Childhood Trauma and Somatic and Mental Illness in Adulthood

Abstract

Background

Childhood trauma is associated with somatic and mental illness in adulthood. The strength of the association varies as a function of age, sex, and type of trauma. Pertinent studies to date have mainly focused on individual diseases. In this study, we investigate the association between childhood trauma and a multiplicity of somatic and mental illnesses in adulthood.

Methods

Data from 156 807 NAKO Health Study participants were analyzed by means of logistic regressions, with adjustment for age, sex, years of education, and study site. The Childhood Trauma Screener differentiated between no/minor (n = 115 891) and moderate/severe childhood trauma (n = 40 916). The outcome variables were medical diagnoses of five somatic and two mental health conditions as stated in the clinical history.

Results

Persons with childhood trauma were more likely to bear a diagnosis of all of the studied conditions: cancer (odds ratio [OR] = 1.10; 95% confidence interval: [1.05; 1.15]), myocardial infarction (OR = 1.13 [1.03; 1.24]), diabetes (OR = 1.16, [1.10; 1.23]), stroke (OR = 1.35 [1.23; 1.48]), chronic obstructive pulmonary disease (OR = 1.45 [1.38; 1.52]), depression (OR = 2.36 [2.29; 2.43]), and anxiety disorders (OR = 2.08 [2.00; 2.17]). All of these associations were stronger in younger persons, regardless of the nature of childhood trauma. Differences between the sexes were observed only for some of these associations.

Conclusion

Childhood trauma was associated with a higher probability of developing mental as well as somatic illness in adulthood. As childhood trauma is an element of individual history that the victim has little to no control over, and because the illnesses that can arise in adulthood in association with it are a heavy burden on the affected persons and on society, there is a need for research on these associations and for the development of preventive measures.

Childhood trauma has been associated with an increased likelihood of being diagnosed with a multiplicity of somatic and mental illnesses in adulthood, especially with cardiovascular disease, metabolic disease and diseases of the respiratory system, as well as mood and anxiety disorders (1–5). The term childhood trauma covers emotional, physical and sexual abuse as well as emotional and physical neglect during childhood and adolescence (6, 7). In Germany, 20–30% of adults report having experienced at least one of these childhood traumas (8, 9).

Weleff and Potter (10) discuss three possible ways in which childhood trauma may be associated with somatic diseases in adulthood:

• poorer health behavior

• poorer mental health

• biological changes..

In contrast to behavioral factors, childhood trauma is clearly anchored in time before the onset of disease later in adulthood and those affected have little or no control over it (11). However, the mediation effects of behavioral factors, such as smoking, alcohol consumption and overweight, found so far have been minor (12). In a study by Shields et al. (13), associations between childhood trauma and being diagnosed with diabetes were still observed even when behavioral factors and depression were taken into account.

Pertinent studies to date have shown that the strength of association varies with the nature of childhood trauma and that age at the time of survey and sex are relevant moderators of the association between childhood trauma and disease in adulthood (1, 4). For somatic diseases in general, stronger associations were observed in women than in men (1, 14). By contrast, an association between abuse experienced as a child and myocardial infarction and stroke was mainly reported in men (15, 16). In diabetes, subtype-specific studies, the majority of which focused on type 2 diabetes, showed that the disease was more frequently diagnosed after childhood trauma, with heterogeneous findings with regard to the nature of the trauma (14, 17). While scarce, studies on gestational diabetes indicate similar associations, especially after experiences of abuse (18). Findings on mental illnesses are also inconsistent (1). Experiences of abuse in childhood were associated with more frequent diagnoses of depression in women (3, 19, 20). However, emotional abuse and emotional neglect were found to be sex-independent factors in relation to depression (3, 8).

Comparability of findings to date is hampered by differences in age and sex distributions and a focus on individual diseases. Thus, the aim of our study was to directly compare the associations of various diseases with childhood trauma within one sample. For this purpose, we used data from the baseline survey of the German National Cohort (NAKO Gesundheitsstudie)—the largest population-based survey in Germany (21, 22). Using a cross-sectional design, associations between childhood trauma and somatic and mental illnesses were investigated as a cross-disease comparative basis for disease-specific association studies. The focus was on:

• Cancer

• Myocardial infarction

• Stroke

• Diabetes

• Chronic obstructive pulmonary disease (COPD)

• Anxiety disorders

• Depression.

Given their increasing incidence rates, the selected diseases have a significant impact on the quality of life of those affected and a high medical need for treatment (11). For instance, cardiovascular disease and cancer rank among the most common causes of death, while mental illnesses cause the most days of incapacity for work and the most cases of early retirement (11).

Methods

Study population

During the NAKO baseline survey between 2014 und 2019, data were collected from > 205 000 adults of the general German population in 18 study centers in 13 federal states (22). For detailed information on sampling, inclusion criteria and study protocols, see Peters et al. (22).

The ethics committees of all study sites approved the study. A written informed consent for participation in the study and data utilization was obtained from all participants. The NAKO health study and all analyses were conducted in accordance with the Declaration of Helsinki.

Interview

Information on the following parameters was obtained by means of a standardized, structured, computer-assisted interview:

• Age at the time of the NAKO baseline survey (continuous and age groups: age <40 years, 40–60 years and > 60 years)

• Sex (male/female)

• Educational attainment of the participants (years of education according to the International Standard Classification of Education; ISCED-97 [23, 24]).

Participants whose years of education could not be calculated were excluded from the analyses (n = 18 719).

In addition, participants stated whether they had “ever been diagnosed by a doctor or psychotherapist” with one of the following illnesses (yes/no):

• Cancer

• Myocardial infarction

• Stroke

• Diabetes mellitus (hereinafter: diabetes)

• Chronic bronchitis or chronic obstructive pulmonary disease (hereinafter: COPD)

• Anxiety disorder or panic attack (hereinafter: anxiety)

• Depression.

To maintain a focus on cases of type 2 diabetes, participants who were <40 years of age at the time of first diagnosis of diabetes were excluded from the diabetes-related analyses (n = 2 903). There was no additional information available in our study that would allow to reliably identify type 1 diabetes and gestational diabetes.

Childhood trauma

Information about childhood trauma was obtained within a self-response module using the Childhood Trauma Screener (CTS), a brief screening version of the Childhood Trauma Questionnaire (CTQ) (6, 7). This instrument covers five types of trauma with one item each: emotional, physical and sexual abuse as well as emotional and physical neglect.

Responses to these items are given on a five-point scale (“not at all” to “very often”). For each of the items, a distinction was made between no/mild and moderate/severe childhood trauma (25). A global value captured whether at least one moderate/severe childhood trauma was reported, regardless of type. The number of childhood traumas was a measure of how many moderate/severe childhood traumas were reported. The distribution of the number and types of childhood trauma is shown in eTable 1. Participants with missing CTS scores were excluded from the analyses (n = 32 782).

Statistical analyses

All analyses were performed using the statistical software “R” (version 4.2.1) and are based on the responses of 156 807 participants with available data on years of education and CTS. The probability of the diseases was predicted using multiple logistic regression analyses with robust standard errors and 95% confidence intervals (95% CI). The false discovery rate (FDR) was used to adjust for multiple testing (q-values) (26).

All analyses were adjusted for age at the time of the NAKO baseline survey, sex, years of education, and study site. The study site was used for adjustment of the multicenter survey. (9, 27, 28). Age, sex and education have repeatedly been shown to be major confounders and moderators of the association between childhood trauma and diseases in adulthood (1, 4, 9, 15).

Questions addressing behavioral variables, such as smoking and alcohol consumption, were only asked at the time of the NAKO baseline survey, but not at the time of diagnosis. Consequently, it was not possible to conclusively determine whether these behavior patterns preceded the onset of the disease. For this reason, no adjustment was made initially.

The global, dichotomized CTS score was used as a predictor to assess associations between moderate/severe childhood traumas in general and the selected diseases. Due to missing values for the diseases, the size of the analysis sample varies slightly. In addition, associations with the various types of trauma were investigated, using the dichotomized CTS items. The number of childhood traumas was used as a continuous predictor to assess the impact of multiple adverse childhood experiences. Sex and age interactions were used to test for moderating effects of sex and age. The defined age groups were used for assessing age interactions. The number of cases with myocardial infarction and stroke was very low among participants aged <40 years (n = 9 and n = 48, respectively). This age group was excluded in the corresponding analyses. Sex- and age-stratified analyses were calculated for descriptive purposes.

Results

Table 1 shows descriptive statistics on the data of the participants and stratified by sex and age group. Women and participants aged 40 years and over reported childhood traumas more frequently. One third of participants with childhood trauma reported multiple types of trauma; the combination of emotional and physical childhood traumas was descriptively observed most frequently (eSupplement Table 1). Women reported diagnoses of cancer, anxiety and depression more frequently. Men more frequently reported diagnoses of myocardial infarction and diabetes. Somatic diseases were reported more commonly with older age groups. In contrast, no differences in the frequency of mental illnesses were apparent in the age groups ≥ 40 years.

Table 1. Characteristics of the participants of the NAKO baseline survey (N = 156 807) and stratified by sex and age groups.

	Complete sample	Men	Women	<40 years	40–60 years	>60 years
Age (years): M (SD)	49.3 (12.3)	49.5 (12.4)	49.0 (12.3)	30.8 (4.8)	49.6 (5.3)	64.7 (3.1)
19–40 years (n = 33 084) (%)	21.1	20.6	21.6	–	–	–
40–60 years (n = 85 833) (%)	54.7	54.3	55.2	–	–	–
60–75 years (n = 37 890) (%)	24.2	25.1	23.2	–	–	–
Sex (% men)	50.5	–	–	49.4	50.1	52.5
ISCED-97 years of education: M (SD)	15.7 (2.3)	15.9 (2.3)	15.5 (2.3)	15.8 (2.2)	15.6 (2.3)	15.7 (2.3)
Childhood trauma (% yes)	26.1	23.5	28.7	17.5	26.8	32.0
Physical neglect (% yes)	10.0	10.0	10.0	4.4	9.7	15.3
Emotional neglect (% yes)	7.5	6.6	8.5	3.6	8.0	9.8
Physical abuse (% yes)	8.1	8.4	7.8	4.8	8.4	10.2
Emotional abuse (% yes)	8.1	6.2	10.0	7.8	8.5	7.3
Sexual abuse (% yes)	6.1	2.6	9.7	4.5	6.7	6.0
Cancer (n = 156 457) (% yes)	6.9	5.7	8.1	1.6	5.8	14.0
Myocardial infarction (n = 156 608) (% yes)	1.4	2.3	0.5	–	0.9	3.7
Stroke (n = 156 527) (% yes)	1.3	1.6	1.0	–	1.0	3.1
Diabetes (n = 154 363) (% yes)	3.8	4.8	2.8	–	2.5	10.0
COPD (n = 156 144) (% yes)	4.7	4.5	4.9	2.4	4.6	6.9
Anxiety (n = 156 321) (% yes)	7.4	5.3	9.6	5.9	8.0	7.5
Depression (n = 155 933) (% yes)	14.0	10.2	17.8	9.9	15.1	15.0

The information was obtained from 156 807 participants in the NAKO baseline survey, unless otherwise specified.

For the age group <40 years, myocardial infarction, diabetes and stroke were not included in the analysis due to insufficient case numbers.

Anxiety, anxiety disorders or panic attacks; COPD, chronic bronchitis or chronic obstructive pulmonary disease; ISCED, International Standard Classification of Education; M, mean; SD, standard deviation

Based on adjusted logistic regression analyses, the strength of association varied between at least one moderate/severe childhood trauma and one disease, depending on the disease in focus (Figure 1, eFigure 1 and Table 2). With somatic diseases, the probability of diagnosis after at least one moderate/severe childhood trauma was increased by 10–45%: The weakest associations were found for cancer (odds ratio [OR] = 1.10; 95% CI [1.05; 1.15], q-value <0.001), myocardial infarction (OR = 1.13; [1.03; 1.24], q-value = 0.028) and diabetes (OR = 1.16; [1.10; 1.23], q-value <0.001). The associations for stroke (OR = 1.35; [1.23; 1.48], q-value <0.001) and COPD (OR = 1.45; [1.38; 1.52], q-value <0.001) were somewhat stronger. The strongest associations were seen for anxiety (OR = 2.08; [2.00; 2.17], q-value <0.001) and depression (OR = 2.36; [2.29; 2.43], q-value <0.001). Accordingly, the probability of diagnosis of one of these mental illnesses was doubled in participants who had experienced moderate/severe childhood traumas.

Figure 1.

Disease frequencies for somatic and mental illnesses with 95% confidence intervals for participants with no/mild or moderate/severe childhood trauma. The frequencies were estimated based on logistic regression analyses which were adjusted for age, sex, years of education, and study site.

Anxiety, anxiety disorder or panic attack; COPD, chronic bronchitis or chronic obstructive pulmonary disease

eFigure 1.

Main effects. The associations between childhood trauma and somatic and mental illnesses are shown. The respective odds ratios are presented with 95% confidence intervals for participants with moderate/severe childhood trauma. Participants with no/mild childhood trauma were used as the reference group. The values are derived from logistic regression analyses which were adjusted for age at the time of the NAKO baseline survey, sex, years of education, and study site.

Anxiety, anxiety disorder or panic attack; COPD, chronic bronchitis or chronic obstructive pulmonary disease

Table 2. Main effects, sex and age interactions of childhood trauma with somatic and mental illnesses among participants in the NAKO baseline survey.

		Main effect				Sex interaction				Age interaction
Diseases	N	OR	[95% CI]	p-value	q-value	OR	[95% CI]	p-value	q-value	Χ²	p-value	q-value
Cancer	156 457	1.10	[1.05; 1.15]	<0.001	<0.001	1.13	[1.03; 1.23]	0.008	0.049	19.49	<0.001	<0.001
Myocardial inf.	156 608	1.13	[1.03; 1.24]	0.011	0.028	1.35	[1.07; 1.71]	0.011	0.050	17.00	<0.001	0.001
Stroke	156 527	1.35	[1.23; 1.48]	<0.001	<0.001	1.11	[0.92; 1.34]	0.289	1.000	5.19	0.075	0.193
Diabetes	154 363	1.16	[1.10; 1.23]	<0.001	<0.001	0.97	[0.86; 1.09]	0.570	1.000	31.04	<0.001	<0.001
COPD	156 144	1.45	[1.38; 1.52]	<0.001	<0.001	1.30	[1.18; 1.44]	<0.001	<0.001	6.65	0.036	0.109
Anxiety	156 321	2.08	[2.00; 2.17]	<0.001	<0.001	1.00	[0.92; 1.08]	0.921	1.000	33.66	<0.001	<0.001
Depression	155 933	2.36	[2.29; 2.43]	<0.001	<0.001	1.14	[1.08; 1.22]	< 0.001	< 0.001	108.05	<0.001	<0.001

Myocardial inf., myocardial infarction; Anxiety, anxiety disorder or panic attack; COPD, chronic bronchitis or chronic obstructive pulmonary disease;

OR, odds ratio; 95% CI, 95% confidence interval; q-value, false discovery rate

The main effect reflects the association between childhood trauma and the respective disease.

An OR >1 indicates a higher probability of diagnosis in persons with at least one moderate/severe childhood trauma.

The interactions indicate whether the association between childhood trauma and probability of diagnosis differs between men and women or between age groups.

The larger the OR or χ, the greater the differences between the sexes or age groups.

The q-value represents the significance level of the association after adjusting for multiple testing, analogous to the p-value.

Example: A higher probability of diagnosis was found for cancer after at least one moderate/severe childhood trauma

(OR = 1.10 [1.05; 1.15], q-value <0.001). This association was stronger for women (OR = 1.13 [1.03; 1.23], q-value = 0.049) and dependent on the age of the participants at the time of the NAKO baseline survey (χ = 19.49, q-value <0.001).

Differences with regard to type of trauma are shown in Figure 2 and eSupplement Table 2. For cancer (OR = 0.96 [0.91; 1.03], q-value = 0.743), myocardial infarction (OR = 0.98 [0.87; 1.11], q-value = 1.000) and diabetes (OR = 1.06 [0.98; 1.14], q-value = 0.617), the probabilities of diagnosis were unchanged after physical neglect. By contrast, for stroke (OR = 1.27 [1.13; 1.44], q-value <0.001), COPD (OR = 1.14 [1.07; 1.23], q-value <0.001), anxiety (OR = 1.24 [1.17; 1.32], q-value <0.001), and depression (OR = 1.23 [1.18; 1.29], q-value <0.001), the probability was increased by 14–27%. The strength of the association with abuse and emotional childhood trauma was comparable for both cancer (12–23% higher probability) and stroke (43–59% higher probability). For myocardial infarction (16–52% higher probability), diabetes (20–50% higher probability) and COPD (65–89% higher probability), the strength of association successively increased from emotional neglect to physical abuse to emotional abuse. For mental illness, the associations after emotional childhood trauma were the strongest (approximately 3 times higher probability). A higher number of childhood traumas was associated with a higher probability for all diseases (eSupplement Table 3). Here again, the strength of the associations was weaker for somatic diseases compared to mental illnesses (8–35% higher probability and 67–83% higher probability, respectively).

Figure 2.

Odds ratios of moderate/severe childhood trauma for somatic and mental illness with 95% confidence intervals for the five types of childhood trauma. The odds ratios are derived from logistic regressions which were adjusted for age, sex, years of education, and study site. In all cases, participants with no/mild childhood trauma were used as the reference group.

Anxiety, anxiety disorder or panic attack; COPD, chronic bronchitis or chronic obstructive pulmonary disease

Only a few sex differences were observed in the associations between childhood trauma and disease probability (eFigure 2, Table 2, eSupplement Tables 3–9). The probability of being diagnosed with cancer (interaction: OR = 1.13 [1.03; 1.23], q-value = 0.049), COPD (interaction: OR = 1.30 [1.18; 1.44], q-value <0.001), and depression (interaction: OR = 1.14 [1.08; 1.22], q-value <0.001) after childhood trauma was higher in women compared to men. These differences were primarily found for physical childhood trauma (eSupplement Tables 5 and 7). Further sex differences were limited to individual types of trauma (eSupplement Tables 5–9).

eFigure 2.

Sex interactions. The associations between childhood trauma and somatic and mental illnesses are shown, stratified by men and women. The respective odds ratios are presented with 95% confidence intervals for participants with moderate/severe childhood trauma. Participants with no/mild childhood trauma were used as the reference group. The values are derived from logistic regression analyses which were adjusted for age at the time of the NAKO baseline survey, years of education and study site.

Anxiety, anxiety disorder or panic attack; COPD, chronic bronchitis or chronic obstructive pulmonary disease

For all diseases it was found that the older the participants were at the time of the NAKO baseline survey, the weaker the association between moderate/severe childhood trauma and probability of diagnosis of diseases in adulthood (eFigure 3, Table 2, eSupplement Tables 3 and 10–15). The lowest dependency on age was found for stroke (interaction: χ = 5.19, q-value = 0.193) and COPD (interaction: χ = 6.65, q-value = 0.109). The strongest age-dependent effects were observed for depression (interaction: χ = 108.05, q-value <0.001). These results were also confirmed by the association with the number of childhood traumas (eSupplement Table 3). After physical abuse, age-dependent associations were observed for almost all diseases (eSupplement Table 13). For anxiety and depression, the age differences were least dependent on the type of trauma (eSupplement Tables 11–15).

eFigure 3.

Age interactions. The associations between childhood trauma and somatic and mental illnesses are shown, stratified by three age groups at the time of the NAKO baseline survey. The odds ratios are presented with 95% confidence intervals for participants with moderate/severe childhood trauma. Participants with no/mild childhood trauma were used as the reference group. The values are derived from logistic regression analyses which were adjusted for sex, years of education and study site. For participants aged <40 years, no values were calculated for myocardial infarction, diabetes and stroke due to insufficient case numbers.

Anxiety, anxiety disorder or panic attack; COPD, chronic bronchitis or chronic obstructive pulmonary disease

Discussion

Based on data of the NAKO health study, we examined the association between childhood trauma and probability of diagnosis of selected somatic and mental illnesses in adulthood. It was found that the associations with mental illnesses were stronger than with somatic illnesses. In addition, it was analyzed whether these associations were dependent on sex, age at the time of the NAKO baseline survey and type of childhood trauma.

In line with the findings of previous meta-analyses, the strongest associations were found between emotional childhood trauma and depression, while the associations between depression and sexual or physical abuse were weaker (20). While no difference between the sexes had been found earlier for diagnoses of depression after emotional childhood trauma (3), we found an increased probability of depression in women after physical abuse in this study. The association between depression and sexual abuse in childhood was stronger in men. Overall, however, our findings confirm the results of a meta-analysis which found only minor differences between the sexes (12). No difference between men and women was observed for the association between anxiety disorder and childhood trauma. In women, an increased probability of COPD was noted after neglect and physical childhood trauma. Only isolated analyses found sex differences for stroke and myocardial infarction.

For all diseases evaluated, the association between childhood trauma and probability of diagnosis was weaker for older participants of the NAKO baseline survey than for younger ones. The age differences were largest for depression and smallest for COPD and stroke. The results could indicate that the closer the onset of the disease is to the time of traumatization, the more strongly childhood trauma is associated with a diagnosis. For instance, the age at onset of depression is typically in early adulthood (29). By contrast, the frequency of COPD increases with increasing age (30). Furthermore, it was reported that childhood trauma was associated with younger age at the time of first diagnosis of depression (1, 3). Given the differences in age at the onset of the disease, the subtypes of diabetes provide a good opportunity to evaluate this hypothesis within one disease group. However, diabetes subtype-specific studies on associations with childhood traumas have rarely been conducted. A second factor which may play a role is the increase in number and severity of comorbidities in older age (31). As a result, the association between childhood trauma and probability of diagnosis of a specific condition may be weaker with older age at onset of disease. Thus, future studies should take comorbidities into account.

In addition to an increase in disease incidence and a younger age at onset of disease, increased severity of the disease course was observed after childhood trauma (1, 3, 4, 13). For example, childhood trauma was associated with increased mortality (32). In Germany, cancer and cardiovascular disease have the highest annual mortality rates, while depression is associated with the highest number of sick days and early retirements (11). Depression is a common comorbidity of mental as well as somatic illnesses (33). In adulthood, the annual financial burden on the health care system is more than doubled by depression (34). According to German insurance data, each depression is responsible for an additional burden of EUR 3000 in the first year of diagnosis and of EUR 1500 in the following years (35). By preventative programs and early support for those affected, it may be possible to lower the probability of later disease (36, 37). But even with existing disease, treatment of childhood trauma could improve the individual quality of life and prognosis of the disease.

Limitations

Limitations include the fact that subjective self-reported data were used to determine the presence of childhood trauma and that each type of trauma was represented by only one item. Some participants had to be excluded because of missing childhood trauma data. However, the missing information was not limited to individual items (9). Given the retrospective nature of the self-reported information, bias due to memory effects or current mental symptoms cannot be ruled out. According to studies, however, both over- and underestimation have been observed for childhood trauma, depending on the current state of health and personality traits; consequently, bias in both directions should be assumed (38, 39).

The medical lifetime diagnosis of somatic and mental illnesses was also recorded based on self-report data. Neither the age at onset of the disease nor its current treatment or current severity were taken into account. There is a need for further analysis with the inclusion of this information. In addition, the spectrum of mental illnesses represented is rather small. The NAKO dataset does not contain information about additional physician-diagnosed mental illnesses. At the time of the NAKO baseline survey, participants were questioned about behavioral factors, such as smoking, alcohol consumption and overweight. It is difficult to relate these factors to the time of first diagnosis. Consequently, it cannot be determined whether these factors actually preceded the onset of disease. We therefore did not include unhealthy behaviors as potential mediators in our analysis. Future studies, taking this temporal relationship between behavioral factors and disease into account, should include differentiated subgroup analyses to test such mediation hypotheses.

Conclusion

The aim of our study was to investigate the association between childhood trauma and somatic and mental illnesses in adulthood. The NAKO health study offers the opportunity to analyze these associations within the same sample and thus to directly compare these diseases with each other. The focus was on conditions with a high personal and social burden of disease. In future research, these findings should be expanded by investigating further disorders and diseases, especially in the area of mental health, and be supplemented by analyses of somatic and psychological comorbidity, and potential mediation effects. In accordance with the aim of this study, an initial cross-disease comparative overview of the investigated associations was provided. Disease-specific studies should build on these findings and deepen the insights gained.