Journal of Affective Disorders 223 (2017) 95–100

Contents lists available at ScienceDirect

Journal of Affective Disorders

journal homepage: www.elsevier.com/locate/jad

Research paper

Impact of childhood adversities on depression in early adulthood: A MARK

longitudinal cohort study of 478,141 individuals in Sweden
⁎
Emma Björkenstama,b, , Bo Vinnerljungc,d, Anders Hjernd,e
a
Department of Public Health Sciences, Division of Social Medicine, Karolinska Institutet, Stockholm, Sweden
b
Department of Community Health Sciences, Fielding School of Public Health and California Center for Population Research, University of California Los Angeles, Los
Angeles, California, United States
c
Department of Social Work, Stockholm University, Stockholm, Sweden
d
Clinical Epidemiology / Department of Medicine, Karolinska Institutet, Stockholm, Sweden
e
Centre for Health Equity Studies, Stockholm University/Karolinska Institutet, Stockholm, Sweden

A R T I C L E I N F O A B S T R A C T

Keywords: Objective: Although the relationship between childhood adversity (CA) and depression is widely accepted, there
Childhood adversity is little information on what proportion of depression is attributable to CA.
Depression Method: We used a Swedish cohort of 478,141 individuals born in 1984–1988 in Sweden. Register-based CA
Antidepressants indicators included parental death, parental substance abuse and psychiatric morbidity, parental criminality,
Epidemiology
parental separation, public assistance recipiency, child welfare intervention, and residential instability.
Cohort study
Estimates of risk of depression, measured as retrieval of prescribed antidepressants and/or psychiatric care with
a clinical diagnosis of depression, between 2006 and 2012 were calculated as Hazard Ratios (HR) with 95%
confidence intervals (CI), using a Cox regression analysis.
Results: All CAs predicted depression in early adulthood. Furthermore, the predictive association between the
CA indicators and depression was graded, with highest HRs observed for 4+ CAs (HR: 3.05 (95% CI 2.83–3.29))
for a clinical diagnosis for depression and HR: 1.32 (95% CI 1.25–1.41) for antidepressant medication after
adjustments were made for important confounding factors. Of the studied CAs, child welfare intervention en-
tailed highest HR for depression.
Conclusion: Regardless of causality issues, children and youth with a history of multiple CA should be regarded
as a high-risk group for depression by professionals in social, and health services that come into contact with this
group.

1. Introduction particularly detrimental to depression (Anda et al., 2002; Björkenstam

et al., 2017; Chapman et al., 2004; Gilman et al., 2003; Gilman et al.,
Depression is a common and potentially debilitating disorder oc- 2002; Sareen et al., 2013; Wirback et al., 2014). Childhood adversities
curring throughout the life-course (Fleisher and Katz, 2001; Patel et al., that have been linked to the development of depression include low
2007). The first onset of depression often occurs in childhood or ado- family socioeconomic status, parental separation, single parenthood,
lescence, although treatment typically does not occur until later in life parental criminality, and parental psychiatric morbidity (Anda et al.,
(Birmaher et al., 1996; Costello et al., 2006; Kessler et al., 2007). De- 2002; Chapman et al., 2004; Gilman et al., 2002; Sareen et al., 2013).
pression has become one of the most common mental health conditions Studies have shown that the associations between different CAs and
in medical and psychiatric practice. It ranks fourth among the leading depression vary depending on type of CA (Anda et al., 2002; Chapman
causes of disability worldwide and is expected to become the second et al., 2004; Gilman et al., 2002). CAs tend to occur in clusters rather
leading cause by 2020 (World Health Organization, 2001). Depression than as single or separate events (Dong et al., 2004), and clustered CAs
is one of the most common mental disorder among adolescents and have a strong graded relationship to depression (Chapman et al., 2004;
young adults (Wittchen and Jacobi, 2005; World Health Organization, Dube et al., 2003; Green et al., 2010b).
2001, 2016). Various pathways through which CA influences depression have
Several studies have pointed out childhood adversity (CA) as been discussed, including both biological and psychological

⁎
Corresponding author at: Department of Public Health Sciences, Division of Social Medicine, Karolinska Institutet, Stockholm, Sweden.
E-mail address: emma.bjorkenstam@ki.se (E. Björkenstam).

http://dx.doi.org/10.1016/j.jad.2017.07.030
Received 28 April 2017; Received in revised form 13 June 2017; Accepted 9 July 2017
Available online 18 July 2017
0165-0327/ © 2017 Elsevier B.V. All rights reserved.
E. Björkenstam et al. Journal of Affective Disorders 223 (2017) 95–100

Table 1
Definitions and classification of childhood adversity.

Childhood adversity Definition ICD classification Data source

Parental death Parental death N/A Causes of Death Register

Parental substance abuse Parental hospitalization for alcohol and/or substance abuse ICD-9: 291, 303, 304.0–304.9, National Patient Register
305.0, 305.2–305.7, 305.9
ICD-10: F10-F19
Parental criminality Parent sentenced to prison, probation, or forensic N/A Register of Court Convictions
psychiatric care
Parental psychiatric Parental hospitalization for mental disorder (excluding ICD-9: 290–319; ICD-10: F00-F99 National Patient Register
morbidity substance-abuse related disorders)
Parental separation Having separated parents N/A Longitudinal Integration Database for Health
Insurance and Labor Market Studies
Household living on public Public assistance during at least one year, where more than N/A Total Enumeration Income Survey
assistance 50% of the yearly income constituted public assistance
Substantial child welfare At least one placement in out-of-home or respite care N/A National Child Welfare Register
intervention before age 12 years.
Residential instability Three or more changes in place of residence N/A Total Population Register

mechanisms (Swartz et al., 2017). Biological explanations suggest that administrative registers (Ludvigsson et al., 2009):
CA contributes to stress-induced brain dysfunction that in turn may lead The Causes of Death Register comprises information on all deaths of
to mental health problems (Evans, 2003; Heim and Binder, 2012; Swedish residents. The National Patient Register (NPR) includes all
Shonkoff et al., 2012; Swartz et al., 2017). Psychological explanations individuals admitted to psychiatric or general hospitals, with complete
on the other hand suggest that CAs lead to emotional dysfunction that coverage for all care since 1987. The NPR only includes clinical diag-
in turn leads to depression (Dube et al., 2003; Kovacs et al., 2008; noses made by physicians. Diagnoses in the NPR are coded according to
McLaughlin et al., 2010; Shapero and Steinberg, 2013). the International Classification of Diseases (ICD). The Prescribed Drug
To date, the vast majority of existing studies has been based on self- Register (PDR) contains patient identities for all dispensed prescribed
reported information, entailing risk for recall bias (Colman et al., 2016; drugs to the entire Swedish population since July 2005.
Hardt and Rutter, 2004; Reuben et al., 2016). One alternative approach Pharmaceuticals in the PDR are grouped according to the Anatomical
for assessing exposure to CA is by the use of register-based data. Among Therapeutic Chemical Classification System (ATC). The Total
others, this method eliminates the possibility of recall bias. To the best Enumeration Income Survey contains data on income and govern-
of our knowledge, few studies have used register-based data to examine mental monetary benefits for all Swedish residents. The Total
associations between cumulative CA and depression (Björkenstam Population Register includes information on age, sex, place of re-
et al., 2017; Dahl et al., 2017). The current study adds to the literature sidence, and other relevant demographic characteristics. The
on CA and depression by using high quality, longitudinal nationwide Longitudinal Integration Database for Health Insurance and Labor
register data of the whole population of Sweden. In addition, although Market Studies integrates existing data from the labor market, educa-
the relationship between CA and depression is widely accepted, there is tional and social sectors. The Register of Court Convictions contains
little information on what proportion of depression is attributable to information on all court convictions in Sweden for individuals 15 years
CA. Lastly, we use a two-step approach for identifying cases of de- of age or older. The National Child Welfare Register has records on out-
pression: use of antidepressive medication, and having a recorded di- of-home care (foster family and residential care) and respite care.
agnosis of depression. With a cohort of approximately 480,000 in- Families were linked together through the Multi-Generation register,
dividuals born between 1984 and 1988 in Sweden, the overall aim was which contains all known relationships between children and parents
to examine the relationship between a wide range of prospectively re- (born in 1932 or later) since 1961.
corded childhood adversities and early adulthood depression. Our
specific aims were to:
2.2. Measures

– investigate the differential associations between single CAs and the

2.2.1. Indicators of childhood adversity
risk of depression
The CA indicators, defined in Table 1, were selected based upon
– examine the effects of cumulative CAs on later depression
prior research demonstrating them to be separately associated with the
– estimate the population attributable fractions (PAF) for depression
risk of depression (Anda et al., 2002; Chapman et al., 2004; Gilman
attributable to CA exposure
et al., 2003; Green et al., 2010b; Najman et al., 2010; Wadsworth and
Butterworth, 2006; Weitoft et al., 2003). We included eight adversities,
occurring between birth and age 14: parental death, parental substance
2. Methods
abuse, parental criminality, parental psychiatric morbidity, parental
separation, household living on public assistance (a proxy for poverty),
2.1. Study population
substantial child welfare intervention (experience of out-of-home or
respite care before age 12), and residential instability. To assess cu-
The study population was defined as all individuals born in Sweden
mulative exposure, the total number of adversities was summed. Each
between 1984 and 1988, recorded in the Medical Birth Register, who
indicator was weighted equivalently in the analyses.
were alive and registered in Sweden on December 31st, 2005 (n =
488,823). We applied certain exclusion criteria: owing to a high pro-
portion of missing data on important variables, those who were adopted 2.2.2. Outcome variables: depression case ascertainment
were excluded (n = 296), as well as those who were granted disability Two indicators of depression were created for the follow-up period
pension before age 23 years (n = 10,386) (mainly persons with severe 2006–2012: (1) first indication of dispensed antidepressive medication
learning disabilities or multiple handicaps). Our final analytical sample (ATC-code N06A) was obtained from the PDR, (2) first entry of in- or
comprised 478,141 individuals. The unique Swedish personal identity outpatient care with a diagnosis of depression (F32-F39 in ICD-10)
number was used to link this cohort to multiple health care and according to the NPR.

96
E. Björkenstam et al. Journal of Affective Disorders 223 (2017) 95–100

2.2.3. Confounders 3. Results

The following confounding factors were considered: Parental
country of birth was categorized as Sweden (both parents Swedish- Cohort characteristics of the study population, by sex and depres-
born), mixed (one Swedish-born), other Nordic (Denmark, Finland, sion are presented in Table 2. Thirty percent of our both females and
Norway or Iceland), other European, and non-European. Highest par- males were exposed to at least one adversity, and 10% had experienced
ental educational attainment, measured when the child was 15 years two or more adversities. The most prevalent CA was parental separation
old, was classified as compulsory school (9 years or less), secondary (25%), followed by parental psychiatric morbidity (5%) and household
school (10–12 years), and university/college (> 12 years). living on public assistance (5%). Approximately 6% of the individuals
received a clinical diagnosis for depression during the follow-up period
(7% of females and 4% of males). All studied CAs were more common
2.3. Statistical analysis in those who were treated with a diagnosis for depression. Nearly 15%
in the entire population retrieved prescribed antidepressant medication
The study cohort was followed from January 1st, 2006 to first during the follow-up period (18% of females and 11% of males).
psychiatric contact that resulted in a clinical diagnosis for depression, Table 3 presents the crude and multi-adjusted HRs with 95% CI for
and/or date of first prescribed antidepressant medication, until death, risk of depression by exposure to CA. All CAs were associated with
or at most until December 31st, 2012. We performed multivariate elevated risk of depression, both in terms of clinical diagnosis and an-
analyses using Cox hazard models. Three regression models were ana- tidepressant medication utilization. The risks for a clinical diagnosis of
lyzed: in Model I, we adjusted for birth year and sex; in Model II ad- depression were most pronounced for child welfare intervention (HR:
ditional adjustments were made for parental country of birth and par- 2.79 (95% CI 2.60–2.99)), parental psychiatric morbidity (HR: 2.11
ental education. In the third model, all CAs were included (95% CI 2.02–2.20)) and parental substance abuse (HR: 2.09 (95% CI
simultaneously. 1.98–2.20)) (Table 3, Model I).
Parental psychiatric morbidity is a crude proxy for genetic liability All studied CAs also entailed greater risk for antidepressant medi-
for psychiatric disorder in offspring, and may increase the risk of ex- cation utilization, with HRs spanning from 1.05 (95% CI 1.01–1.11) for
periencing other CAs as well. Thus, in secondary analyses parental parental death, to 1.24 (95% CI 1.17–1.31) for child welfare interven-
psychiatric morbidity was excluded from the list of CAs and considered tion (Table 3, Model I).
a confounder. All HRs were slightly reduced when additional adjustments were
To calculate population attributable fractions (PAF), the following made for parental country of birth and parental education (Table 3,
formula was used: Model II). However, all studied CAs remained significantly associated
P (RR-1)/1+P(RR-1), where P is the percentage of the CA in the with both depression indicators. When all CAs were included in the
cohort and RR was equal to the relative risk for each CA and depression. same model (Table 3, Model III), the HRs decreased markedly. How-
The PAF estimates the proportion of the outcome in the population that ever, several of the adversities remained substantially associated with
would be reduced if the exposure was eliminated (Rockhill et al., 1998; elevated HR for a clinical diagnosis of depression but more marginally
Rothman, 2012). for the medication indicator.
There was a dose-response relationship between number of CAs and
risk for both depression indicators (Table 3). This graded association

Table 2
Descriptive characteristics of the study population, by sex and depression indicators. Absolute numbers and column percent.

Women Men

Total Clinical diagnosis Medication Total Clinical diagnosis Medication

N (row percent) 232,678 16,502 (7) 40,852 (18) 245,463 10,066 (4) 27,662 (11)
Characteristics
Parental educational level
9 years of education 15,919 (7) 1297 (8) 2842 (7) 16,463 (7) 778 (8) 1935 (7)
10–12 years of education 115,892 (50) 8500 (52) 20,305 (50) 122,349 (50) 5231 (52) 13,765 (50)
> 12 years of education 100,134 (43) 6656 (40) 17,563 (43) 105,924 (43) 4018 (40) 11,877 (43)
Missing 733 (0) 49 (0) 142 (0) 727 (0) 39 (0) 85 (0)
Parental country of birth
Both born in Sweden 194,253 (83) 13,427 (81) 34,179 (84) 204,872 (83) 8163 (81) 22,988 (83)
Mixed 24,332 (10) 2141 (13) 4425 (11) 25,687 (10) 1327 (13) 3016 (11)
Other Nordic 4026 (2) 369 (2) 710 (2) 4357 (2) 236 (2) 486 (2)
EU 3332 (1) 217 (1) 525 (1) 3507 (1) 139 (1) 415 (2)
Non-EU 6735 (3) 348 (2) 1013 (2) 7040 (3) 201 (2) 757 (3)
Childhood adversity
Parental death 5368 (2) 529 (3) 1015 (2) 5752 (2) 353 (4) 658 (2)
Parental substance abuse 7048 (3) 889 (5) 1456 (4) 7362 (3) 643 (6) 932 (3)
Substantial parental criminality 9642 (4) 1085 (7) 1864 (5) 10,178 (4) 749 (7) 1342 (5)
Parental psychiatric morbidity 11,820 (5) 1556 (9) 2433 (6) 12,548 (5) 1099 (11) 1665 (6)
Parental separation 57,915 (25) 5448 (33) 10,577 (26) 60,146 (25) 3468 (34) 7079 (26)
Household living on public assistance 10,537 (5) 1300 (8) 2041 (5) 10,981 (4) 851 (8) 1409 (5)
Child welfare intervention 3032 (1) 519 (3) 648 (2) 3470 (1) 387 (4) 453 (2)
Residential instability 3252 (1) 417 (3) 599 (1) 3202 (1) 245 (2) 383 (1)
Number of childhood adversities
0 158,966 (68) 9358 (57) 27,180 (67) 168,646 (69) 5455 (54) 18,522 (67)
1 51,176 (22) 4451 (27) 9313 (23) 53,183 (22) 2760 (27) 6135 (22)
2 14,636 (6) 1558 (9) 2739 (7) 15,190 (6) 1070 (11) 1897 (7)
3 4839 (2) 645 (4) 944 (2) 5215 (2) 434 (4) 682 (2)
4+ 3061 (1) 9358 (57) 676 (2) 3229 (1) 5455 (54) 426 (2)

97
E. Björkenstam et al. Journal of Affective Disorders 223 (2017) 95–100

Table 3
Associations between childhood adversity and depression indicators in early adulthood. Results from Cox Regression Analyses, presented as hazard ratios (HR) with 95% confidence
intervals (CI).

Clinical diagnosis Medication

Model Ia Model IIb Model IIIc Model Ia Model IIb Model IIIc

Parental death 1.46 (1.36–1.57) 1.36 (1.27–1.47) 1.06 (0.99–1.14) 1.05 (1.01–1.11) 1.05 (1.00–1.10) 1.00 (0.95–1.05)
Parental substance abuse 2.09 (1.98–2.20) 1.97 (1.87–2.08) 1.20 (1.13–1.28) 1.20 (1.15–1.24) 1.19 (1.14–1.24) 1.06 (1.00–1.11)
Substantial parental criminality 1.78 (1.69–1.87) 1.68 (1.60–1.77) 1.16 (1.09–1.22) 1.16 (1.12–1.20) 1.16 (1.12–1.20) 1.07 (1.03–1.12)
Parental psychiatric morbidity 2.11 (2.02–2.20) 2.05 (1.96–2.14) 1.62 (1.55–1.70) 1.22 (1.18–1.26) 1.22 (1.18–1.26) 1.17 (1.13–1.21)
Parental separation 1.58 (1.54–1.62) 1.55 (1.51–1.59) 1.40 (1.36–1.44) 1.08 (1.06–1.10) 1.08 (1.06–1.10) 1.06 (1.04–1.07)
Household receiving public assistance 1.95 (1.86–2.04) 1.88 (1.80–1.98) 1.29 (1.22–1.36) 1.15 (1.11–1.19) 1.16 (1.12–1.20) 1.07 (1.03–1.11)
Child welfare intervention 2.79 (2.60–2.99) 2.57 (2.39–2.76) 1.51 (1.39–1.63) 1.24 (1.17–1.31) 1.23 (1.15–1.30) 1.07 (1.00–1.14)
Residential instability 1.89 (1.74–2.05) 1.81 (1.66–1.97) 1.28 (1.18–1.40) 1.10 (1.03–1.17) 1.09 (1.02–1.16) 1.02 (0.95–1.08)
Total number of childhood adversities
1 1.55 (1.50–1.60) 1.54 (1.49–1.59) 1.08 (1.06–1.10) 1.08 (1.06–1.10)
2 2.01 (1.93–2.10) 1.98 (1.89–2.07) 1.14 (1.11–1.18) 1.15 (1.11–1.18)
3 2.48 (2.32–2.65) 2.42 (2.26–2.59) 1.20 (1.14–1.26) 1.21 (1.15–1.27)
4+ 3.17 (2.94–3.41) 3.05 (2.83–3.29) 1.32 (1.24–1.40) 1.32 (1.25–1.41)

Reference group: no childhood adversity.

a
Adjusted for birth year and sex.
b
Model I with additional adjustments for parental country of birth and parental education.
c
Model II with additional adjustments for all CAs simultaneously.

remained in the multi-adjusted models. For the clinical diagnosis out- Supplementary table 2).
come, the HRs spanned from 1.55 (95% CI 1.50–1.60) for exposure to
one adversity, HR of 2.01 (95% CI 1.93–2.10) for two CAs; HR of 2.48
4. Discussion
(95% CI 2.32–2.65) for three CAs, and finally HR of 3.17 (95% CI
2.94–3.41) for exposure to four or more adversities. For antidepressant
4.1. Main findings
medication utilization, the HRs ranged from 1.08 (95% CI 1.06–1.10)
for one adversity; HR of 1.15 (95% CI 1.11–1.18) for two adversities;
The present study examined the association between childhood
HR of 1.21 (95% CI 1.15–1.27) for three CAs and finally HR of 1.32
adversity and depression in young adulthood, using Swedish data on a
(95% CI 1.25–1.41) for 4+ adversities.
total cohort of 478,141 individuals. Our findings show that accumula-
Estimated population attributable fractions for indicators of de-
tion of childhood adversity is associated with a substantial increase in
pression attributable to CA are presented in Table 4. For the clinical
the risk of depression, both in terms of a clinical diagnosis and in terms
depression outcome, the clinical depression diagnosis that was attri-
of antidepressant medication. The risk of depression grew higher with
butable to parental separation was 11.8% (95% CI 11.0–12.5), and for
increasing number of CAs, and the risk remained after adjustment for
parental psychiatric morbidity 5.2% (95% CI 4.8–5.5). Lower PAF es-
important background variables including parental country of birth and
timates were observed for antidepressant medication. In addition to the
parental education. In our statistical models of estimated PAF, ap-
results presented in Table 4, the clinical diagnosis for depression that
proximately one fifth (18.6%) of all cases of clinical diagnoses of de-
were attributable to experiencing any CA were 18.6% for clinical di-
pression was attributable to experience of CA. However, the similar PAF
agnosis, and 2.7% for medication (data not shown in tables).
figure for having retrieved antidepressant medication was radically
Lastly, we conducted sensitivity analyses in which we excluded
lower (2.7%).
parental psychiatric morbidity from the list of adversities and con-
Our findings indicate a strong positive association between multiple
sidered it as a confounder. In these analyses, the association between
types of childhood adversities and later risk of depression. These as-
CA and depression indicators remained significant (see Supplementary
sociations are similar to those found in prior studies (Anda et al., 2002;
table 1). In the second sensitivity analysis, where we excluded 24,368
Chapman et al., 2004; Dahl et al., 2017; Gilman et al., 2003; Green
individuals whose parents had been treated for a psychiatric disorder,
et al., 2010b; Sareen et al., 2013). The high prevalence of CA in in-
the associations between CA and depression indicators remained (see
dividuals with depression in our study (e.g. 46% in those with a clinical
diagnosis for depression vs. 31% in the total cohort), are in accordance
Table 4 with previous studies (Anda et al., 2002; Chapman et al., 2004; Gilman
Population Attributable Fractions (PAF) with 95% confidence intervals (CI) for depres-
et al., 2003).
sion indicators.
Prior studies in various settings have shown that several of the CA
Childhood adversity Clinical diagnosis Medication indicators s used in our study predict depression in early adulthood
(Anda et al., 2002; Chapman et al., 2004; Dahl et al., 2017; Green et al.,
Parental death 1.02 (0.81–1.23) 0.11 (0.01–0.23)
2010b; Wirback et al., 2014). In particular parental psychiatric mor-
Parental substance abuse 2.84 (2.56–3.12) 0.48 (0.35–0.61)
Substantial parental criminality 2.88 (2.57–3.18) 0.57 (0.4–0.71) bidity and parental substance abuse have been pointed out as key risk
Parental psychiatric morbidity 5.16 (4.80–5.52) 0.93 (0.76–1.1) factors (Anda et al., 2002; Chapman et al., 2004; Green et al., 2010b).
Parental separation 11.77 (11.04–12.51) 1.43 (1.03–1.83) In line with our results, earlier studies have shown that individuals with
Household receiving public assistance 3.76 (3.44–4.10) 1.94 (1.67–2.21) who were in societal care during their childhood constitute a high risk
Child welfare intervention 2.08 (1.87–2.29) 0.25 (0.16–0.34)
Residential instability 1.16 (0.98–1.34) 0.08 (0.00–0.17)
group for depression (Dimigen et al., 1999; Keller et al., 2010;
Total number of childhood Vinnerljung and Hjern, 2014).
adversities The present study further demonstrated that the risk of depression
1 6.80 (6.14–7.46) 0.92 (0.55–1.29) grew higher as the number of adversities increased. A similar dose-re-
2 3.90 (3.53–4.27) 0.56 (0.38–0.75)
sponse relationship has been reported in studies examining retro-
3 2.00 (1.77–2.23) 0.28 (0.17–0.38)
4+ 1.86 (1.66–2.06) 0.30 (0.21–0.38) spectively reported childhood adversities and later depression
(Chapman et al., 2004; Dube et al., 2003; Green et al., 2010a). These

98
E. Björkenstam et al. Journal of Affective Disorders 223 (2017) 95–100

findings stress the importance of simultaneously consider the impact of Although reasons behind child welfare interventions are not available
multiple adversities. in the national register, placements in out-of-home care and respite care
In addition, our findings show that both depression indicators were (both invasive interventions) are usually related to some form of child
attributable to all types of CA. The dose-response pattern between in- maltreatment. Another limitation is that we have not examined the
creasing number of CAs and depression indicators was not found for fluidity of CAs but rather, as done by others (Felitti et al., 1998) treated
PAFs because the size of PAFs partly depends of the prevalence of the them as discrete life events. However, both the consistency of our re-
exposure. Even though the HRs were most elevated for 4+ CAs, the sults with other studies and the large cohort with high quality data
PAFs did not progressively increase, because of the decreased pre- lends confidence to the validity of our findings. Finally, this is an ob-
valence of having experienced an increasing number of CAs. However, servational study with obvious limitations regarding causality issues.
the estimated proportion of depression in terms of clinical diagnosis This is of particular importance when it comes to interpretation of PAF
attributable to increasing number of CAs were substantial. When in- estimates, as the estimation of PAF requires an assumption of causality
terpreting these findings, it is important to bear in mind that, although (Levine, 2007).
the idea with PAF analyses is to tentatively examine the amount of a
disease that theoretically could be avoided if the studied exposures 5. Conclusion
were eliminated, some of the CAs in our study, such as child welfare
intervention and public assistance are indeed intended to ameliorate In conclusion, this study shows that individuals with a history of
the adverse effects of the underlying problems. exposure cumulative adversity are a high-risk group with respect to
There are several possible explanations to the current findings. depression in young adulthood. Given that experience of CA is
Exposure to CA is a major contributor to increased stress levels in common, early and efficient support of disadvantaged children is of
childhood (Shonkoff et al., 2012) and cumulative exposure to childhood great importance for improving their long-term mental health out-
stress may lead to allostatic overload, causing enduring brain dys- comes. Sufficient evidence is already available for governments to
function, which in turn may affect health throughout the lifespan prioritize and invest in early preventative interventions aiming to al-
(Danese and McEwen, 2012; Evans, 2003). There is growing evidence leviate the consequences of early life adversity (see e.g. Ferrer-Wreder
of physiological and anatomical changes in the brains of individuals et al., 2004). Regardless of causality issues, children and youth with a
exposed to CA (Heim and Binder, 2012; Shonkoff et al., 2012; Swartz documented history of multiple CA should be also be regarded as a
et al., 2017). high-risk group for depression by professionals in clinical social and
Many of the effects of CA on children are influenced by social health services that come into contact with this group.
challenges that expose children to cumulative or chronic stress. For
example, studies suggest that CA increases the risk of parental stress, Funding
poorer parenting skills, e.g. greater use of aggressive or punitive dis-
cipline methods and less responsive parenting (Reising et al., 2013). This study was supported by a grant from the Swedish Council for
These circumstances may impair the quality of parent-child relation- Working Life and Social Research (grant number 2013–2729). The work
ships and significantly influence the appearance of depression symp- of professor Vinnerljung and professor Hjern was supported by a grant
toms. from Bank of Sweden Tercentenary Foundation (grant number P10
Lastly, the studied CAs reflect social disadvantage in childhood, 0514). The funders had no role in the analyses interpretation of results
though they may also capture a degree of genetic susceptibility to or the writing of this manuscript.
morbidity. It is well-known that parental psychiatric disorder is a pre-
dictor for offspring psychiatric disorder, including depression Acknowledgements
(Merikangas et al., 2009; Siegenthaler et al., 2012). In an attempt to
take genetic factors into account, we conducted a sensitivity analysis None.
excluding parental psychiatric morbidity from the CAs. The associations
remained strong also in these analyses. Appendix A. Supporting information
With respect to the context of the current study, one would perhaps
expect the impact of CA to be less marked in a Scandinavian context. Supplementary data associated with this article can be found in the
Sweden is known to have a generous universal welfare system for fa- online version at http://dx.doi.org/10.1016/j.jad.2017.07.030.
milies with children, which might ameliorate the negative impact of
adversities on children later mental health morbidity. Nevertheless, all References
indicators of CA were strong correlates of early adulthood depression.
Thus, one might expect that the effect of CA on depression is even Anda, R.F., Whitfield, C.L., Felitti, V.J., Chapman, D., Edwards, V.J., Dube, S.R.,
greater in some other high-income countries. Williamson, D.F., 2002. Adverse childhood experiences, alcoholic parents, and later
risk of alcoholism and depression. Psychiatr. Serv. 53, 1001–1009.
Birmaher, B., Ryan, N.D., Williamson, D.E., Brent, D.A., Kaufman, J., Dahl, R.E., Perel, J.,
4.2. Strengths and limitations Nelson, B., 1996. Childhood and adolescent depression: a review of the past 10 years.
Part I. J. Am. Acad. Child Adolesc. Psychiatry 35, 1427–1439.
Björkenstam, E., Pebley, A.R., Burström, B., Kosidou, K., 2017. Childhood social adversity
This study has several strengths, including the longitudinal popu- and risk of depressive symptoms in adolescence in a US national sample. J. Affect.
lation-based design and use of national registers with high complete- Disord. 212, 56–63.
ness and validity. Previous studies on CA and depression have often Chapman, D.P., Whitfield, C.L., Felitti, V.J., Dube, S.R., Edwards, V.J., Anda, R.F., 2004.
Adverse childhood experiences and the risk of depressive disorders in adulthood. J.
been retrospective and based on self-reported information, entailing
Affect. Disord. 82, 217–225.
risk for recall bias (e.g., underreporting of CA) (Colman et al., 2016; Colman, I., Kingsbury, M., Garad, Y., Zeng, Y., Naicker, K., Patten, S., Jones, P.B., Wild,
Hardt and Rutter, 2004; Reuben et al., 2016)). Despite these strengths, T.C., Thompson, A.H., 2016. Consistency in adult reporting of adverse childhood
experiences. Psychol. Med. 46, 543–549.
our findings should be interpreted in the context of the following lim-
Costello, E.J., Foley, D.L., Angold, A., 2006. 10-year research update review: the epide-
itations: First, the range of childhood adversities is far from exhaustive miology of child and adolescent psychiatric disorders: II. Developmental epide-
and we do not assess the severity, duration, or sequencing of any of miology. J. Am. Acad. Child Adolesc. Psychiatry 45, 8–25.
these adversities. This limitation includes the lack of data on child Dahl, S.K., Larsen, J.T., Petersen, L., Ubbesen, M.B., Mortensen, P.B., Munk-Olsen, T.,
Musliner, K.L., 2017. Early adversity and risk for moderate to severe unipolar de-
maltreatment (abuse and neglect), which has been pointed out as one pressive disorder in adolescence and adulthood: a register-based study of 978,647
strong risk factors for depression. However, substantial child welfare individuals. J. Affect. Disord. 214, 122–129.
intervention may be viewed as a crude proxy for abuse/neglect. Danese, A., McEwen, B.S., 2012. Adverse childhood experiences, allostasis, allostatic

99
E. Björkenstam et al. Journal of Affective Disorders 223 (2017) 95–100

load, and age-related disease. Physiol. Behav. 106, 29–39. Merikangas, K.R., Nakamura, E.F., Kessler, R.C., 2009. Epidemiology of mental disorders
Dimigen, G., Del Priore, C., Butler, S., Evans, S., Ferguson, L., Swan, M., 1999. Psychiatric in children and adolescents. Dialog. Clin. Neurosci. 11, 7–20.
disorder among children at time of entering local authority care: questionnaire Najman, J.M., Hayatbakhsh, M.R., Clavarino, A., Bor, W., O'Callaghan, M.J., Williams,
survey. BMJ 319, 675. G.M., 2010. Family poverty over the early life course and recurrent adolescent and
Dong, M., Anda, R.F., Felitti, V.J., Dube, S.R., Williamson, D.F., Thompson, T.J., Loo, young adult anxiety and depression: a longitudinal study. Am. J. Public Health 100,
C.M., Giles, W.H., 2004. The interrelatedness of multiple forms of childhood abuse, 1719–1723.
neglect, and household dysfunction. Child Abus. Negl. 28, 771–784. Patel, V., Flisher, A.J., Hetrick, S., McGorry, P., 2007. Mental health of young people: a
Dube, S.R., Felitti, V.J., Dong, M., Giles, W.H., Anda, R.F., 2003. The impact of adverse global public-health challenge. Lancet 369, 1302–1313.
childhood experiences on health problems: evidence from four birth cohorts dating Reising, M.M., Watson, K.H., Hardcastle, E.J., Merchant, M.J., Roberts, L., Forehand, R.,
back to 1900. Prev. Med. 37, 268–277. Compas, B.E., 2013. Parental depression and economic disadvantage: the role of
Evans, G.W., 2003. A multimethodological analysis of cumulative risk and allostatic load parenting in associations with internalizing and externalizing symptoms in children
among rural children. Dev. Psychol. 39, 924–933. and adolescents. J. Child Fam. Stud. 22.
Felitti, V.J., Anda, R.F., Nordenberg, D., Williamson, D.F., Spitz, A.M., Edwards, V., Koss, Reuben, A., Moffitt, T.E., Caspi, A., Belsky, D.W., Harrington, H., Schroeder, F., Hogan, S.,
M.P., Marks, J.S., 1998. Relationship of childhood abuse and household dysfunction Ramrakha, S., Poulton, R., Danese, A., 2016. Lest we forget: comparing retrospective
to many of the leading causes of death in adults. The Adverse Childhood Experiences and prospective assessments of adverse childhood experiences in the prediction of
(ACE) Study. Am. J. Prev. Med. 14, 245–258. adult health. J. Child Psychol. Psychiatry 57 (10), 1103–1112.
Ferrer-Wreder, L., Stattin, H., Lorente, C., Tubman, J., Adamson, L., 2004. Successful Rockhill, B., Newman, B., Weinberg, C., 1998. Use and misuse of population attributable
Prevention and Youth Delevopment Programs. Kluwe Academics, New York. fractions. Am. J. Public Health 88, 15–19.
Fleisher, W.P., Katz, L.Y., 2001. Early onset major depressive disorder. Paediatr. Child Rothman, K., 2012. Modern Epidemiology, 3rd ed. Lippincott Williams & Wilkins,
Health 6, 444–448. Philadelphia.
Gilman, S.E., Kawachi, I., Fitzmaurice, G.M., Buka, L., 2003. Socio-economic status, fa- Sareen, J., Henriksen, C.A., Bolton, S.L., Afifi, T.O., Stein, M.B., Asmundson, G.J., 2013.
mily disruption and residential stability in childhood: relation to onset, recurrence Adverse childhood experiences in relation to mood and anxiety disorders in a po-
and remission of major depression. Psychol. Med. 33, 1341–1355. pulation-based sample of active military personnel. Psychol. Med. 43, 73–84.
Gilman, S.E., Kawachi, I., Fitzmaurice, G.M., Buka, S.L., 2002. Socioeconomic status in Shapero, B.G., Steinberg, L., 2013. Emotional reactivity and exposure to household stress
childhood and the lifetime risk of major depression. Int. J. Epidemiol. 31, 359–367. in childhood predict psychological problems in adolescence. J. Youth Adolesc. 42,
Green, J.G., Gruber, M.J., Sampson, N.A., Zaslavsky, A.M., Kessler, R.C., 2010a. 1573–1582.
Improving the K6 short scale to predict serious emotional disturbance in adolescents Shonkoff, J.P., Garner, A.S., Committee on Psychosocial Aspects of, C., Family, H.,
in the USA. Int. J. Methods Psychiatr. Res. 19 (Suppl 1), 23–35. Committee on Early Childhood, A., Dependent, C., Section on, D., Behavioral, P.,
Green, J.G., McLaughlin, K.A., Berglund, P.A., Gruber, M.J., Sampson, N.A., Zaslavsky, 2012. The lifelong effects of early childhood adversity and toxic stress. Pediatrics
A.M., Kessler, R.C., 2010b. Childhood adversities and adult psychiatric disorders in 129, e232–e246.
the national comorbidity survey replication I: associations with first onset of DSM-IV Siegenthaler, E., Munder, T., Egger, M., 2012. Effect of preventive interventions in
disorders. Arch. General. Psychiatry 67, 113–123. mentally ill parents on the mental health of the offspring: systematic review and
Hardt, J., Rutter, M., 2004. Validity of adult retrospective reports of adverse childhood meta-analysis. J. Am. Acad. Child Adolesc. Psychiatry 51 (8–17), e18.
experiences: review of the evidence. J. Child Psychol. Psychiatry, Allied Discip. 45, Swartz, J.R., Hariri, A.R., Williamson, D.E., 2017. An epigenetic mechanism links so-
260–273. cioeconomic status to changes in depression-related brain function in high-risk
Heim, C., Binder, E.B., 2012. Current research trends in early life stress and depression: adolescents. Mol. Psychiatry 22, 209–214.
review of human studies on sensitive periods, gene-environment interactions, and Wadsworth, M., Butterworth, S., 2006. Early life. In: Marmot, M., Wilkinson, R. (Eds.),
epigenetics. Exp. Neurol. 233, 102–111. Social Determinants of Health. Oxford University Press, Great Britain.
Keller, T.E., Salazar, A.M., Courtney, M.E., 2010. Prevalence and timing of diagnosable Weitoft, G.R., Hjern, A., Haglund, B., Rosen, M., 2003. Mortality, severe morbidity, and
mental health, alcohol, and substance use problems among older adolescents in the injury in children living with single parents in Sweden: a population-based study.
child welfare system. Child Youth Serv. Rev. 32, 626–634. Lancet 361, 289–295.
Kessler, R.C., Amminger, G.P., Aguilar-Gaxiola, S., Alonso, J., Lee, S., Ustun, T.B., 2007. Vinnerljung, B., Hjern, A., 2014. Consumption of psychotropic drugs among adults who
Age of onset of mental disorders: a review of recent literature. Curr. Opin. Psychiatry were in societal care during their childhood – a Swedish national cohort study. Nord
20, 359–364. J. Psychiatry 68, 611–619.
Kovacs, M., Joormann, J., Gotlib, I.H., 2008. Emotion (Dys)regulation and links to de- Wirback, T., Möller, J., Larsson, J.O., Galanti, M.R., Engström, K., 2014. Social factors in
pressive disorders. Child Dev. Perspect. 2, 149–155. childhood and risk of depressive symptoms among adolescents–a longitudinal study
Levine, B., 2007. What does the population attributable fraction mean? Prev. Chronic Dis. in Stockholm, Sweden. Int. J. Equity Health 13, 96.
4, 1–5. Wittchen, H.U., Jacobi, F., 2005. Size and burden of mental disorders in Europe–a critical
Ludvigsson, J.F., Otterblad-Olausson, P., Pettersson, B.U., Ekbom, A., 2009. The Swedish review and appraisal of 27 studies. Eur. Neuropsychopharmacol. 15, 357–376.
personal identity number: possibilities and pitfalls in healthcare and medical re- World Health Organization, 2001. World Health Report. Mental Health: New
search. Eur. J. Epidemiol. 24, 659–667. Understanding, New Hope, Geneva, Switzerland.
McLaughlin, K.A., Kubzansky, L.D., Dunn, E.C., Waldinger, R., Vaillant, G., Koenen, K.C., World Health Organization, 2016. Preventing depression in the WHO European Region.
2010. Childhood social environment, emotional reactivity to stress, and mood and Trimbos Institute. Netherlands Institute of Mental Health and Addiction., WHO
anxiety disorders across the life course. Depression Anxiety 27, 1087–1094. Regional Offices for Europe.

100