Association of Depressive Syndrome and Early Deaths
Among Patients After Stem-Cell Transplantation for
Malignant Diseases
By Fausto R. Loberiza, Jr, J. Douglas Rizzo, Christopher N. Bredeson, Joseph H. Antin, Mary M. Horowitz, Jane C. Weeks,
and Stephanie J. Lee
Purpose: The association of depression and in-
creased mortality in the general population, and also
various medical conditions, is well documented. How-
ever, depression is not well studied in the setting of
hematopoietic stem-cell transplantation (HSCT). We ex-
amined the association between depressive syndrome
and survival after HSCT.
Patients and Methods: A total of 193 patients who
received autologous or allogeneic HSCT from Brigham
and Women’s Hospital or Dana-Farber Cancer Institute
were evaluated prospectively. The self-rated Likert-
scaled symptom checklist, the SF-36, and the Spitzer
Quality of Life Index Scale were administered. Out-
comes evaluated included survival and quality of life.
Results: Sixty-seven patients (35%) satisfied the cri-
teria for depressive syndrome. The 1-year probability
of survival for the depressed and nondepressed pa-
tients was 85% (95% confidence interval [CI], 74% to
92%) and 94% (95% CI, 89% to 97%), respectively (P

.04). In multivariable modeling, depressed patients
have a three-fold greater risk of dying than nonde-
pressed patients (95% CI, 1.07 to 8.30; P
.04) be-
tween 6 and 12 months after HSCT after adjusting for
other prognostic factors. Global inferiority in quality of
life was observed in the depressed cohort when last
measured at 24 months after transplantation.
Conclusion: Depressive syndrome after HSCT is as-
sociated with decreased survival, at least from 6 to 12
months after transplantation. Persistence of this asso-
ciation after controlling for possible confounding fac-
tors suggests that depression may be more than simply
a marker for concurrent ill health. This study raises an
interesting hypothesis as to whether psychological or
pharmacologic intervention for depression after HSCT
can improve survival and/or quality of life.
J Clin Oncol 20:2118-2126. ' 2002 by American
Society of Clinical Oncology.
T
HE ASSOCIATION OF depression and increased mor-
tality in the general population is well documented.
1-8
A meta-analysis using a pooled sample representing 19,000
people showed individuals with depression had excess
mortality from infectious, respiratory, nervous, and circula-
tory disorders compared with the general population.
9
Similarly, depression and reduced survival have been linked
in patient populations with various medical conditions such
as heart disease, cancer, and solid organ transplantation.
10-14
Nevertheless, it remains controversial whether depression
(or any psychological de cit) contributes directly to the
increased mortality or is simply a marker for the severity of
underlying chronic illness.
The association of depression and survival is not well
addressed in the setting of hematopoietic stem-cell trans-
plantation (HSCT). Only one study has examined the
relationship between depressed mood and subsequent sur-
vival. Colon et al
15
found that patients with depressed mood
as documented before HSCT had lower survival rates.
Although this study involved a relatively homogenous
population and adjusted for patient age, type of acute
leukemia, stage of disease, and year of transplantation, other
physical and mental patient characteristics were not mea-
sured to allow adjustment for concurrent health status. Thus,
this study could not address whether depression simply
re ects poor health and is consequently a marker for poor
outcome, or is a primary contributor to poor outcome.
Both depression and anxiety are common after
HSCT,
15-23
with the prevalence of depressive symptoms
among transplant recipients estimated to be at least
18%.
18
Clinical depression is a major impediment to high
levels of perceived health, affecting patients quality of
life and, possibly, interfering with complete recovery.
17
Longitudinal studies show that despite steady increases in re-
turn to normal functioning with 75% of HSCT recipi-
ents returning to baseline levels at the end of 2 years, 37%
From the Health Policy Institute and the Department of Medicine,
Hematology and Oncology, Bone Marrow Transplantation Program,
Medical College of Wisconsin, Milwaukee, WI; and the Department of
Adult Oncology, Dana-Farber Cancer Institute, Boston, MA.
Submitted October 30, 2001; accepted January 16, 2002.
Supported in part by grant no. CA75267-03 from the National
Institutes of Health, Bethesda, MD, and the Amy Strelzer-Manasevit
Scholars Program, Minneapolis, MN.
Address reprint requests to Fausto R. Loberiza, Jr, MD, MS,
International Bone Marrow Transplant Registry Health Policy Insti-
tute, Medical College of Wisconsin, 8701 Watertown Plank Rd,
Milwaukee, WI 53224; email: faustol@mcw.edu.
' 2002 by American Society of Clinical Oncology.
0732-183X/02/2008-2118/$20.00
2118 Journal of Clinical Oncology, Vol 20, No 8 (April 15), 2002: pp 2118-2126
10.1200/JCO.2002.08.757

have at least mild depression at the end of the rst year
after transplantation.
19
The biopsychosocial mechanism of how depression may
alter survival independent of physical status is largely
speculative. Randomized studies of pharmacologic inter-
ventions with or without psychotherapy in cancer patients
with depression provide con icting evidence for bene t,
further fueling the controversy.
24-27
The purpose of this
observational study is to identify the association, if any,
between depression and survival after HSCT after controlling
for other potential prognostic factors. Evidence suggesting an
association could prompt further studies aimed at identifying
possible mechanisms and developing and testing interventions.
Even if survival is not directly affected, proper treatment of
depression after transplantation could improve patients quality
of life and return to normal function.
PATIENTS AND METHODS
Study Population
Patients scheduled for autologous or allogeneic HSCT were prospec-
tively recruited from either the Brigham and Women s Hospital (BWH)
or Dana-Farber Cancer Institute (DFCI). Recruitment began at BWH in
August 1996 and at DFCI in January 1997. Eligible patients were
required to be at least 18 years of age and scheduled for either
autologous or allogeneic HSCT within 1 week to 3 months of
enrollment. Excluded were patients who (1) did not speak or read
English, (2) declined participation, (3) did not return baseline question-
naires, or (4) could not be contacted before admission for transplanta-
tion because of logistic reasons. Initial contact of the potential study
patient was made by one of the authors (S.J.L.) to describe the study.
Subsequently, mailed questionnaires, consent forms, and a self-ad-
dressed, stamped envelope were sent to the potential study participants.
Follow-up questionnaires were mailed to surviving participants at 6,
12, and 24 months after HSCT. Results reported here describe data
collected before May 1, 2001, although data collection continues. The
speci c transplantation regimens and protocols have been described in
detail elsewhere.
28
Only patients who returned their 6-month posttrans-
plant surveys are included in this report (n
193). The study protocol
was reviewed and approved by the institutional review boards of BWH
and DFCI.
Data Collection
Patients were surveyed before transplantation and at 6, 12, and 24
months after HSCT using both validated and constructed items.
Constructed items and a symptom checklist were developed after a
review of the literature and input from transplant physicians, nurses,
and patients. The survey was then tested to gauge clarity and ease of
completion on a pilot group of posttransplant patients. The general
description of the population is provided in two previous publica-
tions.
28,29
Baseline questionnaires elicited information on sociodemo-
graphics (age, sex, marital status, race, education, and work status) and
general health status (excellent, very good, good, fair, or poor).
Follow-up surveys collected information on return to work; physician
visits; medication use; and prevalence and severity of symptoms such
as depression, fatigue, anxiety, concentration ability, feelings of isola-
tion, and memory loss. These symptoms were rated on a ve-point
scale labeled as I do not have this symptom, not bothered at all,
bothered a little, bothered a lot, or extremely bothered. All three
follow-up surveys were identical and addressed recovery issues.
Overall survival of the cohort was calculated as of May 1, 2001, using
data from the BWH/DFCI clinical database.
In order to verify the constellation of symptoms representing
depression, a chart review was conducted by a blinded assessor to
determine clinician documentation of depressive symptoms and pre-
scription of antidepressants. Information was abstracted from notes in the
medical record dated within 4 weeks of the 6-month survey completion.
This chart review was considered the gold standard for calculating test
characteristics for our de nition of depressive syndrome.
Measures and Outcome
Because no standardized measure of depression was collected,
depressive syndrome was operationally de ned as being present in any
person who reported being bothered by depression and who had four or
more diagnostic symptoms (anxiety, dif culty concentrating, feelings
of isolation, fatigue, or memory loss). Individuals with three symptoms
but who reported being extremely bothered by depression were
likewise considered as having depressive syndrome. Patients were
classi ed as to the presence or absence of depressive syndrome on the
basis of their 6-month posttransplant questionnaire. Patients ful lling
the above de nition are referred to in this study as depressed to
distinguish them from the nondepressed. Two validated quality-of-
life instruments, the SF-36
30,31
and the Spitzer Quality of Life
Index,
32,33
were administered before transplantation and 6, 12, and 24
months after transplantation. The de ned measure of depressive syn-
drome was correlated with the measure of composite mental health
from the SF-36 (r
.65) to establish convergent validity. The outcome
of interest is survival de ned as time to death. Patients were censored
at time of last follow-up.
Statistical Analysis
Patient-, disease-, and transplantation-related characteristics between
patients with or without depressive syndrome were compared using

2
test or Wilcoxon rank sum test for categorical and continuous variables,
respectively. Baseline quality-of-life domains from the SF-36 and the
Spitzer Quality of Life Index were compared between the two groups
using similar tests.
Data were analyzed using a Cox proportional hazards model.
Because the main focus of the study was to determine whether there
was a survival difference after the diagnosis of depressive syndrome
(which was rst assessed by the 6-month survey); only patients
surviving longer than 6 months and completing surveys were analyzed.
The method built a single model with time to death as the dependent
variable and age, sex, race, religion, marital status, disease type, disease
stage (early, intermediate, advanced),
29
type of transplantation (autol-
ogous, related, unrelated), use of total-body irradiation, and T-cell
depletion as explanatory variables. A main effect term for the presence
or absence of depressive syndrome was forced into the model. The
proportional hazards assumption for all variables was examined using
time-varying covariate and graphical approaches.
34
Construction of
strati ed proportional hazards models or time-dependent covariates
was used whenever nonproportional variables were identi ed. Interac-
tions between presence or absence of depressive syndrome and all
explanatory variables were examined. The nal model included all
patient, disease, and transplantation factors found prognostic of the
outcome plus a term for depressive syndrome. We then evaluated the
effect of chronic graft-versus-host disease, physical and mental com-
2119DEPRESSION AND SURVIVAL AFTER HSCT