Psychosomatic and depressive symptoms from age eight to age eighteen

Schmidt, M H

Even today, when the psychoanalytic position that depression in childhood (i.e., before puberty) could hardly exist (e.g., Malmquist, 1971) has been given up, there are two approaches for defining depressive disorders during that age period: On the basis of an empirical study, Carlson and Cantwell (1980) criticized the concept of masked depression, which nevertheless has been held in serious textbooks (Graham, 1986). This concept proposed that classical depressive symptoms in children are masked by age-related behavior problems (e.g., hyperactivity, aggression and somatic complaints, phobias, underachievement, and delinquency). Ling, Oftedal, and Weinberg (1970) argued this position in dealing with the coincidence of headache and depression in children. It has never been clear how such masking symptoms could be differentiated from nonmasking disorders of the same kind in nondepressive children. Later, this position was dropped in favor of the idea of age-related “associated features” versus “essential symptoms” of childhood depression (Cantwell, 1983) under which somatic symptoms were subsumed as well.

More recent concepts have abandoned the idea of different symptoms in childhood and adult depression. But, in symptom lists like the Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev. [DSMIll-R]; American Psychiatric Association, 1987), somatic problems such as eating disorders, sleeping disorder, fatigue, and weight loss are included. The Tenth Revision of the International Classification of Diseases (ICD-10; World Health Organization, 1991) describes a special concomitant somatic syndrome. Because one of the best longitudinal studies dealing with depressive symptoms (Hoffmann, 1991) did not take somatic complaints into account, we followed this question using data from a German cohort study that independently checked somatic and depressive complaints at ages 8, 13, and 18. In this study, exploratory evaluation was done taking into consideration three possible relations: (a) there is no association between both kinds of symptoms at any age, (b) there is a continuous association between somatic and depressive symptoms, and (c) there are different associations at different age levels.

Subjects and Method

Out of a total population of 1,444 German 8-year-olds born between March and September 1970 and living in Mannheim on March 1, 1978, 361 (25%) were randomly drawn and asked to participate in the investigation. Out of these, 129 (36%) refused to take part in the study, and 16 were excluded due to low intelligence (IQ below 70), chronic diseases or severe handicaps, or because they had moved away from the area during the course of the study. The remaining 216 children formed the random sample.

Each of the 1,444 families of the initial population was asked to fill out a 40-item behavior questionnaire (an adapted version of the Conners scale) and to grant permission to their child’s teacher to fill out the same screening instrument. After the random sample had been drawn and separated, screening data for 733 children were available. In order to increase the number of subjects with behavior problems, the most conspicuous 25% of the children with the highest scores in the teacher and parent questionnaire were selected. Together with the 216 subjects of the random sample, they formed the total field sample of 399 8year-olds. Out of these, 356 (89%) could be reexamined at age 13, 340 (85%) also at age 18 (see Figure 1). Prevalence rates quoted in this chapter are related to subjects of the random sample. All other calculations were made on the basis of the total number of assessed children and adolescents.

Symptoms and child psychiatric disorders were determined by expert rating after a 2-hour parent interview. In the second and third stages of the study, case definition was further supplemented by an additional adolescent interview carried out with the 13- and 18-year-olds. All assessment decisions were made by child psychiatrists and experienced clinical psychologists. Case definitions and diagnoses were based on the evaluation of 28 to 39 symptoms assessed as being absent (0), moderate (1), or severe (2), depending on the information provided by the respective interview items. More methodological details have been described elsewhere (Esser, Schmidt, and Woerner, 1990; Esser et al., 1992). Overall prevalence rates were estimated to be 16.2% at age 8, 17.8% at age 13, and 16.0% at age 18. In the original study (Esser et al., 1990; Esser et al., 1992), diagnoses were classified in four different categories following ICD-9: neurotic and emotional disorders (ICD 300 & 313), conduct disorders associated or not associated with emotional problems (ICD 312), hyperkinetic syndromes (ICD 314), and other specific symptoms and syndromes (ICD 307). In the present study, we summed up symptoms typical of these four diagnoses to sum scores.

Somatic complaints and psychosomatic symptoms were treated equally and thus were categorized in one group, but somatic complaints based on somatic disorder were excluded, of course. Due to the agedependent developmental status, we checked for only 7 psychosomatic symptoms at age 8, 10 symptoms at age 13, and 11 symptoms at age 18 (see Table 1). The definitions of depressed mood and the various psychosomatic symptoms originate from the Mannheimer Elterninterview (Esser et al., 1989).



The overall prevalence rates of psychosomatic symptoms and depressed mood are reported in Table 1. Figures represent severity levels 1 and 2. The two symptoms most frequently found were psychophysiologic headaches and abdominal pain. Depressed mood was reported at each age level, with prevalence rates between 12.1% and 31.0%.

Table 2 informs about sex differences concerning prevalence rates. Information is given separately for degrees 1 (moderate) and 2 (severe). Most of the relevant sex differences were found in symptoms of moderate severity. The only relevant sex difference in symptoms rated severe (2) was found in “overeating,” which was in favor of the girls. Overall sex differences were more obvious in depressive than in psychosomatic symptoms.


Table 3 and 4 show the intercorrelations of the psychosomatic symptoms assessed at ages 13 and 18 and the correlation of psychosomatic symptoms with depressed mood. As there was only one significant intercorrelation, figures of the 8-year-old sample are not demonstrated.

A comparison of Tables 3 and 4 shows that there is no special pattern of correlation between the psychosomatic symptoms, even if there are some more significant intercorrelations than expected at random. It seems important that the correlation coefficients with depression are at least at the same level as the intercorrelations. Furthermore, such problems as falling asleep and food refusal are significantly correlated at both age levels with depressed mood.


Using odds-ratios, Table 5 demonstrates the age-specific role of symptoms for psychiatric disorder at respective age. A convergence toward the 18-year-olds can be stated that makes certain psychosomatic symptoms more meaningful. In contrast, even in 8-year-olds, depressed mood is of significance for psychiatric problems at that age.

By means of correlation analysis, we examined whether the sum of all psychosomatic symptoms is a more powerful indicator of psychiatric disorder than single symptoms at the same age. Table 6 shows that psychosomatic symptoms at age 8, even if more highly correlated with internalizing symptoms, are fairly meaningless. Their value increases by ages 13 and 18-of course, mostly for internalizing symptoms. In contrast, the widespread significance of depression at age 8 concentrates on correlations with internalizing symptoms at ages 13 and 18.


Analyzing the longitudinal data, the specific issues to be investigated included the persistence of symptoms. Table 7 shows in what percentage certain symptoms occurring at a certain age could still be found at a later assessment. For example, in 73.3% of the 8-year-olds with headaches, headaches were still reported at age 13. Of course, the stability for the 10-year interval is lower than that for the 5-year interval. Keeping in mind the expected decrease, the symptoms of headache, overeating, and initial insomnias show remarkable stability. Further, the stability of these symptoms from age 8 to age 13 is similar to that from age 13 to age 18. Other rates show the decreasing importance of somatic symptoms from childhood to late adolescence. The high stability of depressed mood is noteworthy.


Table 8 deals with the predictive value of symptoms for later psychiatric disorder. Figures represent percentages of children and adolescents with psychosomatic symptom(s) and/or depressed mood but without a psychiatric diagnosis, who 5 or 10 years later are assessed as having a psychiatric disorder. For example, 24% of those with headaches at age 8 were diagnosed as psychiatrically disordered 5 years later. It is remarkable that only childhood headache and early adolescent abuse of prescribed pharmaceutics are of importance for later psychiatric disorder. Unexpectedly, childhood depression is highly predictive, whereas adolescent depression is not.

That this is not due to the insignificance of the single symptoms is evident when regarding the sum score of psychosomatic symptoms. This sum score of psychosomatic symptoms was compared with the symptom of depressed mood concerning the predictive value for later psychiatric disorder in undisturbed children (see Table 9). But, even the sum score is not a better predictor than depressed mood.


The following case report demonstrates the development of early depressive symptoms from childhood until early adulthood and their impact on later psychiatric disorder. Anne (name fictitious) lived together with her two elder sisters and her parents in a small suburb of Mannheim. The psychosocial circumstances of the family were strained by chronic disease, unemployment, and, later, by the earning incapacity of Anne’s father.

At age 8, Anne went to her third year of primary school. Her school performance was slightly below average; due to dyslexia, she received special tuition. In the lessons, she was very quiet. Her mother reported that Anne tended to hide herself away and to show depressive reactions (e.g., cry quickly) and that she often was teased by other children. Overall clinical judgment, however, did not result in psychiatric diagnosis (i.e., severity of problem behavior did not fulfill the diagnostic criteria of psychiatric disorder).

At age 13, the assessment revealed compulsive cleanliness, tidymindedness, and slight eating difficulties. As at the previous assessment depressed mood was also obvious. Due to overall clinical judgment, “suspectible psychiatric disorder” was diagnosed.

At age 18, Anne was still living with her parents. She had many friends and a permanent boyfriend whom she planned to marry soon. Her good secondary-school qualification enabled her to train as a decorator, which was her desired profession. After having changed employment, problems occurred, because Anne felt she was being harassed by her colleagues and superiors. She reacted with nightmares, psychosomatic complaints, weight loss (to 42 kg; height = 169 cm), and, again, depressed mood. Her fear of going to work became so strong that at last she quit the job. At assessment, she had started to work in the stocks department of a supermarket, so that she was employed below her educational level. At this age, she clearly fulfilled the criteria of psychiatric disorder: she was given the ICD-10 diagnosis of adjustment disorder with mixed anxiety and depressive reaction (F43.22).

Anne’s course of psychopathologic symptoms shows that depressed mood does not necessarily lead to a manifest disorder. Not until additional distress occurs do symptoms aggravate to psychiatric disorder. In the present case extreme fear of work and symptoms of depression and eating disorder were reactions to severe distress.


One ought to expect that it is possible to estimate reliably the associations between somatic complaints and indicators of psychosomatic problems, which demonstrate more or less physical resonance to affective stress in our sample. As already stated by Esser et al. (1990) the influence of dropouts could mainly be controlled. A standardized assessment was used. Interrater reliabilities were found to be very high (Esser et al., 1990). The main problem of the present study might arise from analysis on the symptom level, not on the diagnosis level. A replication study on the basis of diagnoses has been started. Another source of error could be seen in the fact that the calculations were not carried out for boys and girls separately. Whether sex differences present themselves differently on the diagnosis level than on the symptom level is an issue for further research.

In order to determine the predictive value of certain symptoms, associations between somatic complaints and depressed mood were of special interest in our study. From the three possible models, that of associations changing with age had the highest probability according to our findings. This general result must be specified.

Until early school age, somatic complaints are seen as typical symptoms, but their relevance even increases in adolescence. Their association with depressed mood becomes stronger with age, whereas such an association cannot be seen in early school age. Nevertheless, single somatic symptoms show remarkable intercorrelations, but correlations with depressed mood are of at least the same significance. Associations are mainly found for somatic complaints of the “depression type” (i.e., eating problems, sleeping disorder).

Depression at age 8 proved to be a good indicator of psychiatric disorder at the same age, whereas only some of the somatic complaints-mainly of higher complexity (e.g., “somatization”)-were indicative of psychiatric disorder. This was best validated for symptoms with longitudinal stability, such as headaches, eating and sleeping problems, and depressed mood.

The course of former psychiatrically healthy children and youngsters with certain single somatic and depressive symptoms developing psychiatric disorder at a later age demonstrates the long-term predictive value of those factors for later psychiatric disorder. Again, early depression with or without somatic symptoms was of higher predictive power than somatic symptoms.

It can be summarized that the typical coincidence of somatic symptoms (e.g., eating and sleeping disorders, psychiatric disorder) gains importance only with age, whereas depressed mood at age 8 already has a high predictive value. Therefore, our findings cannot support the hypothesis of childhood depression masked by somatic symptoms.


Out of a sample of 399 8-year-olds from a city of 300,000 inhabitants, 356 could be followed to age 13 and 340 to age 18. Half of the children derived from a representative sample, and the other half derived from a sample of problem children assessed by parents’ and teachers’ questionnaires. Rates of psychosomatic and depressive symptoms at ages 8, 13, and 18 were compared, correlations between both kinds of symptoms were calculated, and the stability of symptoms and their significance for present psychiatric morbidity and later psychiatric disorder were investigated. Changing associations were found between somatic complaints and depressed mood at different age levels, but our findings could not support the hypothesis of childhood depression masked by somatic symptoms.


American Psychiatric Association (1987), Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.). Washington, DC: APA. Cantwell, D. P. (1983), Depression in childhood: Clinical picture and diagnostic criteria. In: Affective Disorders in Childhood and Adolescence: An Update, ed. D. P. Cantwell & G. A. Carlson. New York: Spectrum, pp. 3-18.

Carlson, G. A. & Cantwell, D. P. (1980), Unmasking masked depression in children and adolescents. Amer. J. Psychiat., 137:445-449. Esser, G., Blanz, B., Geisel, B. & Laucht, M. (1989), Mannheimer Elterninterview (MEI). Weinheim, Germany: Beltz. Esser, G., Schmidt, M. H. & Woerner, W. (1990), Epidemiology and course of psychiatric disorders in school-age children: Results of a longitudinal study. J. Child PsychoL & Psychiat., 31:243-263.

Esser, G., Schmidt, M. H., Blanz, B., Fatkenheuer, B., Fritz, A., Koppe, T., Laucht, M., Rensch, B. & Rothenberger, W. (1992), Pravalenz und Verlauf psychischer Storungen im Kindes- und Jugendalter [Prevalence and course of psychiatric disorders in childhood and adolescence]. Zeitschrift fur Kinder- und Jugendpsychiatrie, 20:232-242.

Graham, P. (1986), Child Psychiatry: A Developmental Approach. Oxford, England: Oxford University Press.

Hoffmann, V. (1991), Die Entwicklung depressiver Reaktionen in Kindheit und Jugend. Eine entwicklungspsychopathologische Langsschnittuntersuchung [Depressive reactions in childhood and adolescents: A longitudinal study on developmental psychopathology] (Studien und Berichte 51). Berlin: Max-Planck-Institut fur Bildungsforschung.

Ling, W., Oftedal, G. & Weinberg, W. A. (1970), Depressive illness in children presenting a severe headache. Amer. J. Disease Childhood, 120:122-124.

Malmquist, C. (1971), Depressions in childhood and adolescence. New England J. of Medicine, 284:887-892; 955-961. World Health Organization (1991), Tenth Revision of the International Classification of Diseases (ICD-10). Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzerland: WHO.

Copyright Analytic Press 1999

Provided by ProQuest Information and Learning Company. All rights Reserved

You May Also Like


12 SEXUAL ABUSE OF BOYS BY CLERGY Ponton, Lynn Recent media exposure of sexual abuse of boys by Catholic clergy has engendered a cri…

A psychoanalytic approach to eating disorders: The role of dependency

A psychoanalytic approach to eating disorders: The role of dependency Jeammet, Philippe Although they have been observed for centuri…



A program of comprehensive school-based mental health services in a large urban public school district: The Dallas model

A program of comprehensive school-based mental health services in a large urban public school district: The Dallas model Pearson, Glen …