571

Stressful Life Events in Anxious

and Depressed Children
Douglas E. Williamson, Ph.D., Boris Birmaher, M.D., Ronald E. Dahl, M.D.,
and Neal D. Ryan, M.D.
SUMMARY
Objective: The aim of this study was to examine the occurrence of stressful life events in anx-
ious and depressed children.
Method: Children (612 years of age) with an anxiety disorder (n = 20), depression (n = 45),
and normal controls (n = 11) were assessed using the Life Events Record. Cortisol was as-
sessed from plasma samples collected every 20 minutes from an indwelling catheter in the
one and two hour window around sleep onset.
Results: Depressed children had significantly more events and events that were most likely
independent of the childs behavior, compared to both the anxious and normal control chil-
dren. Independent loss events were significantly more prevalent among the depressed chil-
dren in the preceding year, compared to anxious children, with a trend toward more loss
events compared to the normal controls. For both overall events and independent events, de-
pressed females were significantly more likely to be exposed to stressful environments com-
pared to anxious and normal control females. There were no effects of stress on cortisol
secretion around sleep onset.
Conclusions: The results of this study suggest that stressful life events are significantly
more likely to occur in depressed children, particularly females, compared to anxious chil-
dren, and that these events are predominantly characterized by independent events outside
of the childs control. The results also suggest that loss events may be specific for depression
in children. Interestingly, stress does not apper to impact the HPA axis in children, which is
true for anxious, depressed, and normal control children. The temporal occurrence and sever-
ity, as well as the type of stressful life events as they relate to the onset and maintenance of
anxiety and depression in children, need to be more fully explored.
JOURNAL OF CHILD AND ADOLESCENT PSYCHOPHARMACOLOGY
Volume 15, Number 4, 2005
Mary Ann Liebert, Inc.
Pp. 571580
INTRODUCTION
A
NXIETY DISORDERS are frequently comorbid
with depression and often precede the
onset of a depressive episode particularly
early in development (Birmaher et al. 1996).
Population and offspring studies have shown
that the peak age at onset for anxiety occurs
early in childhood and the peak age-at-onset
for depression occurs in mid-adolescence
(Burke et al. 1990; Weissman et al. 1997). More-
over, recent reports have shown that anxious
or withdrawn behaviors expressed in child-
hood predict the development of depression
Department of Psychiatry, University of Pittsburgh School of Medicine, Western Psychiatric Institute and Clinic,
Pittsburgh, Pennsylvania.
Dr. Joaquim (Kim) Puig-Antich was responsible for the design and implementation of this study prior to his un-
timely death in December 1989. This paper is dedicated to his memory.
This research was supported by grants P01-MH41712 (Dr. Ryan) and K01-MH001957 (Dr. Williamson) from the Na-
tional Institute of Mental Health.
14104C07.pgs 9/16/05 1:31 PM Page 571
572 WILLIAMSON ET AL.
during adolescence (Goodwin et al. 2004). In
light of the significant overlap between anxi-
ety and depression and the posited shared ge-
netic pathway between these two disorders
(Kendler 1996) and the likelihood that signifi-
cant gene-environment interactions play im-
portant etiological roles (Rutter et al. 2001), it
is crucial that the common and divergent
stressful environments associated with these
disorders be better understood.
At least one prospective epidemiological
sample has been reported showing that the
total number of stressful life events during ado-
lescence predict future episodes of depression
and anxiety in young adulthood independent
of initial anxiety and depressive symptoms
(Pine et al. 2002). Studies based on clinical
samples have reported that stressful life events
are significantly elevated in anxious and de-
pressed youths and have suggested that cer-
tain types of events may be specific for each.
One study reported that depression was asso-
ciated with a greater number of overall life
events, compared to a control group of nonde-
pressed, maladjusted children (Berney et al.
1991). Analyses of specific events showed that
events consisting of changes in social relation-
ships appeared to be specific for depression
(Berney et al. 1991). In a study of clinic-
referred children, Goodyer et al. (1985) reported
a significant elevation of moderate to severe
life events (rated on a 4-point scale) in children
and adolescents with conduct disorder, mild
emotional disorder (characterized predomi-
nantly by anxiety symptoms), severe emo-
tional disorder (characterized predominantly
by symptoms of depression), and somatic dis-
orders compared to normal controls. Interest-
ingly, analyses of specific domains of events
revealed that the conduct and mild emotional
disordered youths had significantly elevated
rates of family focused events while the severe
emotional and somatic disordered youths
had significantly elevated rates of exit events
(e.g., parental divorce, death of close relative)
(Goodyer et al. 1985). Consistent with finding
of increased exit events, studies of adults have
reported that stressful life events characterized
by loss (e.g., death) or ending important love
relationships appear to be specific for depres-
sion (Brown and Harris 1978, 1989; Brown et
al. 1986) while events signaling a potential loss
(danger events) appear to be more specific
for anxiety (Finlay-Jones and Brown 1981;
Brown 1993). Consistent with the findings
from adults, increased symptoms of anxiety
have been associated with high-threat events
and increased symptoms of depression with
loss events among children (Eley and Steven-
son 2000).
Other investigations have failed to find either
an elevation of stressful life events or specific
types of events between anxious and de-
pressed youths (Benfield et al. 1988; Goodyer
et al. 1988). For example, Goodyer et al. (1988)
found moderate to severe events to be equally
increased in children with anxious and de-
pressed emotional disorder in the year prior to
onset. The authors did not report on specific
classes of events but did find that poor confid-
ing maternal relationships, parental distress,
and life events were all equally associated
with anxious and depressed emotional disor-
der in the children (Goodyer et al. 1988). Simi-
larly, a study of depressed and nondepressed
inpatient children and adolescents reported no
differences between the two groups on the
total number of stressful life events (Benfield
et al. 1988).
In addition to having the potential to shed
light on the etiology of anxiety and depression,
elucidating the role of stressful life events has
important treatment implications for pediatric
populations. For example, depressed adults
during the course of their treatment have been
shown to be slower to remit when exposed
to increased levels of stressful life events
(McQuaid et al. 2000). Similarly, anxious adults
experiencing chronic stressors were shown to
be less likely to improve during the course of
treatment (Wade et al. 1993). Independent of
treatment, Goodyer et al. reported that ongo-
ing stressors play a role in the persistence of a
depressive episode in children and adolescents
(Goodyer et al. 1997). However, the moderat-
ing role of stressful life events on treatment
outcomes in pediatric populations has not been
systematically explored.
In this study, the occurrence of stressful life
events during the previous year are compared
between anxious, depressed, and normal con-
trol children. It improves upon previous studies
in that children with a clinical diagnosis of anx-
iety or depression are included in addition to a
14104C07.pgs 9/16/05 1:31 PM Page 572
normal control sample. Based on the above-
cited literature, it was hypothesized that anx-
ious and depressed children would have a
similar number of events compared to each
other but significantly more events compared
to normal controls. Events were further classi-
fied into those that were most likely to be inde-
pendent (e.g., death of relative) of or dependent
(e.g., increase in number of arguments with
parents) on the childs behavior. It was hypoth-
esized that anxious and depressed children
would have similar rates of both independent
and dependent events but significantly more of
each compared to normal control children. Fi-
nally, events were classified into those that rep-
resented a loss for the child (e.g., parental
divorce, death of a close relative). It was hy-
pothesized that depressed children would have
significantly more loss events compared to both
the anxious and normal control children. Fi-
nally, exploratory analyses examined whether
stress had any impact on the secretion of corti-
sol around sleep onset across the three groups.
SUBJECTS AND METHODS
The assessment of stressful life events was
part of a larger investigation examining the
psychobiology of depression in children (e.g.,
Puig-Antich et al. 1985, 1989). The details of
the study are briefly summarized here.
Selection of anxious and depressed children
Depressed and anxious children were
screened at the Child and Adolescent Depres-
sion Clinic at New York State Psychiatric Insti-
tute. If the screening indicated that the child
was likely to have either a depressive or
anxiety disorder, the child was selected for a
2-week diagnostic assessment that included
the following: (1) the Kiddie-Schedule for Af-
fective Disorder and Schizophrenia for School-
Aged Children, Present Episode (K-SADS-P;
Chambers et al. 1985); (2) a pediatric examina-
tion including Tanner staging (Marshall and
Tanner 1969, 1970); (3) a second diagnostic as-
sessment 1014 days after the initial diagnostic
assessment identifying symptoms occurring in
the past week only. Diagnostic interviews were
conducted with children and their parent(s), or
SLEs IN ANXIOUS AND DEPRESSED CHILDREN 573
guardian(s) separately and symptoms pooled
across informants in deriving final diagnoses.
Children aged 612 years were included in
the study if after the second diagnostic assess-
ment they continued to meet research diagnos-
tic criteria (RDC) for major depressive disorder
(MDD) or if they met the Diagnostic and Statis-
tical Manual of Mental Disorders, 3rd edition
(DSM-III; APA 1980) for an anxiety disorder.
Within the anxious children, 50% met criteria
for separation anxiety disorder, 60% met crite-
ria for phobia, and 10% met criteria for obses-
sive-compulsive disorder.
Selection of normal controls
Normal control children who never had a
psychiatric disorder were assessed with the
Kiddie-Schedule for Affective Disorders and
Schizophrenia for School-Aged Children, Epi-
demiological version (K-SADS-E; Orvaschel et
al. 1982) through separate interviews with par-
ent(s) or guardian(s). Only children who had
never had any of the Axis I DSM-III diagnoses
in their lifetime were accepted into the normal
control group.
Life events measure
One of the childs parent(s) or guardian(s)
(usually the mother) was administered the Life
Events Record (LER; Coddington 1972) to col-
lect information on life events occurring in the
childs life during the year prior to the inter-
view. In addition to the events listed on the
LER, parents were asked whether the child
had ever been placed in foster care, was physi-
cally or sexually abused, was separated from a
parent or parents for 1 month or more, or had
experienced some other life event not listed on
the LER.
The LER contains life events that are both
clearly negative and stressful (e.g., death of
parent), as well as those that are likely to be ex-
perienced as positive (e.g., outstanding per-
sonal achievement). In order to extract only
those events that were most likely stressful, a
panel of raters separately rated each event as
either stressful, not stressful, or possibly both,
and consensus was reached for all events. Con-
servatively, only those events that were rated
as stressful were included for all analyses.
14104C07.pgs 9/16/05 1:31 PM Page 573
Events identified as most likely not stressful or
ambiguous with regard to whether they were
stressful included: Birth of a sibling; parental re-
marriage; addition of an adult to the family;
mother beginning to work; brother or sister
leaving home; change in peer acceptance; be-
ginning another school year; decrease in the
number of arguments between parents; out-
standing personal achievement; move to a new
school district; decrease in the number of argu-
ments with parents; becoming a full-fledged
member of a church; and change in parents fi-
nancial status. It is possible that some of these
events were experienced as stressful; however,
without additional contextual information sur-
rounding the event category, the negativity or
positivity of these events could not be deter-
mined and were, therefore, not included in the
analyses. All events were also judged on the
likelihood of being influenced by the childs be-
havior, which could be affected by the presence
of psychopathology. Events were classified as
possibly independent of (e.g., death of a grand-
parent) or dependent (e.g., increasing argu-
ments with parents) on the childs behavior.
Additionally, events were judged to represent a
loss event for the child defined as either the
death of a friend or relative or an exit from the
childs life (e.g., parental separation and/or di-
vorce). For ratings of both independence or de-
pendence and loss, consensus for event
categories was reached by a panel of raters. A
complete list of events that were chosen as
being most likely stressful for the child and
their associated categories are presented in the
Appendix.
Cortisol
The details of the between group compar-
isons of baseline cortisol secretion measured
every 20 minutes over a 24-hr period have been
previously reported (Puig-Antich et al., 1989).
In this report, cortisol secreted in the 1-hr and
2-hr window around sleep onset was examined
to determine if there was direct of effect of
stress or diagnosis x stress effect. The impor-
tance of cortisol secretion around sleep onset
has been highlighted by reports suggesting
elevated cortisol among depressed adoles-
cents (Dahl et al., 1991) and hgiher cortisol
near sleep onset predicting recurrences (Rao
574 WILLIAMSON ET AL.
et al., 1996). A total of 61 of the 76 subjects
had cortisol data (MDD, n = 37; ANX, n = 16;
and NC, n = 8).
Data analysis
Demographic variables were compared
using chi-square with continuity correction
or analysis of variance (ANOVA). The rela-
tionship between demographic variables and
life events measures were made using either
Spearman correlations or Mann-Whitney
rank sum tests, as appropriate. Stressful life
events were classified into categories reflect-
ing zero, one, or two or more events, owing
to the low frequency with which more than
two events occurred. Based on these cate-
gories, stressful life events were compared
between the groups using the test of linear
trend in proportions across threat levels be-
tween groups (Cochran 1954). All signifi-
cance tests were calculated using two-sided p
values with
2
level = 0.05.
RESULTS
Demographics
There were a similar number of females/
males for the anxious (11/9), depressed (16/29),
and the normal controls (7/4) (Fishers exact
test, NS). The ages of the anxious (8.98 1.52),
depressed (9.43 1.48 years), and normal con-
trols (8.97 1.09) were also similar (F
2,73
= 0.92,
NS). The racial composition differed between
the three groups with whites/blacks/other un-
evenly distributed among the anxious (4/2/14),
depressed (18/16/11), and normal controls
(6/2/3) (Fishers exact test; p = 0.01). Age and
racial composition were not significantly re-
lated to any of the life event measures consid-
ered below.
Total events
The depressed children were exposed to sig-
nificantly more total life events during the prior
year, compared to both anxious and normal
control children. The anxious and normal con-
trol children had a similar rate of life events
(Table 1). Further analyses showed that there
14104C07.pgs 9/16/05 1:31 PM Page 574
was a gender effect. Depressed females had sig-
nificantly more overall events, compared to
anxious (
2
= 5.44; df = 1; p = 0.02) and normal
(
2
= 9.13; df = 1; p = 0.01) females; however,
there were no significant between-group differ-
ences for total events among the males.
Total dependent events
Dependent events were defined as those that
were most likely to be influenced by the childs
behavior (e.g., suspension from school). There
was a trend for the depressed children to have
an elevated rate of dependent events, com-
pared to the anxious children but a similar rate
compared to the normal control children. The
anxious and normal control children had simi-
lar rates of dependent events.
Total independent events
The depressed children had significantly
more independent events (e.g., deaths) occur
in the past year, compared to both the anxious
and normal control children. The anxious and
normal control children had similar rates of
independent events. Depressed females had
significantly more overall independent
SLEs IN ANXIOUS AND DEPRESSED CHILDREN 575
events, compared to anxious (
2
= 5.44; df = 1,
p = 0.02) and normal (
2
= 9.13; df = 1; p = 0.01)
females. There were no significant between-
group differences for independent events
among the males.
Total loss events
There was a trend for the depressed children
to have greater overall loss events, compared to
the anxious children. The rates of total loss
events in the depressed and anxious children
did not differ from the normal controls (Table 2).
Stratifying the sample by males and females
similarly showed no between-group differences
for total loss events.
Dependent loss events
A single event in this study, separation from
parents for 1 month or more, was categorized as
a dependent loss. There were no differences be-
tween the three groups for this event (Table 2).
Independent loss events
The depressed children had significantly
more independent loss events during the pre-
TABLE 1. TOTAL, DEPENDENT,
a
AND INDEPENDENT
b
STRESSFUL LIFE EVENTS OCCURRING DURING THE PAST YEAR
Number of events
Group Trend
0 1 2+ comparisons test
c
p
Total Events
MDDs n = 45 (%) 22.2 24.4 53.3 M-A
d

2
= 9.59 0.01
Anxious n = 20 (%) 55.0 30.0 15.0 M-N
e

2
= 6.22 0.02
Normal Controls n = 11 (%) 36.4 63.6 0.0 A-N
f

2
= 0.02 NS
Dependent
a
Events
MDDs n = 45 (%) 51.1 31.1 17.8 M-A
2
= 3.55 0.06
Anxious n = 20 (%) 75.0 20.0 5.0 M-N
2
= 0.77 NS
Normal Controls n = 11 (%) 54.5 45.5 0.0 A-N
2
= 0.58 NS
Independent
b
Events
MDDs n = 45 (%) 37.8 35.6 26.7 M-A
2
= 6.70 0.01
Anxious n = 20 (%) 70.0 25.0 5.0 M-N
2
= 7.17 0.01
Normal Controls n = 11 (%) 81.8 18.2 0.0 A-N
2
= 0.74 NS
NS = not significant.
a
Events most likely dependent on childs behavior (see Appendix).
b
Events most likely independent of childs behavior (see Appendix).
c
Chi-Square (with 1 degree of freedom) Test of Linear Trend (Cochran 1954).
d
M-AMDDs versus anxious Children.
e
M-N = MDDs versus Normal Control Children.
f
A-N = Anxious versus Normal Control Children.
MDD, major depressive disorder.
14104C07.pgs 9/16/05 1:31 PM Page 575
ceding year than the anxious children, with a
trend toward higher independent loss events
compared to the normal control children. The
anxious and normal control children did not
differ on the rate of independent loss events
(Table 2). Depressed females had significantly
more independent loss events compared to
anxious females (
2
= 4.72; df = 1; p = 0.03) but
not with normal control females. Among the
males, there were no between-group differ-
ences for independent loss events.
Sleep Onset Cortisol
Among the subset of subjects with sleep
onset cortisol, there was no effect of diagnosis
for either the 1-hr (F
2,58
= 0.62, NS) or 2-hr (F
2,58
= 0.15, NS) period around sleep onset. Across
all three groups, stress was not related to sleep
onset cortisol for either the 1-hr (F
2,58
= 0.09,
NS) or 2-hr (F
2,58
= 0.55, NS) period around
sleep onset. Further analysis showed that there
were no significant diagnosis x sex interac-
tions for either the 1-hr (F
2,58
= 0.41, NS) or 2-
hr (F
2,58
= 0.22, NS) period around sleep onset
(see Table 3).
576 WILLIAMSON ET AL.
DISCUSSION
In this study, depressed children experienced
significantly more stressful life events during
the previous year compared to both anxious
and normal control children. Categorizing
events into those that were most likely influ-
enced by the childs behavior (dependent) and
those that were independent of the childs be-
havior revealed that mainly independent
events were significantly elevated in depressed
children. Events classified as independent loss
events, consisting of deaths and exits from the
childs life, occurred much more frequently in
the depressed children.
The finding that depressed children had sig-
nificantly more events than the anxious chil-
dren is consistent with at least one prior
investigation, albeit not completely analogous.
Berney et al. (1991) reported that depressed
children experienced significantly more stress-
ful life events in the year prior to onset, com-
pared to a group of maladjusted children.
Conversely, Goodyer et al. (1985) reported that
children with mild emotional disorder (pre-
dominantly anxiety symptoms) had similar
TABLE 2. TOTAL, DEPENDENT,
a
AND INDEPENDENT
b
LOSS EVENTS OCCURRING DURING THE PAST YEAR
Number of events
Group Trend
0 1 2+ comparisons test
c
p
Total Loss Events
MDDs n = 45 (%) 60.0 26.7 13.3 M-A
d

2
= 3.58 0.06
Anxious n = 20 (%) 80.0 20.0 0.0 M-N
e

2
= 0.55 NS
Normal Controls n = 11(%) 63.6 36.4 0.0 A-N
f

2
= 0.32 NS
Dependent
a
Loss Events
MDDs n = 45 (%) 93.3 6.7 0.0 M-A
2
= 0.00
g
NS
Anxious n = 20 (%) 95.0 5.0 0.0 M-N
2
= 2.07
g
NS
Normal Controls n = 11(%) 72.7 27.3 0.0 A-N
2
= 1.46
g
NS
Independent
b
Loss Events
MDDs n = 45 (%) 62.2 28.9 8.9 M-A
2
= 3.90 0.05
Anxious n = 20 (%) 85.0 15.0 0.0 M-N
2
= 3.28
g
0.07
Normal Controls n = 11(%) 90.9 9.1 0.0 A-N
2
= 0.00
g
NS
NS = not significant.
a
Events most likely dependent on childs behavior (see Appendix).
b
Events most likely independent of childs behavior (see Appendix).
c
Chi-Square (with 1 degree of freedom) Test of Linear Trend (Cochran 1954).
d
M-A= MDDs versus Anxious Children.
e
M-N = MDDs versus Normal Control Children.
f
A-N = Anxious versus Normal Control Children.
g
Yates Corrected Chi-Square (Trend test incalculable).
MDD, major depressive disorder
14104C07.pgs 9/16/05 1:31 PM Page 576
rates of moderate to severe events prior to
onset, compared to children with severe emo-
tional disorder (predominantly depressive
symptoms). The finding that stressful life events
were significantly more frequent among the
depressed children compared to the normal
control children is consistent with studies that
reported that stressful life events are correlated
with symptoms of depression (Compas et al.
1986; Johnson and McCutcheon 1980; Mullins
et al. 1985; Nolen-Hoeksema et al. 1986;
Swearingen and Cohen 1985), significantly ele-
vated in clinically depressed youth (Goodyer
et al. 1985, 1988), and predict future episodes
of depression in adulthood (Pine et al. 2002). A
direct comparison between this study and prior
reports is made difficult by the fact that differ-
ent methods were used for making diagnoses
and for assessing life events. For example, the
studies by Goodyer et al. (1985) and Berney et
al. (1991) assessed life events using an investi-
gator-based interview schedule with events
occurring in the year prior to onset of the
episode delineated. Our study relied on a re-
spondent-based checklist for assessing events
occurring in the prior year without regard to
onset of the depression or anxiety disorder.
Therefore, the finding of increased stressful
events could very well be a cause, as much as a
result, of the depression. However, the fact
that events independent of the childs behav-
ior were significantly elevated makes this pos-
sibility seem less likely.
The finding that cortisol secretion was not
significantly effected by stress is consistent
with the literature showing the HPA function
in children with psychopathology, particularly
depression, does not appear to be altered
(Dahl et al., 1996; Puig-Antich et al., 1989). It
appears that the HPA axis in depressed and
anxious children is not influenced by exposure
to stressful life events.
There were several limitations to this
study. One of the main limitations was that
events were not classifiable into so-called
danger events, which have been shown to
be associated with anxious symptoms in chil-
dren (Eley and Stevenson 2000) and anxiety
disorders in adults (Brown 1993; Finlay-Jones
and Brown 1981). Interestingly, danger
events associated with anxiety are primarily
SLEs IN ANXIOUS AND DEPRESSED CHILDREN 577
characterized by the risk for a future loss of a
cherished idea or significant other (Brown
1993; Finlay-Jones and Brown 1981). Because
the rating of the potential for a loss requires
detailed event descriptions, they could not
be rated from the event checklist that was
used. Along this line, a significant limitation
of this study revolves around the use of a re-
spondent-based checklist for the assessment
of life events (Williamson et al. 1995, 2003).
Overall, checklists have low reliability, do
not provide accurate dating of event occur-
rence, and do not delineate different levels of
severity within a particular stressor (Cohen
and Wills 1985). The importance of identify-
ing both the temporal occurrence and the
level of severity of an event has been estab-
lished in studies of depressed adults. Brown
et al. (1973) found that markedly threatening
events (e.g., death of a parent), but not less
severe stressors, were significantly elevated
in depressed patients compared to commu-
nity controls across the 16 3-week periods
prior to onset. The issue of event severity and
timing has not been explored with equal
rigor in early-onset anxiety or depression.
Thus, the nature of the timing and severity of
events may be different for depressed chil-
dren compared to children with anxiety dis-
orders. The nature of this relationship
remains muddled, owing to the reliance on
checklists for assessing life events and the
lack of longitudinal studies across develop-
ment. Another limitation of our study was
that only parental reports of life events were
assessed. It is possible that children would
report additional events of which the parent
was unaware. Finally, the sample sizes in-
cluded were rather modest, especially the
normal controls, and could have resulted in
inflating both Type I and II error rates. Ac-
cordingly, these results should be reviewed
as preliminary, pending replication in a larger
sample of children using better measures of
stressful life events (e.g., Williamson et al.
2003).
Life events representing loss or separation
appear to be more prevalent preceding the
onset of a depressive episode in adults (Brown
and Harris 1978, 1989; Brown et al. 1986). Early
loss events, particularly the loss of a parent
14104C07.pgs 9/16/05 1:31 PM Page 577
during childhood, have also been associated
with depression in adulthood (Lloyd 1980).
Other reports have suggested that early loss
combined with subsequent lack of care and the
development of helplessness increases the risk
for depression in adults (Bifulco et al. 1992).
However, this causal pathway does not in-
clude whether the early loss led to the onset of
a childhood depressive episode. The findings
of our study suggest that independent
losses those most likely independent of the
childs behaviorare significantly greater in
depressed children, particularly females, com-
pared to anxious and normal control children.
Accordingly, early loss events resulting in de-
pressive episodes during childhood could
very well set the stage for depressive episodes
later in life, especially in the presence of later
loss events.
It is important to also bear in mind the
clinical implications of this study. The fact
that depressed children have higher levels
of stressful life events might have signifi-
cant effects on treatment outcomes, which
has been shown in depressed adults (Mc-
Quaid et al. 2000). In addition, the increased
level of independent loss events experi-
enced by depressed children, particularly
females, indicates the need for prophylactic
interventions in high-risk children who ex-
perience these types of events. Finally, stud-
ies of anxious adults have implicated the
role of stress in treatment response (Wade et
al. 1993). It is possible that even mild stres-
sors, as experienced by the anxious children
in this study, might play a role in treatment
outcome. The moderating role of stressful
life events in treatment outcome among pe-
diatric populations needs to be more fully
investigated.
CONCLUSIONS
In summary, it appears that depressed chil-
dren experience a greater number of stressful
life events compared to anxious children, in-
cluding independent loss events. Because life
events were assessed without regard to the
onset of the depression, it could not be deter-
mined whether events preceded the onset of
578 WILLIAMSON ET AL.
the depression or were intercurrent. Future
studies are needed that will examine the tim-
ing and severity of stressors to determine if
stressors of all degrees of severity in the full
year prior to onset are elevated in depressed
children.
REFERENCES
American Psychiatric Association: Diagnostic and
Statistical Manual of Mental Disorders, 3rd ed.
Washington (DC), American Psychiatric Associa-
tion, 1980.
Benfield CY, Palmer DJ, Pfefferbaum B, Stowe ML:
A comparison of depressed and nondepressed
disturbed children on measures of attributional
style, hopelessness, life stress, and temperament.
J Abnorm Child Psychol 16:397410, 1988.
Berney TP, Bhate SR, Kolvin I, Famuyiwa OO, Bar-
rett ML, Fundudis T, Tyrer SP: The context of
child depression: The Newcastle childhood de-
pression project. Br J Psychiatry 159:2835, 1991.
Bifulco A, Harris T, Brown GW: Mourning or early
inadequate care? Re-examining the relationship of
maternal loss in childhood with adult depression
and anxiety. Dev Psychopathol 4:433449, 1992.
Birmaher B, Ryan ND, Williamson DE, Brent DA,
Kaufman J, Dahl RE, Perel J, Nelson B: Childhood
and adolescent depression: A review of the past
10 pears, part I. J Am Acad Child Adolesc Psychi-
atry 35:14271439, 1996.
Brown GW: Life events and affective disorder:
replications and limitations. Psychosom Med 55:
248259, 1993.
Brown GW, Andrews B, Harris TO, Adler Z, Bridge
L: Social support, self-esteem, and depression.
Psychol Med 16:813831, 1986.
Brown GW, Harris TO (eds.): Social Origins of De-
pression: A Study of Psychiatric Disorder in
Women. London, Tavistock, 1978.
Brown GW, Harris TO (eds.): Life Events and Ill-
ness. New York, Guilford Press, 1989.
Brown GW, Sklair F, Harris TO, Birley JLT: Life-
events and psychiatric disorders part 1: Some
methodological issues. Psychol Med 3:7487, 1973.
Burke KC, Burke JD, Regier DA, Rae DS: Age at
onset of selected mental disorders in five com-
munity populations. Arch Gen Psychiatry 47:
511518, 1990.
Chambers WJ, Puig-Antich J, Hirsch M, Paez P,
Ambrosini P, Tabrizi MA, Davies M: The assess-
ment of affective disorders in children and ado-
lescents by semistructured interview: Test-retest
reliability of the Schedule for Affective Disorders
and Schizophrenia for School-Age Children, pres-
ent episode version. Arch Gen Psychiatry 42:
696702, 1985.
14104C07.pgs 9/16/05 1:31 PM Page 578
SLEs IN ANXIOUS AND DEPRESSED CHILDREN 579
Cochran WG: Some methods for strengthening the
common
2
test. Biometrics 10:417441, 1954.
Coddington RD: The significance of life events as eti-
ologic factors in diseases of children: I. Asurvey of
professionals. J Psychosom Res 16:205213, 1972.
Cohen S, Wills TA: Stress, social support, and the
buffering hypothesis. Psychol Bull 98:310357,
1985.
Compas BE, Slavin LA, Wagner BM, Vannatta K:
Relationship of life events and social support
with psychological dysfunction among adoles-
cents. J Youth Adolesc 15:205221, 1986.
Eley TC, Stevenson J: Specific life events and
chronic experiences differentially associated with
depression and anxiety in young twins. J Ab-
norm Child Psychol 28:383394, 2000.
Finlay-Jones R, Brown GW: Types of stressful life
event and the onset of anxiety and depressive
disorders. Psychol Med 11:803815, 1981.
Goodwin RD, Fergusson DM, Horwood LJ: Early
anxious/withdrawn behaviors predict lateral in-
ternalizing disorders. J Child Psychol Psychiatry
45:874883, 2004.
Goodyer IM, Herbert J, Tamplin A, Secher SM, Par-
son J: Short-term outcome of major depression: II.
Life events, family dysfunction, and friendship dif-
ficulties as predictors of persistent disorder. J Am
Acad Child Adolesc Psychiatry 36:474480, 1997.
Goodyer I, Kolvin I, Gatzanis S: Recent undesirable
life events and psychiatric disorder in childhood
and adolescence. Br J Psychiatry 147:517523, 1985.
Goodyer IM, Wright C, Altham PME: Maternal ad-
versity and recent stressful life events in anxious
and depressed children. J Child Psychol Psychia-
try 29:651667, 1988.
Johnson JH, McCutcheon SM: Assessing life stress
in older children and adolescents: Preliminary
findings with the Life Events Checklist. In: Stress
and Anxiety. Edited by I.G. Sarason & C.D. Spiel-
berger. Washington (DC), Hemisphere, 1980.
Kendler KS: Major depression and generalized anx-
iety disorder same genes, (partly) different envi-
ronmentsrevisited. Br J Psychiatry 168(Suppl
30):6875, 1996.
Lloyd C: Life events and depressive disorder re-
viewed: I. Events as predisposing factors. Arch
Gen Psychiatry 37:529535, 1980.
Marshall WA, Tanner JM: Variations in pattern of
pubertal changes in girls. Arch Dis Child 44:291
303, 1969.
Marshall WA, Tanner JM: Variations in pattern of
pubertal changes in boys. Arch Dis Child 45:13
23, 1970.
McQuaid JR, Monroe SM, Roberts JE, Kupfer DJ,
Frank E: A comparison of two life stress assess-
ment approaches: Prospective prediction of treat-
ment outcome in recurrent depression. J Abnorm
Psychol 109:787791, 2000.
Mullins LL, Siegel LJ, Hodges D: Cognitive prob-
lem solving and life event correlates of depres-
sive symptoms in children. J Am Acad Child Psy-
chiatry 13:305314, 1985.
Nolen-Hoeksema S, Girgus JS, Seligman ME:
Learned helplessness in children: A longitudinal
study of depression, achievement, and explana-
tory style. J Pers Soc Psychol 51:435442, 1986.
Orvaschel H, Puig-Antich J, Chambers WJ, Tabrizi
MA, Johnson R: Retrospective assessment of
child psychopathology with the K-SADS-E. J Am
Acad Child Psychiatry 21:392397, 1982.
Pine DS, Cohen P, Johnson JG, Brook JS: Adolescent
life events as predictors of adult depression. J Af-
fect Disord 68:4957, 2002.
Puig-Antich J, Dahl R, Ryan N, Novacenko H,
Goetz D, Goetz R, Twomey J, Klepper T: Cortisol
secretion in prepubertal children with major de-
pressive disorder: Episode and recovery. Arch
Gen Psychiatry 46:801809, 1989.
Puig-Antich J, Lukens E, Davies M, Goetz D, Bren-
nan-Quattrock J, Todak G: Psychosocial function-
ing in prepubertal major depressive disorders. I.
Interpersonal relationships during the depressive
episode. Arch Gen Psychiatry 42:500507, 1985.
Rutter M, Pickles A, Murray R, Eaves L: Testing hy-
potheses on specific environmental causes on be-
havior. Psych Bull 127:391324, 2001.
Swearingen EM, Cohen LH: Life events and psy-
chological distress: Aprospective study of young
adolescents. Dev Psychol 21:10451054, 1985.
Wade SL, Monroe SM, Michelson LK: Chronic life
stress and treatment outcome in agoraphobia
with panic attacks. Am J Psychiatry 150:1491
1495, 1993.
Weissman MM, Warner V, Wickramaratne P, Moreau
D, Olfson M: Offspring of depressed parents. 10
years later. Arch Gen Psychiatry 54:932 940, 1997.
Williamson DE, Birmaher B, Anderson BP, Al-
Shabbout M, Ryan ND: Stressful life events in
depressed adolescents: The role of dependent
events during the depressive episode. J Am
Acad Child Adolesc Psychiatry 34:591598,
1995.
Williamson DE, Birmaher B, Ryan ND, Shiffrin TP,
Lusk JA, Protopapa J, Dahl RE, Brent DA: The
stressful life events schedule for children and
adolescents: Development and validation. Psy-
chiatry Res 119:225241, 2003.
Address reprint requests to:
Douglas E. Williamson, Ph.D.
Department of Psychiatry
University of Pittsburgh School of Medicine
Room E-723
Western Psychiatric Institute and Clinic
3811 OHara Street
Pittsburgh, Pennsylvania 15213
E-mail: Williamson DE@upmc.edu
14104C07.pgs 9/16/05 1:31 PM Page 579
580 WILLIAMSON ET AL.
Appendix
STRESSFUL LIFE EVENTS
Events most likely dependent on childs behavior
Suspension from school
Increase in number of arguments with parents
Failure of a grade
Increase in arguments between parents
Separation from parents for one month or more
L
Other (e.g., accident at school)
Events most likely independent of childs behavior
Serious illness requiring hospitalization of parent
Loss of job by parent
Visible congenital deformity
Change in fathers occupation requiring increased absence from home
L
Serious illness requiring hospitalization of child
Parental divorce
L
Jail sentence of parent
L
Death of brother or sister
L
Discovery of being an adopted child
Death of a grandparent
L
Death of a friend
L
Serious illness requiring hospitalization of brother or sister
Death of a parent
L
Acquiring a visible deformity
Sexual abuse
Physical abuse
Foster placement of child
Parental marital separation
L
Pregnancy of unwed teenage sister
Other (e.g., mother assaulted)
Other (e.g., death of a pet)
L
L
Loss event.
14104C07.pgs 9/16/05 1:31 PM Page 580
This article has been cited by:
1. W. Burleson Daviss, Rasim S. Diler, Boris Birmaher. 2009. Associations of Lifetime Depression with Trauma Exposure, Other
Environmental Adversities, and Impairment in Adolescents with ADHD. Journal of Abnormal Child Psychology 37:6, 857-871.
[CrossRef]
2. Rozmin Halari, Mima Simic, Carmine M. Pariante, Andrew Papadopoulos, Anthony Cleare, Michael Brammer, Eric Fombonne,
Katya Rubia. 2009. Reduced activation in lateral prefrontal cortex and anterior cingulate during attention and cognitive control
functions in medication-nave adolescents with depression compared to controls. Journal of Child Psychology and Psychiatry 50:3,
307-316. [CrossRef]
3. Jeremy K. Fox, Leslie F. Halpern, John P. Forsyth. 2008. Mental Health Checkups for Children and Adolescents: A Means to
Identify, Prevent, and Minimize Suffering Associated With Anxiety and Mood Disorders. Clinical Psychology: Science and Practice
15:3, 182-211. [CrossRef]
4. Lourdes Ezpeleta, Roser Granero, Nria Osa, Josep M. Domnech. 2008. Risk factor clustering for psychopathology in socially
at-risk Spanish children. Social Psychiatry and Psychiatric Epidemiology 43:7, 559-568. [CrossRef]
5. Leah R. Reznikov, Lawrence P. Reagan, Jim R. Fadel. 2008. Activation of phenotypically distinct neuronal subpopulations in the
anterior subdivision of the rat basolateral amygdala following acute and repeated stress. The Journal of Comparative Neurology
508:3, 458-472. [CrossRef]