., Pringle M.L.K., Wedge P. (1972): Growing up adopted. National Foundation for Educational Research. London.

Shepherd M., Cooper B., Brown A.C., Kalton G. (1966): Psychiatric illness in general practice. Oxford University Press. London.

Sigvardsson S. (1982a): Alcohol abuse and criminality, a cross-foster-ing study of gene-environment interaction. Umeå University Medical Diss. No. 84.

Sigvardsson S., Cloninger C.R., Bohman M., von Knorring A-L. (1982 b ) : P r e d i s p o s t i o n t o p e t t y c r i m i n a l i t y i n S w e d i s h a d o p t e e s . I I I . Sex differences and validation of the male typology. Arch Gen. Psy-chiat. 39: 1248-1253.

Sjögren T. (1948): Genetic-statistical and psychiatric investigation of a West Swedish population. Acta Psychiat. Neur. Suppl. 52.

Skeels H.M. (1965): Effects of adoption on children from institutions. Department of Health, Education and Welfare. Children 12: 33-34.

78

Skodak M., Skeels H.M. (1949): A final follow-up study of one hund­red adopted children. J. Genetic Psychol. 75: 85-125.

Slater E. (1938): Erbpathologie des manisch-depressiven Irrseins, Die Eltern und Kinder von Manisch-Depressiven. Z. Ges. Neur. Psychiat. 163: 1-147.

Smedby B. (1974): Sjukvårdens storkonsumenter. Storkonsumenter av läkarvård. Soc. Med. Tidskr. 462-468.

Smeraldi E., Negri F., Melica A.M. (1977): A genetic study of affec­tive disorders. Acta Psychiat. Scand 56: 382-398.

Socialstyrelsen (1976): Rätts genetiska undersökningar i faderskaps­ärenden . Liber. Stockholm.

Spitz R.A., Wolff K.M. (1946): Anaclitic depression. Psychoanal. Study of the Child. 2: 313-342.

Stacey M., Dearden R., Pill R., Robinson D. (1970): Hospitals, children and their families: The report of a pilot study. Routledge & Kegan Paul. London.

Stenstedt A. (1952): A study in manic-depressive psychosis: clinical, social and genetic investigations. Acta Psychiat. Neur. Scand. Suppl. 79.

Strömgren E. (1938): Beiträge zur psychiatrischen Erblehre. Munks -gaard. Copenhagen.

Tennant C., Bebbington P., Hurry J. (1980a): Parental death in childhood and risk of adult depressive disorders: a review. Psychol. Med. 10: 289-299.

79

Tennant C., Hurry J., Bebbington P. (1980b): Parent-child separa­tions during childhood: their relation to adult psychiatric morbidity and to psychiatric referral: preliminary findings. Acta Psychiat. Scand. Suppl. 285: 324-331.

Tennant C., Hurry J., Bebbington P. (1982): The relation of child­hood separation experiences to adult depressive and anxiety states. Brit. J. Psychiat. 141: 475-482.

van Theis S. (1924): How foster children turn out. State Charities Aid Association. New York.

Thysell T. (1952): Adoption. Courrier. 2: 7.

Thysell T. (1965): En socialmedicinsk undersökning rörande adoptiv­barn födda av ogifta mödrar. Socialmedicinsk Tidskrift. 1: 24-33.

Tizard B., Rees J. (1975): The effect of early institutional rearing on the behaviour problems and affectional relationships of four-y ear old children. Child Psychol. Psychiat. 16: 61-74.

Tizard B. (1977): Adoption: a second chance. Open Books. London.

Valliant G.E. (1980): Natural history of male psychological health: VIII. Antecedents of alcoholism and "orality". Am. J. Psychiat. 137: 181-186.

Vanggard T., Jorgensen B. (1981): Atypiske endogene depressioner i almenpraksis. Måneds Pract Laegegern 59 : 137-158.

Weissman M., Myers J.K. (1978): Affective disorders in a U.S. urban community. Arch. Gen. Psychiat. 35: 1304-1311.

Westrin C-G. (1973): Low-back sick-listing. A nosological and medical insurance investigation. Scand. J. Soc. Med. Suppl. 7.

80

WHO (1958): Constitution. Annex I. WHO. Geneva.

Winokur G. (1979a): Genetic findings and methodological considera­tions in manic depressive disease. Brit. J. Psychiat. 117: 267-274.

Winokur G., Reich T., Rimmer J., Pitts F. (1970b): Alcoholism III. Diagnosis and familial psychiatric illness in 259 alcoholic probands. Arch. Gen. Psychiat. 23: 104-111.

Winokur G., Cadoret R., Dorzab J., Baker M. (1971a): Depressive disease: a genetic study. Arch. Gen. Psychiat. 24: 135-144.

Winokur G., Rimmer J., Reich T. (1971b): Alcoholism: IV. Is there more than one type of alcoholism? Brit. J. Psychiat. 118: 525-531.

Witmer H.L., Herzog E., Weinstein E.A., Sullivan M.E. (1963): Independent adoptions. Russel Safe Foundations. New York.

Woerner P.I., Guze S.B. (1968): A family and marital study of hy­steria. Brit. J. Psychiat. 114: 161-168.

Woodruff R.A., Guze S.B., Clayton P.J., Carr D. (1973): Alcoholism and depression. Arch. Gen. Psychiat. 28: 97-100.

Yarrow L.J. (1964): Separation from parents during early childhood. In Hoffman M.L., Hoffman L.W. (eds): Review of child development research. Vol. 1. Russel Safe Foundation. New York.

Zuckerman M. (1979): Sensation seeking. Beyond the optimal level of arousal. Lawrence Erlbaum Associates. Hillsdale.

Åmark C. (1951): A study in alcoholism. Acta Psychiat. Scand. Suppl. 70.

Öjesjö L. (1980): Prevalence of known and hidden alcoholism in the revisited Lundby population. Soc. Psychiat. 15: 81-90.