Fluoxetine increases resting energy

body temperature
in humans13
Rachelle

Bross

and

L John

to this

women

weight

loss

to a metabolic

to 3 wk

we

admitted

unit where

of a 1.76-MJ/d

20

nondepressed

they were

formula

diet

and

either

1.8%

(P

<

restriction.

Basal

decreased

0.005)

assigned

60

mg

but

that

body

0.28

0.10

of the

d of

temperature
during

of the
the

period

fluoxetine-treated

commencing

#{176}C
(P < 0.05)

within

control

subjects

of energy

restric-

subjects

<

0.05).

increases

We

energy

3 d of commencing

perature.

Am

propose

that

serotonin

by

diet

reuptake

expenditure

by

J Clin

1995;61:1020-5.

Nutr

increasing

and

inhibition

basal

Subjects

and

Subjects,

diets,

Energy

WORDS

expenditure,

body

tem-

we

also

obtained

serotonin

any

other

endocrine

or significant

subjects
were nonsmokers
who had
body weight
for several
months
and had
in recent weeks. Written
informed
consent

maintained
a stable
taken no medication
was

or
All

for the

study,

which

was

approved

by the

Introduction
Fluoxetine

is a widely

serotonin

pressants,

fluoxetine

depressed

(2) and

by inducing

use
satiety

to weight

(REE),

blocking

the

during
istration

energy
with

by

adaptive

the

reduction

restriction
(6),
a very-low-energy

the

with

that

tnicyclic

(1 1)

(420

and

consumed
kcal)

including

dietary
minerals

80

g protein.

For

a liquid-formula

the

next

weight

loss

thermic
in REE

we

resting

70;

Sandoz

Nutrition,

Minneapo-

started,

all

fashion
capsule

to receive
60 mg fluoxetine
or placebo
in a single
each morning
before
breakfast,
but after any study

effect
that

expen-

of food,

commonly

combined
fluoxetine
weight-reduction

or by
occurs

admindiet
that

normally

engenders
a pronounced
adaptive
reduction
in REE
Urinary
norepinephrine
and normetanephrine
excretion
and serum tniiodothyronine
concentrations
were measured
to
monitor
the altered
sympathetic
nervous
system
activity
and
thyroid
hormone
metabolism
believed
to mediate
the hypometabolic
adaptation
to semistarvation
(6, 7). Because
seroto-

(7, 8).

Am

J Clin Nutr

I From

the School

were

randomly

of Dietetics

assigned

for a minirestriction

antidein both

energy

all

containing
1.76 MJ
70 g protein/d
and
100% of the US
allowances
(12; RDA) for vitamins
and

(Optifast

subjects

3 wk

diet

inhibits

(3, 4) individuals,
apparently
To learn whether
fluoxetine
also

by increasing

increasing

drug

Unlike

is associated
(5).

loss

antidepressant

(1).

nondepressed

early

contributes

used

reuptake

Ethics

Committee
of the hospitals
Department
of Medicine.
Each hospitalization
lasted 26 d. For the first 4-5 d all
subjects
consumed
a liquid-formula
diet consisting
of Ensure
with added Polycose
glucose
polymer
(both from Ross Laboratories,
Columbus,
OH) to provide
150% of basal energy

lis). No other food or drink was consumed

except
mum of 1600 mL water daily. On the day energy

neuronal

monitored

treatments

condition.

expenditure

pharmacology,

and

diabetes,

medical

essential

1020

10),

methods

recommended

diture

(9,

Twenty
moderately
obese,
but otherwise
healthy
women
were admitted
to the Clinical
Investigation
Unit of the Royal
Victoria
Hospital.
Screening
by medical
history,
physical
examination,
chest
X-ray,
electrocardiogram,
complete
blood
count, thyroid
function
tests, standard
biochemical
screening,
and an oral-glucose-tolerance
test indicated
the absence
of

subjects
KEY

in thermoregulation

temperature.

hypertension,

increased

drug treatment.
Urinary
norepinephrine
excretion
and the serum triiodothyronine
concentration
decreased
equally
in both
groups.
Despite
identical
energy
intakes
and equal
nitrogen
balance,
the fluoxetine-treated
subjects
lost weight faster than
the control
subjects
during the final week of energy restriction
(P

body

fluox-

resting
energy
expenditure
then renormal
after
9.8 0.9 d of energy

insignificantly

tion,

within

is involved

basal

and Human

Nutrition

Nutrition
and Food Science Centre, McGill
University,
2 Supported
by a grant from Eli Lilly
Canada,
Inc.

3 Address
Science
H6.90,

the Fonds pour la Formation

and the McGill

Montreal, Canada.

RB received

reprint

Centre,
Montreal,

requests

Royal
Quebec

Received

October

Accepted

for

1995;61:1020-5.

Printed

to

Victoria
H3A

Hoffer,

Hospital,
1A1,

de Chercheurs

partial

a la
Recherche.
LJH is a chercheur-boursier
of the Fonds de la Recherche
en
Sante du Qu#{233}bec.
The Clinical Investigation
Unit of the Royal Victoria
Hospital is supported by a grant from the Fonds pour la Recherche en Sante
du Qu#{233}bec.
Optifast-70
was donated by Sandoz Nutrition
Canada, Inc.
support

from

in double-blind

McGill
687

Pine

et lAide

Nutrition
Avenue

and
West,

Food
Room

Canada.

25, 1994.

publication
in USA.

January
1995

9, 1995.
American

Society

for

Clinical

Nutrition

Downloaded from ajcn.nutrition.org by guest on October 11, 2015

4.4

treatment.
This increased
versed
and fell below

nm

obese

randomly

etine/d or a placebo.
Resting
energy expenditure
of the control
subjects
fell below normal after 5.6 0.6 d of energy restriction, whereas
that of the fluoxetine-treated
subjects
increased
by

and basal

Hoffer

ABSTRACT
Humans
lose weight
when administered
fluoxetine,
an inhibitor
of serotonin
reuptake
by nerve terminals.
To determine
whether
increased
energy
expenditure
contnibutes

expenditure

THERMIC
measurement

was

to walking

on the

completed

on that

hospital

day.

Activity

EFFECT

was

limited

OF FLUOXETINE
and
potassium
verify
adherence

ward.

Basal
Measurements
Body

and

calculations

weight

was

light

bedclothes,

wearing
by using

was

by

Yorba
training

subject

(13).

The

plete

REE

measurement

disregarded

training

before

for the formal

protocol.
surements
were obtained
or third

mornings
measurements

the final
analyzed.

both

oxygen

usually
surement
carbon
The
of the

was

and

calculated

Rougier

REE
and

period.

and

meaevery

Before
each
room

that

state was
next day,

not
but

achieved,
the
this occurred

for
meaonly

oxygen

and

at 55,

voiding.

for

210

mm,

and

165

Control

expenditure

the

Canada)

using

obtained

over
hood.

which
mm

experiments
in REE

was

consumed
a test drink
mL carbonated
water

the ventilated

120,

no change
of glucose

subjects
in 450

due

three

10-mm

the monotony

with

sham

this

protocol.

or

glucose

loads

The

from

to REE

then

expen-

included

to break

by subtracting

contribution

15 mm,
Energy

(14)

sured

by automated

inson

Immunodiagnostics,

of the

total

subjects

period

collection

a preservative)
period.

analyzed
Syncron

daily
CX4

during

the

nephrine

at the

Richmond,
restriction
later

7,

(containing
in a bedside

Samples

14,

in serial

kept

using

of

and

and
24-h

15 mL

21

diet

and

an

automated

for

of

12 mol

Each

HC1/L

as

were

(Beckman
three
times
and

system

analysis.
after

Autoanalyzer
calculated

estimated
(8 mg . kg
measured
to

Urine
Kjeldahl

II (Chauncey,
as nitrogen

fecal
. d)
verify

(0.6

total

nitrogen

with

a predicted

digestion
NY).

intake

minus

g/d)

and

(16, 17). Urine

completeness

perature,

3-4%

within-subject

the analysis
days

of

the

of menstrual

recorded
plotted

in

of fluoxetines
cycle

when

no

should

not

drug

was

for these days were set equal

over the 3 wk of fluoxetine

nary catecholamines
peated-measures
6.01; SAS
and energy
each
time

t test.

to

Cary,
data

NC).
were

value.
were

and the judgment

body

interest

weight,

indexes,

by two-classification
using
PC SAS
Temperature,
expressed

expressed

and

clinical

unre-

(version

body weight,
as changes
from

to determine
P values
reported

are

tem-

and

When
significant
(P
indicated,
a contrast
was

Baseline
subject
were
compared
results

in accor-

respectively,

of physiologic

compared
of variance

baseline
variable
for significance.

All

be

were
analysis

for these contrasts.

ical measurements

chosen
0.10,

SD in REE,
would

hormone

subjects
baseline
or treatment
effects

tailed

was
and

thyroid

Institute,
Inc,
expenditure

against
the
responsible

group
of0.05

balance,

Subject

characteristics
assignment

matched
measured

the

0.05)
done

time
points
in the text are

characteristics
or biochemby the two-sample
twoas mean

SEM.

Body

of

the

20

response
subjects

groups
with no significant
variables
(Table
1). No

weight

decreased
of energy
loss was

was

constant

during

steadily
and equally
restriction.
During
faster
in the fluoxetine

TABLE

Baseline

subject

in both
the final
group

resulted

differences

in

in any

wellof the

adverse
events
occurred.
the baseline
period,
and
groups
week,
(0.25

for the first 2 wk

however,
weight
0.01 compared

characteristics
.

Characteristic

Weight

(kg)

a Tech-

nitrogen

balance

Waist-hip

of urinary

and
losses

excretion
collections,

was

Fluoxetine

Age (y)

by using

urine

values

for REE,

(15)

the

lowest

all other

data

Daily

creatinine

subjects

and

of 10 in each

3 probabilities

nitrogen

Rad,

determined

miscellaneous
of

size

a and

Height (m)
Body mass index (kg/m2)

the sum

day

of the menstrual

that a 5% increase
in REE
(18). Treatment
and control

normeta(Bio

was

first

to the

day

each

collection

norepinephrine

day

energy

during

resulting

HPLC

Dicklast

aliquots.

refrigerator

the

mea-

CA) on the last 2 d of each week during

the energy
period.
Other
aliquots
were
frozen
at -25
#{176}C
for

colonimetrically
was

days

first

for creatinine,
sodium,
and potassium
and CX5 Systems,
Brea,
CA) and

nitrogen

nicon

on

collected

were

II; Becton

on the

bottles

baseline
by

and
was

were

collection

(ARIA

Baltimore)

All urine

(13)

according
to the

Results

energy

as determined

triiodothyronine

radioimmunoassay

baseline

restriction.

and

first,

and
basal

analysis

Random

thyroxine

daily

thermic

outset.

Serum

closed

Mean

each

4 baseline

the temperatures
mean
temperature

sample

dance
was

days

the

mouth

the

Because

by the

administration.

The

production,

to the
from

administered,
each subjects

Statistical

dioxide

the

data

results

determination
a CV < 5%

from

in relation

include

subjects
with the

24 #{176}C.
Only

as kcal/min

under

measured

interruptions
indicated

effect

unit

carbon

Inc, Chambly,

reclined

was

were

this.

morning

from

set at zero

above

the

ways:

to

(Becton-Dick-

according

calculation,
was

Celsius

with

second,

as determined

For the latter

data

results
according
to the Weir equation
(14).
effect
of a glucose
load was assessed
at the end
diet and after 3 wk of energy
restriction.
After

immediately

effect

21

REE measurement,
of 1 12.5 g glucose

(Glucodex;

formal
period

15 mm of each formal
20-mm
Steady
state, defined
as having

dioxide
thermic
baseline

a standard
composed

the

cycle

each

indicated.

in two

and

measured

thermometer

tongue

calculated

temperature

degrees

the

were

measured

basal

temperature

day;

flow.

com-

investigation

low-energy

between

consumption

REE

which

clinical

was

basal

two

at least

unit, three
the baseline

the

previ-

final

protocol

menstrual

calibration,

described

of study
medication
was given,
were obtained
in a semidarkened

achieved.
If a steady
was repeated
the

rarely.

diture

On the
during

maintained

from
were

for

study

was

a digital
under

the

temperature

calorimetry

using

included

to the

during

CA),

Toronto)

recording

Canada).

temperature

excretion

Resting

expenditure

Basal body temperature

First

day of menstruation

S SEM;

,z =

10

group

32.0 3.2

33.0 3.9

91.5

92.8

3.7

1.64 0.02
34.0 1.3
0.86 0.02

ratio
energy

(kJ/min)

(#{176}C)
(study
subjects

day)
per group.

Placebo

group

4.69

0.11

36.5 0.1
16.1 9.5

3.2

1.65 0.01
34.1 1.3
0.83 0.01
4.56

0.15

36.5 0.1
12.2 6.8

Downloaded from ajcn.nutrition.org by guest on October 11, 2015

temperature

for

day

dose

indirect

procedures

sessions

admission

Mills,

and
sodium
to the diets.

by holding

inson,

subjects
breakfast,

(Scale-Tronix;

Don

Linda,

procedure

the

before

scale

Inc,

ventilated-hood

and

ously

second

platform

SensorMedics,

measurement,

with

and

LeGroupe,

measured

(Deltatrac;

morning

voiding,

digital

& Bell-Meditron,

REE

each
after

a high-precision

Ingram

body

subjects

measured

1021

BROSS

1022

AND

with 0.22 0.01 kg/d, P = 0.049),

as shown
in Figure
1.
Average
nitrogen
balance
was slightly
positive
and similar for
both groups during the baseline
period (0.4 0.5 and 0.8 0.3
g N/d for the fluoxetine
and placebo
groups,
respectively),
becoming
negative
during consumption
of the very-low-energy
diet (Figure
2). During
the third week of energy
restriction
daily nitrogen
losses were constant
and not significantly
different in the two groups (-2.87
0.52 g/d in the fluoxetine
group and -3.43
0.37 g/d in the placebo
group; P
0.39).
Cumulative
nitrogen
losses over the entire 21-d period were
=

slightly,
but not significantly
6.6 compared
with 61.1

greater

in the placebo
=
0.21).

group

HOFFER

>
0
0)
a)
0

(71.5

4.8 g; P

Biochemical

REE
closely

a)

responses

0)

was constant
(within-subject
similar
in both treatment

period

(P

0.51;

Table

CV: 2.2 0.3%)

and
groups
during
the baseline

1). During

the

first

week

restriction
a significant
treatment
effect became
0.041),
with REE increasing
in the fluoxetine
but falling

below

the baseline

value

in the placebo

(P =
group

after 5.6 0.6 d of energy

restriction
(P = 0.035;
Figure
3).
After first increasing,
REE in the fluoxetine
group returned
to
baseline,
then fell below it after 9.8 0.9 d of energy restniction. The delay in the reduction
of REE in the fluoxetine
group
is significant
(P
0.001).
By the end of the study REE had
decreased
by 6.7 1.2% in the fluoxetine
group
and by
=

9.7

1.5%

in the

placebo

group.

The thermic effect of an oral glucose load was similar in both

groups under baseline
conditions
(8.1 0.8% of 24-h REE or
4.5 0.4%
of the energy consumed
for the fluoxetine
group,
and 8.7 1.4% of 24-h REE or 4.5 0.5% of energy
consumed
in the placebo
group;
P = 0.71).
After
3 wk of

12

Study
FIGURE

and placebo

2.

Mean

(0)

groups.

SEM)
The

daily

18

21

Day

nitrogen

curves

15

balance

in the fluoxetine

are not significantly

(#{149})

different.

energy
restriction,
the response
to the same glucose
challenge
was unchanged
(7.6 0.8% of 24-h REE or 4.0 0.4% of the
energy consumed
in the fluoxetine
group and 8.8 1.1% of
24-h REE or 4.2 0.5% of energy
consumed
in the placebo
group).
Serum thyroid hormone
concentrations
were similar at baseline

in

both

groups.

During

energy

restriction,

serum

14

0
-1
0)
-

-2

a)
0)
C

-3

(Ti

-4

0)
a)

-5

LiJ
U

-6

-7

a)
0)

-8

-C

C
(Ti
0

12

Study
1.

FIGURE

women

baseline

21

Day
in mean

0
(

SEM)

body

weight

8
12
16
Study
Day

20

of 20

assigned to consume a 1.76-MJ/d weight-reduction

diet
fluoxetine
(#{149})
or placebo (0). Initial body weight was 91.5

kg in the fluoxetine

Weight
third

from

18

randomly

and either
3.7

Change

15

loss
study

by the
week

(P

group

fluoxetine
0.049).

and 92.8
group

was

3.2 kg in the placebo

significantly

faster

group.

during

the

FIGURE
expenditure
REE
kJ/min

was

3.

Change

(REE)
4.69

from

baseline

in the fluoxetine

0.11

in the placebo

kJ/min
group.

in mean

(#{149})
and placebo

in the fluoxetine

SEM)
(0)

group

resting

energy

groups. Baseline
and 4.56 0.15

Downloaded from ajcn.nutrition.org by guest on October 11, 2015

0.003)

of energy
(P

evident
group

ThERMIC
remained
cant

constant,

increase

decreased
did

equally

urinary

(P

Changes

(P

the

0.37),

(P

group

was

but

the

the

middle

however

(P

the

onset

C
(Ti

of

insignificant

0
a)

change

#{176}C
in the

fluoxetine

(Figure

4).

in the placebo
cycle
been
(19,

the

(P

(Ti

period
Thus,

of
basal

illustrated

group

0.002),

reported
20). No

As

in

increased

a)
F-

in

suggesting

the

to occur just before

the
change
in basal
temperfluoxetine
body

administration
temperature

of

12

Study

the

subjects
was significantly
higher
during
second
half of their menstrual
cycles.

the

a)
0)

in the two

treatment
effect
became
decreased
by 0.12

0.10

period

0.87).

in Figures

1). With

an

by 0.28

menstrual

over

fluoxetine-treated
first, but not

1023

15

18

21

Day

the
FIGURE

4. Change

from

temperature

in the fluoxetine

temperature

was

36.5

baseline

(#{149})
and

in

mean

placebo

0. 1 #{176}C
in both

(0)

treatment

SEM)

groups.

basal
Baseline

body
basal

groups.

Discussion
Our

results

indicate

promptly
increases
delays
the reduction
severe

energy

temperature
weight
loss

restriction.

The

observation
is novel,

then

below

It is possible
even without
suggested

observation

that
normal

that

2
thyroid

of 60 mg

also

fluoxetine

fluoxetine/d

increases

promptly

as is the subsequent
after

by a recent

Serum

Fluoxetine

report

baseline
with
no

hormone

increase
agonists

days

body

and

rate of
alone.

duces
may

basal

increases

REE

of REE

to normal

of energy

that

again
was

the REE

of obese

measured
energy

after 14 d of fluoxrestriction
(21).
The

promptly

concentrations

in

restriction.

have returned
to normal
and indeed, this possibility

when
formal

fluoxetine

return

several

that REE would

energy restriction,

unchanged
from
etine
treatment

TABLE

ingestion

of obese
women
and significantly
REE that normally
occurs
during

and slightly,
but significantly
increases
the
above
that associated
with energy
restriction

humans
and

that

the REE
in their

induces

and daily

subjects

and

urinary

was

sustains

catecholamine

peripheral

receptors

hypothalamic
not be relevant

(9,

10),

temperature
to humans.

and

been shown
to increase
the
during
acute
administration

observation

that

a serotonin-active

body
drug

chemistry

Thyroxine

index

and group

Control

fluoxetine

hamsters
serotonin

re-

(22), this
receptors

temperature
(26).
Ours
may

induce

of normal
is the first
a sustained

excretion
Very-low-energy

Clinical

although

in Syrian
Central

mediate
a hyperthermic
response
whereas
peripheral
receptors
mediate
a hypothermic
one (9, 10, 23, 24), perhaps
by increasing heat loss through vasodilation
(25). Therefore,
observations
in Syrian hamsters
could represent
the net effect of a preponderance
of peripheral
effects over central
ones in that species.
By contrast,
the serotonin-releasing
drug fenfluramine
has recently
humans

an

in basal body temperature

is also novel.
Serotonin
change body temperature
by activating
specific
central

diet

Day

Day

diet
14

Day

21

(nmol/L)

Fluoxetine

111 4

119

Placebo

1096

11672

Triiodothyronine

111 3

112 4

1187

1067

(nmol/L)

Fluoxetine

2.0 0.1

1.6 0.i3

1.4 O.i

1.2 0.i

Placebo

2.1 0.1

1.7 0.i

1.5 0.i

1.5 0.i

Fluoxetine

294 31

194 20

265 39

239 28
218 27

200 24

Placebo

158 24

134 20

1155 107

842 933

799 743

1181 ii5

848 773

694 72

Norepinephrine

Normetanephrine

(nmol/d)

(nmol/d)

185

Fluoxetine

1718

Placebo

1574 233

;t SEM.
2.3

Significantly

different

from

treatment

groups

baseline

value:

p < 0.05,

p < 0.01.

Downloaded from ajcn.nutrition.org by guest on October 11, 2015

occurred

as

The temperature
increase
in the fluoxetine
by day 2 of treatment
and was sustained

thermal
nadir
that has
onset
of the luteal
phase
ature

13

0.0007),
excretion

and equal

(Table

group,

temperature

of the

are illustrated

constant

a significant
Basal
temperature

treatment

5, basal

Figure

was

increased

significant

throughout

temperature

placebo

0.025).

(P

FLUOXETINE

insignifiSerum

normetanephnine

conditions

treatment,
0.026).

0. 10 #{176}C
in

clinically
restriction.

groups

and

body

baseline

diet

group

treatment

temperature

under

evident

a transient,

OF

2).

in basal

and

(P

in both

Table

4 and 5. Basal
groups

from

at 7 d of energy

norepinephnine

0.001;

drug

apart

observed

EFFECT

BROSS

1024

AND

Specific
hormonal
measurements
would be required
mine whether
ovulation
and progesterone-mediated

1.0

HOFFER

0.8

genesis

occur

energy

restriction,

a,

Despite

0.6

-c

a
L

0.4

4.

a,

and lean

Day

FIGURE

5. Mean

each subjects

lowest

( SEM)

second

half

Menstrual

difference

recorded

value

higher

in basal

(0)

in the fluoxetine

of the menstrual

24

anything,

28

jects

Cycle

in relation

(#{149})
and placebo

in the fluoxetine

was significantly

of

20

body

temperature

from

to the day of the menstrual

groups.
group

Basal

during

body

the first,

temperature
but not the

human

body

temperature

during

group

jects

ingestion.

at least

until

the

hypometabolic

energy

temperature
of increased

degree

at a time

when

nephnine
and normetanephrine
ing. Although
crude,
these

serum

to suggest

that

13 and

excretion
indicators

a thyroidor sympathetic
to account
for fluoxetines

therefore,

increase

in body

urinary

norepi-

were steadily
provide
no

decreasevidence

nervous
effect.

fluoxetine

REE

point.

control
the

REE

ultimately

groups

control

despite
subjects

and

the fluoxetine-treated
perature
both

a lowered
voluntarily

maintain

the

points.
subjects

Although
in both

complaint

of severely

Such

constancy

indepen-

in

techniques

in

As

in

with
prove

to

This

indicates

that

their

thermal

of

respective

their

the

tem-

subjects
insulation

temperature

set

did not systematically

record
this,
the
noted mild cold intolerance,
a common

energy-restricted

individuals

(28,

29).

The strategy
of increasing
thermal
insulation
will, of course,
reduce
or eliminate
the need for increased
basal thermogenesis

to maintain
a higher temperature
set point.
It is well known that basal body temperature
increases
during
the luteal phase of the normal
menstrual
cycle and this phenomenon

appeared

to persist

during

placebo-treated
subjects
(Figure
perature
of the fluoxetine-treated
a value

equal

to that

of the

luteal

energy

restriction

of the

5). By contrast,
the basal temsubjects
remained
constant
at
phase

of the control

subjects.

week

any

of energy

due

placebo

to

lean

balance

and

was,

if

sub-

the greater

basal

that

daily

faster

the
ideal

than

loss-in

the

fluoxwere
energy

immediately

temperature

energy

after

drug

in the fluoxetine

period

of

adaptation

reduction

to

of

expenditure

conditions

total

of glucose

of medication;

progressive

measured

under

other
daily

effect

greater

the

stage,

significance

a greater
fat

dose

body

at this

factors

for

throughout

we

group

borderline

the thermic

been

despite

only

Even

elevation
body

have

substantial

caused
by physical
may be difficult

treatment

REE

and

activity.

temperature

of basal

is

glycogen,

to reveal

third

that

of only

the previous

sustained

Because

body

temperature

water,

nitrogen

therefore,

and

h after

restriction

both

increased

constancy
we
groups

in

normal

a sustained

ones.

despite
groups

decreased

a continuing

of

their

and

of patient

body

comfort,

we

cannot
rule out the possibility
that energy
expenditure
was
greater
than this under the thermally
unregulated
conditions
that prevailed
most of the time, and might have been greater yet
for the fluoxetine-treated
subjects
because
their temperature
set
point was higher.
A third conjecture
is that fluoxetine
may stimulate
physical

dently of the mechanisms

that govern
the hypometabolic
adaptation
to energy
restriction
by raising
the body temperature
set

in the

hence

the higher

temperature

system-mediated
It is reasonable,

increased

week

semistarvation

is mostly

accounted

REE

might

was

REE.

brings about a 13% increase

in REE in the absence
thermal
insulation
(27). REE in the fluoxetine

increased

favoring
mechanism

(Celsius)

=24

Second,

earlier

each

third

for the fluoxetine-treated

and

have

First,

expenditure

group

that

might

group.

measured

adaptation
to semistarvation
became
established.
The magnitude of the temperature
increase
was in good agreement
with
evidence

the

because

than

small

expenditure-and

Unlike
previous
studies,
this one combined
both REE and
body temperature
measurements
and indicated
a closely similar
time course of increasing
REE and body temperature
in subfluoxetine,

significantly

the
of

be disregarded

It is possible,

REE

energy

administered

By

positive

higher
was

0.07).

higher

continued

days

loss

cannot

severe

of the control

not be expected

loss.

and

more

remained
difference

etine

group

slightly

the possibility

the

cycle.

(31),

during

early

weight

loss

to that

lost weight

of extracellular

of fat

of

weight loss during the final week of energy restriction

was due
a faster dissolution
of adipose
tissue. This is in general
agreement with the observation
that their REE was greater.
Nevertheless,
it remains
true that although
the REE in the fluoxetine

(P
increase
in
administration.

tissue

equal

group

in the

weeks

are

a final

used

we

suggest

fluoxetine-a
drug
that
relevant
to the thermostatic
theory

preceding

holds

a meal

a situation

in

energy

expenditure

to fidgeting,
measurement

(32).

remark,

ciated
with food
(33, 34). If this
such

increments

activity,
such as those due
to perceive
unless
special

that

the

that

increase

ingestion
helps
theory
is correct,

could
the

the

induces
theory

in body
limit
the
a higher

also limit food

thermic

effect

present

observations

early
satiety-could
of weight regulation.
temperature

consumption,

of food

asso-

total amount
eaten
basal
temperature
will

because
increase

in
body

temperature
to the satiety level earlier in the course of a meal.
In agreement
with this prediction,
fluoxetines
anorexiant
effect is relatively
specific
for carbohydrate
(4), the nutrient
which,
when consumed
in customary
amounts,
has the most
immediate
and pronounced
thermic
effect (35).
U
We thank

Mi

Myers

expenditure

measurements,

S Solomon

and I Niarri

for

assisting
MG

Shingler

with

many

for assisting

for the catecholamine

of

the

resting

in patient

measurements.

energy
care,

and

Downloaded from ajcn.nutrition.org by guest on October 11, 2015

cycle

16

12

being

and would

rate

however,

adipose
8

(30),

in the

restriction,

loss

tissues

initial
fluoxetine.

women

to the mobilization

difference

I-

intake

in the placebo

Weight

due

the

or without

energy

those

study.

mostly

0.2

a,

their

than

the

during

with

the fluoxetine-treated

faster

I\

?I I

0.
E

group,

normally

to deterthermo-

EFFECT

THERMIC

OF

FLUOXETINE

19. Halbrecht

References

1025

I. Ovarian

function

and body temperature.

Lancet

1945;2:

668-9.

1. Sommi

RW,

Crismon

second-generation
2.

Cohn

JB,

placebo

ML,

Bowden

antidepressant.

Wilcox

C.

in patients

CL. Fluoxetine:

A comparison

with

a serotonin-specific,

Pharmacotherapy

1987;7:

of

fluoxetine,

major

depressive

disorder.

with

fluoxetine

in obesity.

20.

1-15.

imipramine,
J Clin

and

21.

Psychiatry

Wise

SD.

Clinical

1992;55(suppl):
4.

Nathan

C.

Obesity.
5. Yen
ergic

drug

for weight

loss. Am

Physiological

content:

nitrogen

AA

Fricker

J Clin

Nutr

J Clin

of fluoxetine,

Nutr

22.

the

adults

i-C,

eds.

Res Rev
diets

response

energy

high

24.

25.

I Clin

Nutr

Thermoregulation

in health
action.

and

and disease:

Vol

2. New

serotonin.

In:

physiological

26.

York:

Spectrum

and

Publications

Inc,

DL,

Lesch

arylpiperazines

with

diovascular
ergy

13.

Clark

14.

Weir

J Am

HD,

normal

Munro

HN,

Vol

Fleck

3. New

Hoffer
Metabolic

17.

ments.
Hall

calculating

effects

Young

of very

low

of

University
World

Health

JC. A method

studies. Am

Hauger

rate

with

and body

fluids
protein

RL,

calorie

Blackburn
weight

J Clin

Brodoff
Van

Organization/World
Consultation.

reduction

Organ

Tech

for the rapid

Nutr

Health
Energy

Rep Ser
assessment

1983;37:473-7.

Itallie

Hotter

for nitroge-

Matthews

32.

I Clin

Organization/United
and protein

require-

of sample

size

in dietary

JL,

of peripheral

Life

Sci

Fleishaker

disorder:

Annu

IC,

Clin

Exp

Hubbell-Alberts
by

E.

nortriptyline

neuroendocrine

and

and platelet

3):S19-31.

Rev Physiol

A, Mickelsen

Minneapolis:

treated with

Alcohol

subtypes

1993;16(suppl

seroto-

Sci 1992;48:419-23.

tryptophan.

receptor

metabolism.

Hendler

R.

1954;16:125-34.

0, Taylor

The University

Very

eds. Obesity.
Yang

Bistrian
from

New

low

calorie

York:

M-U.

BR,

very

Diet

lB

HL.

The biology

of Minnesota

and

diets.

Lippincott
weight

In:

Press,

Bjorntorp

Co,

P,

1992:683-707.

loss.

Blackburn

low

GL.

calorie

N Engl

Composition

protein

only

GL, Bray GA, eds. Management

restriction.

Ravussin

MA:

Littleton,

34.

E, Lillioja

PSG

a respiratory
JR.

J Biol

Rampone

Food
Med

Al,

KS,

J Med

of weight
and

of obesity

Publishing

loss

mixed
Co,

diets.

by severe
Inc,

1985:

expenditure

chamber.

J Clin

intake

Christin

L, Bogardus

in man:
Invest

C. Deter-

methods

and

results

1986;78:1568-78.

as a mechanism

of temperature

regulation.

1948;20:545-52.
Med

Halliday
carbohydrate

1983;65:307-12.

TE,

energy

Reynolds

thermogenesis.
protein,

S, Anderson

of 24-hour

33. Brobeck

35. Nair

1985;724.

Life

intake of chickens

dietary

Rausch

J, Henschel

TB,

U,

using

DE.

diets.

inhibition.

63-72.

Yale

Expert

BN,
BN,

minants
GL,

1992;

of fenfluramine-induced

in mice.

depressive

EF. Energy

A, Brozek

caloric

1969:424-525.

VR,

of

J Obes

decreases brain temper-

K. Activation

Neuropharmacol

In: Blackburn

special

metabolism.

and

of serotonin
in major

resulting

1949;109:1-9.

ed. Mammalian

Press,

BR,

rate

Int

1977;297:1158-61.

31.

metabolic

metabolic

of tissues
HN,

10th

1989.

resting

An assessment

subjects.

serotonin2

GC, Fisher H. Ethanol

Clin

Brodoff

1991;53:21-6.

J Physiol

Academic

and Agriculture

Nations
18.

Nutr

for

the

body
:803-5.

1950.

30.

Press,

of

central

hypothermia

fluoxetine,

SM,

DuBois

1984;73:750-8.

Food

Stahl

Keys

en-

allowances.

basal

J 1977;1

Psychopharmacology

E. Antagonism

of

J, Horisaka

induces

Lu M, Wagner

28.

29.

of long-term

GH.

in obese

hamsters.

PJ, Macko

Y, Yamada

receptors

27.

and car-

1991;43:527-52.

dietary

Academy

A. Analysis

Bistrian

Rev

of predictions

J Clin

In: Munro

temperature,

in Syrian

of human starvation.

1980;77:557-61.

metabolism.

York:

U,

Invest

Am

to protein

nous constituents.
16.

Sugimoto

markers.

Serotonin-selective

behavioral,

Reappraisal

methods

TA.

Recommended

National
U.

men.

New

reference

Assoc

DC:

young

Pigott

Pharmacol

Council.

Hoffer

JB.

CS,

LK. Accuracy

Diet

Research

ed. Washington,

15.

in humans.

JR, Johnson

needs.

National

Aulakh

neuroendocrine,

effects

11. Mahalko
12.

KP,

JF, Tomkin

of fluoxetine

a measure

adinazolam

Essman

regulation

1978:1-67.

Murphy

A, Fowler

Res 1992;16:852-6.

M. Energy-metabolism
Am

Andrews

in using
Br Med

PJ. Food

Hypotheses
D, Garrow

intake

regulation

by

diet-induced

1991;34:7-12.
iS.

or fat meals

Thermic

response

in lean and obese

to isoenergetic
subjects.

Clin

Sci

Downloaded from ajcn.nutrition.org by guest on October 11, 2015

MB.

ed. Serotonin

pharmacological
10.

Sulpizio

Downregulation

R, Waller

WB,

sleep

Apfelbaum

1991;53:826-30.

9. Myers

and REM

fenfluramine,

diet.

CM,
effects

1988;48:1239-47.

on a very-low-calorie

Problems
clinic.

1992;16:391-5.
Gao B, Duncan WC Jr, Wehr TA. Fluoxetine

nin2

and low
and

ID.

1978;22:1439-46.

of

expenditure,

Cooke

in an infertility

JC, Murphy

hyperthermia:

1990;3:49-74.
with

GA,

106:321-9.
23.

a seroton-

adaptive

Weston
recordings

thermogenic

ature

1992;55(suppl):177S-80S.
in

Nutr

Nutr

BN,

1992:751-61.

Very-low-calorie

R, Melchior

in obese

Co,

on triiodothyronine,

Am

J, Rozen

adaptation

J Clin

P, Brodoff

pharmacology

restriction.
III.

impact

balance.

Bjorntorp

mechanisms

to energy

R, Bonde

protein
8.

Preclinical

rates

Hendler

In:

JB Lippincott

RW.

PS.

metabolic
7.

agonists.

York:

Fuller

6. Shetty

Am

1815-45.

Serotonin

New

U,

studies

EA,

Stinson
the

1985;46:26-31.
3.

Lenton

temperature