Obesity among outpatients with major
George I. Papakostas, Timothy Petersen, Dan V. Iosifescu, Alana M. Burns,
Andrew A. Nierenberg, Jonathan E. Alpert, Jerrold F. Rosenbaum and Maurizio Fava
Depression Clinical and Research Program, Massachusetts General Hospital Harvard Medical School, Boston, MA, USA
Studies focusing on the prevalence of obesity in major depressive disorder (MDD), or the impact of excess
body fat on the treatment of MDD are lacking. The aim of the present work is to systematically study
obesity in MDD outpatients. A total of 369 MDD outpatients enrolled in an 8-wk trial of 20 mg fluoxetine
had height and weight measured at baseline. We then examined: (1) the prevalence of being overweight or
obese, (2) the relationship between obesity and a number of demographic and clinical variables, and, (3)
the relationship between relative body weight and obesity with clinical response. We found that more
than 50% of patients were overweight [body mass index (BMI) o25 kg/m2], while 20% were obese (BMI
o30 kg/m2). Obese patients presented with worse somatic well-being scores than non-obese MDD
patients, but they did not differ with respect to depression severity, anxiety, somatic complaints, hope-
lessness or hostility. Greater relative body weight, but not obesity, predicted non-response. In conclusion,
greater relative body weight was found to place MDD outpatients at risk for fluoxetine resistance.
Received 24 November 2003; Reviewed 31 March 2004; Revised 2 May 2004; Accepted 16 May 2004
Key words: Fluoxetine treatment, major depressive disorder, obesity
Obesity is a major public health concern. An esti-
mated half of the current US population is overweight
[National Task Force on the Prevention and Treatment
of Obesity (NTFPTO), 2000], defined as a body mass
index (BMI) of 25 kg/m2or greater, while the preva-
lence of obesity in the general population, defined as a
BMI of 30 kg/m2or greater, has been estimated at 20%
for men and 25% for women (Flegal et al., 1998). In
addition, the prevalence of obesity has increased
more than 50% from 1960 to 1994 (Flegal et al., 1998).
Although the adverse impact of obesity on medical
illness and all-cause mortality has been well-charac-
terized (Katzmarzyk et al., 2002; NTFPTO, 2000;
Pi-Sunyer, 1993; Raman, 2002), less is known about
the relationship between obesity and depression. In
fact, studies specifically reporting on the prevalence
of obesity in major depressive disorder (MDD) or
on the impact of excess body fat on the treatment of
MDD are lacking. Given the increasing prevalence of
obesity in the general population, studies are needed
to better define the role of obesity in MDD, and
specifically on treatment response with standard anti-
depressants such as the selective serotonin reuptake
inhibitors (SSRIs). The purpose of the present study
was to systematically study excess body weight and
obesity in MDD outpatients, with a focus on the
treatment of MDD.
A total of 384 outpatients, aged 18–65 yr, who
met criteria for a current major depressive episode
(MDE) according to the Structured Clinical Interview
for DSM-III-R – Patient Edition (SCID-P; Spitzer et al.,
1989), who were medication-free for at least 2 wk,
with a baseline 17-item Hamilton Depression Rating
Scale (HAMD-17; Hamilton, 1960) score of o16 were
enrolled into an 8-wk, fixed-dose, open-label trial of
20 mg fluoxetine conducted at the Massachusetts
General Hospital (MGH) Depression Clinical and
Research Program (DCRP). Patients were recruited
from November 1992 to January 1999 with the use
Address for correspondence: Dr G. I. Papakostas, Massachusetts
General Hospital, Department of Psychiatry, Depression Clinical
and Research Program, 15 Parkman Street, WACC 812, Boston,
MA 02114, USA.
Tel.: (617) 726-6697Fax: (617) 726-7541
International Journal of Neuropsychopharmacology (2005), 8, 59–63. Copyright f 2004 CINP
of radio advertisements, newspaper advertisements
or were referred from colleagues. Institutional Review
Board (IRB)-approved written informed consent was
obtained from all study participants. Patients who
were non- or partial-responders to this open trial were
enrolled in a 4-wk, double-blind, triple-dummy, ran-
domized study comparing high dose fluoxetine with
augmentation of fluoxetine with either desipramine
or lithium. The results of the double-blind study are
reported elsewhere (Fava et al., 2002). The present
study focuses on the first phase of the trial.
Exclusion criteria included pregnant women and
women of childbearing potential who were not using
a medically accepted means of contraception, lactating
women, patients with serious suicidal risk or serious,
unstable medical illness, patients with a history of
seizure disorder, patients with the DSM-III-R diag-
noses of organic mental disorders, substance use dis-
orders, including alcohol, active within the last year,
schizophrenia, delusional disorder, psychotic dis-
orders not elsewhere classified, bipolar disorder, or
antisocial personality disorder, patients with a history
of multiple adverse drug reactions or allergy to the
study drugs, patients with mood-congruent or mood-
incongruent psychotic features, current use of other
psychotropic drugs, patients with clinical or labora-
tory evidence of hypothyroidism, patients whose
depression had failed to respond in the past to a trial
of either higher doses of fluoxetine (60–80 mg/d), or
to the combination of fluoxetine and desipramine,
or the combination of fluoxetine and lithium, patients
who had failed to respond during the course of their
current MDE to at least one adequate antidepressant
trial, defined as 6 wk or more of treatment with either
>150 mg imipramine (or its tricyclic equivalent) or
>60 mg phenelzine (or its monoamine oxidase in-
During the screen visit, all enrolled patients signed
an IRB-approved written informed consent form. A
medical and psychiatric history, physical examination,
serum chemistries, haematological measures, electro-
cardiogram (EKG), and urine pregnancy test were
then performed. The 31-item of the Hamilton Rating
Scale for Depression (HAMD-31) was also adminis-
tered during the screen visit. The screen visit was
conducted by experienced psychologists or psy-
chiatrists. In our group, training in the use of instru-
ments such as the HAMD-31 and SCID-P is done by
peer review of videotaped interviews. Our inter-rater
reliability for the use of the SCID-P was recently esti-
mated as k=0.80 (Fava et al., 2000). At the conclusion
of the screen visit, all enrolled patients were asked to
return 1 wk later for the baseline visit.
Visits subsequent to the screen occurred at baseline
and then every other week for a total of 8 wk. The
HAMD-31 was administered during all study visits.
In addition to the HAMD-31, the self-rated Symptom
Questionnaire (Kellner, 1987) which contains sub-
scales on depression (SQ-D), anxiety (SQ-A), anger/
hostility (SQ-H), somatic symptoms (SQ-SS), and so-
matic well-being (SQ-SWB) along with the self-rated
Beck Hopelessness Scale (BHS; Beck & Steer, 1988)
were also administered during the baseline visit.
Patients who returned for their baseline visit were
started on a 20 mg, fixed-dose regimen of fluoxetine.
A responder was defined as having a 50% or greater
reduction in HAMD-17 score from baseline to end-
point. An intent-to-treat (ITT) analysis with the last
observation carried forward was used to define the
severity of depression at end-point, in which the last
recorded HAMD-17 score substituted the end-point
score for patients who prematurely discontinued the
study. BMI was defined as weight (in kg)/height2
(in m2). A total of 369 patients had both height and
weight measured at baseline, allowing for the calcu-
lation of baseline BMI.
The National Institutes of Health Clinical Guidelines
on the Identification, Evaluation, and Treatment of
Overweight and Obesity in Adults (NIH, 1998) define
overweight as a BMI equal to or greater than 25 kg/m2
and obesity as a BMI equal to or greater than 30 kg/
m2, with healthy weight corresponding to a BMI be-
tween 19 and 25. Defining overweight as a minimum
BMI of 25 kg/m2is also consistent with recommend-
ations of the WHO (1998). Appropriate parametric and
non-parametric tests were used to compare differences
in variables between obese and non-obese patients.
With the use of separate logistic regressions we then
tested for the relationship between (1) relative body
weight (BMI as a continuous variable), (2) overweight
status, (3) obesity, or (4) change in weight during
the 8-wk trial and clinical response, controlling for
gender and the severity of depression at baseline
(HAMD-17 total score). We chose to control for gender
because of a recent study showing a gender-based
discrepancy in the relationship between body weight
and MDD (Carpenter et al., 2000).
In total, 369 (96.0%) of the original 384 outpatients
had both height and weight recorded at baseline. The
sample consisted of 199 women (53.9%) and 170 men
60G. I. Papakostas et al.
(46.1%). The mean age for the entire sample in years
was 39.8¡10.4 yr. In total, 312 (84.5%) out of 369
patients completed the study. Of these, 202 (54.7%)
patients responded to treatment. The mean length
of time in the study for responders was 7.6¡1.2 vs.
6.5¡2.7 for non-responders.
The mean baseline BMI for the entire sample was
26.5¡5.2 kg/m2. The distribution of BMI for the entire
sample is presented in Figure 1. Of all 369 patients
with BMI measured at baseline, 190 patients were
overweight (51.4%). 94 of 199 women were over-
weight (47.2%) and 96 of 170 men (56.5%). There were
74 patients who were classified as obese (20.0%). Fifty
out of 199 women (25.1%) and 24 out of 170 men
(14.1%) were obese. Demographic and clinical charac-
teristics of obese vs. non-obese MDD patients are
presented in Table 1.
A logistic regression revealed that greater relative
x2=3.843, coefficient/S.E.=1.960, 95% CI 1.000–1.076).
There was a trend towards statistical significance
for poorer outcome in patients who were overweight
(p=0.067). The presence of obesity did not signifi-
cantly predict outcome (p=0.16). The mean BMI in
responders and non-responders was 25.9¡5.2 kg/m2
vs. 27.1¡7.0 kg/m2. There was no statistically signifi-
cant change in weight during the trial (81.1¡24.7
vs. 81.3¡24.6 kg). Change in weight did not predict
More than half of the present sample of outpatients
with MDD were overweight, while 20% of patients
were obese. Nearly 25% of women and 14% of men
were found to be obese. These figures reflect the
national average (Flegal et al., 1998; NTFPTO, 2000),
with the exception of the somewhat lower prevalence
of obesity among men from the present sample com-
pared to the national average (14% vs. 20%). These
results are also in line with studies looking at the
incidence of obesity in bipolar disorder reported
between 21% (McElroy et al., 2002) to 35.4% (Fagiolini
et al., 2003).
Carpenter et al. (2000) were the first to report on
the relationship between body weight and MDD.
In an epidemiological study involving more than
40000 subjects nationwide, the authors reported that
greater relative body weight was associated with
an increased risk for past-year MDD and suicidal
ideation among women while lesser relative body
weight was associated with an increased risk for
past-year MDD, suicidal ideation and suicide at-
tempts among men. Shortly thereafter, Roberts et al.
(2000) found that obesity, defined as a BMI at the
85th percentile or higher, predicted MDD after a 1-yr
follow-up. This finding was soon replicated for
longer follow-up periods (Roberts et al., 2003). While
these reports suggest an increased risk of depression
in obese patients, our study suggest that MDD out-
patients are not more likely to be obese than their
non-depressed counterparts. In addition, while obese
MDD patients presented with worse somatic well-
being scores than non-obese MDD patients, they
did not differ on the basis of depression severity,
Body mass index
Figure 1. The distribution of MDD patients according to
body mass index.
Table 1. Demographic and clinical characteristics of obese vs.
non-obese MDD patients
Duration MDE (yr)
Age onset (yr)
Beck Hopelessness Scale
Cigarettes (per day)
SQ, Symptom Questionnaire.
Obesity in MDD61
or in the severity of a number of depressive symptoms
including anxiety, somatic complaints, hopelessness
However, our study suggests that greater BMI is
associated with an increased risk of non-response
to treatment in MDD. Recently, Fagiolini et al. (2003)
reported a shorter time to recurrence during the
maintenance phase of treatment in obese than non-
obese outpatients with bipolar I disorder. That a di-
chotomous definition of high or normal BMI such
as obesity or being overweight did not significantly
predict treatment response in our trial is in line
with the aforementioned epidemiological study by
Carpenter et al. (2000) that found a link between
greater relative body weight (BMI continuous) and
MDD, but not between obesity (dichotomous) and
MDD. Thus, it may be that a definition of obesity as
a minimum BMI of 30 kg/m2may not be best suited
for the purposes of studying any adverse effects of
excess weight on mood or the treatment of depression.
One limitation of the present study is the absence
of data on body fat distribution, which is an inde-
pendent predictor of health risk (NIH, 1998). Another
limitation is that of sampling bias. Clinical trials have
a number of inclusion and exclusion criteria and as a
result, patients in clinical trials do not directly reflect
the typical outpatient population. This may be par-
ticularly true in the present study, since we excluded
patients with severe/unstable medical illness. As a
result, given the relationship between excess body fat
and poor health status, many patients excluded on this
basis may have been overweight or obese. An ad-
ditional limitation is the lack of data on the treatment
history of patients enrolled in the study which may
have shed further light on the inter-relationship be-
tween relative body weight and treatment response
in depression. Thus, the degree to which these find-
ings generalize to a more heterogeneous population
of depressed patients including those with severe
severe/unstable medical illness remains to be deter-
mined. The final limitation is the absence of a control
group which would help clarify to what degree the
adverse impact of excessive body weight on outcome
to pharmacotherapy with fluoxetine is mediated
through decreasing drug or placebo response rates.
While some epidemiological studies suggest an in-
creased risk of MDD in obesity, the prevalence of
obesity in the present sample of outpatients with
MDD does not appear to differ from the general
population. In addition, while obese MDD patients
presented with worse somatic well-being scores than
non-obese MDD patients, they did not differ with
respect to depression severity, anxiety, the number
of somatic complaints, hopelessness or hostility at
baseline than non-obese patients. However, greater
relative body weight was found to place MDD out-
patients at risk for fluoxetine resistance regardless
of the severity of depression at baseline. Studies with
less stringent inclusion/exclusion criteria or focusing
on the medically ill may yield different results.
Supported by NIMH grant no. R01-MH-48-483-05
(M.F.), the American College of Neuropsychophar-
macology/GlaxoSmithKline Fellowship in Clinical
Neuropsychopharmacology (G.I.P.), and the Harvard
Medical School/Kaplen Fellowship in Depression
Statement of Interest
Beck AT, Steer RA (1988). Manual for the Beck Hopelessness
Scale. San Antonio: Psychological Corp.
Carpenter KM, Hasin DS, David AB, Myles FS (2000).
Relationships between obesity and DSM-IV major
results from a general population study. American Journal of
Public Health 92, 251–257.
Fagiolini A, Kupfer DJ, Houck PR, Novick DM, Frank E
(2003). Obesity as a correlate of outcome in patients with
bipolar I disorder. American Journal of Psychiatry 160,
Fava M, Alpert JE, Nierenberg AA, Russell JM, O’Boyle M,
Camilleri A, Harrison WM (2000). A validation study of a
computerized management system for the diagnosis and
treatment of depression. Report presented at the American
Psychiatric Association Annual Meeting.
Fava M, Alpert J, Nierenberg A, Lagomasino I, Sonawalla S,
Tedlow J, Worthington J, Baer L, Rosenbaum JF (2002).
Double-blind study of high-dose fluoxetine versus lithium
or desipramine augmentation of fluoxetine in partial
responders and nonresponders to fluoxetine. Journal of
Clinical Psychopharmacology 22, 379–387.
Flegal KM, Carroll MD, Kuczmarski RJ, Johnson CL (1998).
Overweight and obesity in the United States: prevalence
and trends, 1960–1994. International Journal of Obesity
and Related Metabolic Disorders 22, 39–47.
62G. I. Papakostas et al.
Hamilton M (1960). A rating scale for depression. Journal
of Neurology Neurosurgery and Psychiatry 23, 56–62.
Katzmarzyk PT, Craig CL, Bouchard C (2002). Adiposity,
adipose tissue distribution and mortality rates in the
of Obesity and Related Metabolic Disorders 26, 1054–1059.
Kellner R (1987). A Symptom Questionnaire. Journal of
Clinical Psychiatry 48, 268–274.
McElroy SL, Frye MA, Suppes T, Dhavale D, Keck Jr. PE,
Leverich GS, Altshuler L, Denicoff KD, Nolen WA,
Kupka R, Grunze H, Walden J, Post RM (2002). Correlates
of overweight and obesity in 644 patients with bipolar
disorder. Journal of Clinical Psychiatry 63, 207–213.
National Task Force on the Prevention and Treatment
of Obesity (NTFPTO) (2000). Overweight, obesity, and
health risk. Archives of Internal Medicine 16, 898–904.
National Institutes of Health (NIH) (1998). Clinical
Guidelines on the Identification, Evaluation, and
Treatment of Overweight and Obesity in Adults: The
Evidence Report. Obesity Research 6 (Suppl. 2), S51–S209.
Pi-Sunyer FX (1993). Medical hazards of obesity. Annals of
Internal Medicine 119, 655–660.
Raman RP (2002). Obesity and health risks. Journal of the
American College of Nutrition 21, S134–S139.
Roberts RE, Deleger S, Strawbridge WJ, Kaplan GA (2003).
Prospective association between obesity and depression:
evidence from the Almeida county study. International
Journal of Obesity and Related Metabolic Disorders 27,
Roberts RE, Kaplan GA, Shema SJ, Strawbridge WJ (2000).
Are the obese at greater risk for depression? American
Journal of Epidemiology 152, 163–170.
Spitzer RL, Williams JBW, Gibbon M, First M (1989).
Structured Clinical Interview for DSM-III-R – Patient Edition
(SCID-P). New York: New York State Psychiatric Institute,
Biometrics Research Department.
WHO (1998). Obesity: preventing and managing the global
epidemic. In: Report of a World Health Organization
Consultation on Obesity, Geneva, 3–5 June 1997. Geneva,
Switzerland: World Health Organization, pp. 1–276.
Obesity in MDD63