Very-low-calorie

diets with
impact on triiodothyronine,
balance1

high and low protein

energy expenditure,

content:
and nitrogen

Rosa

Hendler,

MD,

andAlfonsA

Bonde,

III, PhD

gesting

that

other

weightboss.

or more

complex

AmfClinNutr

KEY

WORDS

bolic

rate,

Obesity,

processes
l988;48:l239-47.

govern

weight-reducing

dietary-induced

thermogenesis,

diet,

endogenous

nitrogen

expenditure,

mete-

rate (RMR)
decreases
in proportion
to boss of
mass
(LBM)
or some similar
index of metabolic size (ie, surface
area, weight#{176}75, etc). However,
energy expenditure
decreases
during
weight
boss more than

Popular
use of very-low-calorie
diets (VLCDs)
for
treatment
of obesity
has raised
important
questions
about the factors
influencing
the safety and efficacy
of
various
diet regimens
available.
Since the rash of deaths
associated
with commercial
liquid-protein
weight-loss
products
(1), attention
has been focused
on the signifiof negative
weight
loss.

nitrogen

balance

generally

however,

the

precise

and other macronutrients

loss remains
unclear.
Weight
loss from VLCD
energy

deficit

influence

Two

regimens

or the difference

is the result
between

total

weight
of the

1239-47.

Printed

in USA.

1988 American

RMR
and dietary-induced
together
account
for

the

boss nor

of total energy expenditure

that
to adaptive
metabolic
responses

poto

thermogenup to 70-85%

of
(8, 9). DIT was shown to be an imporof the adaptive
response
to over- and

in small

animal

studies

(10),

where

1 From
the Department
of Internal
Medicine,
Yale
School ofMedicine,
New Haven, CT.
2 Supported
in part by grants from the National
Institutes

expen-

and intake.
This deficit and the weight
loss that
diminish
progressively
with continued
dieting betotal energy
expenditure
decreases.
Some of this
decreased
efficacy is readily understood,
for example,
totel meal-stimulated
thermogenesis
is decreased
in proportion
to the degree
of caloric
restriction
and resting
1988;48:

neither

it varies

proportionally
with food intake
to minimize
weight
change.
On this basis DIT was suggested
to be a component of an adaptive
response
to underfeeding
in man as
well (1 1). Attention
was also focused
on thyroid
and
sympathetic
nervous
system
responses
because
of their

diture
results
cause

Am J Clin Nuir

component

underfeeding

7) although

for this adaptation

to weight
involved
are understood.

contribute

underfeeding
are
esis (DIT),
which
total expenditure

tant

calculations(6,

components

tentially

protein

during

by these

factors
responsible
the mechanisms

observed

ofdietary

on N balance

body

predicted

It is now clear that inclusion

of high-biological-quality
protein
in VLCD
regimens
minimizes
the risk of sudden
cardiac
death (2, 3) as does limiting
the duration
of severe caloric restriction.
It was shown (4) that weight loss
may increase
the protein
intake required
to maintain
N
balance
above the Recommended
Dietary
Allowances

daily

energy

during

metabolic

lean

(RDA);

balance,

metabolism

triiodothyronine

Introduction

cance
during

protein

University
of Health

(RR-l25)
and the Cambridge
3 Address
reprint requests

Quest Foundation,
Pacific Grove, CA.
to R Hendler,
Yale University
School of
Medicine, 333 Cedar Street, New Haven, CT 06510.
Received August 13, 1987.
Accepted
for publication
January
5, 1988.
Society

for Clinical

Nutrition

1239

Downloaded from www.ajcn.org by on August 11, 2009

ABSTRACt
Optimal
composition
of reducing
diets
remains
controversial.
Seventeen
obese inpatients
received
440 kcal/d,
either 41% protein
plus 55% carbohydrate
(CD) or 95%
protein
(PP), for 3 wk. There were no significant
diet effects (all data CD vs PP) in weight loss
(8.88 1 .0 1 vs 8.74 0.79 kg), loss of lean mass (2. 10 0.35 vs 1 .6 1 0.39 kg), metabolic
rate reduction
(15.3 2.8 vs 13.0 5.2%),
or meal-stimulated
thermogenesis
(26.6-37.9
vs
29.0-26.
1 net kcal/3
h [time NS also]). Triiodothyronine
(T3) responses
differed (2.35 0.11
to 1.57 0.l4vs2.43
0.1 1 to 1.47 0.12 nmol/L,p
< 0.01)asdid
freeT3(3.4
0.2 to 2.6
0.2 vs 3.2 0.2 to 2.0 0.2 pmol/L,
(p < 0.01));
thyroxine
declined
similarly
in both groups.
Subjects
fed CD gained no advantage
over subjects
fed PP. Regression
analyses
revealed
no
relationship
between
thyroid
hormones,
energy deficit, or bean mass with nitrogen
losses,
sug-

HENDLER

1240
TABLE

1
characteristics

Physical

ET AL

of subjects
Percent

Diet

CD(n

9)
8)

PP(n
*

Age

Height

28.0 2.8
33.6 2.6

1.71 0.03
1.67 0.03

Body weight

IBW*

Body mass
index

Lean body
mass

kg

kg/mi

kg

137.2
125.5

14.2
10.2

141 5
133 4

well

ually

body

conventional

known
robes as regulators
of thermogenesis
and
protein
metabolism.
The objective
of the present
study was to determine
the influence
of diet composition
on metabolic
adaptation to weight loss, especially
changes
in energy expenditure and conservation
of LBM.

introduced

resting

obese

subjects

glucose

values

healthy

plasma

(75-g)
oral
renal function

417-mmol
liver, and

informed

were

in writing

and

consent

Investigation
Committee
Medicine
in accordance

dated in Tokyo,

Japan,

normal

All subjects

before

forms

studied.

All had nor-

response

to a

test (12).
Thyroid,
as serum
electrolytes
and

within

any medications.

consent

tab protocol

were

and a normal

glucose-tolerance
tests as well

electrocardiograms

were taking

the study.

were

limits.

No sub-

gave voluntary
The experimen-

by the Human
Yale University
School
of
Helsinki
Declaration
as up-

of the
with the

approved

1975.

Subjects
were admitted
to the Yale Adult General
Clinical
Research
Center (CRC) for
1 mo. During the month
before
admission,
all consumed
normal,
weight-maintaining,
mixed
diets. Throughout
the hospitalization
subjects
were weighed

each morning,

in a hospital

gown,

voiding.
Urine
measurements

was collected
of N, sodium,

daily

were initially

diets detailed

of

the CRC.

Diets

After admission

for at beast 3-5 d before

and

ate weight-maintaining

diets

the VLCD,

containing

50%

all

carbohy-

placed

below)

before

breakfast

and after

in 24-h lots for subsequent

and potassium
excretion.
Sub-

on a weight-maintenance

for 3-5 d. Thereafter,

diet (all

they were randomly

to a VLCD, either the mixed diet (Cambridge

Diet,
CD) or the pure protein diet (PP). Group characteristics(all
NS
between groups) are described
in Table 1.
On the last day ofthe weight-maintenance
period and on the
14th,

measured

and

2 1st days

of the

and blood samples

intravenous

VLCD

period,

were obtained

for measurements

catheter

the

was

RMR

thermic

effect

weekly.
of food,

To examine

DIT

was

hormones,

of weight

standing

the study

were measured
measured
twice weekly;
uric acid, and complete
and electrocardiograms

After completion

to guard

boss. Blood

possible

evaluated

against

pressure

changes

at the

end

of

undesirable

conse-

and pulse responses

to

serum electrolytes
were
plasma proteins,
liver function,
serum
blood counts
were measured
weekly;
were done weekly.
twice

daily;

of the VLCD

period,

VLCD

preparation

(donated

by

International,
Inc, Monterey,
CA), nominally
contaming
41% protein,
55% carbohydrate,
and 4% fat (by manufacturers
analysis).
The CD is based on nonfat
milk, soy proteins,
and fructose
and contains vitamin and mineral
supplements needed to meet the US RDA for adubts(4).
PP contained
95 1% protein,
2 1% carbohydrate,
mined from standard
tables (14). This

from

lean roasted

meats

and was supplemented

US RDA for adubts(4).
was
on
Daily fluid
erages) was
mmol/d
in
required
to

K intakes

(chicken

and

or turkey),

with vitamins

3 1% fat as deter-

diet was made

tuna,

and minerals

for N-balance
calcubased on Kjebdahl
total N determinations
(15) peraliquots
ofeach
diet prepared
as served
to subjects.
intake (dietary and noncaboric,
noncaffeinated
bevat least 2.0 L/d throughout.
Na intake
was -90
each diet initially and was increased
individually
as
counteract
occasional
postural
hypotension.
Total

on CD and PP were 69 and 55 mmob/d,

(15%

protein,

ice cream,
54%

banana,

carbohydrate,

tables [l4J and measured

respectively,

K gluconate
or K chloride)
known
normal
serum
K bevels (> 3.8

mmol/L) during the VLCD period.

The DIT test meal was a room temperature,
ofmilk,

to meet the

N content

Dietary

including
supplements
(oral
to be required
to maintain

enate

entirely

or egg whites

and sugar
and

31%

350-mL
containing
fat;

from

homog6 1 8 kcal
standard

portions).

Measurements

the weight-maintenance
period and on the 21st day of the
VLCD period. All subjects were under close medical
supervision throughout

CD is a commercial

from an indwelling
of thyroid

insulin, and circulating

substrates.
In addition,
cardiovascular
and circulating
plasma catechobamine
responses
to upright
were determined

(NS).

Cambridge

lations
formed

assigned

quences

kcal mixed, balanced

diet
were instructed
to continue

subjects

were grad-

Energy
expenditure
assessments
for RMR and DIT were
made by indirect
calorimetry
with the ventilated-hood
technique with a Beckman
Metabolic Cart(SensorMedics
Corporation, Anaheim,
CA). Metabolic
rate was calculated
from measured oxygen consumption
with adjustments
for respiratory
quotient
(RQ) and urinary
N excretion
(16, 17). The ventibated-hood
method
was used because
it allows the prolonged
3h measurement
required
to evaluate
DIT with only minimal
subject
discomfort.
RMR
was measured
between
0700
and
0900 after an overnight
fast while subjects
lay quietly awake
in bed. Subjects
were allowed
to listen to the radio or watch
television
during the test but conversation
was not permitted.

Downloaded from www.ajcn.org by on August 11, 2009

Seventeen

mal fasting

in the

from

subjects

30% protein,
and 20% fat. During
this base-line
period
intake was adjusted
individually
to prevent
weight boss.
During the 2l-d VLCD period, the CD and PP subjects consumed an average of436
2 and 439 5 kcab/d, respectively

Subjects

posture

which

drate,
calorie

Methods

7th,

to a 1200-1500
foods,

after discharge

subjects

jects

60.9 4.8
52.5 3.7

weight (13).

bdeal body

jects

46.4 4.0
45.1 3.1

HIGHRMR

was calculated

from

45 mm

AND LOW-PROTEIN

of measurement

while

the

subjects
were in a stable, resting state.
To measure
DIT, subjects were removed
mediately
after RMR
determination
and

from the hood imgiven the room temperature

test meal to consume
within 5-10 mm. Then they
were placed back in the hood and metabolic rate was measured
continuously
for 3 h. Total energy expenditure
for the 3 h was
summarized

in six 30-mm

periods.
Although
5 h (18), these

sponse to food was to extend

nated after 3 h because most subjects
reliable

hood

resting

longer.

metabolic

rate

became

The net thermogenic

response

mm.

The net area of hormonal

similarly

the

and

concentrations

beast 30 mm

after

(19). Supine

placement

ofan

lay quietly;

2, 5, and

10 mm

ofstanding.

measured

concurrently

Analytical

techniques

increment

responses

were

were

eval-

plasma catewere drawn at

intravenous

cath-

were then drawn

pressures

and

after

pulse

were

with blood sampling.

Radioimmunoassay

techniques

were

used

to

determine

plasma insulin (20) and total and free serum triiodothyronine

(T3) (2 1) (Corning
Medical,
Medfield, MA). Serum
thyroxine
(T4) was determined
by a competitive
protein-binding
method
(22).

Plasma

catechobamines

(epinephnne

and

norepineph-

tine) were measured

by a radioenzymatic
procedure
(23) (Upjohn Diagnostics,
Kalamazoo,
MI). Plasma glucose was determined by the glucose oxidase
technique
with a glucose analyzer
(Beckman
Instruments,
Sunnyvale,
CA), plasma free fatty
acids by microfluorometry
(24), and $-hydroxybutyrate
enzymatically
from perchboric
acid extracts of whole blood (25).
Daily
24-h urine collections
were analyzed for total N by the
Kjebdahb technique
(15), for creatinine by the picric acid coborimetric
procedure
(26), and for Na and K by flame photomCt!-) (27).

The composition
ofweightbost
during the VLCD period was
estimated
from changes in weight and N balance. N balance
was the difference
between N intake and boss. N bosses were
calculated

from

measured

total

urinary

N plus

35.69

mmol/d

estimated
fecab N boss (28) and 0.357 mmob/kg.d
for unmeasured
integumentab
loss (29). All subjects
reported
decreased

stool volume and frequency

and none experienced
diarrhea
during the VLCD period.
Starting
bean body mass was estimated
from 24-h urinary
creatinine
according to the relationship
ofForbes
and Bruimng
(30). Losses of LBM during VLCD periods were calculated
from cumulative
N balance
assuming
71.38 mmob (1 g) of N
is equivalent
to 6.25 g body protein
or 3 1 .25 g LBM.
Data are presented
as the means SEM. Statistical analyses
were performed
with the Students
t tests for paired
and unloss

paired
actions

data and two-way

between

analyses

variables

(31).

and lean body

mass

ofvariance

to evaluate

inter-

Results
Weight

Weekly

nitrogen

losses

The total weight lost by CD and PP subjects

ofVLCD
therapy
was 8.88 1 .0 1 kg and 8.74

after 3 wk
0.79 kg,

balance

Subjects

Week

Week

Week

mol/wk
On CD
1
2
3
4
5
6
7
8
9

iSEM
to posture
standing

indwelling

Blood

180

of concentration

samples

samples

in

by

TABLE

-4.95
-2.13
-4.43
-2.33
-1.90
-1.85
-3.44
-1.58
-2.77

-2.82

0.40

-2.42
-1.19
-1.70
-1.11
-1.06
-0.71
-2.01
-0.78
-0.25
1.24 0.23

I .03
-1.32
-1.10
-0.62
-0.48
-1.16
-2.15
+1.35
-0.26
-0.75
0.32
-

On PP
10
11
12
13
14
15
16
17

-3.12
1.47
-4.20
-1.49
-2.71
-3.01
-3.38
-2.32

iSEM

-2.71

respectively.

These

-0.54
-0.82
-2.28
-0.27
-2.31
-1.47
-1.68
-0.49

0.33

weight

-1.23

bosses

+0.10
-0.29
I .08
+0.69
-0.43
+0.81
-0.52
+2.83
+0.26 0.43

0.29

were

not

significantly

different
expressed
in absolute
terms or as percent
of mitial weight.
Table 2 contains
individual
weekly
N losses.
Both dietary regimens
resulted
in a similar pattern
ofnegative
N
balance,
highest
during
the first 2 wk then diminishing
in the third week. Note that we observed
substantial
subject-to-subject
variability
throughout
the 21-d VLCD period. Although
only one of nine CD subjects
vs four of
eight PP subjects
achieved
positive
N balance
in the third
week;
this was not a significant
difference.
Subjects
on
CD and PP lost 2. 10 0.35 and 1.6 1 0.39 kg (p = NS),
respectively,
ofLBM
3.4 and 3.0% oftheir

Fasting

hormones

during

initial

3 wk ofdieting,
(p = NS),

LBM

representing
respectively.

and substrates

Table 3 shows group means

of thyroid
function
tests
by week. In each group, the effect of 3 wk ofdieting
was
significant
(p < 0.05)
on total
T4-binding
capacity
(TTBC),
estimated
free T4 (EFT4),
T3, and free T3
whereas
the small decline
in T4 levels was not significant.
Total serum T3 did not differ between
the diet groups
and
decreased
rapidly
during the VLCD
period by ---24% (p
< 0.0001
after only 1 wk). The decline
was also reflected
in free T3 (p < 0.01). There was a small nonsignificant
difference
in free T3 between
groups
after dieting
(0.8
0.2 vs 1.2 0. 1 pmol/L,
CD vs PP, respectively,
p
= 0.08).
Table 4 shows other fasting
levels of hormones
and
substrates
before and after 3 wk ofeach
diet. The significant (p < 0.003)
fasting
plasma
insulin
decline
shown

Downloaded from www.ajcn.org by on August 11, 2009

eter while subjects

mean

in units

multiplied
by minutes.
Sympathetic
nervous
system
responses
uated
weekly by comparing
supine and

chobamine

re-

were termi-

multiplying

and substrate

but are expressed

thermic

too restless for

if kept in the
to the test meal

determinations

(kcab/3 h) was calculated

by dividing
metabolic
rate by the fasting
value,
calculated

the
tests

1241

VLCDS

1242

HENDLER

TABLE

Thyroid

function

tests

Subjects

WeekO

On CD
14 (nmol/L)

94.0

TTBC(nmol/L)t

Weeki

5. 1

90. 1 6.43
36017

38021

EFT4(pmol/L4
T3(nmob/L)

181
2.37 0.11

211
1.67 0.11

FT3 (pmol/L)II

3.4 0.2

2.6 0.2

On PP
T4 (nmol/L)

84.9 6.4
39918
17 1
2.43 0.1 1
3.2 0.2

TTBC(nmol/L)t

EFT4(pmol/L4
T3(nmol/L).
Fr3 (pmol/L)II
S

ET AL

96.5
35826
21
1.83
2.5

Week2

Week3

83.7 6.4
33713

8 1 . 1 6.4
30513

211

9.0
1
0.17
0.2

221

1.55 0.08
2.5 0.2

1.57 0.14
2.6 0.2

96.5

83.7
34521
22
1.47
2.0

9.0
35826
22 1
1.60 0.12
2.3 0.2

5.2
1
0.12
0.2

Thyroxine.

t Total T4-binding

capacity.

II FreeT3.
represents
49 and 43% decreases,
CD and PP, respectiveby (diet effect NS). This decrease
in insulin
is consistent with the observed
15% reduction
in fasting plasma
glucose
concentration
observed
after the VLCD
period
(p < 0.02 with no group
effect).
These
changes
were accompanied
by the expected
increases
in circulating
free
fatty
acids
(p < 0.002)
and
/3-hydroxybutyrate
(p
<

0.003).

Although

ketone

responses

to

weight

loss

tended
to be greater
in the PP subjects,
consistent
with
our previous
report (32), this did not reach significance.
Supine plasma
norepinephrine
bevels declined
after dieting in both groups although
only the PP reached
statistical significance
(CD, 22%, p = 0.26; PP, 33%, p < 0.02;
no significant
group effect). Supine
plasma
epinephrine
concentrations
tended
to decline
with these diets although
this effect was only significant
when the groups
were combined
(p < 0.05).
Energy

expenditure

Table
for each

5 shows weekly
postabsorptive
metabolic
rate
subject.
The mean decreases
(both p < 0.002)
in
metabolic
rate were 15.3 2.8 and 13.0 5.2% in the
CD and PP groups,
respectively
(p = NS between
groups).
Figure
1 shows clearly
that the net thermogenic
reTABLE 4
Effects ofdieting

on fasting

plasma

substrate

and hormone
CD

Glucose(mmol/L)
Insulin (pmol/L)
Freefattyacids(zmol/L)
$-hydroxybutyrate
(mmol/L)
Norepinephrine
(nmol/L)
Epinephrine
(pmol/L)

sponse
to the standardized
test meal was undiminished
by weight loss with either VLCD treatment.
In fact, the
CD group response
seemed
to increase
(p = 0.07) after
weight boss.
Hormonal
Glucose,

and substrate
responses
to a meal
insulin,
and free fatty acid responses
to the
test meal are shown
in Figure 2. The glycemic
response
to the meal was similarly
affected
in both treatment
groups,
showing
peak responses
delayed
by 60 to 90 mm
and the net areas above the base line ofthese
curves tending to increase
(CD, 183 23 to 266 45 mmol.
mini
L, p = 0. 13 and PP, 156 20 to 355 63 mmol.
min/L,
p < 0.02) at the end of the diet period.
The net areas of
the insulin
response
curves
after the test meal
were
slightly
lower after the diets (CD, 76 100 8600
to
65 300 5000
pmol
min/L
and PP, 102 600 27 300
to 70 300 16 500 pmol.min/L,
bothp
=
NS), which
is
consistent
with the sluggish plasma glucose
response.
No
.

group effects were evident

in postprandial
free fatty acid
levels or blood ketones
(results
not shown),
which both
suppressed
similarly
in each group after the test meal.
Both groups
of subjects
had significant
(p < 0.02) increases in peak venous
plasma
norepinephrine
levels after the DIT test meal before and after weight loss (CD,

concentrations

before

CD after dieting

5.11 0.17

4.330.28

194 36
8605
0. 12 0.03
1 .23 0. 19
82 16

100
110070
1 .35
0.96
66

36
70
0. 1 3
I1

PP before

PP after dieting

5.11 0.17
15 1 3
68565
0. 10 0.02
1.38 0. 12
98 22

4.330.17
8622
105859
1 .7 1 0.38
0.93 0.11
7 1 11

Downloaded from www.ajcn.org by on August 11, 2009

t Estimated
free T4.
Triiodothyronine.

HIGHTABLE

AND

LOW-PROTEIN

VLCDS

individual

Subjects

Week

resting

metabolic

Week

rate

Week

Plasma

Week

3
-J

kca//min

3
4
5
6
7
8
9

1 SEM

1.21
1.14
1.58
1.39
0.99
1.40
1.38
1.03
1.48
1.30 0.07

1.42
1.31
1.74
1.76
1 . 12
1.48
1.52
1.03
1.39
1.42 0.08

Glucose

7.0

7.0

6.5

6.5

6.0

6.0
5.5

/$

On CD
2

After

Before

Postabsorptive

1243

1.12
1.13
1.59
1.21
0.94
1.51
1.41
1.0 1
1.36
1.25 0.08

1.15
1.03
1.57
1.24
1.03
1.17
1.32
0.94
1.34
1.20 0.06

4.5
4.0

30

60

120

90

150

180

Plasma

1200

1000

1000

600

600

400

400

1.09
1.06
1.30
0.52
0.93
1.52
1.47
1.18
1.130.11

120

150

180

30

60

120

90

150

30

60

90

120

150

180

30

60

90

120

150

180

180

Plasma

Free Fatty Acids

1200

1200

1000

1000

600

600

800
0

E
E

400
200

200
C

30

60

90

120

150

180

Minutes

from
1 .29

1.80 0.40 nmob/L

and PP, from
0. 1 1 nmob/L,
p = NS between
weight boss lowered
basal norepineph-

1.39 0.29 to
0.09
to 2.06

groups).

Although

FIG 2. Plasma
standardized

glucose,

after

meal. Presented

0.50

response

significantly
affected

60

90

120

150

by weight

blood

maintenance
The decrease

After

fell from

changes

(CD:

between

the net area

DIT

test

meal

groups

90

120150

180

and

of the
was

was

not

Un-

Minutes

responses

declined

6%

to posture
from

weight-

after the 3-wk VLCD

period.
in both groups
as CD subjects
4/79 4 to 1 1 5 6/76 5 mm Hg and
from 120 5/79 3 to 1 1 2 4/68 3
blood pressure
responses
also declined
VLCD
regimens
as reflected
by systolic

standing

3 vs -7

1 1 4 mm Hg, p
nephrine
responses

weight

norepinephrine

pressures

12 1

after

-2

60

the

bevels (NS)
was similar

PP subjects
fell
mm Hg. Postural
slightly
after the

30

to the

boss.
and

Supine

180

Minutes

u0

acid responses

1.

Likewise,
after

different

Cardiovascular
30

free fatty

PP

1.0

and

as in Figure

0. 1 5 nmob/L).

norepinephrine

,J

insulin,

rime bevels as described

above,
the peak increment
after
the DIT test meal was not significantly
changed
from responses
before
the
VLCD
period
(CD:
before
0.41
0. 12, after 0.48 0.09 nmol/L;
PP: before
0.76 0.08,

Before

Minutes

10 mm before
3 mm Hg,p
<

0.02).
to

boss or the dietary

Figure

standing

regimen

and

after

dieting

NS; PP: 0 2 vs
3 shows that norepi=

were

unchanged

by

used.

FIG I. Thermogenic

response
to a standardized
618 kcal meal before and after 3 wk ofVLCD.
Metabolic
rate (left) and the net thermogenic response
within 3 h after the meal (right) are shown for the CD
(closed circles and bars) and the PP regimens
(open circles and bars).
Values are :i SEM.

Discussion

can

The results ofthis study show that

be achieved
by obese subjects

similar weight losses

on VLCD
of either

Downloaded from www.ajcn.org by on August 11, 2009

1.10
1.06
1.35
0.71
0.95
1.50
1.56
1.03
1.160.10

90

200
-

1.21
1.09
1.43
0.75
1.08
1.71
1.56
1.06
1.240.11

60

800

800
0

On PP
1.26
1.10
1.57
0.86
1.20
1.68
1.65
1.06
1.300.11

30

Insulin

1200

200

10
11
12
13
14
15
16
17
iSEM

1244

HENDLER
Before

tal plasma
protein
content
was unaffected
by either
regimen,being65
1 and62
1 g/Lbeforevs65
2and6l
2 g/L after 3 wk ofCD
and PP, respectively.
Similarly,
albumin
values
did not change
during
either
therapy.

After

2.0

1.5

ET AL

aC

The

consequence
suggested
deficiency.

1.0

fall in TTBC

drate

Minutes

FIG 3. Changes
mm ofstanding.

After

in plasma
norepinephrine
levels after 2, 5, and 10
CD shown as shaded bars and PP as open bars SEM.

as a

in these

diets

(such

as CD)

is probably

insufficient

(CD)

or pure

protein

(PP).

Slightly

im-

33, 34). Although

nificantly
superior

we could not show the protein

in N sparing,
longer studies

groups
of subjects
undoubtabby
would.
that the small difference
in total K intake
plements)
between
the two groups
(CD
found
these N-balance
data also through

association

of K depletion

and excessive

>

diet sigin larger

Further,
(diet plus

note
sup-

PP) may

con-

the

well-known

N losses

during

weight
loss (35). Thus, although
none ofour
subjects
on
either
diet became
hypokalemic,
the ability
of our diets
(particularly
the PP diet) to conserve
N may have been
improved
ifK intake
had been somewhat
higher.

Like Yang
uab variability
17 subjects

termed,
subjects

week than
responses.

et al (34) we observed
considerable
in N balance.
By the third week,
were

close

an adaptive
N balance

to

individ5 of our

N equilibrium,
as Yang
et al
response.
On the other hand, in
was more
negative
in the third

in the second,
perhaps
reflecting
maladaptive
Our adaptive
subjects
appear to have achieved

N equilibrium

more

quickly

during

or Yang (34) previously

of 30-40
d were required

mum
adaptive
subjects.
overemphasized

However,
as it may

the

reported,
for even

third

week

than

when a minithe best of the

this difference
should
not be
merely
be because
N intake

was not measured

but rather was estimated
(from standard food tables or manufacturers
product
data) in these
previous
studies.
We found that similar
estimates
of N
content
of diets
in this investigation
consistently
exceeded
measured
N content
by --5%. This error would
ordinarily
be trivial
but would
be accentuated
by the bal-

ance calculations
for these subjects
so near equilibrium.
It is important
to recognize
that total N balance
may
not reflect organ- or tissue-specific
N losses. Many investigators
have examined
plasma
proteins
as supplemental
indices
of visceral
or rapidly
exchanging
proteins.
Although
every subject lost N during
weight reduction,
to-

these

diets

(--

ment

only

includes

tissues.
to weight

boss on

14%) is similar
to previously
reported
data
(38-40). Unlike
RMR
the stimulation
ofmetabolism
by
the test meal (DIT)
was essentially
unaffected
by weight
boss or the diet used. This observation
that DIT is relatively
unresponsive
to weight
boss supports
the findings
of other
investigators
(41, 42). Note
that our measure-

the first three

postprandial

hours

and

it is well known
that meal effects actually
extend
well beyond this, perhaps
as much
as an additional
3-4 h (18).

Thus, bate effects may have been undetected.

In contrast
to the unchanged
thermic
effect

ofa

meal,

weight
reduction
clearly
influences
the hormonal
and
substrate
profiles
of postprandial
blood.
Because
the
same
test meal was administered
throughout,
the sluggish postprandiab
response
of plasma
glucose
after dieting probably
reflects
some degree
ofdelayed
absorption.

Even though
boss actually
cose

drate

levels after
restriction

resulted
during

the glycemic
response
was delayed,
weight
increased
the area above basal plasma
gluthe DIT
because

in the greater
caloric

net insulin
the insulin

restriction,

secretory
response

meal. This may reflect

the carbohydrate-free

increment
insulin

(p

<

carbohydiet (PP)

0.05). As expected

bevels

response
(indicated
curve above
basal)

decreased.

The

by area under
tended
to decrease
(NS unless
groups
combined,
then p < 0.05). Although
the PP group
response
fell somewhat
more than
did the CD (14 vs 3 1%), consistent
with the smaller glucose response,
this group difference
did not reach statistical significance
(p = 0.2 1). Fasting
norepinephrine
bevels
declined
with weight
loss on both diets. However,
the
plasma
concentration
responses
to a standardized
meal
was not affected by dieting.
In addition
to catechobamine
responses
to a meal, the
norepinephnne
and hemodynamic
responses
to standing changed
only slightly with weight loss on either diet.
This contrasts
sharply
with our previous
experience
when all subjects
on pure
protein
400 kcal/d diets deveboped postural
hypotension
(32). The difference
between
these two studies
is that Na intake
was increased
as required
in the present
study to counteract
any hypotensive responses
but intake was fixed in the previous
study
thus many subjects
experienced
substantial
negative
Na
balance.

Downloaded from www.ajcn.org by on August 11, 2009

nutrients

we (32)

be interpreted

undetected
protein
bosses from critical
The decrease
in RMR in response

Standing

proved
N balance
(NS) was seen with PP as LBM bosses
in the CD and PP subjects
accounted
for 24 and
18%
(NS) oftotal
weight
lost, respectively.
This LBM boss fobbowed the same pattern
found
by other
investigators
(4,

(34)
four

should

to raise thyroid-binding
capacities
above
levels seen with
carbohydrate-free
diets.
Our data confirm
this finding
because
the decrease
in TTBC
was similar
with both diet
treatments.
It must
be emphasized
that these
measures
are indirect
and cannot
totally
preclude
the existence
of

0.5

mixed

probably

of decreased
total
carbohydrate
intake
as
by Kelleher
et al (36), rather
than as protein
Hoffer
et al (37) suggested
that total carbohy-

HIGH-

AND

LOW-PROTEIN

1245

VLCDS

00
-1.5

C.)
C

-3.0-

#{149}

r=O.14
p = 0.6

C
C)
0

z
.----&----

>

-7.5

-.

0#{149}

0
I

-9.0
0

0.25

0.50

0.75

Change
FIG 4. Relationship
is

between changes
combined,
with CD
by dashed lines.

for all subjects

indicated

1 .00

1.25

1.50

1.75

2.00

in Free T3 (pmoVL)

in free T3 after 3 wk ofVLCD

with cumulative
N balance.
Regression
indicated
by closed circles, PP by open circles, and 95% confidence

Ofaib the changes

in thyroid
status resulting
from these
diet treatments,
only the decrease
in T3 (CD, 33%; PP,
39%) and free T3 (CD, 23%; PP, 38%) probably
are rebevant. The slightly greater decreases
in T3 and FF in the
PP subjects
is consistent
with a proposed
robe for dietary
carbohydrate
in supporting
T3 levels during
weight loss
(40, 43, 44). Kaptein
et al (45) reported
that initial
T3
levels predicted
the subsequent
fall in T3 that occurred
during 40 d of400 kcal/d pure protein
feeding.
Our data
seem to support
this although
in our shorter
study the
relationship
failed
to reach
statistical
significance
(r
= 0.47,
p = 0.055).
The decrease
in RMR
observed
cannot
be explained
solely by weight loss, which was only 6.5 and 7.0% in CD
and PP subjects,
respectively,
or by LBM,
which
decreased
by only 3.4 and 3.0%, respectively.
Thus,
our
data are consistent
with the hypothesis
that adaptive
changes
in RMR may help to minimize
the energy deficit
and weight loss during underfeeding
and also suggest that
DIT does not participate
in such a response.
Note that
the final RMR and DIT measurements
were made while
subjects
were still actively
losing weight,
which
would
seem to maximize
such adaptive
effects over what might
be observed
on a weight-maintenance
diet at a reduced,
stable weight after weight loss.
Our data were also examined
for evidence
linking the
sympathetic
nervous
system
with changes
in DIT or

shown
limits

However,
plasma
norepinephrine
concentration,
and incremental
areas do not correlate
significantly with the thermogenic
response
after the test meal
or with the changes
in RMR
before or after dieting
in
either group.
Although
this seems to minimize
the involvement
of the sympathetic
nervous
system
(SNS) in
regulation
of these processes,
more-sensitive
measures,
such as norepinephrine
turnover,
could reveal relationships
not visible
in plasma-concentration
responses.
Such a dissociation
of plasma
concentration
and turnover responses
was reported
(46), leaving open the possibility ofa robe for the SNS in thermic
adaptive
responses
during weight boss.
From
a strong
correlation
of T3 and RMR
changes
during
high- and low-carbohydrate
VLCD
therapies,
we
(41) suggested
that changes
in T3 may mediate,
in part,
a change
in RMR during weight loss. Because
ofthis
experience,
we examined
the relationship
ofthese
variables
in the present
data set. Although
T3, FT, and RMR all
decreased
during the study, there was no relationship
between RMR and T3 or FT or between
their incremental
changes
during
weight
loss. However,
note that calorie
intake was substantially
higher in that study (800 kcal/d)
than in this present
work.
Thus,
it is possible
that the
relationship
was not detected
because
energy intake was
simply too bow.
Recent
reports (33, 45) suggested
links between
T3 and
RMR.
peaks,

Downloaded from www.ajcn.org by on August 11, 2009

-6.0

1246

HENDLER

sion
data

balance.
We thank Andrea Belous, Aida Groszmann,
Ralph Jacob, and Clara
Wong for their expert technical
assistance.
We are particularly
grateful
to the nurses ofthe General Clinical Research
Center for the care given
to our patients,
and to Dr Robert Gelfand
for reviewing
the manuscript
and want to acknowledge
Dr Robert S Sherwin
for his thoughtful
contributions to the design ofthe study.

References
1. Isner JM, Sours HE, Paris AL, Ferrans
VJ, Roberts
WC. Sudden,
unexpected
death in avid dieters using theliquid-protein
modifiedfast diet: observations
in 17 patients and the role ofthe prolonged
QT interval.
Circulation
1979;60:l401-b2.
2. Van Itallie TB, Yang MY. Cardiac dysfunction
in obese dieters: a
potentially
lethal complication
ofrapid
massive weight loss. Am J
Gin Nutr 1984; 39:695-702.
3. Sours HE, Frattali
VP, Brand CD, et al. Sudden
death associated
with very low calorie weight reduction
regimens.
Am J (in
Nutr
198 l;34:453-6
1.
4. Hoffer LI, Bistrian BR, Young VR, Blackburn
GL, Matthews DE.
Metabolic
effects ofvery low calorie weight reduction
diets. J Gin
Invest l984;73:750-8.
5. Committee
on Dietary
Allowances,
Food and Nutrition
Board,
National
Research
Council.
Recommended
dietary
allowances.
9th ed. Washington,
DC: National
Academy
Press, 1980.
6. Kbeiber M. The fire of life: an introduction
to animal energetics.
Huntington,
NY: Krieger RE, 1975.
7. Boer JO, van Es AJH, Rcovers
LCA, van Raaij JMA, Hautvest
JGAJ. Adaptation
of energy metabolism
of overweight
women to
low-energy
intake,
studied
with whole body calorimeters.
Am J
Gin Nutr 1986;44:585-95.
8. Horton
ES. Introduction:
an overview
ofthe assessment
and regulation of energy balance
in humans.
Am J Gin Nutr l983;38:
972-7.
9. Danforth
E Jr. Diet and obesity.
Am J Gin Nutr 1985;4l:l
13245.
10. Rothwell
NJ, Stock MJ. Regulation
ofenergy
balance.
Ann Rev
Nuts 1981; 1:235-56.
1 1. James WPT, Trayhurn
P. Thermogenesisand
obesity. Br Med Bull
l98l;37:43-8.
12. National
Diabetes
Data Group. Classification
and diagnosis
of diabetes mellitus and other categories
ofglucose
intolerance.
Diabetes
1979;28: 1039-57.
13. Metropolitan
Life Insurance
Company.
New weight standards
for
men and women.
Stat Bull Metropol
Life InsurCo
1959;40:l-4.
14. Watt BK, Mcml
AL. Composition
of foods. Washington,
DC:
United
States Department
of Agriculture,
1979. (USDA
handbook #8.)
15. Monro HN, Fleck A. Analysisoftissues
andbody
fluids for nitrogenous constituents.
In: Munro HN, ed. Mammalian
protein metabolism. New York: Academic
Press, Inc., 1969: Vol 3. 424-525.
16. Lister G, Hoffman
JIE, Rudolph
AM. Oxygen
uptake in infants
and children:
A single
method
for measurement.
Pediatrics
l974;53:656-62.
17. J#{233}quier
E. Long term measurement
ofenergy
expenditure
in man:
direct or indirect calorimetry.
In: Bjorntorp
R, Cairella M, Howard
AN, eds. Recent advances
in obesity research
III. London,
UK:
John Libbey,
1981:130-5.
18. Ebiner JR. Acheson
KJ, Doerner
A, et al. Comparison
of carbohydrate utilization
in man using indirect calorimetry
and mass spectrometry
after an oral load of 100 g naturally-labeled
3C glucose.
BrJ Nutr l979;41:4l9-29.
19. Cryer PE. Diseases
of the adrenal
medullae
and sympathetic
ncrvous system. In: Febig P, Baxter JD, Broadus
AE, eds. Endocrinology and metabolism.
1st ed. New York: McGraw-Hill,
1981:51150.
20.

Lazarow
A. Immunoassay
for insulin: two antibody
1963; 12:115-26.
Ekins RP, Ellis SM. The radioimmunoassay
of free thyroid
hormones in serum. In: Robbins
J, Braverman
LE, eds. Thyroid
research. Amsterdam:
Excerpts
Medics,
1976:597-600.

Morgan

CR,

system. Diabetes
21.

Downloaded from www.ajcn.org by on August 11, 2009

during weight boss. We performed

regresanalyses
for weekly
and overall
N balance
with T3
for each diet and for all 17 subjects
combined.
Similarby,
we examined
potential
correlations
between
N
balance
during
weight
boss with
initial
LBM,
body
weight,
and other
thyroid
indices
as well as between
changes
in these variables.
Additionally,
we examined
the influence
ofenergy
expenditure
(RMR and DIT) before and after weight
boss on N balance.
A regression
analysis
for all subjects
over the entire 21-d VLCD
period of N balance
with changes
in free T3 is shown
in
Figure
4, which shows no significant
correlation
between
these variables.
Of all variables
only initial
LBM and
weight
boss were significant
correlates
of body
N bosses
during dieting (r = 0.67, p < 0.01 and r = 0.59, p < 0.02,
CD and PP, respectively).
No other regression
analyses
performed
on these data yielded
significant
correlations
with N sparing
during
weight
boss. However,
our study
was only 2 1 d bong and therefore
these data may not be
representative
ofchanges
that
might
have
been
manifest
after the third week.
It is possible
that our failure to observe
a relationship
between
T3 and N economy
was due to the restricted
range of N bosses we observed.
Perhaps
changes
in T3
would have correlated
with N losses if our subjects
had
experienced
larger N deficits and/or
the study had been
of longer duration.
However,
other investigators
(44),
whose subjects
had considerably
larger N losses than our
subjects
had, were also unable
to show an effect of
changes
in T3 on N balance.
T3 remains
an attractive
potential
control
of N economy,
however,
because
of its
well known effects in hyperthyroidism
and hypothyroidism. It is very important
not to overlook
interindividual
variability
and the well-known
technical
problems
of Nbalance
studies
(47), which
may require
large subject
populations
to provide
definitive
data on the controls
of
adaptation
to weight
loss with this approach.
We agree with Yang and Van Itallie (33) that the only
definitive
clarification
ofthis
extremely
important
question will be from prospective
studies
in which T3 is manipulated
experimentally
during low-calorie
dieting.
In conclusion,
these two isocaloric
VLCDS,
providing
44 g and 96 g protein/d,
were similarly
effective
in promoting
weight loss and conserving
total body
N. These
short-term
(3-wk) results should
not be taken to imply
that N balance
will be achieved
during
any protracted
VLCD
regimen
in a particular
individual.
Neither
diet
had any effect on DIT although
both produced
similar
decreases
in postabsorptive
metabolic
rate. T3 declined
significantly
in both diet groups;
however,
this change
was not clearly linked to changes
in metabolic
rate or N
N conservation

ET AL

HIGH22.
23.
24.

25.

26.

27.
28.

29.

31.
32.

33.

34.

Selligson
H, Selligson
D. Measurements
ofthyroxine
by competitive protein binding.
Gin Chim Acta 1972; 38:199-205.
Passon PG, Peuler JD. A simplified
radiometric
assay for plasma
norepinephrine
and epinephrine.
Anal Biochem
1973;5b:6l8-31.
Miles J, Glassock
R, Aitkens
J, Gerich
J, Haymond
M. A microfluorometric
method
for the determination
offree fatty acids in
plasma. J Lipid Res l983;24:96-9.
Williamson
DH, Mellanby
J, Krebs HA. Enzymatic
determination
of D (-)-$-hydroxybutyric
acid and acetoacetate
in blood.
Biochem J 1972;82:90-6.
Tietz NW, Pruden
EL, Siggaard-Andersen
0. Electrolytes,
blood
gasses, and acid-base
balance.
In: Tietz NW, ed. Textbook
of cmical chemistry.
Philadelphia:
WB Saunders,
1986:1 172-3 16.
Willard HH, Merritt LL, Dean JA. Instrumental
methods
of analysis. Princeton:
D Van Nostrand,
1965:309-56.
Brown MR, Klish WJ, Hollander
J, Campbell
MA, Forbes GB. A
high protein,
low calorie liquid diet in the treatment
ofvery obese
adolescents:
long term effect on lean body mass. Am J Clin Nutr
l983;38:20-31.
Joint Food and Agriculture
Organization/World
Health Organization Expert
Committee.
Energy and protein
requirements.
Geneva: WHO,
1973. (WHO technical
report series #522.)
Forbes GB, Braining
GJ. Urinary
creatinine
excretion
and lean
body mass. Am J Gin Nutr l976;29:1359-66.
Snedecor
GW, Cochran WG. Statistical
methods.
Ames, IA: Iowa
State University
Press, 1980.
DeHaven
J, Sherwin
R, Hendler
R, Feig P. Nitrogen
and sodium
balance and sympathetic
nervous
system activity in obese subjects
treated
with a low-calorie
protein
or mixed diet. N EngI J Med
l980;302:477-82.
Yang MU, Van Itallie TB. Variability
in body protein
loss during
protracted,
severe caloric restrictions:
robe oftriiodothyronine
and
other possible determinants.
Am J Gin Nutr 1984;40:6l
1-22.
Yang MU, Barbosa-Saldinar
F, Pi-Sunyer
X, Van Itallie TB. Metabolic effects ofsubstituting
carbohydrate
for protein in a low cabrie diet: a prolonged
study in obese patients.
Int J Obes 1981;5:
231-6.

35. Sapir

control

1247

VLCDS
DO,

Chambers

of ammonium

NE, Ryan

JW. The

excretion

during

role of potassium
in the
starvation.
Metabolism

l976;25:2l 1-20.
36. Kelbeher
PC, Phinney
SD, Sims EAH, et al. Effects of carbohydrate-containing
and carbohydrate-restricted
hypocaloric
and eucaloric diets on serum concentrations
of retinol-binding
protein,
thyroxin-binding
prealbumin
and
transferrin.
Metabolism
l983;32:95-lOl.
37. Hoffer LI, Bistrian BR, Young VR, Blackburn
GL, Wannemacher
RW. Metabolic
effects ofcarbohydrate
in low-calorie
diets. Metabolism 1984; 33:820-5.
38. Bray GA. Effect of calorie
restriction
on energy expenditure
in
obese patients.
Lancet 1969;2:397-8.
39. Welle SL, Campbell
RG. Decrease
in resting metabolic
rate during
rapid weight loss is reversed
by low dose thyroid
hormone
treatment. Metabolism
l986;35:289-91.
40. Hendler
RG, Walesky
M, Sherwin
R. Sucrose substitution
in prevention
and reversal
of the fall in metabolic
rate accompanying
hypocaloric
diets. Am J Med 1986;8 I :280-4.
41. Bessard T, Schutz Y, J#{233}quier
E. Energy expenditure
and postprandial thermogenesis
in obese women
before and after weight loss.
Am J Gin Nutr 1983; 38:680-93.
42. Shetty PS, Jung RT, James WPT, Barrand
MA, Callingham
BA.
Postprandial
thermogenesis
in obesity. Gin Sci 198l;60:5l9-25.
43. Azizi F. Effect ofdietary
composition
on fasting induced
changes
in serum thyroid hormonesand
thyrotropin.
Metabolism
1978;27:

935-42.
44.

45.

46.

Spaulding
SW, Chopra
U, Sherwin
RS, Santokh
SL. Effect of caloric restriction
and dietary composition
on serum T3 and reverse T3
in man. J Gin Endocrinol
Metab l976;42:l97-200.
Kaptein
EM, Fisler JS, Duda MJ, NicoloffJT,
Drenick
EJ. Relationship
between the changes in serum thyroid hormone
levels and
protein
status during prolonged
protein
supplemented
odoric
deprivation.
Gin Endocrinol
1985;22:l-l5.
Schwartz
PS, Jaeger LF, Silverstein
5, Veith RC. Sympathetic
ncrvous system activity and the thermic
effect offeeding
in man. Int J
Obes

47.

Hegsted

1987;ll:141-9.

DM.

Balance

studies.

J Nutr

1976; 106:307-11.

Downloaded from www.ajcn.org by on August 11, 2009

30.

AND LOW-PROTEIN