Scholarly article on topic 'Major depressive disorder and subthreshold depression in prepubertal children from the Danish National Birth Cohort'

Major depressive disorder and subthreshold depression in prepubertal children from the Danish National Birth Cohort Academic research paper on "Clinical medicine"

Share paper
Academic journal
Comprehensive Psychiatry
OECD Field of science

Academic research paper on topic "Major depressive disorder and subthreshold depression in prepubertal children from the Danish National Birth Cohort"

Available online at


Comprehensive Psychiatry 70 (2016) 65 - 76



Major depressive disorder and subthreshold depression in prepubertal children from the Danish National Birth Cohort

Rikke Wesselhoefta'*, Einar R. Heiervangb, Per Kragh-S0rensenc, Merete Juul S0rensend, Niels Bilenberge

aResearch Units of Department of Child and Adolescent Mental Health Odense and E-mental Health Odense, Mental Health Services in the Region of Southern

Denmark, University of Southern Denmark, Denmark bInstitute of Clinical Medicine, University of Oslo and Oslo University Hospital, Oslo, Norway cInstitute of Clinical Research, University of Southern Denmark, Denmark dRegional Centre for Child and Adolescent Psychiatry, Aarhus University Hospital, Risskov, Denmark eDepartment of Child and Adolescent Mental Health Odense, Research Unit (University function), Mental Health Services in the Region of Southern Denmark,

University of Southern Denmark, Denmark


Background: Subthreshold Depression (SD) is an impairing condition that might convert into Major Depressive Disorder (MDD). Still, the characteristics of childhood SD are largely unknown.

Purpose: We aim to examine how SD in children differs from MDD regarding symptom profile, comorbidity, functional impairment and associated life stressors. We will examine the frequency of childhood SD and MDD and compare previous mental health between groups. Basic procedures: This is a two-phase nested case-control study within the Danish National Birth Cohort (DNBC) (N = 4500). Mothers completed the online Development and Well-Being Assessment (DAWBA) regarding their child aged 8-10 years. Main findings: A total of 3421 children participated (response rate 76%); 35 children were diagnosed with MDD and 55 with SD. Anhedonia, irritability and worthlessness/guilt were more frequent in MDD than SD, and anhedonia was rare in non-depressed children (1.2%).

Comorbid anxiety and conduct disorders were equally common in the groups. Children with MDD had higher functional impairment caused by the depressive condition than children with SD, but overall functional impairment was the same.

Life stressors, including maternal depressive symptoms, were equally frequent for children with MDD and SD. Emotional problems and functional impairment at age 7 years predicted later SD and MDD. SD was twice as common as MDD in the DNBC (point prevalence 1.0% vs. 0.5%). Principal conclusions: Children with SD and MDD display striking similarities. They differ mainly by the number of depressive symptoms and depression-related functional impairment.

Both SD and MDD at age 8-10 years were preceded by emotional symptoms and functional impairment at age seven years. This indicates a continuity of childhood depressive conditions but also a window of opportunity for prevention of MDD.

© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (

1. Introduction

Major Depressive Disorder (MDD) is a frequent and distressing mental disorder [1], representing one of the top five contributors to the global burden of disease in Europe and North America [2,3]. Childhood onset MDD shows a high recurrence [4] and adolescents with MDD have a 40% risk of a recurrent depressive episode within two years [5]. Therefore, youth with MDD have a high risk for

* Corresponding author.

multiple depressive episodes during the crucial period of development. MDD is associated with more suicidal attempts if onset occurs in youth compared to adulthood [6]. Furthermore, youth with MDD are seven times more likely to complete suicide than non-depressed youth [7] and suicide is on the top five leading causes of death among adolescents [8,9].

Children and adolescents who have a family history of depression [10] or are exposed to environmental stressors [11] show an increased risk for depression. However, adolescent depression is more genetic in origin compared

0010-440X/© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (

to childhood onset depression [12,13] and environmental stressors play a larger role in triggering childhood depression [12,14]. Also, the prevalence and incidence rates of depression rise prominently in adolescence [15,16], indicating that childhood rather than adolescence is a possible 'window of opportunity' for prevention of MDD [17].

Preventive interventions are recommended to target individuals presenting with subclinical level symptoms (indicated prevention) or individuals exposed to important risk factors (selective prevention) [18]. A subclinical level of depressive symptoms, also called Subthreshold Depression (SD), is a precursor condition to MDD in adults [19,20] and in adolescents [21 -23]. A worldwide study finds that adults with SD and MDD display similar risk factor profiles and health impairment [24], while a recent meta analysis finds the same mortality for the two conditions [25]. Also, a prospective study finds that children with subclinical psychiatric conditions have a risk for adverse outcomes in adulthood that is similar to that of children with clinical psychiatric disorders [26]. The authors recommend that children with subclinical psychiatric conditions are targeted in preventive interventions to reduce future impairment, distress and societal costs [26].

The phenotypic characteristics of childhood SD are however poorly described [27], making it difficult to recognize and intervene towards this condition. The definitions used for SD vary between studies, but the DSM-IV research criteria for minor depressive disorder have been widely accepted [28]. Although MDD and SD differ mainly on the number of depressive symptoms present (> 5 or 2-4, respectively) from the same pool of nine depressive symptoms [28], the possible combinations are multiple. Also, some symptoms may be more associated with threshold than subthreshold depression. The symptom of markedly diminished interest or pleasure (anhedonia) predicts severity, non-response and non-remission in depressed adults [29], as well as longer time to remission in depressed adolescents [30]. Anhedonia may therefore be a 'severity marker' associated with more severe depressive conditions or indicating a severe course. Similarly, depressed individuals with insomnia have an increased risk of suicidal behavior [31,32], and insomnia and irritability are found to be the most treatment resistant symptoms in pediatric MDD [33]. Only three studies have so far compared the symptom profiles for childhood MDD and SD [27]. Our study will contribute to this gap in the literature.

In order to initiate selective preventive interventions towards children at risk for MDD, we need to know more about the environmental stressors that these children experience. Risk factors for depression differ throughout development, but childhood seems to be a particularly sensitive period, where exposure to stressors initiates a developmental trajectory, rendering the individual vulnerable to development of depression also as an adult [34]. The majority of putative risk factors seem to be disorder-specific in youth, and living with a single parent or staying at a foster home is mainly associated with depression [35]. In addition, experiencing Stressful Life Events (SLE) increases the risk

for onset of depression [36] and for recurrent depression [37], while it lowers the response to treatment [38]. When children experience physical health problems like epilepsy, asthma and other chronic illnesses, they have an increased risk for developing depression [39,40]. Similarly, having comorbid or previous anxiety disorders [41], or living with a depressed mother [42], increases the risk for childhood MDD. These stressors are also found to increase the risk for adolescent SD converting into MDD [43]. This suggests that full-blown MDD is likely to develop in children with SD, who are exposed to a certain amount of stressors.

In order to identify (and eventually prevent) these children at risk for developing MDD, we need more knowledge of the phenotypic presentation of childhood SD and the exposure to stressors, before the incidence of depression increases significantly with puberty onset. The aim of this study was to describe childhood SD and MDD in a population-based Danish birth cohort. Focusing on depressive symptoms, functional impairment, comorbid disorders, and associated life stressors, we compared children with SD and MDD with each other and with non-depressed children. Furthermore, we examined the frequency of SD and MDD in the birth cohort, and the predictive value of poor mental health scores at age seven years for depressive disorders at age 8-10 years.

We hypothesized that children with SD would display less depressive symptoms, comorbidity and impairment than children with MDD, but still higher than non-depressed children [27]. We expected that anhedonia would be more frequent in children with MDD than with SD, serving as a severity marker. Based on the literature, we expected SD to be a more frequent condition in children than MDD [27]. Finally, we expected a dose-response relationship between experienced life stressors and depressive symptoms, assuming that the developmental pathways for SD and MDD would be identical, with the important difference that children with MDD would have experienced more cumulative stressors at the time of evaluation.

2. Material and methods

2.1. Participants

More than 100,000 pregnant women were recruited to the Danish National Birth Cohort (DNBC) from 1996 to 2002 [44]. Through interviews and questionnaires, they have repeatedly reported on the physical and psychological wellbeing of themselves and their child. In order to assess a pre-pubertal sample, we included DNBC children born in the years 2000-2003. The study is approved by the Danish Data Protection Agency ( 2010-41-4477).

2.2. Study design and measures

The study had a two-phase case-control design nested within the DNBC (Fig. 1) [45]. Phase 1 was a screening

Danish National Birth Cohort (DNBC), Major Depressive Disorder (MDD), Subthreshold Depression (SD)

Fig. 1. Flowchart.

phase, where mothers responded to a mental health screening questionnaire regarding their child as part of the seven year DNBC follow up assessment (N = 21,906). Phase 2 was a diagnostic assessment phase divided into a pilot study and a main study, where 4500 participants were invited and children with depressive disorders were identified. The pilot study (N = 500) was performed in order to identify optimal screening criteria for depressive disorders. For the main study, a random comparison group was sampled (N = 1500), before a group of screen-positive children (N = 2500) was sampled. This sampling procedure was chosen to minimize selection bias by adhering to the 'basic rules' for control selection outlined by Rothman et al. ("Controls should be selected such that the exposure distribution among them will estimate without bias the exposure distribution in the source population") [45].

The Strengths and Difficulties Questionnaire (SDQ) [46] parent version was included in the DNBC 7-year follow up and served as the Phase 1 screening measure. The SDQ is a brief mental health questionnaire for children aged 4-17 years, covering four problem areas (emotional, hyperactivity/inattention, conduct and peer relationships), one area of strength (prosocial behavior), and distress and social impairment [46,47]. An SDQ total difficulties score is calculated by summing the four problem subscales, and an impact score is calculated by summing the distress and impairment items. The SDQ has been validated in different

cultures, with good psychometric properties [48-50]. A child was considered screen-positive when presenting at age 7 either with an emotional subscale score or a total difficulties score > 90th percentile, or with an emotional subscale score or a total difficulties score > 75th percentile combined with an elevated SDQ impact score.

The Phase 2 assessment included mother reports of a second SDQ and the online version of the Development and Well-Being Assessment (DAWBA) [51] regarding their child. The DAWBA is a diagnostic assessment that covers most current child psychopathology and is validated in several settings [48,52]. It includes structured questions about symptoms and impairment related to DSM-IV criteria [53], and also open-ended questions for qualitative responses. To reduce burden on informants, only DAWBA sections for disorders most commonly reported to be comorbid with depression were assessed in this study: anxiety disorders (separation anxiety, social phobia, specific phobia, generalized anxiety, panic disorder, PTSD, obsessive-compulsive disorder), conduct disorder and oppositional defiant disorder.

Three physicians trained in child and adolescent psychiatry assigned DSM-IV diagnoses after reviewing the full DAWBA information. They were trained and attended regular supervision by one of the authors (ERH), who has documented reliability with the developer of the instrument [48]. Substantial inter-rater agreement was required before the independent rating (see Results) [54,55]. All cases with

potential depressive disorder or complex psychopathology were discussed at consensus meetings. The category SD was applied for children presenting with at least one core depressive symptom and at least one additional depressive symptom, with considerable impairment present for minimum two weeks, but not meeting full MDD DSM-IV criteria. This corresponds to the DSM-IV research criteria for minor depressive disorder, except that previous MDD was not considered an exclusion criterion [28]. From here on, the terms depressed and depressive children refer to the group including both children with MDD and SD.

The DAWBA depression section exhibits skip rules. Hence, informants are initially asked to report on depressive core symptoms within the previous month. If core symptoms are reported, informants are presented the full depression section covering all depressive symptoms, associated impairment, and open-ended questions for more detailed descriptions in their own words.

The DAWBA parent version also includes a background section covering potential life stressors; e.g. poor physical health, school problems, past year stressful life events and family stress. Parent self-report anxiety and depressive symptoms were assessed using the 10-item Everyday Feelings Questionnaire (EFQ) [56]. The EFQ has been validated in both epidemiological [56] and clinical [57] populations. The DAWBA background section was not available in Danish when the pilot study was carried out, which explains variation in some reported results related to stressors.

2.3. Data analysis

The case groups comprised children diagnosed with SD and MDD. If a child from the random sample comparison group had developed a depressive disorder in Phase 2, he/she was transferred to the appropriate case group. Attrition analyses were performed comparing gender and birth year between responders and non-responders.

The exposure odds ratios (ORs) for depressive symptoms, comorbid disorders and gender, were compared in binary logistic regression analyses. Rates of additional depressive symptoms for the SD and MDD groups were compared using Pearson's chi-squared test and Fisher's exact test.

SDQ scores at diagnostic assessment (Phase 2) were treated as exposure variables and dichotomized at the 90th percentile of the random sample. SDQ reported overall functional impairment was compared between depressed and non-depressed children and between children with MDD and SD using binary logistic regression. Multiple logistic regressions were performed, adjusting each SDQ score for the others (exempting total score that depends on the subscale scores).

Functional impairment due to a depressive disorder was measured by the DAWBA. Information on four areas of function (family, friendships, learning/school, and leisure) as well as distress (all coded 0,1,2,3) was obtained from parents, and treated as a sum score (0-15). Comparisons were made between children with SD and MDD using two-sample t-test with equal variance.

Stressful life events (SLEs) were coded as dummy variables and analyzed on a crude level. In separate analyses, the cumulative effect of SLEs was estimated (zero, one, two or more), with no SLE as the baseline. EFQ scores were dichotomized at the 90th percentile of the distribution for mothers of random sample children.

The exposure ORs for life stressors were compared between MDD, SD and non-depressed in binary logistic regression analyses. Furthermore, life stressor sum scores were created and one-way ANOVAs were performed comparing both sum scores and logarithmic transformation of sum scores between groups. Life stressor sum scores were further compared post hoc pairwise using Bonferroni correction for t-test.

The DNBC point prevalence of SD and MDD was calculated based on the frequency of these conditions in the random sample group.

SDQ scores at age seven (Phase 1) were treated as exposure variables and dichotomized at the 90th percentile of the random sample. Comparisons were made for depressed vs. non-depressed groups and for MDD vs. SD, using both unadjusted binary logistic regression, and multiple logistic regressions adjusted for other SDQ scores (except for SDQ total score).

Interrater reliability was analyzed using Fleiss' Kappa [55]. STATA 12 or 13 was used for statistical analyses (www

3. Results

Of the 4500 DNBC mothers invited, 3421 filled out the SDQ and DAWBA online (response rate 76%) (Fig. 1). Participating children had a mean age of 8.98 years (SD 0.75, range 8-10 years) and 52.8% were boys. Attrition analyses showed no difference between responders and non-responders for child gender. However, mothers of children born in 2002 and 2003 were less likely to respond than mothers of children born in 2001.

Of the 1500 random sample mothers invited, 1187 responded (response rate 79%). From this random sample group 6 children met criteria for MDD, while 12 children met criteria for SD. These children were therefore transferred to the respective case groups (leading to a non-depressed group of N = 1169). In total, 35 children were diagnosed with MDD and 55 children were diagnosed with SD.

Interrater reliability between the three DAWBA raters was tested prior to independent DAWBA rating using Fleiss' Kappa (K) [55]. Agreement among raters ranged from substantial for anxiety disorders (K = 0.64) to almost perfect for MDD (K = 0.82) and oppositional/conduct disorders (K = 0.91), according to Landis and Koch [54].

3.1. Depressive symptoms

Table 1 presents the frequency of depressive core symptoms and self-harm symptoms in children with MDD, SD and non-depressed children. All symptoms were more

Frequency of depressive core symptoms, self-harm symptoms, comorbidity and male gender.

MDD SD Non-depressed MDD vs. SD vs. MDD vs. SD

(N = 35) (N = 55) (N = 1169) non-depressed non-depressed

Frequency Frequency Frequency Crude 0R Crude 0R Crude OR

(%) (%) (%)

Core symptoms Depressed mood 33 (94.3) 52 (94.5) 192 (16.4) 84.0 (20.1-352.8)*** 88.2 (27.3-285.3)*** 1.8 (0.2-18.4)

Irritable mood 31 (88.6) 25 (45.5) 129 (11.0) 62.5 (21.7-179.8)*** 6.7 (3.8-11.8)*** 11.1 (2.3-54.3)**

Anhedonia 22 (62.9) 17 (30.9) 14 (1.2) 139.6 (58.8-331.5)*** 36.9 (17.0-80.3)*** 4.0 (1.3-12.3)*

Self-harm symptoms Talked about self-harm 16 (45.7) 19 (34.5) 24 (2.1) 40.1 (18.4-87.3)*** 25.1 (12.6-50.0)*** 1.5 (0.5-4.5)

Newly self-harm 3 (8.6) 4 (7.3) 3 (0.3) 36.4 (7.1-187.2)*** 30.4 (6.6-140.0)*** 3.7 (0.3-43.4)

Ever self-harm 6(17.1) 7 (12.7) 5 (0.4) 48.1 (13.9-166.6)*** 33.9 (10.4-110.7)*** 2.6 (0.5-12.9)

Comorbid disorders Anxiety disorder 8 (22.9) 8 (14.5) 23 (2.0) 14.8 (6.1-36.0)*** 8.5 (3.6-20.0)*** 2.5 (0.6-10.5)

Conduct or Oppositional 8 (22.9) 5 (9.1) 11 (0.9) 31.2 (11.6-83.7)*** 10.5 (3.5-31.5)*** 4.4 (1.0-19.7)a

Defiant Disorder

Gender Male gender 24 (68.6) 36 (65.5) 604 (51.7) 2.0 (1.0-4.2) 1.8 (1.0-3.1)* 0.8 (0.3-2.6)

Logistic regression; *p < 0.05; **p < 0.01; ***p < 0.001, ap = 0.06.

frequent for children with SD or MDD than for non-depressed (p < 0.001). For both depressive groups, depressed mood was the most frequent core symptom, followed by irritability and anhedonia. Anhedonia was very rare in the non-depressed group (1%) but a frequent symptom for children with MDD (63%) and somewhat frequent for SD (31%). When we adjusted for all other clinical features listed in Table 1, anhedonia was the item that best predicted MDD status (adjusted OR 36.92; CI 9.80-139.10; p < 0.001). Using the same approach, depressed mood was the item that best predicted SD status (adjusted OR 34.67; CI 10.17-118.18; p < 0.001).

Comparing SD and MDD, irritable mood (p < 0.01) and anhedonia (p < 0.05) were more common in children with MDD (Table 1). Due to DAWBA skip rules, questions about additional depressive symptoms (appetite/weight, sleep problems, psychomotor agitation or retardation, fatigue/ loss of energy, worthlessness/guilt) were not presented to mothers not confirming any depressive core symptom. Therefore, only 269 of the mothers of non-depressed children reported this information and the frequencies of additional depressive symptoms were only compared for children with SD and MDD. The most frequent of all depressive symptoms for both children with MDD and SD were: depressed mood (MDD 94.3%, SD 94.5%), sleep problems (MDD 91.4%, SD 83.6%) and worthlessness/guilt (MDD 91.4%, SD 72.7%). The frequencies of additional depressive symptoms in children with MDD vs. SD were: appetite/weight changes 46% vs. 42%, sleep problems 91% vs. 84%, psychomotor agitation/retardation 80% vs. 66%, fatigue/loss of energy 74% vs. 60%, poor concentration 86% vs. 73%. Worthlessness/guilt was the only additional depressive symptom that significantly differed between MDD and SD groups, being more frequent in children with MDD (p < 0.05).

Self-harm was frequent in depressed children, with no significant difference between MDD and SD. According to the mothers, 45.7% of children with MDD and 34.5% of

children with SD had talked about self-harm within the last four weeks, and 8.6% of children with MDD and 7.3% with SD had recently self-harmed (Table 1).

3.2. Comorbid disorders

As expected, children with SD or MDD more often met criteria also for anxiety disorders and conduct or oppositional disorders, compared to non-depressed children (p < 0.001) (Table 1). Comparing MDD and SD children, there was a trend towards more conduct or oppositional disorders in the MDD group compared to the SD group (p = 0.06), and this group also had a significantly higher SDQ conduct subscale score (p < 0.01) (Table 2).

3.3. Distress and functional impairment

Overall functional impairment was measured by the SDQ in Phase 2. As expected, children with SD or MDD had higher impact scores compared to non-depressed children, also when adjusted for other SDQ problem areas, but there was no difference between the two depressive groups (Table 2). However, looking at the impairment caused specifically by the depressive symptoms (reported in the DAWBA depression section), the mean impact score was higher for MDD than for SD children (10.25 vs. 7.67; t -4.35; p < 0.0001) at the time of assessment.

3.4. Life stressors

The frequency of various life stressors as outlined in the DAWBA background section was compared between depressed and non-depressed children, and between the two depressive groups (Table 3). The analyses included 60 depressive and 1111 random sample children (due to missing data on the background section).

Children with MDD or SD were more likely to have a poor physical health than non-depressed children (p < 0.001), and children with MDD more frequently suffered from convulsions or epilepsy than non-depressed

a» D CI

■a a

on- dju

iff as

on-N nmea

Dn D ea

8 Al (_J

den- SD entile

non R cen

.n O cer

s. e p vs de h

D rud 0th

I Al s a»

ep D (CI

i Pi no O nti

^ (U £

Du D dju th

S 3 £

M a * 8 o ¡3

rudepth Cru 90t

t^ ^ on -H

-H -H r^ ^ nn CN


(N (N ""sf in -H

^ ^ ^d ^ ^ C^ r^

<N (N -h ^t ^t -H -H


m m oo cn un

© © ^ © © © ©

© 00 © t— O O t—

m ^ 00 un O

nn CN nn ^ -H

I I I I I .7 .2 .9 .0 .4

0. 4. 0. 1. 0.

^ m "i r- ^ m -H

r- o oo m . m -H

<N | -H | 00 | |

I r- I (N I O (N .6 1. .0 1. .6 1. 4.

oo cn -h r^ m CN

un un m m m

Un in © © -sT nn CN

^ ^ 00 2 1 2 8 2 1

© CN CN -H © CN N ^ H O N 00

6. 3. 1.

CN ^ ^ f^ un 6. 4. 2. 5. 8. 2. 2.

-h m m m o

m o m m © ©

oo m m © 0. 8. 0. 6.

m oo m cn oo

^^ l^n

.8 .6 .1 .4 .2 .9 .0

un \o on on r^ rn cK

.7 .2 .2 .1 .8 .9 .9


rn 00 ^ CN ^ © VO ^ ^ VO Un <N

ity vi

le le al

al c le c le al

subs al c subs c bs re

tc subs ial ic su or c

d n sr e so ac p tal ot

o C e P ro P Imp T


. ; (U .01;othe

! & -a

I * (U

I * te

dar .05; dju tan 0. A

(p < 0.01). Both depressive conditions were associated with 'other' serious disease (p < 0.01), and language/speech problems were more frequent in children with SD (p < 0.001) compared to non-depressed children.

Children with MDD or SD were more likely to receive special support in school (p < 0.001) and still, a large proportion of their mothers reported that more school help was needed (p < 0.001). A previous change of schools was more common both for children with MDD and SD (MDD; p < 0.01, SD; p < 0.05). Learning disabilities were more frequent in children with SD compared to non-depressed, an association not observed for the MDD group.

Both depressive groups had more often experienced stressful life events (SLEs) like parental separation or divorce within the past year compared to the non-depressed group (Table 3). Serious disease or hospitalization was only associated with MDD (p < 0.01), while loss of friendship was associated with SD (p < 0.01).

Compared to non-depressed children, children with MDD had more often been exposed to a single SLE (OR 4.76; CI 1.58-14.33; p < 0.01), or to multiple SLEs (OR 11.12; CI 3.57-34.64; p < 0.001) within the previous year. Children with SD did not differ from non-depressed children regarding exposure to a single SLE, but they had more often been exposed to multiple SLEs (OR 4.78; CI 2.16-10.58; p < 0.001) than non-depressed.

Family stressors were more frequent for the depressed groups compared to the non-depressed. Marital problems and poor parent mental health were more often reported by mothers of depressed children (p < 0.01 and p < 0.001, respectively) compared to mothers of non-depressed children. In addition, mothers of children with MDD or SD reported equal and higher levels of anxiety and depressive symptoms on the EFQ scale compared with mothers of non-depressed children (p < 0.001) (Table 3). Yet, the parents of children with MDD were more often unemployed or burdened by financial problems compared to parents of children with SD or no depression.

A life stressor sum score was created based on the stressors listed in Table 3 (health problems, school problems, SLEs, family stresses) and the maternal EFQ total score. There was a statistically significant difference in mean sum scores between groups as determined by one-way ANOVA (F(2,1168) = 58.72, p = 0.000). Life stressor sum score means were; Non-depressed 12.41 (sd 6.43), SD 20.08 (sd 8.37), MDD 24.45 (sd 11.72). Sum score means were also compared pairwise using Bonferroni corrected t-test. All pairs showed a statistically significant difference in means (SD vs. non-depressed p = 0.000; MDD vs. non-depressed p = 0.000; SD vs. MDD p = 0.042).

The life stressor sum score was not completely normally distributed and therefore a logarithmic transformation was performed. The difference between groups remained statistically significant (F(2,1158) = 29.84, p = 0.000). However, after log transformation of the stressor sum score, the pairwise comparison of means between MDD and SD was

Table 3

Frequency and odds ratio of life stressors.

MDD SD Non-depressed MDD vs. SD vs. MDD vs. SD

(N = 22) (N = 38) (N = 1111) non-depressed non-depressed

Frequency Frequency Frequency Crude 0R Crude 0R Crude OR

(%) (%) (%)

Health problems Poor physical health 5 (22.7) 6 (15.6) 29 (2.6) 11.0 (3.8-31.8)*** 7.0 (2.7-18.0)*** 1.6 (0.4-5.9)

Language/speech 1 (4.5) 4 (10.5) 16 (1.4) 3.3 (0.4-25.7) 8.1 (2.6-25.4)*** 0.4 (0.0-3.9)

Vision/hearing 1 (4.5) 0 (0) 27 (2.4) 1.9 (0.3-14.7) NA NA

Movement/ coordination 1 (4.5) 2 (5.3) 10 (0.9) 5.2 (0.6-42.8) 6.1 (1.3-28.9)* 0.9 (0.1-10.0)

Enuresis/encopresis 5 (22.7) 3 (7.9) 103 (9.3) 2.9 (1.0-8.0)* 0.8 (0.3-2.8) 3.4 (0.7-16.1)

Convulsions/epilepsy 2(9.1) 0(0) 13 (1.2) 8.5 (1.8-39.9)** NA NA

Other serious disease 3 (13.6) 5 (13.2) 31 (2.8) 5.5 (1.6-19.6)** 5.3 (1.9-14.4)** 1.0 (0.2-4.9)

School problems Learning disabilities 2(9.1) 15 (39.5) 104 (9.4) 1.0 (0.2-4.2) 6.3 (3.2-12.5)*** 0.2 (0.0-0.8)*

Special class, special 7 (31.8) 12 (31.6) 71 (6.4%) 6.8 (2.7-17.3)*** 6.8 (3.3-14.0)*** 1.0 (0.3-3.1)

school or extra support

School shifts (> 1) 6 (27.3) 6 (15.8) 78 (7.1) 5.0 (1.9-13.1)** 2.5 (1.0-6.1)* 2.0 (0.6-7.2)

More school help needed 9 (40.9) 14 (36.8) 88 (7.9) 8.1 (3.4-19.4)*** 6.8 (3.4-13.6)*** 1.2 (0.4-3.5)

Stressful Life Events (past year) Accident/serious injury 1 (4.5) 0(0) 15 (1.4) 3.5 (0.4-27.6) NA NA

Severe disease/ 3 (13.6) 2 (5.3) 21 (1.9) 8.2 (2.3-29.8)** 2.9 (0.7-12.8) 2.8 (0.4-18.5)


Death (parent/sibling/ 3 (13.6) 2 (5.3) 82 (7.4) 2.0 (0.6-6.8) 0.7 (0.2-3.0) 2.8 (0.4-18.5)


Loss of friendship 4 (18.2) 10 (26.3) 106 (9.5) 2.1 (0.7-6.3) 3.4 (1.6-7.2)** 0.6 (0.2-2.3)

Economic crisis 3 (13.6) 3 (7.9) 55 (5.0) 3.0 (0.9-10.6) 1.7 (0.5-5.5) 1.8 (0.3-10.0)

Separation/divorce 4 (18.2) 7 (18.4) 58 (5.2) 4.0 (1.3-12.3)* 4.1 (1.7-9.7)** 1.0 (0.3-3.8)

Other serious events 10 (45.5) 13 (34.2) 169 (15.2) 4.6 (2.0-10.9)*** 2.9 (1.5-5.8)** 1.6 (0.5-4.7)

Family stresses (current) Parental unemployment 6 (27.3) 2 (5.3) 78 (7.1)a 4.9 (1.9-12.9)** 0.7 (0.2-3.1) 6.8 (1.2-37.1)

Parental work situation 8 (36.4) 15 (39.5) 269 (24.5)b 1.8 (0.3-4.2) 2.0 (1.0-3.9)* 0.9 (0.3-2.6)

Financial problems 10 (45.5) 6 (15.8) 165 (15.0)b 4.7 (2.0-11.1)*** 1.1 (0.4-2.6) 4.4 (1.3-14.9)

Marital problems 9 (40.9) 13 (34.2) 167 (15.2)b 3.9 (1.6-9.2)** 2.9 (1.5-5.8)** 1.3 (0.5-3.9)

Parental physical health 4 (18.2) 8 (21.1) 145 (13.2)b 1.5 (0.5-4.4) 1.8 (0.8-3.9) 0.8 (0.2-3.2)

Sickness (mother/others) 6 (27.3) 15 (39.5) 251 (22.9)b 1.3 (0.5-3.3) 2.2 (1.1-4.3)* 0.6 (0.2-1.8)

Parental psychological health 10 (45.5) 11 (28.9) 93 (8.5)b 9.0 (3.8-21.4)*** 4.4 (2.1-9.2)*** 2.0 (0.7-6.1)

Maternal mood EFQ > 90th percentile 12 (54.5) 13 (35.1) 82 (7.5)c 14.8 (6.2-35.3)*** 6.7 (3.3-13.6)*** 2.2 (0.8-6.5)

Gender Male gender 16 (72.7) 26 (68.4) 566 (50.9) 2.6 (1.0-6.6) 2.1 (1.0-4.2)* 0.8 (0.3-2.6)

EFQ: Everyday Feelings Questionnaire; NA: not applicable due to none affected in the SD group. *p < 0.05; **p < 0.01; ***p < 0.001; aN = 1098; bN = 1097; cN = 1093.

no longer statistically significant different (SD vs. non-depressed p = 0.000; MDD vs. non-depressed p = 0.000; SD vs. MDD p = 0.790).

3.5. Frequency of pre-pubertal depressive disorders in DNBC

Twice as many children from the random sample group (N = 1187) were diagnosed with SD (N = 12) than with MDD (N = 6). This leads to a DNBC SD point prevalence of 1.0% (CI 0.6-1.8) and an MDD point prevalence of 0.5% (CI 0.2-1.1).

3.6. Mental health at age seven

Children who developed MDD or SD at age 8-10 years were more likely to have high impairment (measured by SDQ) already at age seven compared to non-depressed children (Table 4). This association persisted, when adjustments for other SDQ subscale scores were performed (MDD p < 0.001; SD p < 0.01). Similarly, an emotional

subscale score at or above the 90th percentile was associated with later MDD and SD, when controlling for other subscale scores (p < 0.01 and p < 0.001 respectively). In addition, children who later developed SD were more likely than non-depressed children to have a high hyperactivity subscale score at age seven years (p < 0.001). When we compared SDQ scores at age seven years between children who later developed depressive disorders, only the functional impairment score differed between groups (Table 4), being significantly higher for children with later MDD compared to children with later SD (p < 0.05).

4. Discussion

We identified 90 children in the Danish National Birth Cohort with pre-pubertal depressive disorder. Due to the diagnostic criteria, children with MDD clearly presented more depressive symptoms (at least five) than children with SD (at least two and less than five). However, the specific

S ^ g > u ?S

Al Q % §

rv « ^ a

on O o

I Al § o

Jm ^ ü si S


-H (N (N (N m ^ Ö Ö Ö Ö Ö ^

-sf -sf m f^ oo

Ö Ö Ö Ö ^ -sf


in IN m m Ö Ö Ö Ö Ö

rn o <N (N oo

- i S- C- S- C-

Tt in ^ t^ OO-H r-j Ö rn

-H -xf o m o in ON -xf- -xf- Ö ON

(N ^ O f-^ ^ 00

■"sf m oo -sf (N f-^ (N in (N I I I I I

N h N n^t <N ^ <N "sT ^ m-^moooo


rn (N in ö oo ö in

r- -H 00 -H fN f^

CN CN -H f-j CN -H in

00 OO h m in >o

-sf -sf (N (N t-^ ^ -sf

-H ^ h 'O (N M

CN -H -H f-j CN ^

O ""sf «n (N m

en (N (N f-^ <N rn

<s o & ■

S W ü c^ Ph

•ö o o

0 v §

„ & -ö

* aj -a * S

'S v &

depressive symptoms actually presented by children with MDD and SD were remarkably similar. In both groups, depressed mood was the most frequent core symptom, followed by irritability and anhedonia. When looking at all depressive symptoms, the order of the most frequent symptoms was also identical, and self-harm was as common in both depressive groups

Still, three depressive symptoms were more frequent in children with MDD than children with SD: anhedonia, irritability and worthlessness/guilt. Anhedonia was the core symptom best predicting MDD status, supporting that anhedonia is a marker of depression severity also in this age group, as observed in preschool children [58], adolescents [30] and adults [29] . Also, anhedonia, depressed mood and worthlessness/guilt are the depressive symptoms most predictive of conversion of adolescent SD into MDD [21,59]. In our study, depressed mood was the symptom best predicting SD status, whereas worthlessness/guilt was significantly higher in children with MDD than SD Irritability was significantly more common in MDD than SD, but it was a quite common symptom in the non-depressed population (11. 0%), rendering anhedonia a more efficient predictive marker (1. 2% in the non-depressed population) If a child displays diminished interest and pleasure, it should initiate further examination, because the child might have MDD or be at high risk for developing MDD in the near future

Anxiety disorders and oppositional/conduct disorders were equally frequent comorbid conditions in children with SD and MDD . Compared to SD, there was a trend for more frequent conduct and oppositional disorders in MDD, which was supported by a significantly higher SDQ conduct subscale score in the MDD group Some studies have reported similar levels of comorbid disorders for childhood SD and MDD [60,61], while others have found less comorbid disorders in SD than in MDD [62,63].

Overall functional impairment measured by the SDQ was high in both children with SD and MDD and did not differ between depressive groups . This is an important finding, underlining the severity of SD and its comorbid conditions and the influence on everyday functioning of the child However, distress and functional impairment caused by the depressive condition reported on DAWBA were significantly higher for children with MDD than with SD . This suggests that functional impairment due to a childhood depressive disorder could follow a dimensional pattern of severity, where SD is placed between MDD and non-depressed . This is equivalent to the dose-response relationship between depressive condition severity and functional impairment level observed in adults [64] . Lewinsohn and colleagues also showed that the level of psychosocial impairment increased as a direct function of the number of depressive symptoms in adolescents [65]

We found that children with MDD and SD experience more life stressors than non-depressed children, when looking at life stressors individually and as a sum score

However, interpretation of these results should be conservative, considering the small sample sizes in our depressed groups. In line with previous studies poor physical health and epilepsy were associated with depressive disorders [39,40]. Children with depressive disorders had more special education at schools, although their mothers still reported an unmet need for school support. However, due to the cross-sectional design of the study, it cannot be ruled out that school problems could be a consequence of the depressive condition, rather than a risk factor for development of depression. This is a topic for future longitudinal risk factor studies. Children with SD had more learning disabilities than non-depressed children and also compared to children with MDD. This finding is supported by a meta-analysis indicating that learning disabilities cause depressive symptoms, but not MDD [66]. It is possible that children with learning disabilities develop emotional symptoms due to a challenging school situation that are not prone to further development into MDD.

The experience of stressful life events (SLEs) was more common for children with depressive disorders than non-depressed. Children with MDD had more often experienced severe disease or hospitalization, whereas children with SD had more often experienced loss of a friendship. Both depressive groups were more exposed to parental divorce or separation in the previous year compared to non-depressed children, which is in line with studies of adolescents demonstrating that parental divorce or paternal absence increases the risk for depressive symptoms [35,67,68]. The exposure to SLEs indicated a dose-response relationship, where an association with multiple SLEs was observed for children with SD, but at a lower level than the association for children with MDD. However, the summed experience of life stressors did not significantly differ between children with MDD and SD. Therefore our hypothesis that a dose-response relationship would exist between experienced life stressors and depressive symptoms could not be confirmed. Still, there were trends in this direction and testing the hypothesis on larger depressed sample sizes is recommended.

Families of depressed children were characterized by more marital problems, poor parental psychological health, unemployment, and financial problems than families of non-depressed children. Given that our comparison group was a random sample (children with hyperactivity, anxiety and conduct disorder remained in the group), this is reliable and important information. The mothers of children with SD and MDD had higher levels of anxious and depressive symptoms according to the EFQ, compared with mothers of non-depressed children. Although we cannot draw causal implications from the present cross-sectional design, several studies find that children with clinically depressed first-degree relatives have a higher risk of developing SD [69] and MDD [10]. Our study might also reflect that children seem to display depressive symptoms parallel to depressive symptoms displayed by their mother [42]. However, it is also possible that shared risk factors increase

the risk for depression in both parents and children, or that the burden of caring for depressed children acts as a stressor and increases internalizing symptoms in parents. This should be explored further when working with children and families clinically.

In the DNBC random sample, we found an MDD point prevalence of 0.5% and SD point prevalence of 1.0. The prevalence of MDD and SD in pre-pubertal children is rarely compared but generally, studies find that SD is more prevalent than MDD [27]. A study of 9-year-old girls reported point prevalence estimates of 1.8% (MDD) vs. 2.5% (SD) [70]. As opposed to our study, this study included child reports at clinical assessment, which might explain the higher rates [71]. Two studies that both used the DAWBA but focused on all psychiatric disorders reported very low point prevalence rates in this age group of MDD and SD in Great Britain (0.27% vs. 0.07%) [52] and Norway (0.07% vs. 0.11%) [48]. A Brazilian study using DAWBA assessed 6-year-olds and showed a 1.3% prevalence of depressive disorders, but it was not specified how major and minor depression was defined [72]. Cohort participants usually present higher socio-economic status and health status than non-participants [45], which has also been demonstrated for the DNBC [73]. We therefore believe that the point prevalence rates reported in this study are likely to represent an underestimation of the Danish population.

Children diagnosed with MDD or SD at age 8-10 years showed functional impairment and emotional symptoms already at age seven. Functional impairment is a significant risk factor for depression also in adolescents, where Lewinsohn et al. showed that 19.4% of study participants with functional impairment developed MDD vs. 7.3% of those without functional impairment [65]. This indicates that emotional symptoms are rather persistent in children, and that early identification of children with functional impairment and emotional symptoms may represent an important window of opportunity for prevention of MDD.

4.1. Strengths and limitations

This study collected data from a reasonably large population-based sample of almost 3500 pre-pubertal children. Large-scale population-based studies are necessary in order to address subclinical conditions of rare disorders. The response rate of 76% was good compared to other Nordic population-based studies [48,74]. The comparison group was sampled randomly prior to sampling of children who were screen-positive for psychiatric symptoms, making it representative to the source population [45]. Consequently, our comparison group includes children with non-depressive psychiatric disorders like externalizing disorders and anxiety disorders as opposed to a comparison group of children that are screen-negative for psychiatric symptoms. Our findings are therefore conservative, yet reliable.

The use of parent information only is a limitation that might lead to an underestimation of the prevalence, because

children and adolescents usually report more internalizing symptoms with superior predictive value compared to their parents [71]. Another limitation is that mothers with depressive symptoms could falsely report increased depressive or behavioral symptoms about their child due to their own condition, although maternal reporting bias is estimated to be smaller for general population samples [75,76]. The association could also be in the opposite direction where a depressive child causes emotional symptoms and distress in the mother. Finally, the statistical comparisons of life stressors are numerous and include cells with few cases. Therefore conclusions about specific risk factors should be cautious.

5. Conclusion

Studies of adolescents and adults show that subjects with SD have an elevated risk for developing MDD. Our study gives no reason to believe that this would be any different in children.

We present the phenotypic characteristics of children with SD and MDD. The similarities ofthese two groups are striking concerning depressive symptomatology, frequency of self-harm symptoms, comorbidity, experienced life stressors and overall functional impairment. In fact, the number of depressive symptoms and the level of depression-related functional impairment represent the main differences between childhood SD and MDD.

Anhedonia stands out as being very rare in non-depressed children and it is a severity marker that should always alert the clinician. Furthermore, children with SD and MDD experience multiple life stressors, and we should keep the possibility of reducing these in mind for treatment purposes (MDD) as well as selective prevention purposes (SD).

Children with MDD and SD at age 8-10 years show emotional symptoms and resulting impairment already at age seven. Furthermore, the incidence rates of depression rise sharply after puberty. We therefore believe that childhood would be an important time window to identify subjects with SD and target them with indicated interventions aiming to prevent the deterioration of their condition.


This study was funded by grants from the Psychiatric Research Fund of the Region of Southern Denmark, the Lundbeck Foundation and the University of Southern Denmark. The funders had no involvement in any aspect of the study.

The Danish National Research Foundation has established the Danish Epidemiology Science Centre that initiated and created the Danish National Birth Cohort. The cohort is furthermore a result of a major grant from this Foundation. Additional support for the Danish National Birth Cohort is obtained from the Pharmacy Foundation, the Egmont

Foundation, the March of Dimes Birth Defects Foundation, the Augustinus Foundation, and the Health Foundation.

The DNBC 7-year follow-up is supported by the Lundbeck Foundation (195/04) and the Danish Medical Research Council (SSVF 0646).

The research secretaries Tina Ravn and Bente Anthony are acknowledged for their extraordinary help and support with the conduction of the study.

The child and adolescent psychiatrists Claudia Drechsler Christensen, Sanne Kloppenborg and Morten 0rnstrup from the Mental Health Services in the Region of Southern Denmark contributed significantly by rating the DAWBA interviews of the study.

David Gyllenberg, University of Turku, is acknowledged for his qualified feed back on a previous draft of this paper.

Inger Kristine Meder at the Danish National Birth Cohort, Statens Serum Institut, Copenhagen, is acknowledged for qualified help with the logistics.

It is due to the unique and on-going contribution of children and mothers of the Danish National Birth Cohort that this study has been achievable.


[1] Ferrari AJ, Somerville AJ, Baxter AJ, Norman R, Patten SB, Vos T, et al. Global variation in the prevalence and incidence of major depressive disorder: a systematic review of the epidemiological literature. Psychol Med 2013;43:471-81.

[2] World Health Organization. The global burden of disease: 2004 update. Geneva: World Health Organization; 2008.

[3] Murray CJ, Vos T, Lozano R, Naghavi M, Flaxman AD, Michaud C, et al. Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990-2010: a systematic analysis for the global burden of disease study 2010. Lancet 2012;380:2197-223.

[4] Kovacs M. Presentation and course of major depressive disorder during childhood and later years of the life span. J Am Acad Child Adolesc Psychiatry 1996;35:705-15.

[5] Birmaher B, Ryan ND, Williamson DE, Brent DA, Kaufman J, Dahl RE, et al. Childhood and adolescent depression: a review of the past 10 years. Part I. J Am Acad Child Adolesc Psychiatry 1996;35: 1427-39.

[6] van Noorden M, Minkenberg S, Giltay E, den Hollander-Gijsman M, van der Wee N, Zitman F. Pre-adult versus adult onset major depressive disorder in a naturalistic patient sample: the Leiden routine outcome monitoring study. Psychol Med 2011;41:1407-17.

[7] Gould MS, King R, Greenwald S, Fisher P, Schwab-Stone M, Kramer R, et al. Psychopathology associated with suicidal ideation and attempts among children and adolescents. J Am Acad Child Adolesc Psychiatry 1998;37:915-23.

[8] World Health Organization. Health for the world's adolescents: a second chance in the second decade: summary; 2014.

[9] Minino AM, Murphy SL. Death in the United States, 2010. NCHS Data Brief 2012;99:1-8.

[10] Weissman MM, Wickramaratne P, Nomura Y, Warner V, Pilowsky D, Verdeli H. Offspring of depressed parents: 20 years later. Psychiatry 2006;163:1001-8.

[11] Mayer L, Lopez-Duran NL, Kovacs M, George CJ, Baji I, Kapornai K, et al. Stressful life events in a clinical sample of depressed children in Hungary. J Affect Disord 2009;115:207-14.

[12] Thapar A, McGuffin P. A twin study of depressive symptoms in childhood. Psychiatry 1994;165:259-65.

[13] Rice F, Harold G, Thapar A. The genetic aetiology of childhood [34 depression: a review. J Child Psychol Psychiatry 2002;43:65-79.

[14] Harrington R. Adolescent depression: same or different? Arch Gen Psychiatry 2001;58:21-2. [35

[15] Wesselhoeft R, Pedersen CB, Mortensen PB, Mors O, Bilenberg N. Gender-age interaction in incidence rates of childhood emotional disorders. Psychol Med 2015;45:829-39. [36

[16] Costello EJ, Copeland W, Angold A. Trends in psychopathology across the adolescent years: what changes when children become [37 adolescents, and when adolescents become adults? J Child Psychol Psychiatry Allied Discip 2011;52:1015-25 .

[17] Merry SN, Hetrick SE, Cox GR, Brudevold-Iversen T, Bir JJ, McDowell H . Psychological and educational interventions for [38 preventing depression in children and adolescents Cochrane Database Syst Rev 2011:CD003380.

[18] Munoz RF, Cuijpers P, Smit F, Barrera AZ, Leykin Y. Prevention of [39 major depression. Annu Rev Clin Psychol 2010;6:181-212.

[19] Rodriguez MR, Nuevo R, Chatteiji S, Ayuso-Mateos JL. Definitions and factors associated with subthreshold depressive conditions: a [40 systematic review. BMC Psychiatry 2012;12:181.

[20] Cuijpers P, Smit F Subthreshold depression as a risk indicator for

major depressive disorder: a systematic review of prospective studies . [41

Acta Psychiatr Scand 2004;109:325-31.

[21] Pine DS, Cohen E, Cohen P, Brook J. Adolescent depressive symptoms as predictors of adult depression: moodiness or mood [42 disorder? Psychiatry 1999;156:133-5.

[22] Hill RM, Pettit JW, Lewinsohn PM, Seeley JR, Klein DN. Escalation

to major depressive disorder among adolescents with subthreshold [43

depressive symptoms: evidence of distinct subgroups at risk J Affect Disord 2014;158:133-8.

[23] Lewinsohn PM, Shankman SA, Gau JM, Klein DN The prevalence [44 and co-morbidity of subthreshold psychiatric conditions Psychol Med 2004;34:613-22.

[24] Ayuso-Mateos JL, Nuevo R, Verdes E, Naidoo N, Chatterji S . From [45 depressive symptoms to depressive disorders: the relevance of thresholds. Psychiatry 2010;196:365-71. [46

[25] Cuijpers P, Vogelzangs N, Twisk J, Kleiboer A, Li J, Penninx BW Differential mortality rates in major and subthreshold depression: [47 meta-analysis of studies that measured both. Psychiatry 2013;202: 22-7.

[26] Copeland WE, Wolke D, Shanahan L, Costello EJ. Adult functional [48 outcomes of common childhood psychiatric problems: a prospective, longitudinal study. JAMA Psychiatry 2015;72:892-9.

[27] Wesselhoeft R, Sorensen MJ, Heiervang ER, Bilenberg N. Subthresh-old depression in children and adolescents—a systematic review. J [49 Affect Disord 2013.

[28] American Psychiatric Association. Diagnostic and statistical manual of mental disorders, fourth edition, text revision. Washington, DC: [50 American Psychiatric Association; 2000.

[29] Vrieze E, Demyttenaere K, Bruffaerts R, Hermans D, Pizzagalli

DA, Sienaert P, et al. Dimensions in major depressive disorder and [51

their relevance for treatment outcome. J Affect Disord 2014;155: 35-41.

[30] McMakin DL, Olino TM, Porta G, Dietz LJ, Emslie G, Clarke G, et al. Anhedonia predicts poorer recovery among youth with selective [52 serotonin reuptake inhibitor treatment-resistant depression. J Am Acad Child Adolesc Psychiatry 2012;51:404-11.

[31] McCall WV, Blocker JN, D'Agostino Jr R, Kimball J, Boggs N, [53 Lasater B, et al. Insomnia severity is an indicator of suicidal ideation during a depression clinical trial. Sleep Med 2010;11:822-7.

[32] Bernert RA, Kim JS, Iwata NG, Perlis ML. Sleep disturbances as an [54 evidence-based suicide risk factor. Curr Psychiatry Rep 2015;17: 554. [55

[33] Tao R, Emslie GJ, Mayes TL, Nakonezny PA, Kennard BD. Symptom improvement and residual symptoms during acute antidepressant [56 treatment in pediatric major depressive disorder. J Child Adolesc Psychopharmacol 2010;20:423-30.

Shanahan L, Copeland WE, Costello EJ, Angold A. Child-, adolescent-and young adult-onset depressions: differential risk factors in development? Psychol Med 2011;41:2265-74. Shanahan L, Copeland W, Costello EJ, Angold A. Specificity of putative psychosocial risk factors for psychiatric disorders in children and adolescents. J Child Psychol Psychiatry Allied Discip 2008;49:34-42. Kessler RC. The effects of stressful life events on depression. Annu Rev Psychol 1997;48:191-214.

Harkness KL, Theriault JE, Stewart JG, Bagby RM. Acute and chronic stress exposure predicts 1-year recurrence in adult outpatients with residual depression symptoms following response to treatment. Depress Anxiety 2014;31:1-8.

Bulmash E, Harkness KL, Stewart JG, Bagby RM. Personality, stressful life events, and treatment response in major depression. J Consult Clin Psychol 2009;77:1067-77.

Ekinci O, Titus JB, Rodopman AA, Berkem M, Trevathan E. Depression and anxiety in children and adolescents with epilepsy: prevalence, risk factors, and treatment. Epilepsy Behav 2009;14:8-18. Hysing M, Elgen I, Gillberg C, Lundervold AJ. Emotional and behavioural problems in subgroups of children with chronic illness: results from a large-scale population study. Child Care Health Dev 2009;35:527-33. Sorensen MJ, Nissen JB, Mors O, Thomsen PH. Age and gender differences in depressive symptomatology and comorbidity: an incident sample ofpsychiatrically admitted children. J Affect Disord 2005;84:85-91. Weissman MM, Wickramaratne P, Pilowsky DJ, Poh E, Hernandez M, Batten LA, et al. The effects on children of depressed mothers' remission and relapse over 9 months. Psychol Med 2014;44:2811-24. Klein DN, Shankman SA, Lewinsohn PM, Seeley JR. Subthreshold depressive disorder in adolescents: predictors of escalation to full-syndrome depressive disorders. J Am Acad Child Adolesc Psychiatry 2009;48:703-10. Olsen J, Melbye M, Olsen SF, Sorensen TI, Aaby P, Andersen AM, et al. The Danish National Birth Cohort—its background, structure and aim. Public Health 2001;29:300-7.

Rothman KJ, Greenland S, Lash TL. Modern epidemiology. Lippincott: Williams & Wilkins; 2008128-47.

Goodman R. Psychometric properties of the strengths and difficulties questionnaire. J Am Acad Child Adolesc Psychiatry 2001;40:1337-45. Goodman R, Scott S. Comparing the strengths and difficulties questionnaire and the child behavior checklist: is small beautiful? J Abnorm Child Psychol 1999;27:17-24.

Heiervang E, Stormark KM, Lundervold AJ, Heimann M, Goodman R, Posserud MB, et al. Psychiatric disorders in Norwegian 8- to 10-year-olds: an epidemiological survey of prevalence, risk factors, and service use. J Am Acad Child Adolesc Psychiatry 2007;46:438-47. Bourdon KH, Goodman R, Rae DS, Simpson G, Koretz DS. The strengths and difficulties questionnaire: U.S. normative data and psychometric properties. J Am Acad Child Adolesc Psychiatry 2005;44:557-64. Goodman A, Goodman R. Strengths and difficulties questionnaire scores and mental health in looked after children. Psychiatry 2012; 200:426-7.

Goodman R, Ford T, Richards H, Gatward R, Meltzer H. The development and well-being assessment: description and initial validation of an integrated assessment of child and adolescent psychopathol-ogy. J Child Psychol Psychiatry Allied Discip 2000;41:645-55. Ford T, Goodman R, Meltzer H. The British child and adolescent mental health survey 1999: the prevalence of DSM-IV disorders. J Am Acad Child Adolesc Psychiatry 2003;42:1203-11. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington, DC: American Psychiatric Association; 1995.

Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-74.

Fleiss JL. Measuring nominal scale agreement among many raters. Psychol Bull 1971;76:378-82.

Uher R, Goodman R. The everyday feeling questionnaire: the structure and validation of a measure of general psychological well-being and distress. Soc Psychiatry Psychiatr Epidemiol 2010;45:413-23.

[57] Mann J, Henley W, O'Mahen H, Ford T. The reliability and validity of the everyday feelings questionnaire in a clinical population. J Affect Disord 2012;148:406-10.

[58] Luby JL, Mrakotsky C, Heffelfinger A, Brown K, Spitznagel E. Characteristics of depressed preschoolers with and without anhedonia: evidence for a melancholic depressive subtype in young children. Psychiatry 2004;161:1998-2004.

[59] Georgiades K, Lewinsohn PM, Monroe SM, Seeley JR. Major depressive disorder in adolescence: the role of subthreshold symptoms. J Am Acad Child Adolesc Psychiatry 2006;45:936-44.

[60] Goodyer I, Cooper PJ. A community study of depression in adolescent girls. II: the clinical features of identified disorder. Psychiatry 1993; 163:374-80.

[61] Gonzalez-Tejera G, Canino G, Ramirez R, Chavez L, Shrout P, Bird H, et al. Examining minor and major depression in adolescents. J Child Psychol Psychiatry Allied Discip 2005;46:888-99.

[62] Sihvola E, Keski-Rahkonen A, Dick DM, Pulkkinen L, Rose RJ, Marttunen M, et al. Minor depression in adolescence: phenomenology and clinical correlates. J Affect Disord 2007;97:211-8.

[63] Jonsson U, Bohman H, von Knorring L, Olsson G, Paaren A, von Knorring AL. Mental health outcome of long-term and episodic adolescent depression: 15-year follow-up of a community sample. J Affect Disord 2011;130:395-404.

[64] Cuijpers P. Minor depression: risk profiles, functional disability, health care use and risk of developing major depression. J Affect Disord 2004;79:71-9.

[65] Lewinsohn PM, Rohde P, Seeley JR. Major depressive disorder in older adolescents: prevalence, risk factors, and clinical implications. Clin Psychol Rev 1998;18:765-94.

[66] Maag JW, Reid R. Depression among students with learning disabilities: assessing the risk. J Learn Disabil 2006;39:3-10.

[67] Culpin I, Heron J, Araya R, Melotti R, Joinson C. Father absence and depressive symptoms in adolescence: findings from a UK cohort. Psychol Med 2013:1-12.

[68] Storksen I, Roysamb E, Moum T, Tambs K. Adolescents with a childhood experience of parental divorce: a longitudinal study of mental health and adjustment. J Adolesc 2005;28:725-39.

[69] Lewinsohn PM, Klein DN, Durbin EC, Seeley JR, Rohde P. Family study of subthreshold depressive symptoms: risk factor for MDD? J Affect Disord 2003;77:149-57.

[70] Kashani JH, McGee RO, Clarkson SE, Anderson JC, Walton LA, Williams S, et al. Depression in a sample of 9-year-old children, prevalence and associated characteristics. Arch Gen Psychiatry 1983; 40:1217-23.

[71] Rothen S, Vandeleur CL, Lustenberger Y, Jeanpretre N, Ayer E, Gamma F, et al. Parent-child agreement and prevalence estimates of diagnoses in childhood: direct interview versus family history method. Methods Psychiatr Res 2009;18:96-109.

[72] Petresco S, Anselmi L, Santos IS, Barros AJ, Fleitlich-Bilyk B, Barros FC, et al. Prevalence and comorbidity of psychiatric disorders among 6-year-old children: 2004 Pelotas birth cohort. Soc Psychiatry Psychiatr Epidemiol 2014;49:975-83.

[73] Jacobsen TN, Nohr EA, Frydenberg M. Selection by socioeconomic factors into the Danish National Birth Cohort. Epidemiol 2010;25: 349-55.

[74] Rask CU, Elberling H, Skovgaard AM, Thomsen PH, Fink P. Parental-reported health anxiety symptoms in 5- to 7-year-old children: the Copenhagen Child Cohort CCC 2000. Psychosomatics 2012;53: 58-67.

[75] Najman JM, Williams GM, Nikles J, Spence S, Bor W, O'Callaghan M, et al. Mothers' mental illness and child behavior problems: cause-effect association or observation bias? J Am Acad Child Adolesc Psychiatry 2000;39:592-602.

[76] van der Toorn SL, Huizink AC, Utens EM, Verhulst FC, Ormel J, Ferdinand RF. Maternal depressive symptoms, and not anxiety symptoms, are associated with positive mother-child reporting discrepancies of internalizing problems in children: a report on the TRAILS study. Eur Child Adolesc Psychiatry 2010;19:379-88.