Você está na página 1de 10

Development and Psychopathology 23 (2011), 11391147

# Cambridge University Press 2011


doi:10.1017/S0954579411000526

Associations among depression, perceived self-efficacy, and


immune function and health in preadolescent children

MARY T. CASERTA, PETER A. WYMAN, HONGYUE WANG, JAN MOYNIHAN, AND THOMAS G. OCONNOR
University of Rochester School of Medicine and Dentistry

Abstract
Experimental animal studies and adult research consistently show that stress exposure and/or psychological symptoms are associated with poorer health and
immune functioning. The application to children is not yet clear, however, and we lack developmental models for studies in this area. The objective of this paper
was to test the hypothesis that self-reported self-efficacy and depression, two markers of psychological well-being in children, would predict immunity and rate of
illnesses. The data are based on a prospective study of 141 healthy, normally developing children aged 713 years who were recruited from an ambulatory pediatric
setting. Children completed self-efficacy and depression measures and had blood obtained for IL-6 plasma levels and natural killer cell functional assays on three
occasions, 6 months apart. Parents maintained weekly child illness diaries over 1 year using a thermometer to record fever. Parent psychiatric symptoms and income
were used as covariates. Results indicated that, across the three occasions of measurement collected over the 1-year period, higher perceived self-efficacy was
significantly associated with lower plasma interleukin 6 concentrations. There was no overall main effect of depressive symptoms on immune measures; however,
for older girls, higher depression was associated with elevated natural killer cell cytotoxicity and an increased rate of total illnesses and febrile illnesses. The
findings provide some of the first evidence that psychological processes are associated with immunity and health in a normally developing sample of
preadolescents. Furthermore, the pattern of results suggests a modified model of a link between psychological well-being and immunological processes in children.
These results build on and expand research on the notion of allostatic load and develop a groundwork for developmental studies in this area.

Research findings suggest that depression and psychosocial dicators of psychological well-being in children, depression
stress may have an etiological role in diverse diseases in and self-efficacy, and markers of immune functioning and ill-
adulthood, including cardiovascular disease, chronic pain, nesses in a short-term longitudinal study of typically develop-
stroke, diabetes mellitus, susceptibility to infectious diseases, ing 7- to 13-year-old children.
and HIV disease progression (Cohen, Tyrrell, & Smith, 1991; Clinical and epidemiological studies provide robust evi-
Holahan et al., 2010; Ickovics et al., 2001; Miller, Chen, & dence that psychosocial stress and depression are associated
Cole, 2009; Rozanski, Blumenthal, & Kaplan, 1999). These with changes in adult immune function (Musselman, Evans,
associations provide a basis for the allostatic load model, a & Nemeroff, 1998; Padgett & Glaser, 2003). Specifically, a re-
leading framework for linking stress, health, and develop- cent meta-analysis involving over 400 subjects showed that
ment (Juster, McEwen, & Lupien, 2010). The pathophysiolog- levels of the proinflammatory cytokine interleukin 6 (IL-6)
ical mechanisms leading to such varied disease states are not and tumor necrosis factor-alpha were elevated in subjects
fully elucidated, but considerable evidence points to inflam- with major depressive disorder compared to normal controls
mation as one likely mediator (Juster et al., 2010; Karelina (Dowlati et al., 2010). In addition, elevated concentrations of
& DeVries, 2011). Whether the health impacts of depression IL-6 were reported in subjects with higher levels of stress asso-
and stress and the underlying mechanisms are apparent in ciated with elevated blood pressure and more severe symptoms
childhood is not yet clear and represents a major develop- during an upper respiratory infection (Brydon & Steptoe, 2005;
mental question with sizable public health implications. Cohen, Doyle, & Skoner, 1999). Chronic psychosocial stress
The current study is part of an ongoing effort to translate re- and depression may increase immune activation via greater glu-
search on stress exposure and health outcomes from adults to cocorticoid resistance (Pace & Miller, 2009; Pace et al., 2006).
children. Specifically, we examined the links between two in- Other studies suggest that depression may suppress immune
activation, as shown in previous studies linking depression to
This work was supported by Grant RO1 HD 38938 from the National Insti- reductions in lymphocyte proliferative responses to mitogens,
tute of Child Health and Development and in part by General Clinical Re- natural killer (NK) cell activity, and cellular immunity to vi-
search Grant 5 MO1 RR00044 from the National Center for Research Re- ruses (Irwin, 2002; Kronfol, 1983; Schleifer, Keller, & Bartlett,
sources, NIH. 2002). These apparently contradictory findings have not yet
Address correspondence and reprint requests to: Mary Caserta, Depart-
ment of Pediatrics, University of Rochester Medical Center, 601 Elmwood
been reconciled, and underscore the need for further study.
Avenue, Box 690, Rochester, NY 14642; E-mail: mary_caserta@urmc. Research translating the adult findings on depression or
rochester.edu. stress and immunity to pediatric samples is now underway
1139
1140 M. T. Caserta et al.

(Worthman & Panter-Brick, 2008), but to date findings are Marshall, & Skwerer, 1990). Small sample size and varying
limited, particularly with respect to specific mechanisms. health status of participants may explain the discrepant find-
Nevertheless, several reports link some aspect of general fam- ings so far reported. In any event, epidemiological data under-
ily stress to one or the other marker of immune functioning. score the substantial role of age and sex on rates of depression
Wolf, Miller, and Chen (2008) found that parent-reported (Angold, Costello, & Worthman, 1998) and point to these as
perceived stress and depressive symptoms were associated likely moderators of a link with physical health. As a result, we
with increased levels of the T-helper cell type 2 markers consider age and sex as moderators of the link between depres-
IL-4 and eosinophilic cationic protein in their children. An- sive symptoms and immune markers.
other indicator of family stress, harsh family climate in early In addition, it may be that a number of psychological states
childhood, was linked to a pro-inflammatory phenotype dur- predict inflammation, as suggested, for example, by the adult
ing adolescence (Miller & Chen, 2010). In a healthy ambula- work on optimism and positive emotional style (Cohen,
tory sample, we previously reported that parental psychiatric Doyle, Turner, Alper, & Skoner, 2003). We therefore also in-
symptoms predicted elevated NK cell activity in their chil- clude in this study a measure of perceived self-efficacy, or pos-
dren and increased CD8 T cell responses to chronic viral itive beliefs in ones ability to enact actions in order to achieve
stimulation with cytomegalovirus (Caserta et al., 2008). desired goals (Bandura, 1977). Perceived self-efficacy is a pre-
The above findings support the broad hypothesis that stress dictor of positive socialemotional and behavioral functioning
may compromise health in children, but further research is among at-risk or high-stress youth (Cowen et al., 1997; Mas-
needed to address several developmental questions not yet tack- ten & Coatsworth, 1998; Wyman, Cowen, Work, & Parker,
led in existing studies. One concerns the point in development 1991). This study extends prior work on behavioral outcomes
at which reliable associations may be detected. Animal studies to immunity and health.
demonstrate that the effect of early stress exposure may have In summary, in the present study, as part of a develop-
lasting impact on inflammatory markers (Hennessy, Deak, & mental study of an allostatic load model, we sought to extend
Schiml-Webb, 2010) and that there are reliable links between work on the childrens psychological well-being and health,
behavior and immunity (Granger, Hood, Dreschel, Sergeant, and the mechanisms involved. Pediatric research has included
& Likos, 2001), but the immune outcomes are, as in most hu- a range of immune markers, some linked to particular disease
man studies, limited to adulthood. Accordingly, it is not clear states of the clinical samples studied; no particular or optimal
when in development a reliable impact on immune mechanisms set of immune markers is easily discernible from existing
can be detected. It may be that reliable links between stress ex- studies. Here, we include a common marker of inflammation,
posure and immune outcomes are not detectable in younger IL-6, and a frequently assessed indicator of immune activa-
children, perhaps because of greater immunological reserve tion, NK cell cytotoxicity. To examine if the association
that underlies resilience or because stress exposure effects are with immune markers translate to health outcomes, we in-
detectable only after prolonged, that is, more chronic, stress ex- clude a measure of illness, based on detailed diary data. These
posure. The current study seeks to address this question by test- immune and health data were examined in relation to chil-
ing whether one of the most reliable associations in adult psy- drens self-reports over a 1-year period.
choneuroimmunology, an association between IL-6 and self-
reported depressive symptoms, is found in young children.
Methods
A further developmental question for research is the iden-
tification of the risk phenotype that may associate with child
Subject enrollment and protocol
immune function. Major depressive disorder, the focus of
much of the adult work, is rare in preadolescent (especially Children ages 510 and their primary caregivers were re-
prepubertal) children (Angold, Costello, Erkanli, & Worth- cruited and enrolled between July 1, 2001 and June 30,
man, 1999). Therefore, if inflammation were specifically as- 2003 from a population participating in a longitudinal study
sociated with depression, then the link may not be detectable of childhood viral infections in the Division of Pediatric In-
in preadolescent samples. In fact, studies of immune function fectious Diseases at the University of Rochester Medical Cen-
and depressive symptoms in children and adolescents have ter. The children had initially been identified during visits to
yielded inconsistent findings. For example, whereas one the emergency department or other pediatric services of the
study of 812-year-olds reported that major depression was Golisano Childrens Hospital at Strong in Rochester, NY.
associated with lower NK cytotoxity and increased concana- One child per family (chosen based upon inclusion criteria
valin A mitogen responses (Barlett, Schleifer, Demetrikopou- and previous participation in the prior study) was enrolled,
los, & Keller, 1995), a different study of adolescents found and all children were well at the time of enrollment. Children
major depression associated with increased NK cell activity, with chronic diseases adversely affecting the immune system
increases in circulating lymphocytes and lymphocyte subsets, (e.g., those on chronic corticosteroid therapy) were excluded.
and with lower concanavalin A mitogen responses (Schleifer The Research Subjects Review Board of the University of Ro-
et al., 2002). Still other studies found few differences on mea- chester approved this study. All caregivers provided informed
sures of immunity associated with depression (Birmaher et al., consent, and children provided verbal assent. A $45 honorar-
1994; Shain et al., 1991; Targum, Clarkson, Magac-Harris, ium was given to the family following each visit.
Stress exposure and/or psychological symptoms and health in children 1141

The full study protocol spanned a 3-year period, and con- Laboratory assays
sisted of seven visits approximately 6 months apart. At each visit,
NK cell cytotoxicity assays were performed on whole blood
blood samples were collected from children, and each childs
samples as previously reported (Bromelow, Galea-Lauri,
vital signs, including height and weight, were measured. The
OBrien, & Souberbielle, 1998; Caserta et al., 2008). The per-
enrolled parent supplied demographic information and com-
cent specific lysis was calculated as 100(cpm experimental
pleted several measures of family stress and perceived personal
cpm spontaneous)/(cpm maximum cpm spontaneous) for
distress, including psychiatric symptoms. At each visit, the par-
each dilution of whole blood. The percent specific lysis at
ent also provided an interim medical history of illnesses and
each dilution was log transformed and linear regression was
other health conditions between visits, and reported any symp-
used to estimate the dilution that corresponded to 20% lysis,
toms of illness in the child in the preceding two weeks (see
which was then used as an overall measure of NK cell cyto-
Caserta et al., 2008). For the first 2 years of the study (Visits
toxicity for each visit.
14), children engaged in quiet activities in a separate room
Plasma samples were assayed in duplicate using the Quan-
while their parents were interviewed. Beginning at the fifth visit
tikine High Sensitivity ELISA kit for human interleukin 6
and continuing for the next year (Visits 57), children com-
(IL-6) concentrations (R and D Systems, Minneapolis, MN)
pleted self-report measures of depression symptoms and self-
according to the manufacturers instructions (Eisenberger, In-
efficacy (described below). The fifth visit was chosen to initiate
agaki, Mashal, & Irwin, 2010; Kiecolt-Glaser et al., 2011).
the collection of child self-report data because it was presumed
that the youngest children in the study (the youngest child was
age 5 years at study Visit 1) would not have been reliable re- Child psychosocial measures
porters on all of the self-report measures at the earlier waves. The
present study focuses on this 1-year period during which child- At Visits 57, the children completed the Childrens Depres-
report measures of depression and self-efficacy were obtained. sion InventoryShort Form (CDI-SF; Kovacs, 1992). The
self-report CDI-SF was designed for children from age 7 to
17 years and consists of 10 questions covering cognitive, affec-
Illness diary tive, and behavioral symptoms associated with depression. For
At enrollment, parents were given a diary form and digital each item, the child selects which of three descriptors span-
thermometer and instructed to record each week their childs ning minimal to significant symptomotology best describes
health status, including temperatures associated with every him/her in the past 2 weeks. The items originated from the
instance of illness. Strict definitions of illnesses were not pro- longer, 27-item CDI (Kovacs, 1992) and were selected based
vided to parents. Rather, parents were asked to record all epi- on a backward stepwise internal consistency reliability analy-
sodes they considered illnesses, all associated symptoms, and sis. Higher CDI-SF scores reflect more depression symptoms.
their childs corresponding temperature. Diaries were col- The 10 items had an alpha reliability coefficient of 0.80.
lected via mail on a monthly basis. If a diary was not returned, At Visits 57, children also completed the Perceived Self
then a study nurse contacted the family by phone to collect Efficacy Scale (PSE; Cowen et al., 1991). This 14-item self-
the information. If a family could not be reached, the diary report measure assesses childrens judgments about their abil-
was updated at the next clinic visit. All of the illness records ity to achieve their desired goals in common problem situa-
were reviewed by a pediatric infectious disease specialist who tions, spanning school challenges, peer and family conflicts,
was not part of the study team and who was kept unaware of and being in new, unfamiliar situations. For each item, chil-
all subject characteristics. That specialist coded each episode dren rate their degree of certainty in achieving a desired out-
of illness to determine if the episode was consistent with an come on a 3-point Likert scale (not at all sure to very sure).
illness and to determine if multiple records from a 14-day The measure is designed to assess a combination of self-effi-
time period reflected a single (e.g., a single viral upper re- cacy beliefs and outcome expectations (Bandura & Cervone,
spiratory tract infection recorded as a cold and sore throat) 1983). The PSE scale has demonstrated adequate internal
or multiple illnesses (e.g., viral respiratory tract infection consistency and construct validity: higher scores are associ-
and otitis media). Physical trauma, elective or orthopedic sur- ated with positive behavioral and socialemotional function-
gery, mental health/behavioral problems, constipation, and ing among children in general populations and among those
well-childcare visits were not coded as illnesses. Allergies, exposed to chronic psychosocial adversity (Wyman et al.,
asthma, eczema, and contact dermatitis were included. Over 1999). An internal consistency coefficient of 0.59 was found
the 3-year study, 1065 incidents were coded as separate ill- for the present study. Higher PSE scores reflect greater per-
nesses and 330 were reported as febrile illnesses, with 230 ceived self-efficacy.
temperatures recorded (Caserta et al., 2008). Fever was de-
fined as a temperature .388C. If the families reported fever
Covariates
but a temperature was not provided (i.e., temperature data
were missing), then the illness was recorded as a febrile ill- Parent psychiatric symptoms were ascertained at each visit
ness. If families reported a fever but the recorded temperature using the 51-item Brief Symptom Inventory (BSI; Derogatis,
was 388C, then the illness was recorded as without fever. Dellapietra, & Kilroy, 1992). The items assess nine symptom
1142 M. T. Caserta et al.

clusters including depression, anxiety, and psychoticism. For immune function in two ways. First, we included both main
each item, the parent provides a severity rating of 0 to 4 based effects and interaction terms in the regression model. Second,
on the prior month. The total symptom score was used, which we examined separately the effects of depression on immune
has been shown to be sensitive to changes in psychological function by estimating models for four subgroups defined by
status due to mental disorders and socialinterpersonal events. age (older vs. younger) and sex in order to test our hypothesis
Higher BSI total scores indicate more psychiatric symptoms. that the adverse effects on health from depression would be
We used the BSI as a covariate in models testing associations greatest in older girls in our sample.
between childrens self-evaluations and health because of The general estimating equation (Zeger & Liang, 1986)
previous analyses indicating that parent-reported symptoms with a log link function and Poisson errors was used to test
were a predictor of increased illnesses and altered immune the association of depression and self-efficacy with parent-
function in children (Caserta et al., 2008). In addition, we reported child illnesses, as appropriate for the distribution of
used family income as another covariate: total yearly family these count data. Depression and self-efficacy reported by
income was assessed at enrollment and at subsequent visits children at Visit 5 were the independent variables of primary
parents were asked to report any changes in income. Child interest in analyses predicting number of illnesses in the fol-
age and sex were also included as covariates and moderators lowing year. In addition to childrens age and sex, the other
(see below). covariates in the model were: total illnesses reported by par-
ents in the previous year (Visits 35) and the two indicators of
family stress at Visit 5: parent psychiatric symptoms (BSI)
Statistical analysis
and annual household income. To account for seasonal ef-
Analyses were performed with SAS (version 9.2, SAS Insti- fects on childrens illnesses, the number of parent-reported
tute Inc, Cary, NC). Descriptive analyses were conducted to illnesses between Visits 5 and 7 was used as the dependent
summarize the data using means, standard deviation, fre- variable so that it covers 1 year. Actual length of reporting
quency, and percentage as appropriate; and correlations time was treated as the offset in the model to justify the var-
were examined among the study variables. Mixed models iation of this variable among subjects. By using a prospective,
were used to evaluate the relationships among the repeatedly longitudinal model (i.e., Visit 5 CDI-SF and PSE predicting
measured variables of CDI-SF, PSE, BSI, and income. Pat- illnesses over the next year), the potential confounding effects
terns of attrition were examined and the assumption of miss- of illness on childrens depression and perceived self-efficacy
ing completely at random was tested. Plasma IL-6 concentra- were reduced. To examine whether the association was mod-
tion and NK cell function were log-transformed for normality erated by age and sex, interaction analyses were also per-
model assumption and back transformed to the original formed and models were estimated separately for the four
metric for presentation. groups of children stratified by age (using the median age,
To evaluate whether depression and self-efficacy were as- 9.3 years) and sex; the association of depression and self-
sociated with the two immune function measures, plasma IL- efficacy with febrile illness was analyzed in a similar manner.
6 concentration and NK cell function, longitudinal data anal- Whether or not the child had at least one reported febrile ill-
ysis was performed using mixed models with random subject ness between Visits 5 and 7 was used as dependent variable to
effects to account for within-subject correlation over time. fit a logistic model using the general estimating equation,
Plasma IL-6 concentration and NK cell function assessed at with a logit link and binary error, with the same set of covari-
Visit 5 through Visit 7 were the dependent variables. Depres- ates and independent variables as in the model for total ill-
sion (CDI-SF) and self-efficacy (PSE) prior to each visit were nesses. Moderation effects of age and sex were also tested
used as time-varying predictors. In addition to child age and and age/sex subgroup analyses were conducted. Mediation
sex, two additional covariates were added to the modelsan- analysis (Baron & Kenny, 1986) was performed to test if
nual household income and parent psychiatric symptoms plasma IL-6 concentration and NK cell function mediated
(BSI)to account for well-established effects of those two the relationship between depression and child illnesses.
family variables on child health and to better ascertain the
unique contributions of childrens self-reported depression
Results
and self-efficacy. Also included in each model was immune
function at the prior visit (Visit 4) corresponding to the de-
Cohort description
pendent variable (plasma IL-6 concentration or NK cell func-
tion) and whether an illness had been reported for the child in Of 170 enrolled childcaregiver dyads, 141 were retained at
the preceding 14 days. Race/ethnicity was not found to be as- the fifth visit when child depression and self-efficacy mea-
sociated with either of the immune responses and therefore sures were first collected (Table 1). Ninety-three percent of
was not included as a covariate. Body mass index (BMI) the enrolled caregivers were female (mean age 35 years,
was added to the model testing plasma IL-6 concentrations range 2173) and 89% were either the target childs biolog-
to account for its association with plasma IL-6 concentration ical or adoptive parent. A total of 120 subjects completed all
(Shelton & Miller, 2010). We tested for moderation effects of seven visits (71% of enrolled). The reasons for leaving the
gender and age on the relationship between depression and study were noncompliance with follow-up visits (62%),
Stress exposure and/or psychological symptoms and health in children 1143

Table 1. Subject characteristics BSI score and income were not associated with self-efficacy;
BSI was positively correlated with child depression (correla-
Sample tion coefficient .28, p , 0.01); income was marginally neg-
Characteristics Self-Efficacy Depression atively correlated with child self-reported depression scores
(n 141) Mean (SD) Mean (SD) (correlation coefficient 2.16, p .05). Income was nega-
Childs race tively associated with parent psychiatric symptoms (BSI total;
White 67 (47%) 28.22 (4.35) 12.28 (2.28) correlation coefficient 2.19, p .03). Neither depression
Black 40 (28%) 29.08 (3.67) 13.38 (4.11) nor self-efficacy scores were associated with child sex. Base-
Hispanic 9 (6%) 28.78 (3.03) 13.44 (5.08) line CDI-SF was correlated with child age, with older chil-
Asian 4 (3%) 28.25 (6.13) 11.50 (1.91)
dren having lower scores (correlation coefficient 2.22,
Biracial 21 (15%) 29.48 (4.73) 12.29 (2.49)
Childs gender p .01).
Male 76 (54%) 29.08 (4.05) 12.92 (3.43) Over the three visits (57), 372 plasma IL-6, and 353 NK
Female 65 (46%) 28.23 (4.31) 12.32 (2.77) cell assays from 138 individuals were collected for analysis.
Childs age Two subjects plasma IL-6 concentrations were excluded
7.19.3 years 70 (50%) 28.70 (4.38) 13.34 (3.81)
from analysis due to extreme outlying values. The intraclass
9.413.1 years 71 (50%) 28.68 (3.99) 11.96 (2.13)
Household income correlation coefficients for plasma IL-6 concentration and
$25K 52 (37%) 29.23 (4.21) 13.46 (4.08) NK cell function were .35 and .44, respectively, indicating
$26K55K 48 (34%) 28.96 (4.59) 12.10 (2.51) moderate stability. Plasma IL-6 concentrations had an overall
$56K95K 27 (19%) 27.93 (3.12) 12.63 (2.26) range of 0.1211.50 pg/ml and NK cell function ranged from
.$95K 14 (10%) 27.21 (4.19) 11.50 (1.74)
1.20 to 22.35 (dilution corresponding to 20% lysis). The total
Note: Age and income at time of first assessment of depression and self-
number of illnesses recorded between Visits 5 and 7 was 190
efficacy. Age is presented at the median split (see details on age interactions). and the number of febrile illnesses was 38.

Childrens self-report measures and immune responses:


self-withdrawal due to lack of interest (24%), and relocation Perceived self-efficacy
away from area (14%). Sample attrition was unrelated to
any measured child or parent characteristic, family income, Children reporting higher perceived self-efficacy had lower
or family stress and personal distress measures completed plasma IL-6 concentrations. After controlling for childrens
by caregivers at enrollment. The assumption of data missing age, sex, BMI, parent psychiatric symptoms, family income,
completely at random was tested and found to be valid. and previous IL-6 concentrations (Visit 4), for each 1-U in-
crease in PSE score, the log10 of plasma IL-6 concentration
decreased by 0.01 pg/ml (95% confidence interval [CI]
Preliminary analyses 20.020.00; Table 3). The model indicated that children
Table 1 includes basic demographic data on the sample. Chil- with higher BMI also had elevated plasma IL-6 concentra-
drens self-reported depression symptoms were moderately tions over the 1-year period (for each 1-U increased in BMI
stable over the three visits spanning one year (Visits 57), the log10 of plasma IL-6 concentrations increased by 0.01
as shown by an intraclass correlation coefficient of 0.49 for pg/ml (95% CI 0.010.02); in addition, occurrence of ill-
CDI-SF scores (subject as the class). The intraclass correla- ness in the 2 weeks before the study visit was associated
tion coefficient for PSE scores was 0.26. Intercorrelations with an 0.08-U increase in log10 of plasma IL-6 (95% CI
among the study variables at baseline were tested with Fish-
ers Z transformation test and are presented in Table 2. Chil- Table 3. Adjusted mean differences in the log10 of IL-6
drens depression symptoms and perceived self-efficacy levels for study variables
scores were not significantly correlated. Parent self-reported
IL-6 Estimate

Variables Estimate 95% CI p


Table 2. Pearson correlations among variables at Visit 5
Depression 0.00 20.01, 0.01 .83
CDI Efficacy BSI Income Age Self-efficacy 20.01 20.02, 20.00 .03
BMI 0.01 0.01, 0.02 ,.01
CDI Parent BSI 0.02 20.04, 0.09 .51
Efficacy 0.03 Income 20.004 2 0.02, 0.01 .58
BSI 0.28* 20.07 Sick prior 14 days 0.08 0.02, 0.14 .01
Income 20.16* 20.12 20.19*
Age 20.22* 0.03 20.09 0.11 Note: CI, confidence interval; BMI, body mass index; BSI, Brief Symptom
Inventory. For each variable, the estimates correspond to the change in log10
Note: CDI, Child Depression Inventory; BSI, Brief Symptom Inventory. interleukin 6 (IL-6) level of 1-U increase. Analyses were adjusted for child
*p .05. age, sex, and prior IL-6 levels (Visit 4).
1144 M. T. Caserta et al.

0.020.14). Estimates reported in Table 3 are based on log10 cell scale, each 1-U increase in depression increased NK
transformation because of the distribution of IL-6, but that function by 0.27. For younger females, older males, and
makes interpretation of effect size difficult; accordingly, we younger males, no association was found between self-
also report, for illustrative purposes only, nontransformed es- reported depression and NK function.
timates of the raw IL-6 scale for significant predictors: for
self-efficacy, for each 1-U increase in PSE score, the plasma
Childrens self-report measures and health
IL-6 concentration decreased by 0.03 pg/ml; for each 1-U in-
crease in BMI, the plasma IL-6 concentrations increased by In the full cohort of children, neither PSE nor CDI-SF mea-
0.01 pg/ml; occurrence of illness in the 2 weeks before the sured at Visit 5 was associated with total illnesses over the
study visit was associated with an 0.32-U increase in plasma subsequent 1-year period (controlling for initial level of ill-
IL-6. nesses), nor with the likelihood of having a febrile illness.
Perceived self-efficacy was not associated with NK cell However, analyses stratified by age and sex showed that
function ( p .27) or with illnesses ( p .13) or febrile ill- higher depression scores were associated with increased ill-
nesses ( p .07). Further analyses assessing interactions by nesses in the following year only among older girls (Table 5):
age and sex did not detect any evidence of moderation. each 1-U increase in depression symptoms was associated an
None of the two-way or three-way interactions were statistical increased rate of total illnesses of 18% (rate ratio 1.18, 95%
significant ( p . .05). CI 1.071.30). Febrile illnesses, a subset of total child ill-
nesses, showed the same effect: for older girls, higher depres-
sion scores predicted an increased risk of having at least one
Childrens self-report measures and immune responses:
febrile illness during the next year (odds ratio 1.62, 95% CI
Depressive symptoms
1.052.49). There was no significant association between
Over the 1-year period, there was not a main effect association depression and illnesses found in the other groups. There
between depression and IL-6. Moderation analyses indicated was not a significant association between self-efficacy and ill-
no interactions: the lack of significant association between de- nesses in the sample as a whole or when the sample was stra-
pression and IL-6 was consistent across age and sex and Age tified by age and sex.
Sex subgroups (younger female, p .82, younger male, p Mediation analyses were limited to older girls because it
.94; older female, p .25, older male, p .81). was only in the subsample of girls that we obtained links be-
Childrens NK cell function was not associated with de- tween depression and illness and depression and NK cell
pression scores in the full cohort over the 1-year study, but function. However, we found no evidence that the relation-
there was evidence of moderation by age (using the median ship between depression and illnesses were mediated by
split) and sex. A stratified subgroup analyses showed that de- NK cell function.
pression was positively associated with NK function in older
girls but not in any of the other three groups (Table 4). Further
Discussion
analyses of the older girl subsample indicated that, after con-
trolling for self-efficacy, parent psychiatric symptoms, family In a prospective, 1-year longitudinal study of a normative,
income, being sick in the prior 14 days, and previous NK healthy sample of preadolescent children, we found modest
function (Visit 4), each 1-U increase in depression (CDI- support for a link between psychological well-being and im-
SF) score increased the log10 of NK function by 0.02 U mune functioning and illness. The strongest link, observed in
(95% CI 0.000.04). For the nontransformed (raw) NK the entire cohort of children, was a negative association be-
tween self-efficacy and IL-6. Two further associations were
also detected, but only in older girls: depression was associ-
Table 4. Adjusted mean differences in the log10 of NK ated with increased NK cell function and with higher rates of
function for depression in females and males stratified illness.
by age The findings extend research on stress and health and allo-
static load in children in several important ways. First, rather
NK Function than rely on parent reports of stress, we included childrens
self-reports of well-being, indexed by efficacy and depressive
Depression Estimates by Group Estimate 95% CI p symptoms as predictors of immunity and health. These two
indicators of psychological well-being are widely researched
Female age 9.3 years 0.00 20.02, 0.02 .91
Male age 9.3 years 0.01 20.01, 0.02 .36 in the developmental and clinical psychology literature as
Female age . 9.3 years 0.02 0.00, 0.04 .02 causes and effects of psychosocial stress and as predictors
Male age . 9.3 years 20.02 20.03, 0.01 .37 of long-term behavioral adjustment. We provide some of
the only evidence so far that these psychological measures
Note: For each variable, the estimates correspond to the change in log10 nat- also predict health outcomes in normally developing chil-
ural killer (NK) cell function of 1-U increase. Analyses were adjusted for
child age, sex, parent psychiatric symptoms (Brief Sympton Inventory), in- dren. The self-report data also provides an extension to the
come, and prior NK function. prior research on stress and childrens health, which has to
Stress exposure and/or psychological symptoms and health in children 1145

Table 5. Estimated rate ratios for the association between depression and total illnesses and febrile illnesses
for females and males stratified by age

Total Illnesses Febrile Illnesses

Depression Estimates by Group RR 95% CI p RR 95% CI p

Female age 9.3 years 0.95 0.84, 1.07 .39 0.89 0.60, 1.30 .54
Male age 9.3 years 1.07 0.98, 1.16 .12 1.06 0.89, 1.26 .53
Female age . 9.3 years 1.18 1.07, 1.30 ,.01 1.62 1.05, 2.49 .03
Male age . 9.3 years 1.01 0.84, 1.22 .87 1.34 0.85, 2.12 .21

Note: Outcome was febrile illness in a 1-year period after assessment of depression and self-efficacy. Analyses were adjusted for child age, sex,
parent psychiatric symptoms (Brief Sympton Inventory), income, and illnesses in the previous year. RR, rate ratio.

date been dominated by parent reports of child stress exposure that we were not able to investigate in our sample. In any event,
and, in some cases, inferred from parent self-report of their it may be noteworthy that a link with NK cell function was
own psychological symptoms. The finding that childrens found only in older girls, who may, by virtue of the age and
own views of their ability to cope with stress predicted lower sex effects on depression, present the most adultlike in terms
levels of IL-6 extends a considerable adult literature on psy- of risk for depression. That a limited or focused effect of de-
chological outlook and immunity. For example, positive pression in older girls was also detected for illnesses also war-
emotional style, assessed by measures of extraversion, agree- rants consideration and implies, at a minimum, that age and sex
ableness, and positive relationship style is associated with may emerge as modifiers of the links between stress/psycho-
less colds following respiratory viral challenge (Cohen, Al- logical symptoms and immunity in pediatric samples. It is
per, Doyle, Treanor, & Turner, 2006; Cohen et al., 2003). noteworthy that sex effects on the link between depression
In addition, increased optimism was inversely associated and immune markers have also been reported in adult studies
with an IL-6 response to acute psychosocial stress (Brydon, (Brummett et al., 2010; Steptoe, ODonnell, Badrick, Kumari,
Walker, Wawrzyniak, Chart, & Steptoe, 2009), whereas pes- & Marmot, 2008). Further work that incorporates the hormonal
simism was associated with increased levels of IL-6, C-reac- and neurophysiological changes associated with puberty (e.g.,
tive protein, and fibrinogen (Roy et al., 2010). Coping, ap- Dahl & Gunnar, 2009) is needed to revise and expand a devel-
praisal and other psychological resources and dispositions opmental model of psychoneuroimmunology.
need also to be incorporated into models of stress and health A nonfinding also deserves attention for what it may suggest
in pediatric samples. It was significant that the prediction about developmental influences and the detection of allostatic
from self-efficacy to IL-6 was independent of parental symp- load in young people, namely, the lack of association between
toms, providing further support for the need to consider psy- depression and IL-6. This lack of association in the current
chological processes within the child as distinct from general sample does contradict adult work that shows a reliable but
markers of stress exposure as reported on by parents. small negative association between depressive symptoms and
Second, results from the study begin to piece together a de- IL-6 (Dowlati et al., 2010; and tumor necrosis factor-alpha,
velopmental model of psychoneuroimmunology. On the one which was not assessed in the current study). That raises the
hand, children who report greater ability to achieve desired re- possibility that, for reasons not yet clear, associations between
sults in common situations showed a healthier, and presumably behavioral/psychological well-being and IL-6 either may be
more adaptive immunological profile and reduced levels of IL- less robust in children or slower to develop, perhaps requiring
6; evidence of positive psychological appraisal and coping was chronic stress or depression, a possibility implicit in the adult
associated with reduced generalized inflammation, a prediction work and in the allostatic load model but not yet adequately
that comports easily with existing animal and adult work. On tested. In other words, our findings on a preadolescent sample
the other hand, in older girls (only), we found that depression do not mimic in every way the patterns so far reported in adults.
was associated with increased NK function or cytotoxicity. Finally, we did not find evidence that changes in immune
That is congruent with our earlier finding that psychosocial markers mediated the effects of depression on health out-
stress as reported by parents predicted increased NK cell func- comes. In fact, we found, in older girls, depression leading
tion (Caserta et al., 2008). The extent to which the increased to both increased NK cell function and increased illness. It
NK cell function associated with depression is compatible may be that the impact on immunity and on observable health
with or contradicts adult studies is confounded by the inconsis- outcomes are under different time course constraints; it is also
tency across adult studies, with some reporting a decrease (Park likely that health outcomes are multiply determined and that
et al., 2006) but others reporting an increase (Ravindran, Grif- any specific immune marker may not carry sufficient impact
fiths, Merali, & Anisman, 1998; Seidel et al., 1996) in NK cells to alter a systemic illness.
in depression. These discrepancies may be resolved by atten- Several limitations of the study should be noted. Detailed
tion to depression subtypes (Ravindran et al., 1998), something data on pubertal development were not available in this study;
1146 M. T. Caserta et al.

more precise measurement of pubertal changes may help age and sex, replication of these findings is necessary before
clarify the moderating influence of sex and age on depres- drawing firm conclusions. Finally, we proposed directional
sion and IL-6 and illnesses. The young age of the sample hypotheses concerning within-individual change, derived
may also confound interpretations of age and sex; indeed, from adult and animal work, that changes in depression
we did not detect strong age and gender effects on levels and self-efficacy would be associated with immune markers,
of depressive symptoms that are found in pubertal or adoles- the dependent variable. It is plausible that the direction of
cent samples. In addition, the interpretation of depression effects is reciprocal, and the time course for charting the
symptoms should be considered in light of the generally dynamic interplay between psychological well-being and
nonclinical nature of symptoms and impairment; arguably, health and immune markers requires further study. These
a strength of this study, which would increase its generaliz- limitations are offset by several strengths of the paper, in-
ability, is that we detected associations within a healthy sam- cluding the longitudinal design with multiple occasions of
ple, outside of the clinical extremes. A variety of other po- measurement of psychological and immune data, and de-
tential markers of immunity that could have been assessed tailed health diary data verified, to a considerable extent,
were not included. We targeted those that had a sizable adult by medical investigators on the project.
literature, and so had a limited focus; analyses of a wider ar- Clinical implications of the findings are not yet clear, but
ray of pro- and anti-inflammatory markers will be needed to promising. Perhaps the most obvious is the need to consider
develop a comprehensive developmental model of psycho- physical well-being in those children who report psycholog-
neuroimmunology. Furthermore, although the study sample ical distress, in both assessment and intervention. If further re-
size was adequately powered for main effects analyses and search also endorses links between psychological well-being
compares favorably with studies in the area, the power to de- and physical health in children, then this would provide a con-
tect multiway interactions was limited and the reliability of ceptualbiological basis for instituting an interdisciplinary
interactions is notably less robust than main effects. There- model for child health assessment, something that is not
fore, despite the strong a priori case for analyzing data by now widely established.

References
Angold, A., Costello, E. J., Erkanli, A., & Worthman, C. M. (1999). Pubertal Cohen, S., Alper, C. M., Doyle, W. J., Treanor, J. J., & Turner, R. B. (2006).
changes in hormone levels and depression in girls. Psychological Medi- Positive emotional style predicts resistance to illness after experimental
cine, 29, 10431053. exposure to rhinovirus or influenza a virus. Psychosomatic Medicine,
Angold, A., Costello, E. J., & Worthman, C. M. (1998). Puberty and depres- 68, 809815.
sion: The roles of age, pubertal status and pubertal timing. Psychological Cohen, S., Doyle, W. J., & Skoner, D. P. (1999). Psychological stress, cyto-
Medicine, 28, 5161. kine production, and severity of upper respiratory illness. Psychosomatic
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral Medicine, 61, 175180.
change. Psychological Review, 84, 191215. Cohen, S., Doyle, W. J., Turner, R., Alper, C. M., & Skoner, D. P. (2003).
Bandura, A., & Cervone, D. (1983). Self-evaluative and self-efficacy mecha- Sociability and susceptibility to the common cold. Psychological Sci-
nisms governing the motivational effects of goal systems. Journal of Per- ence, 14, 389395.
sonality and Social Psychology, 45, 10171028. Cohen, S., Tyrrell, D. A., & Smith, A. P. (1991). Psychological stress and sus-
Baron, P. M., & Kenny, D. A. (1986). The moderatormediator variable dis- ceptibility to the common cold. New England Journal of Medicine, 325,
tinction in social psychological research: Conceptual, strategic, and statis- 606612.
tical considerations. Journal of Personality and Social Psychology, 51, Cowen, E. L., Work, W., Hightower, A., Wyman, P., Parker, G., & Lotyc-
11731182. zowski, B. (1991). Toward the development of self-efficacy in children.
Barlett, J. A., Schleifer, S. J., Demetrikopoulos, M. K., & Keller, S. E. (1995). Journal of Clinical Child Psychology, 20, 169178.
Immune differences in children with and without depression. Biological Cowen, E. L., Wyman, P. A., Work, W. C., Kim, J. Y., Fagen, D. B., & Mag-
Psychiatry, 38, 771774. nus, K. B. (1997). Follow-up study of young stress-affected and stress-
Birmaher, B., Rabin, B. S., Garcia, M. R., Jain, U., Whiteside, T. L., William- resilient urban children. Development and Psychopathology, 9, 565577.
son, D. E., et al. (1994). Cellular immunity in depressed, conduct disor- Dahl, R. E., & Gunnar, M. R. (2009). Heightened stress responsiveness and
der, and normal adolescents: Role of adverse life events. Journal of the emotional reactivity during pubertal maturation: Implications for psycho-
American Academy of Child & Adolescent Psychiatry, 33, 671678. pathology. Development and Psychopathology, 21, 16.
Bromelow, K. V., Galea-Lauri, J., OBrien, M. E., & Souberbielle, B. E. Derogatis, L. R., Dellapietra, L., & Kilroy, V. (Ed.). (1992). Screening for
(1998). A highly sensitive whole blood natural killer cell assay. Journal psychiatric disorder in medical populations. Amsterdam: Elsevier.
of Immunological Methods, 217, 177184. Dowlati, Y., Herrmann, N., Swardfager, W., Liu, H., Sham, L., Reim, E. K.,
Brummett, B. H., Boyle, S. H., Ortel, T. L., Becker, R. C., Siegler, I. C., & et al. (2010). A meta-analysis of cytokines in major depression. Biolog-
Williams, R. B. (2010). Associations of depressive symptoms, trait hos- ical Psychiatry, 67, 446457.
tility, and gender with C-reactive protein and interleukin-6 response after Eisenberger, N. I., Inagaki, T. K., Mashal, N. M., & Irwin, M. R. (2010). In-
emotion recall. Psychosomatic Medicine, 72, 333339. flammation and social experience: An inflammatory challenge induces
Brydon, L., & Steptoe, A. (2005). Stress-induced increases in interleukin-6 feelings of social disconnection in addition to depressed mood. Brain,
and fibrinogen predict ambulatory blood pressure at 3-year follow-up. Behavior, and Immunity, 24, 558563.
Journal of Hypertension, 23, 10011007. Granger, D. A., Hood, K. E., Dreschel, N. A., Sergeant, E., & Likos, A.
Brydon, L., Walker, C., Wawrzyniak, A. J., Chart, H., & Steptoe, A. (2009). (2001). Developmental effects of early immune stress on aggressive, so-
Dispositional optimism and stress-induced changes in immunity and cially reactive, and inhibited behaviors. Development and Psychopathol-
negative mood. Brain, Behavior, and Immunity, 23, 810816. ogy, 13, 599610.
Caserta, M. T., OConnor, T. G., Wyman, P. A., Wang, H., Moynihan, J., Hennessy, M. B., Deak, T., & Schiml-Webb, P. A. (2010). Early attachment-
Cross, W., et al. (2008). The associations between psychosocial stress figure separation and increased risk for later depression: Potential media-
and the frequency of illness, and innate and adaptive immune function tion by proinflammatory processes. Neuroscience and Biobehavioral
in children. Brain, Behavior, and Immunity, 22, 933940. Reviews, 34, 782790.
Stress exposure and/or psychological symptoms and health in children 1147

Holahan, C. J., Pahl, S. A., Cronkite, R. C., Holahan, C. K., North, R. J., & Roy, B., Diez-Roux, A. V., Seeman, T., Ranjit, N., Shea, S., & Cushman, M.
Moos, R. H. (2010). Depression and vulnerability to incident physical ill- (2010). Association of optimism and pessimism with inflammation and
ness across 10 years. Journal of Affective Disorders, 123, 222229. hemostasis in the Multi-Ethnic Study of Atherosclerosis (MESA). Psy-
Ickovics, J. R., Hamburger, M. E., Vlahov, D., Schoenbaum, E. E., Schuman, chosomatic Medicine, 72, 134140.
P., Boland, R. J., et al. (2001). Mortality, CD4 cell count decline, and de- Rozanski, A., Blumenthal, J. A., & Kaplan, J. (1999). Impact of psycholog-
pressive symptoms among HIV-seropositive women: Longitudinal anal- ical factors on the pathogenesis of cardiovascular disease and implica-
ysis from the HIV epidemiology research study. Journal of the American tions for therapy. Circulation, 99, 21922217.
Medical Association, 285, 14661474. Schleifer, S. J., Bartlett, J. A., Keller, S. E., Eckholdt, H. M., Shiflett, S. C., &
Irwin, M. (2002). Psychoneuroimmunology of depression: Clinical implica- Delaney, B. R. (2002). Immunity in adolescents with major depression.
tions. Brain, Behavior, and Immunity, 16, 116. Journal of the American Academy of Child & Adolescent Psychiatry,
Juster, R.-P., McEwen, B. S., & Lupien, S. J. (2010). Allostatic load biomark- 41, 10541060.
ers of chronic stress and impact on health and cognition. Neuroscience Schleifer, S. J., Keller, S. E., & Bartlett, J. A. (2002). Panic disorder and im-
Biobehavioral Reviews, 35, 216. munity: Few effects on circulating lymphocytes, mitogen response, and
Karelina, K., & DeVries, A. C. (2011). Modeling social influences on human NK cell activity. Brain, Behavior, and Immunity, 16, 698705.
health. Psychosomatic Medicine, 73, 6774. Seidel, A., Arolt, V., Hunstiger, M., Rink, L., Behnisch, A., & Kirchner, H.
Kiecolt-Glaser, J. K., Gouin, J.-P., Weng, N.-P., Malarkey, W. B., Bevers- (1996). Increased CD56 natural killer cells and related cytokines in ma-
dorf, D. Q., & Glaser, R. (2011). Childhood adversity heightens the im- jor depression. Clinical Immunology and Immunopathology, 78, 8385.
pact of later-life caregiving stress on telomere length and inflammation. Shain, B. N., Kronfol, Z., Naylor, M., Goel, K., Evans, T., & Schaefer, S.
Psychosomatic Medicine 73, 1622. (1991). Natural killer cell activity in adolescents with major depression.
Kovacs, M. (1992). The Childrens Depression Inventory. New York: Multi- Biological Psychiatry, 29, 481484.
Health Systems. Shelton, R. C., & Miller, A. H. (2010). Eating ourselves to death (and de-
Kronfol, Z. (1983). Cancer and depression. British Journal of Psychiatry, 142, spair): The contribution of adiposity and inflammation to depression.
309. Progress in Neurobiology, 91, 275299.
Masten, A. S., & Coatsworth, J. D. (1998). The development of competence Steptoe, A., ODonnell, K., Badrick, E., Kumari, M., & Marmot, M. (2008).
in favorable and unfavorable environments. Lessons from research on Neuroendocrine and inflammatory factors associated with positive affect
successful children. American Psychologist, 53, 205220. in healthy men and women: The Whitehall II study. American Journal of
Miller, G., Chen, E., & Cole, S. W. (2009). Health psychology: Developing Epidemiology, 167, 96102.
biologically plausible models linking the social world and physical Targum, S. D., Clarkson, L. L., Magac-Harris, K., Marshall, L. E., &
health. Annual Review of Psychology, 60, 501524. Skwerer, R. G. (1990). Measurement of cortisol and lymphocyte sub-
Miller, G. E., & Chen, E. (2010). Harsh family climate in early life presages populations in depressed and conduct-disordered adolescents. Journal
the emergence of a proinflammatory phenotype in adolescence. Psycho- of Affect Disorders, 18, 9196.
logical Science, 21, 848856. Wolf, J. M., Miller, G. E., & Chen, E. (2008). Parent psychological states pre-
Musselman, D. L., Evans, D. L., & Nemeroff, C. B. (1998). The relationship dict changes in inflammatory markers in children with asthma and
of depression to cardiovascular disease: Epidemiology, biology, and healthy children. Brain, Behavior, and Immunity, 22, 433441.
treatment. Archives of General Psychiatry, 55, 580592. Worthman, C. M., & Panter-Brick, C. (2008). Homeless street children in
Pace, T. W., & Miller, A. H. (2009). Cytokines and glucocorticoid receptor Nepal: Use of allostatic load to assess the burden of childhood adversity.
signaling. Relevance to major depression. Annals of the New York Acad- Devopment and Psychopathology, 20, 233255.
emy of Sciences, 1179, 86105. Wyman, P. A., Cowen, E. L., Work, W. C., Hoyt-Meyers, L., Magnus, K. B.,
Pace, T. W. W., Mletzko, T. C., Alagbe, O., Musselman, D. L., Nemeroff, C. & Fagen, D. B. (1999). Caregiving and developmental factors differen-
B., Miller, A. H., et al. (2006). Increased stress-induced inflammatory re- tiating young at-risk urban children showing resilient versus stress-
sponses in male patients with major depression and increased early life affected outcomes: A replication and extension. Child Development,
stress. American Journal of Psychiatry, 163, 16301633. 70, 645659.
Padgett, D. A., & Glaser, R. (2003). How stress influences the immune re- Wyman, P. A., Cowen, E. L., Work, W. C., & Parker, G. R. (1991). Devel-
sponse. Trends in Immunology, 24, 444448. opmental and family milieu correlates of resilience in urban children
Park, E. J., Lee, J. H., Chae, J. H., Lee, K. H., Han, S. I., & Jeon, Y. W. who have experienced major life stress. American Journal of Community
(2006). Natural killer T cells in patients with major depressive disorder. Psychology, 19, 405426.
Psychiatry Research, 144, 237239. Zeger, S. L., & Liang, K. Y. (1986). Longitudinal data analysis for discrete
Ravindran, A. V., Griffiths, J., Merali, Z., & Anisman, H. (1998). Circulating and continuous outcomes. Biometrics, 42, 121130.
lymphocyte subsets in major depression and dysthymia with typical or
atypical features. Psychosomatic Medicine, 60, 283289.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.