Journal of Psychosomatic Research 53 (2002) 907  911
Depression in Type 1 diabetes in children
Natural history and correlates
Margaret Grey*, Robin Whittemore, William Tamborlane
Yale School of Nursing, 100 Church Street South, New Haven, CT 06519, USA
Abstract
The combination of diabetes and depression in children and depression, and adolescents up to three-fold greater, than youth
adolescents is largely unstudied. The purpose of this article is to without diabetes. Correlates of depression and diabetes include
review the literature on the natural history and correlates of co- gender, poorer metabolic control, and family behaviors. Very little
morbid diabetes and depression in children and adolescents. is known about treatment in these youth, and more studies are
Children with diabetes have a two-fold greater prevalence of indicated. © 2002 Elsevier Science Inc. All rights reserved.
Keywords: Type 1 diabetes; Children; Adolescents; Depression
Introduction image, peer group pressure, autonomy from the parents and
identity formation [7]. De Groot et al. [8] propose several
Depression is a serious health problem and one that possible mechanisms for an association between depression
affects a large percentage of those individuals suffering with and poor glycemic control, including neuropsychological
a chronic illness. According to Connell et al. [1], several impacts of depression on memory and diabetes self-care
factors contribute to the increased prevalence of depression knowledge, persistent subthreshold depression s effect on
among those with a chronic illness, including the effect of adherence and the effect of negative attitudes associated
the disease on physical functioning and activity, social with diabetes self-care. Depression may be associated with
relationships, quality of life and morale. Lifetime prevalence attitudinal or personality variables such as low self-esteem,
rates of major depression among adults with Types 1 and 2 pessimism, poor concentration and loss of interest in daily
diabetes range between 14.4% and 32.5% [2,3]. Depression activities. Thus, adults with diabetes and depression have
has a profound adverse influence on quality of life and poorer daily functioning and quality of life [9], poorer
overall functioning and has additional repercussions with adherence to the diabetes regimen [3], poorer metabolic
individuals with diabetes because of its association with control [10,11] and an increased risk for both micro- and
poor management [4]. This association is poorly understood macrovascular complications than adults with diabetes but
but has been hypothesized to operate through depression- not depression [12,13].
induced abnormalities in neuroendocrine and neurotrans- Adults tend to have the view that childhood is an idyllic
mitter function [5], through decreased compliance with time, and that depression is not as common as it is in
diabetes management, or because of as-yet-unknown com- adulthood. In general, this assumption would be correct.
plex behavioral physiologic interaction [6]. If these hypo- Nonetheless, the combination of diabetes and depression in
theses can be supported, then new approaches to children, and especially adolescents, is important because:
interventions, combining pharmacologic and cognitive (1) it is associated with 10-fold increase in suicide and
behavioral approaches, may be helpful. For adolescents, suicidal ideation (and children and adolescents who take
the challenge of diabetes is combined with the devel- insulin have a ready method of performing suicide if they
opmental tasks of adapting to puberty and a changing body wish to) [14,15]; (2) recurrence and course may be more
severe than in adults, with some studies showing that
depression tends to be more severe, take longer to resolve
* Corresponding author. the initial episode and is more likely to recur than in youth
0022-3999/02/$  see front matter © 2002 Elsevier Science Inc. All rights reserved.
PII: S 0022-3999(02)00312-4
908 M. Grey et al. / Journal of Psychosomatic Research 53 (2002) 907 911
without diabetes [16]; (3) it may be associated with poorer
metabolic control in diabetes, which may lead to compli-
cations and other poorer outcomes [6]; and (4) several
studies have suggested that youth with diabetes and
depression may be likely to have other comorbid condi-
tions, such as eating disorders, adjustment disorders or
anxiety disorders [6,16,17].
Natural history
The prevalence of depression in children and adolescents
without diabetes ranges from 0.4% to 8.3%, and increases
markedly from childhood into adolescence [18 20]. A
number of risk factors have been identified that increase
the likelihood of depression during childhood, including
female gender, family dysfunction and stressful experiences.
Fig. 1. Point prevalence of CDI scores 13 and duration of diabetes in
One stressful experience that may increase risk for depres-
102 adolescents with Type 1 diabetes (CDI scores 13 or greater indicate
sion in the child or adolescent is diabetes.
clinical depression).
There are few population-based studies of children and
adolescents with diabetes, but the prevalence of depression
in this population ranges from two- to threefold that of young adults, males may find the diabetes to have a more
peers without diabetes [21]. These rates are significantly profound effect on work and social life than females.
higher in youth with diabetes as indicated by a recent study In a recent analysis conducted with a cohort of adoles-
by Kokkonen and Kokkonen [22] in which the prevalence cents in our clinic who were participating in a trial of
of depression was reported to be 12% in Scandinavian intensive management, we found that the overall prevalence
children aged 8 12 years and 18% in adolescents. How- of depressive symptoms was 17%. Note that this was a
ever, these rates may vary according to the duration of somewhat selected sample because it was a group recruited
illness, in addition to geographic variations. Kovacs et al. for a study of intensive management; thus, the true preval-
[23 26], in a series of studies, followed an onset cohort of ence may be higher or lower. Further, we found that
children, 8 13 years of age, for 6 years. The authors used duration of diabetes was significantly correlated with
the Children s Depression Inventory (CDI) [27] to measure depressive symptoms in a U-shaped distribution, as shown
depressive symptoms. Although the sample size decreased in Fig. 1. We found that depressive symptoms were more
significantly over the 6 years of follow-up, they [26] found common in the earlier years postdiagnosis, less common
that levels of depressive symptoms remained relatively between 4 and 9.9 years after diagnosis and rose again after
stable, but the presence of depressive symptoms at dia- 10 years.
gnosis was associated with a higher prevalence of depres-
sion later. In contrast, Grey et al. [28] compared youth with
newly diagnosed diabetes with a cohort of age- and gender- Correlates
matched peers, and found that the youth with diabetes
reported significantly higher depressive symptomatology Researchers working in the area of childhood depression
than those without diabetes at the time of the diabetes have identified several correlates of depression in children
diagnosis and then again at the end of the second year. without diabetes [30]. As noted earlier, age is important,
They characterized this second period of depression as since adolescents are significantly more prone to depression
being associated with the end of the physiologic hon- than children at earlier ages. Gender is also important
eymoon period and the necessity of learning to live with [31,32]. Prior to adolescence, most studies suggest that
diabetes lifelong. there are no gender differences in prevalence. In adoles-
Jacobson et al. [29] also followed an onset cohort of cents, however, girls are consistently found to have more
young adults (ages 19 26 years) for 10 years. Using the depression than boys [31,32]. Maternal depression is also
depression subscale of the SCL-90, they found that depres- associated with a higher incidence of depression in the
sion was not significantly different among those with dia- children [7]. Family stress or dysfunction appears to pre-
betes and a comparison group who had a moderately severe dispose children and adolescents to depression [7]. And,
acute illness. In addition, they found that males had signific- finally, any stressor, such as health status, illness or injury,
antly more depression (and another psychiatric symptoms) can be associated with depression in children.
over time. While the finding that males had more depression In children and adolescents with diabetes, some of the
than females was unexpected, the authors suggest that in same correlates have been found. For example, Jacobson
M. Grey et al. / Journal of Psychosomatic Research 53 (2002) 907 911 909
et al. [29] found that boys with diabetes were significantly
more likely to be depressed after 10 years than girls with
diabetes or a sample of youth who had a severe acute illness.
On the other hand, in the study of LaGreca et al. [32] of
42 adolescents with Type 1 diabetes, they found girls to be
significantly more depressed that boys.
A few anecdotal reports suggested that poorer metabolic
control of diabetes was associated with depression in youth
with diabetes. Although this finding is not consistent,
studies appear to suggest that approximately 20% of the
variance in metabolic control is statistically explained by
depression in youth [21,26,32 35]. With the exception of
the work of LaGreca et al., gender differences were not
explored in these studies.
In our cohort of 102 adolescents with Type 1 diabetes,
we have begun to explore correlates of depression. First we
looked to see if depression was associated with age, gender
Fig. 2. Comparison of CDI scores in youth with diabetes with positive
and socioeconomic status, and found no correlations. Then
history of psychosocial problems or no history of psychosocial problems
we examined the data for relationships with acute compli-
(CDI scores 13 or greater indicate clinical depression).
cations, such as weight gain or severe hypoglycemia, and
again found no correlations. We also looked for correlations
with the presence of microvascular complications (as was Other outcomes that have been associated with depres-
true in adults), and again found none. Finally, we examined sion in teens in at least one study include: Poorer quality of
the data for correlations with family variables, using the life [35], poorer adherence to the diabetes regimen of
FACES II scale [36] and the Diabetes Family Behavior glucose monitoring, insulin injections and diet [6,32], lower
Scale [37]. We found that those teens who reported lower self-esteem [15], an increased incidence of disordered eating
family adaptability (r = .41, P < .001), lower family cohe- [6] and suicide/suicidal ideation [14].
sion (r = .43, P < .001) and less warmth and caring dia- Even less is known about the mechanism responsible
betes family behaviors (r = .44, P < .001) were more likely for the relationship between depression and diabetes in
to have depressive symptoms than those with higher family youth. The usual assumption has been that loss of self-
functioning. Interestingly, guidance and control in relation- esteem and poor adjustment to the chronic illness were
ship to diabetes family behaviors were not associated with responsible. Fear of diabetes-related complications and
depression (r =.23, ns) in this group of teens. In addition, we helplessness to avoid them may lead to poorer self-care.
examined whether a history of psychosocial problems Our data suggest that preexisting family functioning may
(pervasive or persistent problems related to adjustment be useful in understanding some of these phenomena. But
disorders, family conflict/stress, rebellious or aggressive new hypotheses about the potential effects of immune
behavior, mood disorders or compulsive behavior as functioning and/or glutamic acid decarboxylase (GAD)
reported to the clinician at intake by the adolescent or their autoantibodies effects on the synthesis of the major
parent) was associated with depression. As shown in Fig. 2, inhibitory neurotransmitter (gamma-aminobutyric acid,
youth who had a history of psychosocial problems were GABA) in the beta cell and central nervous system are
significantly more likely to report depressive symptoms than beginning to emerge, leading some to conclude that the
youth who did not report such a history. mechanism may be physiological. Since disturbances in
We used multiple regression to examine factors that GABAergic function are hypothesized to contribute to the
were associated with depressive symptoms at baseline in neurobiology of major depression [38,39], and approxi-
the adolescent study. Initially, the model included all of the mately 70% of adolescents with Type 1 diabetes produce
variables listed above, and stepwise regression yielded the GAD autoantibodies [40], these autoantibodies may play a
model shown in Table 1. The bestfitting model found that role in the development of both diabetes and depression. It
reports of lower warm and caring family behaviors, lower may also be that counterregulatory hormones, such as
family adaptability and a history of psychosocial difficult- cortisol released during times of stress, that impair glucose
ies remained significantly associated with depressive tolerance also make diabetes control more difficult.
symptoms, with about 24% of the variance explained. In Despite the myriad causes, little is known about how to
other words, youth who had a history of psychosocial treat these high-risk youth [41]. Indeed, optimal treatment
difficulties either individually or in the family and who of depression in youth without diabetes remains contro-
reported that they felt less warm and caring or adaptability versial. There have been very few well-controlled studies,
from their families in their diabetes management were and there have been no longer-term studies of drug
most likely to have higher depression scores. treatment  most of the studies simply look at short-term
910 M. Grey et al. / Journal of Psychosomatic Research 53 (2002) 907 911
Table 1
characterize the nature of the depression in these youth
Multiple regression results for CDI scores at baseline
and begin to explicate why depression has a temporal
Variable b SE b t P R2 change
relationship with diabetes duration in youth. Few studies
DFBS  warmth/caring 0.38 0.02 3.81 .000 .15 include control or comparison groups, so it is difficult to
FACES  adaptation 0.36 0.01 4.24 .005 .13
know if the correlates of depression in children without
History of psychosocial 0.20 0.30 2.02 .046 .06
diabetes are true in children with diabetes, except to
problems
compare with psychiatric populations in the literature. We
DFBS  guidance/control 0.05 0.02 0.47 .65 
need to understand the underlying mechanisms that explain
R2 =.28; adjusted R2 =.24; F = 15.27; P < .001.
the relationship between depression and diabetes and
whether pharmacologic treatment can alter this mechanism.
effects (6 8 weeks). Typically, such studies eliminate We also need a better understanding of family characteristics
youth with chronic conditions in the screening process. that support metabolic control and psychosocial health.
Further, most studies of treatment examined drug treatment Many clinicians believe that it is all about  compliance ;
compared to placebo, but comparisons of drug treatment that when a patient is depressed, they just cannot take care
with and without family or behavioral treatment are also of themselves adequately. But it may be that, just as insulin
needed. In addition, while there have been studies of the resistance explains a certain amount of the difficulties in
impact of family interventions on youth with diabetes metabolic control in teens, there is a physiologic explanation
[42,43], as well as on youth with depression, no studies for the poorer metabolic control in these youth as well as for
of the impact of family interventions on depression in the depression. Finally, controversy exists about the best
youth with diabetes were found. Insights into interventions methods for measuring and studying depression. Better
for families with difficulties may also be found in studying definitions of depression and its symptoms and more
resilient youth  those who face family conflict but who reliable methods to measure them need to be developed.
escape depression and other psychosocial effects. Interviews may be the most accurate for diagnosis of
depressive disorders, but they are not practical for large-
scale studies. Studies are needed that examine the optimal
Summary timing of preventive interventions, such as coping skills
training provided at high-risk times such as diagnosis and
The comorbidity of diabetes and depression in children before 10 years of diabetes duration, and the impact of
and adolescents is a significant problem, affecting up to interventions for diabetes on depression. In addition, inter-
20% of youth with diabetes compared to less than 7% of vention studies examining the impact of treatment of
youth without diabetes. Depression alone carries significant depression on metabolic control of youth with diabetes are
potential for disability, but when combined with diabetes, urgently needed.
the comorbidity carries the potential for serious long-term
consequences; depression is associated with poorer meta-
bolic control, and thus, such youth may be more at risk for
References
long-term complications. Poor metabolic control in youth
with diabetes and depression may be associated with both
[1] Connell CM, Fisher EB, Houston CA. Relationships among social
physiologic and behavioral factors, as depression may make
support, diabetes outcomes, and morale for older men and women.
adherence to the diabetes care regimen more difficult and
Behav, Health Aging 1992;1:63 75.
hypersecretion of cortisol may impair response to insulin. It
[2] Gavard JA, Lustman PJ, Clouse RE. Prevalence of depression in
is unclear whether diabetes may make adherence to inter- adults with diabetes: an epidemiological evaluation. Diabetes Care
1993;16:1167  78.
ventions for depression more difficult. There is no question
[3] Lustman PJ, Anderson RJ, Freedland KE, de Groot MK, Carney RM,
that youth who are depressed have poorer psychosocial
Clouse RE. Depression and poor glycemic control: a meta-analytic
outcomes and are more likely to have other psychiatric
review of the literature. Diabetes Care 2000;23:934  42.
disorders, such as adolescent adjustment disorders, eating
[4] Mazze RS, Lucido D, Shamoon H. Psychological and social correlates
disorders and anxiety disorders. Finally, these youth may be of glycemic control. Diabetes Care 1984;7:360 6.
[5] Lustman PJ, Griffith LS, Gavard JA, Clouse RE. Depression in adults
at high risk for suicide or suicidal ideation. Suicide and
with diabetes. Diabetes Care 1992;15:1631 9.
severe eating disorders may lead to death. It is important to
[6] Littlefield CH, Craven JL, Rodin GM, Daneman D, Murray MA,
note that family adjustment and support are critical corre-
Rydall AC. Relationship of self-efficacy and bingeing to adherence
lates of depression in youth with diabetes, suggesting that
to diabetes regimen among adolescents. Diabetes Care 1992;15(1):
family approaches, such as behavioral therapy for families, 90  4.
[7] Delamater A. Psychological aspects of diabetes mellitus in children.
may be effective in preventing or treating depression in
In: Lahey B, Kazdin E, editors. Advances in clinical childhood psy-
these youth.
chology. New York: Plenum, 1986. pp. 128  42.
On the basis of these few studies, we can conclude that
[8] de Groot M, Jacobson AM, Samson JA, Welch G. Glycemic control
very little is known about depression in youth with diabetes.
and major depression in patients with type 1 and type 2 diabetes
More longitudinal studies are needed to more carefully mellitus. J Psychosom Res 1999;46:425  35.
M. Grey et al. / Journal of Psychosomatic Research 53 (2002) 907 911 911
[9] Lustman PJ, Velozo CA, Eubanks B, Montag JA, Cole DM. Effects of [26] Kovacs M, Iyengar S, Goldston D, Stewart J, Obrosky DS, Marsh
psychiatric disorder on rehabilitation efforts and ability to return to J. Psychological functioning of children with insulin-dependent dia-
work following job-related injury. Work 1991;1:37 43. betes mellitus: a longitudinal study. J Pediatr Psychol 1990;15:
[10] Sachs G, Spiess K, Moser G, Prager R, Kunz A, Schernthaner G. 619  32.
Glycosylated hemoglobin and diabetes self-monitoring (compliance) [27] Kovacs M. The Children s Depression Inventory (CDI). Psychophar-
in depressed and non-depressed type 1 diabetics. Psychother, Psycho- macol Bull 1985;21:995  8.
som Med Psychol 1991;41:306  12. [28] Grey M, Cameron ME, Lipman TH, Thurber FW. Psychosocial status
[11] Van der Does FEE, De Neeling JND, Snoek FJJ, Kostense PJ, Goo- of children with diabetes over the first two years. Diabetes Care
tenhuis PA, Bonter LM, Heine RJ. Symptoms and well-being in rela- 1995;18:1330 6.
tion to glycemic control in type 2 diabetes. Diabetes Care 1996;19: [29] Jacobson AM, Hauser ST, Willett JB, Wolfsdorf JI, Duorak R, Her-
204 10. man L, deGroot M. Psychological adjustment to IDDM: 10-year fol-
[12] Carney RM, Rich MW, Freedland KE, Saini J, Tevelde A, Simeone C, low-up of an onset cohort of child and adolescent patients. Diabetes
Clark K. Major depressive disorder predicts cardiac events in patients Care 1997;20(5):811  8.
with coronary artery disease. Psychosom Med 1988;50:627  33. [30] Birmaher B, Ryan ND, Williamson DE, Brent DA, Kaufman J, Dahl
[13] Jacobson AM, Rand LI, Hauser ST. Psychologic stress and glycemic RE, Perel J, Nelson B. Childhood and adolescent depression: a review
control: a comparison of patients with and without proliferative dia- of the past 10 years, Part I. J Am Acad Child Adolesc Psychiatry
betic retinopathy. Psychosom Med 1985;47:372  81. 1996;35:1427 39.
[14] Goldston D, Kovacs M, Ho V, Parrone P, Stiffler L. Suicidal [31] Vila G, Roberts JJ, Clemencon CN, Vera L, Crosnier H, Rault G, Jos
ideation and suicide attempts among youth with insulin-dependent J, Mouren-Simeoni MC. Eating and emotional disorders in adolescent
diabetes mellitus. J Am Acad Child Adolesc Psychiatry 1994;33: girls with insulin-dependent diabetes mellitus. Eur Child Adolesc
240  6. Psychiatry 1995;4:270  9.
[15] Goldston DB, Kelley AE, Reboussin DM, Daniel SS, Smith JA, [32] LaGreca AM, Swales T, Klemp S, Madigan S, Skyler J. Adolescents
Schwartz RP, Lorentz W, Hill C. Suicidal ideation and behavior and with diabetes: gender differences in psychosocial functioning and
noncompliance with the medical regimen among diabetic adolescents. glycemic control. Child Health Care 1995;24:61  78.
J Am Acad Child Adolesc Psychiatry 1997;36:1528  36. [33] Grey M, Genel M, Tamborlane WV. Psychosocial maladjustment in
[16] Kovacs M, Obrosky DS, Goldston D, Drash A. Major depressive latency-aged diabetics: determinants and relationship to control. Pe-
disorder in youths with IDDM: a controlled prospective study of diatrics 1980;65:69 72.
course and outcome. Diabetes Care 1997;20(1):45  51. [34] Grey M, Cameron ME, Thurber FW. Coping and adaptation in chil-
[17] Kovacs M, Mukerji P, Iyengar S, Drash A. Psychiatric disorder and dren and adolescents with diabetes. Nurs Res 1991;40:144 9.
metabolic control among youths with IDDM. Diabetes Care 1996; [35] Grey M, Boland EA, Yu C, Sullivan-Bolyai S, Tamborlane WV. Per-
19(4):318  23. sonal and social factors associated with quality of life in adolescents
[18] Anderson JC, McGee R. Comorbidity of depression in children and with IDDM. Diabetes Care 1998;21(6):909  14.
adolescents. In: Reynolds WM, Johnson HF, editors. Handbook of [36] Olsen DH, McCubbin HI, Barnes H, Larsen A, Muzen M, Wilson M.
depression in children and adolescents. New York: Plenum, 1994. Family inventories. St. Paul (MN): Family Social Science, 1982.
pp. 581  601. [37] McKelvey J, Waller DA, North AJ, Marks JF, Schreiner B, Travis LB,
[19] Fleming JE, Boyle MH, Offord DR. The outcome of adolescent de- Murphy JN. Reliability and validity of the Diabetes Family Behavior
pression in the Ontario Child Health Study follow-up. J Am Acad Scale. Diabetes Educ 1993;19:125 32.
Child Adolesc Psychiatry 1993;32:28 33. [38] Sanacora G, Mason GF, Rothman DL, Behar KL, Hyder F, Petroff OA,
[20] Lewinsohn PM, Clarke GN, Seeley JR, Rohde P. Major depression in Berman RM, Charney DS, Krystal JH. Reduced cortical gamma-
community adolescents: age at onset, episode duration, and time to aminobutyric acid levels in depressed patients determined by proton
recurrence. J Am Acad Child Adolesc Psychiatry 1994;33:809  18. magnetic resonance spectroscopy. Arch Gen Psychiatry 1999;56:
[21] Kokkonen J, Taanila A, Kokkonen ER. Diabetes in adolescence: the 1043  7.
effect of family and psychologic factors on metabolic control. Nord J [39] Sanacora G, Mason GF, Krystal JH. Impairment of GABAergic trans-
Psychiatry 1997;51:165 72. mission in depression: new insights from neuroimaging studies. Crit
[22] Kokkonen K, Kokkonen ER. Mental health and social adaptation in Rev Neurobiol 2000;14:23  45.
young adults with juvenile-onset diabetes. Nord J Psychiatry 1995;49: [40] Borg H, Gottsater A, Landin-Olsson M, Fernlund P, Sundkvist G.
175  81. High levels of antigen-specific islet antibodies predict future beta-cell
[23] Kovacs M, Brent D, Feinberg TF, Paulauskas S, Reid J. Children s failure in patients with onset of diabetes in adult age. J Clin Endo-
self-reports of psychologic adjustment and coping strategies during crinol Metab 2001;86:3032  8.
the first year of insulin-dependent diabetes mellitus. Diabetes Care [41] Birmaher B, Ryan ND, Williamson DE, Brent DA, Kaufman J. Child-
1986;9:472 9. hood and adolescent depression: a review of the past 10 years, Part II.
[24] Kovacs M, Feinberg TF, Paulauskas S, Finkelstein R, Pollack M, J Am Acad Child Adolesc Psychiatry 1996;35(12):1575  83.
Crouse-Novak M. Initial coping responses and psychological charac- [42] Wysocki T, Greco P, Harris M, Bubb J, Elder C, Harvey L, McDonnell
teristics of children with insulin-dependent diabetes mellitus. J Pediatr K. Behavior therapy for families of adolescents with IDDM: main-
1985;106:827  34. tenance of treatment effects. Diabetes 1997;46(Suppl. 1):351.
[25] Kovacs M, Goldston D, Obrosky DS, Bonar LK. Psychiatric disorders [43] Satin W, LaGreca AM, Zigo MA, Skyler JS. Diabetes in adolescence:
in youth with IDDM: rates and risk factors. Diabetes Care 1997;20: effects of a multifamily group intervention and parent simulation of
36  44. diabetes. J Pediatr Psychol 1989;14(2):259  75.