Ketone Body Production and Disposal: Effects of Fasting,
Diabetes, and Exercise
Edmond 0. Balasse and Franqoise Fery
Laboratory of Experimental Medicine, University of Brussels, 115 Blvd. de Waterloo,
B-1000 Brussels, Belgium
1. INTRODUCTION
Under usual circumstances, glucose and free
fatty acids (FFA) are the major blood-borne energy
fuels. The ketone bodies (KB) mainly represented
by acetoacetic (AcAc) and P-hydroxybutyric
(POHB) acids are formed from partial oxidation of
FFA in the liver and represent a "third" substrate
that has a major importance as fuel under certain
physiological and pathological conditions which
are listed in Table I. All these situations are charac-
terized by decreased carbohydrate utilization, a
low insulin/glucagon ratio, and increased FFA
concentration resulting from accelerated lipolysis.
The physiological significance of a dual form
of transport of endogenous fat for energy supply
(FFA and KB) has been the subject of many specu-
lations. It is recognized that the main physiologic
function of ketone bodies is to serve as a fat-de-
rived substrate which can be utilized by the brain'
an organ that does not take up FFA. Thus, under
conditions of carbohydrate deprivation, KB will
restrict glucose disposal by the brain and allow for
the conservation of proteins which would other-
wise be catabolized in excessive amounts for glu-
cose synthesis. Although KB represent primarily a
cerebral substrate, they can be utilized by several
other tissues such as heart, skeletal muscle, intes-
tine, and kidney,* and there must exist in the
whole organism mechanisms for dispatching the
KB flow among organs while maintaining some
priority for the brain. Another potential advantage
of the provision of KB as a second form of fat-de-
rived fuel is that they may serve to decrease the
need for transport of FFA which are poorly soluble
in body fluids and may become toxic at high
concentration^.^ The advantage of water solubility
DiabetesNetabolism Reviews, Vol. 5, No. 3, 247-270 (1989)
0 1989 by John Wiley & Sons, Inc.
of KB is, however, counterbalanced by their much
greater acidity per carbon transported when com-
pared with long-chain fatty acids, and KB repre-
sent a potential threat for the acid-base equilib-
rium. The chemical and physiological buffering
mechanisms can maintain plasma pH in the nor-
mal range provided that the ketoacid concentra-
tion does not exceed 7-10 mM. It can thus be
anticipated that under physiological conditions,
appropriate regulatory processes will maintain ke-
tonemia below this level.
This review will deal with global aspects of the
regulation of KB metabolism in man under a
variety of conditions, including postabsorptive
state, fasting of variable duration, diabetic ketosis,
and physical exercise. We will attempt to quantify
KB production and uptake under these experimen-
tal conditions and determine to what extent over-
production and decreased utilization contribute to
the genesis of the ketotic states.
Basically, two different methods can be used
for this purpose. The hepatic catheterization tech-
nique measures the net production of the substrate
by the liver which is assumed to represent the only
significant site of release of ketones into the blood
stream. In human studies, the ketone balance is
generally measured across the splanchnic bed
rather than across the liver owing to the difficulty
of obtaining portal blood samples. Since the extra-
hepatic splanchnic tissues extract significant
amounts of ketone^,^*^ this method should under-
estimate hepatic production rate. On the other
hand, the isotopic technique measures the total
rate of appearance of ketones from all possible
sources into all body compartments, whether or
not the substrate is delivered into the circulation.
The contribution of our laboratory to the study
CCC 0742-4221/89/040247-24$04.00
248
Table I. Most Frequent Causes of Ketosis in Humans
Fasting
High fat, low carbohydrate diet
Strenuous exercise
Postexercise (Courtice-Douglas effect)
Diabetes
Alcoholic ketosis
Ketotic hypoglycemia in children
of the physiology and pathophysiology of ketones
is essentially based on a tracer methodology which
was developed about 20 years ago for animal
studies6,' and eventually adapted for use in human
experiments. We have generally employed the
primed constant infusion technique with either
14C-AcAcor 14C-~( -)-POHB as tracer. Although
less aggressive than the catheterization technique,
the isotope infusion method poses several meth-
odological problems which will be briefly dis-
cussed in the following section.
11. METHODOLOGICAL PROBLEMS
ASSOCIATED WITH THE USE OF
*'C-KETONE BODIES FOR MEASURING
KETONE BODY TURNOVER
There are essentially three points of method-
ological controversy which need to be briefly dis-
cussed.
A. Isotopic Disequilibrium Between
AcAc and POHB
The turnover of KB can be measured using
either 14C-AcAcor 14C-~( -)-POHB. Since the two
ketone bodies are interconvertible, the label al-
ways becomes distributed in the two ketones.
However, complete isotopic equilibrium between
AcAc and POHB is not necessarily reached even
after prolonged isotope infusion. As shown in
Figure 1, the ratio between the specific activities of
the two compounds depends on the label used and
on the prevailing KB concentration. Whatever the
tracer used, the greatest disequilibrium exists at
low KB levels, probably because the fractional
disposal rate of KB exceeds the rate of intercon-
version. When ketonemia exceeds 2 mM, the spe-
cific activities of the two ketones are identical if the
tracer used is 14C-POHB.It means that under these
conditions, the rate of interconversion of individ-
ual KB is much faster than their irreversible dispo-
sal and KB turnover can be calculated as that of a
single compound. On the other hand, when KB
KETONE BODY PRODUCTION AND DISPOSAL
are traced with I4C-AcAc,the two specific activities
never become equal even at high ketonemia.
Therefore l4C-POHB is certainly the tracer of
choice under most usual circumstances. When
disequilibrium exists, calculation of total KB turn-
over is generally performed using the specific
activity of total ketones: (14C-AcAc+ ''C-POHB)/
(AcAc + POHB). This mode of calculation has
been criticized on theoretical grounds,' but many
studies from different laboratories, including ours,
have validated this calculation procedure in hu-
mans and A dual tracer technique for
measuring individual KB kinetics and rates of
interconversion recently has been proposed. l3,I4
B. Measurements of KB Transport in the
Non-Steady State
When experimental manipulations provoke
rapid changes in KB concentration, it is obvious
that the changes in production and disposal of KB
do not occur at the same rate and should therefore
be calculated separately. In most studies per-
formed under non-steady state conditions, the
rates of appearance (Ra) and disposal ( R d ) of KB
have been calculated using the equations proposed
by Steele" which take into account an "opera-
tional" volume of distribution (V) in which the
changes in concentration and in specific activity
take place. The value of V to be used is inversely
related to the rate of change. To some extent, V can
be determined from experiments in which known
variations in the inflow rate of KB are induced by
infusions of exogenous ketones."-12 Values of V
derived from this technique or determined empiri-
cally16 generally range between 10 and 20% of
body weight. The equations of Steele imply that
the kinetics of KB can be reasonably described by a
o i i i l i b i
KETONEMIA (mM)
Figure 1. Specific activity ratio between p-hydroxybu-
tyrate and acetoacetate as a function of total ketone body
concentration in fasting subjects exposed to a constant
infusion of "C-p-hydroxybutyrate (solid circles) or I4C-
acetoacetate (open circles).
BALASSE AND FERY
single compartment model, an assumption that
has been criticized by some authors. However,
several studies have provided experimental val-
idation of the use of Steele’s equation in measur-
ing KB transport under non-steady state condi-
tions.10-12
C. Isotopic Artifacts in Measuring KB Turnover:
Possibility of ”Pseudoketogenesis”
Landau17 has recently drawn attention to a
possible artifact in measuring KB production using
l4C-labe1ledketone bodies. According to this pro-
cess, acetoacetyl-CoA formed from the last four
carbons of a long chain fatty acid in the course of
its undergoing @oxidation in peripheral tissues
could dilute the specific activity of plasma AcAc
via reversal of the 3-oxoacid-CoA transferase reac-
tion. There could also be loss of label by carbon
exchange between acetoacetyl-CoA and acetyl-
CoA in the reversible reaction catalyzed by
acetoacetyl-CoA thiolase. These two mechanisms
could participate in a process of ”pseudoke-
togenesis,” leading to an artifactual overestimation
of KB turnover.” The amplitude of this phenome-
non in humans in vivo is currently unknown.
However, Nosadini et al., measuring tracer dilu-
tion across human forearm, have observed a signi-
ficant “ketogenesis” by muscle without net pro-
duction of unlabelled ketone bodies. This “pseu-
doketogenesis” calculated for the entire muscle
mass represents 9% of the overall KB production
rate in normal subjects and 25% in diabetic pa-
tients.”
111. PRODUCTION AND DISPOSAL OF
KETONE BODIES IN NORMAL
FASTING SUBJECTS
In normal overnight-fasted adult subjects, the
concentration of total KB is very small, usually
Table 11. Concentration and Rates of Transport of Ketone Bodies in Humans
Overnight
Concentration (mM)
Acetoacetate
p-Hydroxybutyrate
Total ketones
Production rate of total ketones
mmol/min
g124 h
Urinary loss of total ketones
(percent of production)
Source: Representative data taken from Refs. 9 and 39.
249
around 0.1-0.4 mM (1-4 mg/dl), but ketogenesis is
far from negligible, amounting to about 0.2-0.4
mmol/min, which corresponds to 30-60 g/24 h
(Table 11). Data in the literature are in good agree-
ment with these figures.’0~11~”~21 As fasting pro-
ceeds (Figure 2), there is a rapid rise in the plasma
levels of FFA and KB and in KB production until a
new steady state is attained after about 5 days. At
that time, KB concentration and turnover rates
amount, respectively, to 7-10 mM (70-100 mg/dl)
and 1.5-2.5 mmol/min. Thus, when shifting from
the overnight-fasted to the fully fasted state, the
production rate of KB is multiplied by a factor of
4-5 and concentration by a factor of 40-50 (Table
11). This discrepancy suggests that peripheral up-
take is a saturable phenomenon, an important
observation that will be discussed later.
These data on production rates of KB during
prolonged fasting are higher than those reported
in more obese patients by another group using a
tracer methodology.22As mentioned before, our
values might be overestimated by interference of
“pseudoketogenesis,” but this artifact should
equally affect all tracer methods. In fact, total KB
production rates as high as 2.5 mmol/min are not
unrealistic, since the constant infusion of cold
AcAc or POHB at this rate in normal postabsorp-
tive subjects results in lower and not higher con-
centration than those observed spontaneously
during protracted starvation.’ Very high produc-
tion rates (up to 46 pmols/kg/min, i.e., -90 g/24 h)
have been observed recently at a ketonemia of
-3.5 mM in 5-year-old children fasted for only
20-22 h. l4
A. Mechanisms Involved in Ketone Homeostasis
Once full starvation ketosis has been attained
in a given subject, ketonemia remains remarkably
steady with time, individual levels varying usually
Diabetic
fast Starvation ketoacidosis
0.06 1.5 3
0.12 6.0 12
0.18 7.5 15
0.25 1-2 3
36 140-280 430
<1 2-5 10-15
250
N 3, KETONE BODY PRODUCTION
E
m o
f
P L A S M A KETONE BODY CONC.
0 8 * O
ONF
,
5
' I
Figure 2. Total ketone-body production and concentration in 41 subjects who fasted for
different periods of time. Solid circles: nonobese subjects; open circles: moderately obese
patients. Reproduced from Ref. 31 with permission.
between 7 and 10 mM. As discussed later, aggra-
vation by exercise of the negative caloric balance of
the fully fasted state is unable to accelerate ke-
togenesis.z Thus, very efficient mechanisms main-
tain fasting hyperketonemia below a critical level
of acidemia compatible with the maintenance of
blood pH within normal limits. The components of
this regulatory mechanism will be analyzed here-
after.
1 . Antilipolytic Effects of Ketone Bodies
It has been shown that KB inhibit adipose
tissue lipolysis in vitro2* and in The in
vivo effect that is observed during infusion of
ketones has the following characteristics: (1) it
persists as long as hyperketonemia is maintained
but is followed by a marked rebound in the FFA
levels above baseline values at cessation of the
infusion,26(2) it is obtained with both AcAc and
pOHB,25,26 (3) it is independent of changes in pH,
sodium AcAc and AcAc acid being equally po-
tent,26and (4)it is independent of insulin release
and has been observed in insulin-dependent dia-
betic patient^.^' The effect is potent enough to
inhibit, for example, the lipolytic response to exer-
cise. Indeed, as shown in Figure 3, the increase in
total KB concentration to levels exceeding 4 mM by
means of an AcAc infusion considerably reduces
the FFA and glycerol responses to work without
interfering with the hypoinsulinemic action of ex-
ercise.
KETONE BODY PRODUCTION AND DISPOSAL
0
0
0 0
g 0
-- 0 0
"
Ib I
15
- - - '
210 DAYS
2. lnsulinotropic Effects of Ketone Bodies
The first demonstration of a stimulatory effect
of KB on insulin secretion has been provided by
Madison et al. in dogs.28 This property of KB
subsequently has been confirmed in humans29and
has been documented in ~ i t r o . ~Two ' important
conditions have to be met in order to demonstrate
the insulinotropic effect of KB in humans. Firstly,
the concentration has to be raised at relatively high
levels exceeding 2 mM.29Secondly, KB level has to
be increased abruptly. During a constant infusion
of KB in a normal subject, insulinemia is increased
only during the initial 5-15 min and eventually
returns to its baseline value despite persistant or
even increasing hyperketonemia3'; in many infu-
sion studies, no effect of KB on insulin levels could
be d ~ c u m e n t e d , ~ , ' ' simply
, ~ ~ , ~ ~because of the lack
of early sampling. The C-peptide response might
be easier to detecP3 owing to the slower clearance
rate of this compound or because of insulin trap-
ping by the liver.29The evanescent character of the
insulin response to prolonged hyperketonemia is
probably related to the hypoglycemic action of
ketones which counteracts the action of KB. In-
deed, as shown in Figure 4,a sustained increase in
insulin levels can be observed during KB infusion
in normal dogs provided that glycemia is clamped
at its baseline value by an appropriate glucose
infusion.M Thus, according to a proposition ini-
tially made by Madison et a1.,28it is likely that
during fasting ketosis KB are instrumental in main-
taining sufficient insulin secretion to restrain lipol-
ysis and ketogenesis below certain limits.
BALASSE AND FERY 251

6-
5-
a$ 1 - Na - ACAC
P piiizii SALl NE

I
I
I
I
4- I I
I
I
3- II I
2- I I
1-
0- I
I I
I I
I I
I
I
I I
I I I I
I
I I
I !

I I

I ***t
I I
I I
I I

I
= //H+-===- - - - -xr+r
---
1
+ / I 1 1 I I I -d
-90 -30 0 30 60 90 120 -90 -30 0 30 60 90 120

T I ME (min.)

Figure 3. Inhibition of the lipolytic effect of exercise by infusion of ketone bodies in

normal overnight-fasted subjects. On the right, exercise (50% VOz max) was performed
under saline infusion (control experiments; n = 6 ) . On the left, the same exercise was
performed under constant infusion of sodium acetoacetate (1.9 & 0.1 mmol/min/l.73 m2;
n = 7). Asterisks refer to values significantly different from corresponding mean basal
level. Note that the rise in FFA and glycerol levels is considerably reduced during ketone
infusion without modification of the insulin response to work.

3. Antiketogenic Effects of Ketone Bodies percent inhibition in ketogenesis observed in these

In addition to their antilipolytic and insuli-
notropic effects, KB have been shown to increase
the s&tsitivityof adipose tissue to the antilipolytic
effect of insulin.35These three properties of KB
provide the basis for a regulatory feedback mecha-
nism whereby KB would tend to inhibit their own
production rate. The data shown in Figure 5 dem-
onstrate that this regulatory process is operative in
fasting humans. Indeed, the experimental eleva-
tions of KB concentration induced by AcAc infu-
sion in starved obese subjects reduces FFA concen-
tration and depresses ketogenesis. The average
experiments amounted to about 33%.32 These data
have been confirmed recently by Keller et al.36
B. Peculiarity of Ketone Body Kinetics and Its
Possible Physiological Significance
With regard to FFA, the main form of endoge-
nous fatty acid transport, it is generally agreed that
the plasma concentration is proportional to the
production rate, i.e., that the MCR remains fairly
constant within the range of concentrations en-
countered under physiological and pathological
condition^.^^
252 KETONE BODY PRODUCTION AND DISPOSAL

Na.Acetoacelato n :7 Na.Clcetoacetale
n =0

8
Blood
Ketone t1 I ACAC I ACAC [6

I c LOO

)rmoles/ml 0.2
0
Blood 0.12

Glucose
compensation I
mgIkg.min 20, ,
-30
, 1
0
, , ,
30
,

min
, ,
60
, ,
-30
,
0
,a
hx,:,I,z, 30
min
60

Figure 4. Effect of an experimental hyperketonemia induced by a sodium acetoacetate

infusion on the arterial blood or plasma concentration of glucose, glycerol, and FFA in
normal dogs. I N levels were measured both in arterial and portal venous plasma. On the
left, infusion of sodium acetoacetate alone; on the right, infusion of sodium acetoacetate
plus glucose in order to prevent hypoglycemia. Note that ketone bodies stimulate insulin
secretion persistently provided that hypoglycemia is prevented. Reproduced from Ref. 34
with permission.

Ketone bodies behave in a totally different
manner as shown in Figure 6, in which we col-
lected kinetic data from 26 normal or slightly obese
subjects submitted to a fast for periods ranging
from 15 hours to 20 days with KB concentration
varying from 0.2 to 11.7 mM. This figure also
contains data on patients with diabetic ketosis,
which will be discussed later.
Considering the relationship between KB pro-
duction rate and concentration (Figure 6A), it
appears that it follows an exponential-type pat-
tern. It means that the response of concentration to
a given increase in production is considerably
amplified as ketonemia rises. In the steady state,
overall disposal rate is equal to R,. Tissular (meta-
bolic) disposal rate represents the difference be-
BALASSE AND FERY 253

BLOOD
KETONES
p moles/ ml

KETONE
SPECIFIC
ACTIVITY
dpm/pmole

L2
INFLOW 4 1
RATE OF 3 -
TOTAL
KETONES 2 -
mmolcs/mi n
1 - E NOOG
0 -

PLASMA 1.4 6I -.
F FA
--.-.
pmoles /ml
1.2
1.0
] I
I
I
I
I
-.m-
PLASMA
GLUCOSE - -- - _ _
b---
mg / 100 ml *7 O0 1 --• 2a
I ~ ' 120
' . ~ 180
- ' ' 240
' ' ' ?
0 60
MINUTES

Figure 5. Sample experiment in a 14-day-fastedobese subject illustrating the inhibition

of endogenous ketone production by the infusion of exogenous ketones. Data from Ref.
32.

tween overall disposal rate and urinary loss, which
represents only a small fraction of Ra under physi-
ological conditions. According to Figure 6B, the
relationship between concentration and metabolic
disposal affects the shape of a saturation curve.
This indicates that tissues as a whole have a
limited capacity to take up ketones. Thus, the rise
in ketonemia associated with progressive fasting is
primarily due to increased ketogenesis, but the
phenomenon is amplified by a gradual reduction
in removal capacities.
According to Figure 6B, the maximal disposal
rate of KB by tissues is around 2.5 mmol/min.
Comparing panels A and B of Figure 6, it can be
considered schematically that during fasting R a
increases to a maximal value of -2.7 mmol/min in
slight excess of the maximal metabolic uptake, the
difference representing urinary excretion.
In Figure 6C we calculated the relationship
between MCR (Rdlconcentration) and concentra-
tion and observed that MCR decreases with in-
creasing concentration in a curvilinear manner. Up
to a concentration of 1-2 mM, the curve has a very
steep slope, a much slower decrement of MCR
being observed above 5 mM. This pattern is com-
patible with the concept that some tissues have a
254 KETONE BODY PRODUCTION AND DISPOSAL

KETONE BODY PRODUCTION(rnmo1s/rnin/1,73m2)

3
1

Figure 6. Relationship between production rate

and concentration (panel A), concentration and
disposal rate (panel B), and concentration and
metabolic clearance rate (panel C) of total ketone
bodies in normal fasted subjects (open symbols)
and in insulin-dependent diabetic patients (closed
symbols). Ketone bod kinetics was measured
h\
'J
using as infusion of C-acetoacetate (circles) or
0
m Z
- '4C-p-hydroxybutyrate (triangles). Note that the
E two tracers yield similar turnover values. Data
U E 0
2 -05. 0 from Ref. 39.
0
I-
Y
0

0 2 4 6 8 lo 12 14
PLASMA KETONE BODY CONC I m M )

O i i ' -
6 01 ' 10
PLASMA KETONE BODY CONC. ( m M )
12 ' 14
BALASSE AND FERY
high affinity for ketones at low concentration but
rapidly become saturated as concentration rises,
whereas other tissues have a capacity to clear
ketones that is more limited at low concentration
but also much less depressed by an increase in
ketonemia. Analysis of data from the literature on
extraction of KB by individual tissues support this
concept. Figure 7 represents the extraction ratio of
total KB (AcAc + POHB) for two important organs
consuming KB (brain and muscle). An extraction
ratio (arteriovenous difference/arterial concentra-
tion) has basically the same significance as a meta-
bolic clearance since it corresponds to a metabolic
clearance per unit blood flow. Both parameters
represent the capacity of the organ to remove
ketones from blood. Muscle has a high extraction
ratio (-50%) at low concentration, but it decreases
to less than 5% at high KB levels, thus indicating
that uptake of KB by muscle is a saturable process.
It is noteworthy that other tissues such as gut
behave in a similar mannerq7On the other hand,
the brain extracts about 10-15% of incoming KB
whatever the concentration. The curvilinear rela-
tionship between MCR and concentration (Figure
6C) observed in the whole organism is thus com-
patible with the idea that muscle represents the
major site of KB consumption at low ketonemia,
whereas at high ketonemia KB are preferably con-
0.3
11
BRAIN
0.2
0 rendered hyperketonemic by a constant infusion
2 0.6,
I
I:
0.5- MUSCLE
z Po
,=
0
V
0.4-
0.3-
I-
5 0.2.
0.1 -
OJ
o i i j i S 6 i i j
KETONEMIA (mM)
Figure 7. Extraction ratio of total ketone bodies by
brain and muscle as a function of ketonemia in normal
subjects undergoing a fast of variable duration. Com-
piled from data in Refs. 1, 41, 50-52, 58, 68-72.
255
sumed by the brain. Owen and Richard,38were the
first to propose the existence of such a redistri-
bution of KB uptake between muscle and brain
during progressive fasting, this phenomenon al-
lowing KB to serve as a cerebral fuel replacing
glucose during starvation.
The fact that at high ketone levels, when
overall Rd is close to saturation, a minute increase
in ketogenesis results in a marked rise in concen-
tration has important implications to the function-
ing of the mechanisms of ketone homeostasis
during progressive fasting. Indeed, the efficacy of
the negative feedback effect of KB on their own
production is progressively amplified as hyperke-
tonemia progresses until it reaches its full effec-
tiveness when tissular uptake is saturated.39
C. Role of Insulin on Peripheral Utilization of
Ketone Bodies In Vivo
As discussed above, KB concentration regu-
lates KB utilization according to a saturation-type
relationship. Whether peripheral uptake is also
modulated by hormonal factors such as insulin
should also be considered. Since the late 1 9 2 0 ~ , ~
numerous studies have attempted to elucidate the
role of insulin in regulating peripheral KB uptake
in vivo, but results have been di~cordant.~ The
difficulty in approaching this problem in vivo lies
in the fact that the main action of insulin on KB
metabolism is to depress ketogenesis and therefore
reduce ketonemia. This will necessarily enhance
the MCR of ketones (see before) even if insulin has
no direct action on disposal. One way of circum-
venting this problem is to analyze the effects of
insulin on KB uptake in postabsorptive subjects
of exogenous KB in amounts that largely exceed
endogenous production. Under these conditions,
we observed in dogs7 that the combined adminis-
tration of supraphysiological amounts of insulin
and glucose enhanced the uptake and oxidation of
KB, but it cannot be excluded in these experiments
that part of the apparent increase in uptake was
due to the inhibition by insulin of the small endog-
enous component of overall ketone turnover. A
more precise technique was recently developed by
Keller et al., who analyzed in humans the effect of
physiologic amonts of insulin on KB uptake during
a ketone-body clamp at 2 mM and showed a
significant, although discrete (25%), increase in
MCR.36
Obviously the depressing effect of fasting on
the MCR of KB and on their extraction by various
256
s\
w I
P= 2004
w pic methods, but this is not an absolute rule. As
* 01 I I I I I 1
0 42 6 8 1012
PLASMA KETONE BODY CONC ( m M )
Figure 8. Negative relationship between the metabolic
clearance rate and the concentration of total ketone
bodies in normal overnight-fasted subjects rendered
hyperketonemic with constant infusions of variable
amounts of sodium acetoacetate. Total ketone body
kinetics was evaluated either with I4C-acetoacetate
(closed circles) or ''C-/l-hydroxybutyrate (open circles).
Note the similarity with endogenous ketosis (Figure 6,
panel C).
tissues including muscle is much too large to be
explained solely by the decrease in insulin concen-
tration. Other factors such as the rise in FFA
(which might compete with ketones) could play a
contributory role. However, according to our stud-
ies the main factor seems to be the hyperketone-
mia itself. Indeed, according to Figure 8, a negative
curvilinear relationship between MCR and concen-
tration similar to that observed in fasting ketosis
can be documented in normal postabsorptive sub-
jects rendered artificially hyperketonemic by infu-
sion of KB in variable amounts, a condition which
is not associated with insulin lack or elevated FFA
levels. This observation should be kept in mind
before considering the existence of a KB removal
defect as a contributory factor in hyperketonemic
states such as starvation or diabetes.
IV. PRODUCTION AND DISPOSAL OF
KETONE BODIES IN DIABETIC KETOSIS
There is no doubt that diabetic hyperketone-
mia is primarily caused by an increased production
of KB. Several factors contribute to this overpro-
duction. The low levels of insulin and the excess in
antiinsulin hormones stimulate lipolysis and aug-
KETONE BODY PRODUCTION AND DISPOSAL
ment the hepatic FFA load. On the other hand,
these hormonal changes enhance the ketogenic
capacity of the liver, so that an increased propor-
tion of the hepatic FFA uptake is converted into
KB. It has been demonstrated by hepatic catheteri-
zation studies that in severe hyperketonemic states
(whether due to fasting or diabetes), the hepatic
conversion of FFA to ketones approximates 80-
90%.41-43
In the literature, estimates of total KB produc-
tion in diabetic ketosis vary considerably among
laboratories depending especially on the technique
used. In general, values obtained by hepatic cath-
eterization are lower than those provided by isoto-
shown in Figure 6A, we found that uncontrolled
diabetic patients with a ketonemia of 12-14 mM
produce about 3 mmol/min of total KB. There are
very few other data obtained at comparable KB
levels. For lower degrees of diabetic hyperketone-
mia (2.5-7 mM), the available data are lower,",45
similar,46or highe9l than ours. It is amazing to
note that there is considerable difference in results
obtained by two independent groups employing
sophisticated multicompartment analysis of data
collected after pulse injections of tracer^.^^,^^ At the
present time, there is no way to reconcile all these
heterogenous results, and every laboratory trusts
its own data.
An important area of discussion has been to
determine whether diabetic ketosis results from
overproduction alone or from a combination of
overproduction and underutilization. For instance,
Owen et al., using the hepatic vein catheterization
technique, have observed that ketotic diabetic pa-
tients produce no more ketone bodies than non-
diabetic subjects undergoing a 3-day fast.43Since
hyperketonemia of diabetics can greatly exceed
that of starvation, these authors concluded that a
removal defect was essential in causing diabetic
ketosis. Miles et al., studying the development of
ketosis at cessation of an i.v. insulin infusion in
type I diabetic patients, noticed that overpro-
duction of KB was the primary event in causing
hyperketonemia. However, the fall in the MCR of
KB observed in their patients during progression
of hyperketonemia suggested the participation of a
removal defect." Nosadini et aL21and Hall et al.,45
using pulse injections of tracers, came to a similar
conclusion but observed that the reduced clear-
ance concerned mainly POHB. Finally, Sherwin et
al.47infused a given amount of POHB to normal
and diabetic patients and observed that the rise in
KB level was twofold greater in type I diabetics,
BALASSE AND FGRY
suggesting a role for insulin in influencing ketone
disposal.
Thus most authors agree that diabetic hyper-
ketonemia is associated with an important reduc-
tion in KB clearance which participates in the
development of ketosis, and it is generally be-
lieved that this phenomenon is related to the
insulin deficiency characterizing the diabetic
~ t a t e .However,
~ , ~ ~ in most studies the data have
not been correctly interpreted because the main
characteristic of KB kinetics, that is, the marked
dependency of MCR on concentration, has been
disregarded. In our mind, the only way to search
for a removal defect in diabetes,is to compare the
MCR of KB in diabetics with that of normal con-
trols presenting a physiological ketosis of fasting,
the comparison being made at an identical KB
c~ncentration.~~ The results of such a comparison
are provided in Figure 6, where data on both types
of ketcais have been displayed. It appears that at
any given KB concentration, at least in the range
observed during starvation (up to 12 mM), the
kinetics of KB is identical in the two groups. Thus,
as for fasting ketosis, the ketosis of diabetes is
primarily caused by an increased production of
KB, but the phenomenon is amplified by a pro-
gressive limitation in the ability of tissues to re-
move ketones from blood as the concentration
rises. The inverse relationship between the meta-
bolic clearance and the plasma levels of ketones
which underlies this process represents a general
characteristic of ketones that applies to both types
of ketosis. A maximal metabolic disposal rate of
about 2.5 mmols/min is attained in both groups at
concentrations of 10-12 mM, which corresponds
to the highest KB levels encountered during pro-
longed fasting, and there is no evidence for a KB
removal defect specific to diabetes.
Therefore, it seems obvious that the much
higher KB levels that can be observed in decom-
pensated diabetic patients must result from higher
rates of ketogenesis. Unfortunately, there is very
little data on KB production in decompensated
diabetics with KB levels exceeding those observed
during long-term starvation. Values as high as 400
g/24 h have been r e p ~ r t e dAt. ~any
~ ~rate,
~ the fact
that diabetic patients with ketoacidosis have much
higher levels of FFA and glucagon than normal
subjects undergoing protracted starvation49sup-
ports the idea that these patients produce unphys-
iologically high amounts of ketones. Thus, the
regulatory feedback mechanisms that maintain,
during starvation, KB production very close to the
maximal disposal rate are not functioning properly
257
in type I diabetics probably because it implies an
intact p-cell function. It is not necessary to postu-
late that in diabetic ketoacidosis ketogenesis is in
great excess of that of prolonged fasting ketosis,
because, according to the exponential relationship
between production and concentration (Figure
6A), a small excess in production can account for a
large increase in concentration.
V. KETONE BODY METABOLISM DURING
MUSCULAR EXERCISE
When exploring the kinetics of a metabolic
fuel in vivo, it is always quite informative to
include studies on the effects of muscular exercise
on the turnover of this substrate. Indeed, the
abrupt increase in muscular energy demand asso-
ciated with work challenges a variety of regulatory
processes whose functions might appear much
more clearly than at rest. Surprisingly few studies
have been performed in this field with regard to
KB m e t a b o l i ~ m . ~ - ~ ~
A prominant feature of KB response to exer-
cise is its great dependence on the initial degree of
hyperket~nemia.’~ Therefore, the effect of exercise
will be analyzed for the whole range of ketonemia
observed during transition from the postabsorp-
tive to the fully fasted state. Information will also
be provided on KB transport during the postexer-
cise recovery period. Finally, the mechanisms re-
sponsible for the abnormal response encountered
in insulin-deprived diabetics w ill be explored. All
our exercise studies consisted of a walk on a
treadmill for 2 h at a moderate intensity corre-
sponding to 50% of the maximal aerobic capacity
(VO, ma)^.'^,^^
A. Exercise in Normal Subjects Submitted to a
Fast of Variable Duration
1 . The Effects of Exercise Depend on the Duration of
Exercise and Initial Degree of Fasting
Figure 9 shows the effects of exercise on KB
metabolism and other parameters in subjects
fasted for 18 h (group A), 2-3 days (group B), and
3-5 days (group C). In overnight-fasted subjects
(total KB concentration of -0.2 mM) there is a
progressive rise in R, which is approximately dou-
bled at the end of the exercise. This causes an
increase in concentration. The MCR is stimulated
by about 40%, indicating that contracting muscles
increase their capacity to extract ketones from
blood. The rise in disposal rate (Rd) results from
258 KETONE BODY PRODUCTION AND DISPOSAL

Figure 9. Rate of transport of total ketone bodies, plasma concentration of FFA and IRI,
and IN-to-IRG ratio at rest and during exercise (50% VOz max) in 3 groups of normal
subjects with an increasing degree of fasting hyperketonemia. Group A: n = 10, fast of 18
* *
-+ 2 h; group B: n = 5, fast of 62 10 h; group C: n = 6, fast of 112 13 h. Asterisks refer
to values significantly different from corresponding mean basal level ( p < 0.05 or less by
paired f-test). Reproduced from Ref. 23 with permission.
BALASSE AND FERY
the combined effect of a rise in concentration and
in MCR. The increase in ketogenesis is related to
the rise in FFA levels and to an increase in the
ketogenic capacity of the liver, which is stimulated
by the fall in insulin/glucagon ratio. The modifica-
tions observed in R,, Rd, and MCR are those
expected for a muscular fuel and are qualitatively
comparable to those observed for other muscular
fuels such as FFA%and glucose.55Needless to say,
the KB concentration is too small in ovemight-
fasted subjects to contribute significantly to mus-
cular energy needs.
A different response is observed if the same
exercise is performed in subject fasted for 3-5 days
with an initial total KB level of -5.5 mM (group C).
Under these conditions, ketogenesis is depressed
during early exercise and tends eventually to re-
turn to its baseline value. Ketonemia follows these
changes. Interestingly, the MCR remains un-
changed and R d decreases as a result of the fall in
concentration. The overall effect of exercise per-
formed in starved subjects is thus to reduce the
rate of production, the rate of disposal, and the
concentration of KB. The lipolytic effect of exercise
is, however, maintained. The decreases in insulin
concentration and in insulin/glucagon ratio ob-
served at exercise in the overnight fasting state are
abolished after starvation. In subjects fasted for
2-3 days (group B) with an average KB concentra-
tion of about 3 mM, the response of KB kinetics is
intermediate between those of the two extreme
groups. This relatively complex kinetic response to
exercise needs to be further analyzed and inter-
preted, especially for AR, (hepatic response) and
-il
Y
E
0 0:5 1.0 1.5
Figure 10. Relationship between the rate of production of total ketones (R.) and FFA
levels at rest (r = 0.75; p < 0.001) and during exercise ( r = 0.86; p < 0.001) in normal
subjects fasted for periods ranging from 16 hours to 5 days. Slopes of regression lines are
significantlydifferent ( p < 0.005) as tested by covariance analysis. Reproduced from Ref.
23 with permission.
259
AMCR (muscular response). Indeed, these are the
two basic independent parameters influenced by
exercise, from which depend entirely the changes
in concentration and in R d .
2. Significance of Changes in R,
In order to explain why the changes in Ra
depend both on the duration of work and on the
initial degree of hyperketonemia (Figure 9), it is
necessary to first identify the three main factors
which modulate ketogenesis at work.
Decrease in Splanchnic Blood Flow. Exercise is
known to be associated with an abrupt and persis-
tent reduction in splanchnic blood flow which
should approximate 50% under our experimental
condition^,^^ whatever the duration of fasting pre-
ceeding exercise.57 These circulatory changes
should lower hepatic FFA uptake and depress
ketogenesis. Our observations (Figure 10) that the
relationship between FFA concentration and ke-
togenesis is different at exercise and at rest is in
agreement with this concept.
Increase in FFA Levels. The progressive increase
in FFA levels occurring during exercise tends to
compensate the depressing effect of splanchnic
circulatory changes on FFA load and ketogenesis.
lncrease in the Ketogenic Capacity of the Liver.
Exercise is known to be associated with a stimula-
tion of the hepatic conversion of FFA to ketones,58
this effect being at least partly related to the fall in
EXERCISE
0 65 l:o 1.5 2.0 2.5
FFA(mM)
260
the insulin/glucagon ratio. It should be recalled,
however, that the hepatic fractional conversion of
FFA to ketones at rest increases with the initial
degree of hyperketonemia, this relationship ap-
plying to both fasting and diabetic ketosis (Figure
11).It amounts to about 30% at low ketonemia and
80 to 90% when ketonemia exceeds 4 mM.
For each group studied (Figure 9), the time
sequence of the exercise-induced changes in R, can
be explained by the interaction between the three
factors modulating ketogenesis at work. In the
most ketotic group (C), insulin/glucagon ratio is
unaffected by work, and the fractional conversion
of FFA to ketones by the liver is probably barely
stimulated by exercise as it should already be near
maximum at rest. Under these conditions, ke-
togenesis is influenced almost exclusively by the
hepatic FFA uptake. During the early phase of
work, when FFA levels are not yet elevated, the
FFA uptake is likely to be depressed by the reduc-
tion in hepatic blood flow, accounting for the
decrease in Ra. With progression of work, FFA
concentration increases and is almost doubled by
the end of exercise. This increase should approxi-
mately counterbalance the negative effect of the
reduced hepatic blood flow on FFA uptake. Thus,
after its initial decrease, R, will increase parallel to
FFA levels and tend to return to its preexercise
value. In group A, the time course and amplitude
of changes in FFA load should be comparable to
that of group C, but in this case, after a lag period
of about 10 min, ketogenesis is stimulated by an
increase in fractional conversion of FFA to ketones
loo 1
OJ
I , , 1
0 1 2 3
Figure 11. Hepatic conversion of FFA into ketone bodies as a function of ketonemia in
fasting (circles) and diabetic (triangles) ketosis. Compiled from data in Refs. 41-43, 52,
73-76.
KETONE BODY PRODUCTION AND DISPOSAL
due at least partly to the fall in insulin/glucagon
ratio. As expected, the pattern observed in group B
is intermediate to that observed in the two ex-
tremes.
According to Figure 9, it could be assumed
that the stimulatory effect of exercise on R, is
declining continuously during the transition from
the postabsorptive to the fully fasted state. In fact,
this is not the case as shown on Figure 12, which
represents all individual responses at a given time
period of exercise (90-120 min) as a function of
basal ketonemia. This analysis reveals that the
relationship between AR, and basal ketonemia is
discontinuous: AR, increases with concentration
below 0.6 mM, this relationship being reversed
above 2.5 mM. As expected, a similar pattern is
observed for AKB, which depends mainly on AR,,
and for ARd, which is mainly influenced by con-
centration. The change in FFA is positively cor-
related to initial ketonemia (and to AR,) in the low
KB range (<1mM), but beyond this concentration
AFFA is independent of ketonemia and averages 1
mM. The change in insulin and A [insulin/glu-
cagon] are negatively correlated with basal ke-
tonemia. The discontinuity observed in the ke-
togenic response as initial ketonemia rises can be
explained as follows: in the low range of KB levels,
there is a positive correlation between AR, and
AFFA which suggests that the degree of stimula-
tion of ketogenesis is mainly determined by the
intensity of the lipolytic effect of exercise. On the
other hand, when ketonemia exceeds 2.5 mM,
AFFA is independent of the initial degree of ketosis
4 5 6 7 8 9 1 0 1 1 1 2
KETONEMIA (mM)
BALASSE AND FERY 261

The maximal stimulation of Rd at work (+0.8

mmol/min/l.73 m2) is attained when basal ke-
tonemia approximates 0.6 mM (Figure 12), a value

/,-.,
J

t 0.8
usually observed after 24-36 hours of fasting.
A Ra
(mmols/ toil Assuming that the totality of the extra amount of
ketones taken up by tissues at work is selectively
min / l.73m2 )
0 ............. !....... 3.c. utilized by working muscles, it can be calculated
- 0.4 that its oxidation accounts for no more than 7% of
the increase in muscular oxygen consumption.
A Rd >, This is a maximal figure because part of ARd should
be consumed by tissues other than muscle in
....?............\.-:--- . . .. -.
proportion to the increase in ketonemia. It is thus
-0.4 J clear that KB do not significantly contribute to
meet the extra energy needs of working muscles
A FFA t 1.5, during fasting, whatever the degree of hyperke-
(mM)
tonemia.
..........................

t 0.6
----. ....*.:* 3. Significance of Changes in MCR
A G L Y C E R O L t0.4
(mM) +0.2$ f -
The changes in MCR represent an index of the
ability of exercise to stimulate the capacity of
working muscles to extract ketones from blood. As
shown in Figure 9, the stimulatory effect of exer-
cise on this parameter is present at the low ketone
body levels characterizing overnight-fasted sub-
jects. Interestingly, the effect is maximal at the
beginning of exercise and becomes less marked as
work progresses. This is probably related to the
--108j . I
rise in KB concentration associated with work
0 1 2 3 4 5 6 7 since, as discussed before, the MCR of KB is very
B A S A L KETONE5 ( m M ) much dependent on the concentration, especially
at low degrees of ketosis (Figure 6C). When exer-
Figure 12. Influence of basal ketone concentration on cise is performed after more prolonged fasting,
metabolic and hormonal response to a 2-h execise in 21 MCR is less and less stimulated by work as ke-
normal subjects fasted for periods ranging from 16 hours
to 5 days. In each subject, A represents the difference tonemia rises until the effect completely disap-
between mean value recorded during the last 30 min of pears after several days of fasting. Since the uptake
work (t = 90-120 min) and mean baseline level. Studies of KB by muscle is a saturable process, these data
were performed with a primed constant infusion of suggest that muscle becomes unable to stimulate
14C-p-hydroxybutyrate.Reproduced from Ref. 23 with its capacity to utilize KB at work when saturation is
permission.
already attained at rest.

and the gradual reduction in the ketogenic re-

sponse to work probably reflects the progressive
decrease in the ability of exercise to enhance the
fractional conversion of FFA to KB. The finding
that the A concentration and ARd curves parallel
that of AR, (Figure 12) can be explained simply by
the fact that ketogenesis is the main determinant of
ketonemia, which in turn modulates Xd. It should
be noted, however, that at low KB levels the
stimulatory effect of exercise on MCR attenuates
the influence of AR, on concentration and ampli-
fies the impact of the change in concentration on
Rd*
4. Balance Between Hepatic and Muscular Effects of
Exercise on Ketone Body Metabolism
On the whole, our data suggest that the effect
of exercise on hepatic production of ketones (AR,),
on one hand, and on their muscular extraction
(AMCR), on the other hand, behave as two well-
coordinated processes. Indeed, both parameters
decrease progressively as ketonemia rises above 1
mM (Figures 9 and 12), and, consequently, AR, is
positively correlated with AMCR (Figure 13). This
parallelism between hepatic and muscular re-
sponses can be viewed as having homeostatic
262
601
,Oi
I **I
-2oJ
I
1 I
-40 0 40 80 120 160 200
A Ra(%)
Figure 13. Correlation between AMCR and AR, in-
duced by a 2-h exercise. Panel A: mean changes re-
corded during last 30 min of work (T = 0.70; p < 0.001).
Panel B: integrated changes over entire exercise period (Y
= 0.77; p < 0.001). All data are expressed as percent of
mean preexercise value. Reproduced from Ref. 23 with
permission.
properties, since it tends to minimize changes in
ketonemia at work. We would like to hypothesize
that KB themselves might be responsible for this
coordinated response. Indeed, owing to their insu-
linotropic and antilipolytic action mentioned be-
fore, KB might play an important role both in the
suppression of the insulin lowering effect of exer-
cise and in the limitation of the lipolytic response
to work observed at high ketonemia (Figure 12),
and KB levels might thus be instrumental in re-
straining the exercise-induced stimulation of ke-
togenesis. Such a negative feedback control has
been shown to be operative at rest.32On the other
hand, at increasing KB levels, muscular uptake of
ketones could be self-limited because in vitro ex-
p e r i m e n t ~have
~ ~ shown that, in muscular tissue,
high acetoacetate levels inhibit the activity of succi-
nyl-CoA transferase, the first enzyme involved in
KB utilization.
In conclusion, considering both the hepatic
and the muscular metabolism of KB at exercise, it
is clear that no single metabolic response applies to
the entire range of KB concentrations observed
KETONE BODY PRODUCTION AND DISPOSAL
during the transition from the postabsorptive to
the fully fasted state. With regard to ketogenesis,
there exists during the progression of fasting an
initial phase during which the increased energy
expenditure associated with exercise leads to a
process of ”accelerated starvation’’ and a second
phase characterized by the predominance of ho-
meostatic mechanisms that tend to maintain ke-
togenesis below certain limits. Indeed, exercise
appears unable to stimulate ketogenesis above the
maximal values attained during prolonged fasting
at rest. This is in sharp contrast with the response
of the two other substrates, FFA and glucose, for
which the highest rates of production observed in
circumstances of either health or disease are those
encountered with intensive exercise.%,@With re-
gard to the muscular behavior, the progressive
disappearance of the stimulatory effect of exercise
on the MCR of KB allows for the preservation, at
high ketonemia, of the preferential utilization of
KB by nonmuscular tissues, especially the brain.
B. Postexercise Ketosis
Postexercise ketosis refers to the abrupt in-
crease in KB levels occurring at cessation of a
prolonged exercise. It was first described in 1936
by Courtice and Douglas,61who noticed that “ke-
tosis never develops during an uninterrupted 10 or
even 12 miles walk but appears very soon after
resuming rest.”
In order to analyze the mechanisms responsi-
ble for the postexercise hyperketonemia, we deter-
w e d KB turnover during a 2-hour exercise at 50%
V02 max and during 60 min following the discon-
tinuation of work.
As shown in Figure 14,cessation of exercise is
associated with a marked increase in total KB
levels. A maximal increase in total KB concentra-
tion of 0.7 m M was observed after 30 min of
recovery, this rise representing about three times
that observed during the 2-hours of work. Tran-
sient rises in FFA (but not glycerol) and insulin
concentrations are observed during the early post-
exercise period. The rise in FFA has been attrib-
uted to an abrupt increase in the outflow of FFA
from adipose tissue owing to the release of the
sympathetic vasoconstrictor tone.62In this hypoth-
esis, part of the FFA formed during exercise would
have been trapped in the adipose tissue as a
consequence of vasoconstriction. The rise in insu-
lin might be due to a similar circulatory phe-
nomenon in the pancreas. It should be noted in-
cidentally that postexercise ketosis is the only
BALASSE AND FBRY 263

I I
I I k0.8 Ra
0.6(mmolslm in)
0.4
0.2
I
Rd 0.81
(mmolslmin) 0.6
0.4
0.2

FFA
(mM)

I* I
IRI

0 60 120 180
TIME ( m i n )
Figure 14. Mechanism of postexercise hyperketonemia in seven normal overnight-
fasted subjects. The rise in ketonemia at cessation of work results from both an increase in
production and a decrease in the metabolic clearance rate of ketones. Postexercise values
have been compared with last exercise value by a paired t-test. Asterisks refer to p < 0.05
or less. Studies were performed with a primed constant infusion of ''C-/l-hydroxybuty-
rate.

physiologic ketosis associated with elevated insu-
lin levels.
The KB turnover changes observed (Figure 14)
indicate that the rise in KB concentration occurring
during the early recovery phase was the result of a
rise in R, above its last exercise value lasting for
about 30 min with no equivalent rise in Rd. As a
matter of fact, a fall in Rd contributing to the
initiation of the postexercise hyperketonemia was
observed during the first 10 min of the recovery
phase. The relative steadiness in Rd prevailing
during the postexercise period in the face of a
marked elevation in ketonemia was the result of a
40-50% decrease in MCR accompanying cessation
of exercise. Thus, both an increase in production
and a decrease in removal contribute to postexer-
cise ketosis.
The postexercise hyperketonemia shown in
Figure 14 is of relatively short duration and of
moderate amplitude. It is known that the impor-
tance of the postexercise ketosis is dependent on a
variety of factors, among which the duration and
intensity of exercise, the composition of preceding
diet, and the degree of physical training play
important roles.61,63The influence of intensity on
the evolution of several plasma substrates includ-
ing KB is shown in Figure 15 for 3 overnight-fasted
subjects exercised for 2 hours at increasing inten-
264
I aggravates hyperglycemia and hyperketonemia in
3 l . poorly controlled diabetic patients." It is usually
1:
01 I I hypothesis.
I I
I I
I I
t , , r ~ l 3 l
0 60 120 180 240 300
MINUTES
Figure 15. Changes in the plasma concentration of
total ketone bodies and other substrates observed dur-
ing and after a 2-h exercise in 3 normal overnight-fasted
subjects exercised respectively at 40, 50 and 60% of V 0 2
max.
sity. The 2-hours exercise at the highest intensity
(60% VO2 max) is associated with a marked in-
crease in ketonemia during exercise (+1 mM)
followed by a marked postexercise ketosis lasting
for more than 3 h. In this overnight-fasted subject,
a ketonemia of 4 mM (usually observed after a
total fast of 3-4 days) is achieved in only 5 hours.
Hypoglycemic levels (3 mM) are reached at the end
of exercise and during postexercise. The hypogly-
cemic episodes associated with intensive and pro-
longed exercise are usually asymptomatic probably
because they are associated with high ketone body
levels which ensure an adequate supply of energy
to the brain.
A physiological role for postexercise hyperke-
tonemia might be to favor the replenishment of
muscle glycogen stores, a phenomenon that occurs
after exercise even in the absence of glucose feed-
ingMThe demonstration on rat muscle in vitro
that ketones inhibit glycolysis and increase the
KETONE BODY PRODUCTION AND DISPOSAL
conversion of glucose to glycogen65supports this
hypothesis.
C. Exercise in Insulin-deprived Type I
Diabetic Patients
It is well established that muscular exercise
assumed that this hyperketonemic effect is related
to an exaggerated stimulation of k e t o g e n e ~ i s , ~ ~ * ~ ' * ~ ~
but there is little experimental data to support this
In order to analyze this question (Figures 16
and 17), we measured the rate of transport of total
KB in 6 type I diabetic patients withdrawn from
i.v. insulin 4-6 hours before the study. Their
preexercise total KB concentrations varied between
3.15 and 6.33 mM and glucose levels range be-
tween 14.4 and 20.8 mM. The KB turnover studies
were performed during a preexercise period of 30
min followed by a 2-hour exercise at about 50%
VOn max. Two patients could not exercise longer
I l ~
than 90 min. The diabetics were compared with a
~
control group consisting of six normal subjects
presenting a similar range of physiologic ketosis
induced by prior fasting. The changes in substrate
and hormone concentration associated with exer-
cise in the two groups are compared in Figure 16.
Signficantly greater increases in total KB levels and
glucose concentrations were observed in the dia-
betics from the 30th minute of exercise onward,
thus confirming the deleterious effect of exercise
on diabetic control. On the other hand, there were
no significant differences among groups for any of
the other metabolic or hormonal parameters
tested. In particular, lipolysis was equally stimu-
lated in both groups as shown by FFA and glycerol
changes. As shown in Figure 17, the time course of
the changes in R, in diabetic patients and control
subjects is characterized by an initial fall lasting for
about 20 min followed by a secondary rise. The
mechanisms of this biphasic pattern have been
discussed earlier for the control subjects, and they
also apply to the diabetic patients. Although this
biphasic evolution of R, is observed at all degrees
of hyperketonemia examined in this study, it
should be noted that the higher the initial hyperke-
tonemia, the less R, is stimulated in late exercise,
so that the changes in R, integrated over the
working period are negatively correlated with
basal ketonemia (Figure 18). As expected, this
relationship is very similar to that shown in
Figure 12 for the subjects in whom ketonemia
BALASSE AND FERY 265

CONTROLS(n-6)
0.
Fl
&
7.
I I
6. I I * ?

*
KETONES 5.
(mM) -
.. -.
.. .I
....-. .*-- -*.
.
L.

....- - - t--------
..-.-.
I I
I I I I
I I
I I*

y I *---*&
.l
- __ -..-.-. I
....t - - - - - - - -
I I
I

I I
I I
I I
I I

I I
I
I
I
I
-*.-.+----.
I
I I
I
I I
I I
I
I I

, , I
-30 0 60 ’ 120 ’ 180 -30 0 60 120 180
MINUTES

Figure 16. Substrate and hormone concentrations during and after exercise in six ketotic
type 1 diabetic patients and six fasted normal subjects presenting a similar degree of
ketosis. Asterisks indicate statistically significant changes from mean basal value ( p <
0.05). Reproduced from Ref. 77 with permission.
266 KETONE BODY PRODUCTION AND DISPOSAL

Diabetic
patients
Control
subjects c - +‘ij A
t10
1-I I
pz=I
I I
- -
0 \ I

- to2 I I I I
a %- 1.5 -1 a

r n d
azo

- -02 6
I* I

.-c E
0
0-
t2
+I
j ””.;....\
.........
\
d

- -02
I;:
E E
l I -
E
I-a\
- 3 0 0 3 0 60 90 - 3 0 0 3 0 60 90 a
TIME (minl a -4
-5 J
Figure 17. Changes from mean basal value induced by I I I I I I I I

exercise in the concentration and turnover rate of total

ketone bodies induced by exercise in six ketotic type I
diabetic patients and six normal subjects pre-
senting a similar degree of ketosis (same patients as in
Figure 16). Asterisks refer to significant changes from t4.
mean basal value (p < 0.005). Reproduced from Ref. 77 .E t 3 .
with permission. E
0 +2.
QI
\
d
+I-
exceeds 0.6 mM. According to Figure 18, the ke- EE - oI j -
tone body concentration at which the overall stim- -
ulatory effect of a 90-min exercise on ketogenesis
is reversed to an inhibitory action approximates a
4 mM in diabetic patients as well as in normal -4 ‘
-5
fasted subjects. Despite the fact that the neg-
ative relationship between the exercise-induced
j;:l, , , , , l ,
0 1 2 3 4 5 6 1
changes in R, and basal ketonemia are similar
B a s a l K e t onaemia I mmol /I)
in normal subjects and diabetic patients (Fig-
ure 18), the response of concentration to exercise Figure 18. Changes in concentrations (A) and in rates
differs between groups. In the control subjects, the of appearance (B) and disposal (C) of ketone bodies
change in concentration is negatively correlated to recorded at the 90th minute of exercise as a function of
basal ketonemia, which suggests that in this group basal ketonemia. Solid circles and solid lines: normal
control subjects with fasting-induced hyperketonemia.
the changes in concentration are mainly influenced Open circles and dotted lines: insulin-deprived type I
by the changes in R,. On the contrary, in diabetic diabetic patients. All correlation coefficients are signifi-
patients, the change in concentration tends to be cantly different from zero (p < 0.05 or less) except for
positively correlated to basal ketonemia, suggest- diabetic patients in panel A. Covariance analysis indi-
ing that the hyperketonemic action of exercise in cates that the slopes and elevation of the regression lines
are not significantly different for the two groups (p >
uncontrolled diabetic patients cannot be accounted 0.05) except for panel A, where slopes are significantly
for by an exaggerated stimulation of ketogenesis. different (p < 0.01). Reproduced from Ref. 77 with
The hyperketonemic effect of exercise in ketotic permission.
BALASSE AND FERY
diabetic patients must therefore originate in some
defect in the removal mechanisms for ketones
causing in diabetic patients a greater imbalance
between production and uptake. This is not incon-
sistent with the observation that Rd is depressed at
high ketosis in diabetic patients as it is in normal
subjects. Simply, the mechanism responsible for
this inhibition is different in the two groups: in
fasted subjects-the fall in Rd can be accounted for
by the fall in concentration; in diabetic subjects Rd
decreases despite a rise in concentration which
implies a reduction in the ability of tissues to take
up ketones.
Studying the same problem with an hepatic
catheterization technique, Wahren et al. came to
the conclusion that diabetics exhibit an exagger-
ated increase in ketone body production at work
as compared with normal overnight-fasted sub-
j e c t ~ . ~In
’ , ~our opinion, this comparison is mean-
ingless because the two groups exhibit very dif-
ferent basal ketone body levels. Indeed, as dis-
cussed before, the pattern of ketone body response
to work is highly dependent on initial ketonemia.
Thus, when insulin-deprived, insulin-depen-
dent diabetic patients with a variable degree of
hyperketonemia are submitted to moderate exer-
cise of 120 min duration, their ketone body metab-
olism is modified in a manner very similar to that
of control fasted subjects matched for ketonemia.
In both groups, at low ketonemia, exercise stimu-
lates the rate of production and the rate of disposal
of ketones. These effects are progressively atten-
uated as basal ketonemia rises and are even re-
versed in markedly ketotic patients. However,
ketotic diabetic patients exhibit an abnormal re-
sponse in increasing their ketonemia above basal
level during the second hour of exercise, a phe-
nomenon not observed in the control subjects.
Contrary to the prevailing ~ p i n i o n , ~ ‘ , this
~ , ~ ’does
not seem to result from an exaggerated increase in
ketogenesis, but from a slight removal defect pos-
sibly related to insulinopenia. It should be empha-
sized that this conclusion applies to the specific
experimental conditions used in this study regard-
ing the duration and intensity of exercise.
VI. SUMMARY
Turnover studies performed during progres-
sive fasting in normal subjects indicate that the
production rate and the concentration of KB rise
markedly during the early phase of fasting and
start reaching a plateau after about 5 days. In
addition to increased production, a reduction in
267
the metabolic clearance rate of KB contributes to
the hyperketonemia. This reduced metabolic clear-
ance rate reflects essentially the progressive satu-
ration of muscular ketone uptake that occurs with
increasing ketonemia. The hormonal and meta-
bolic environment of fasting plays only a minor
role in this process, since a fall in KB metabolic
clearance similar to that observed during fasting is
observed if hyperketonemia is artificially induced
in the postabsorptive state by the infusion of
exogenous ketones. As extraction of KB by muscle
becomes limited during ongoing fasting, KB are
preferentially taken up by the brain to serve as a
substrate replacing glucose.
The remarkable stability of ketonemia during
prolonged fasting is maintained through the oper-
ation of a negative feedback mechanism whereby
KB tend to restrain their own production rate. The
antilipolytic and insulinotropic effects of KB are
instrumental in this process. This homeostatic
mechanism maintains ketogenesis only slightly
above the maximal metabolic disposal rate, the
difference corresponding to urinary excretion,
which is always below 10% of total turnover under
physiologic conditions.
When type I insulin-deprived diabetic patients
are compared at the same KB concentration with
control subjects with fasting ketosis, the character-
istics of KB kinetics are comparable in the two
groups. The maximal KB removal capacity is iden-
tical in the two situations, and it is not possible to
identify a ketone removal defect specific to dia-
betes. Thus, these data favor the concept that
excessive production of KB represent the main
factor leading to uncontrolled hyperketonemia. It
should be realized that a production exceeding
only slightly that prevailing during prolonged fast-
ing is sufficient to cause a progressive build-up in
concentration, leading to uncontrolled diabetic ke-
tosis.
In the overnight-fasted state, a prolonged ex-
ercise (2 h) performed at moderate intensity (50%
VO’ max) stimulates the capacity of muscle to
extract ketones from blood as evidenced by a
stimulation of the metabolic clearance rate. Simul-
taneously, KB production is accelerated and these
two processes contribute to increase muscle
energy supply, but the effect is quantitatively
negligible. Qualitatively, however, this response is
similar to that observed for glucose and FFA, the
two major blood-borne muscular fuels.
On the other hand, when the same exercise is
performed at high KB levels after prolonged fast-
ing, the muscular and hepatic responses are abol-
268
ished. This has useful consequences for bodies'
economy in that KB are conserved for utilization
by the brain and ketogenesis is maintained within
the range of values observed during prolonged
fasting at rest, despite the intense catabolic envi-
ronment characteristic of exercise. Thus, in no
circumstances do KB represent a significant fuel
for muscular work.
The above description is a schematic outline of
the effects observed at termination of a 2-hour
exercise period under two specific experimental
conditions. In fact, ketogenesis during exercise is
influenced by various factors acting in opposite
directions. Indeed, on one hand, exercise reduces
markedly the splanchnic blood flow which tends
to lower the hepatic FFA load and restrain ke-
togenesis; on the other hand, exercise stimulates
lipolysis and reduces the insulin/glucagon ratio,
two processes that should enhance ketogenesis.
These ketogenic and antiketogenic factors interact
in a complex manner depending on the duration of
exercise and initial degree of fasting so that the
overall picture of the exercise-induced changes in
KB transport is necessarily complex.
Turnover measurements have also been ap-
plied to the analysis of postexercise ketosis. It
could be demonstrated that this phenomenon is
caused by the combination of an increased produc-
tion and a marked decrease in the metabolic
clearance rate of KB.
Finally, we have attempted to identify the
mechanisms responsible for the aggravation by
exercise of the hyperketonemia of insulin-deprived
type I diabetic patients. In general, the changes in
KB production observed during exercise in diabetic
patients are similar to those observed in normal
fasted control subjects matched for ketonemia,
but, different from the results for fasted controls,
ketotic diabetics increased their ketonemia above
preexercise levels during the last 60 min of a
2-hour exercise. Contrary to the prevailing opin-
ion, this phenomenon is not related to an exagger-
ated stimulation of ketogenesis but rather to a
slight removal defect possibly related to the
marked insulinopenia.
Acknowledgments
This work was supported by the Fonds de la Re-
cherche Scientifique Medicale (grant No. 3.4546.88) and
by the Fondation Erasme. The authors are indebted to
Mrs. C. Demesmaeker for excellent secretarial assis-
tance.
KETONE BODY PRODUCTION AND DISPOSAL
References
1. Owen OE, Morgan AP, Kemp HG, Sullivan JM,
Herrera MG, and Cahill GF Jr: Brain metabolism
during fasting. 1 Clin Invest 46:1589-1595, 1967.
2. Robinson AM, and Williamson DH: Physiological
role of ketone bodies as substrates and signals in
mammalian tissues. Physiol Rev 60:143-147, 1980.
3. Havel RJ: Caloric homeostasis and disorders of fuel
transport. N Engl J Med 2871186-1192, 1972.
4. Balasse EO, Bier DM, and Havel RJ: Early effects of
anti-insulin serum on hepatic metabolism of plasma
free fatty acids in dogs. Diabetes 21280-288, 1972.
5. Windmueller HG, and Spaeth AE: Identification of
ketone bodies and glutamine as the major respira-
tory fuels in vivo for postabsorptive rat small intes-
tine. 1 Biol Chem 253:69-76, 1978.
6. Mayes PA, and Felts JM: Determination of 14C-
labelled ketone bodies by liquid-scintillation count-
ing. Biochem 1 102230-235, 1967.
7. Balasse EO, and Havel RJ: Evidence for an effect of
insulin on the peripheral utilization of ketone bodies
in dogs. J Clin Invest 50:801-813, 1971.
8. Barton RN: Isotopic study on ketone body kinetics:
Invalidity of calculations based upon specific radio-
activity of total ketone bodies. Metabolism 29:392-
394, 1980.
9. Balasse EO: Kinetics of ketone body metabolism in
fasting humans. Metabolism 28:41-50, 1979.
10. Beylot M, Beaufrere B, Normand S, Riou JP, Cohen
R, and Mornex R Determination of human ketone
body kinetics using stable-isotope labelled tracers.
Diabetologia 29:90-96, 1986.
11. Keller U, Sonnenberg GE, and Stauffacher W: Vali-
dation of a tracer technique to determine non-steady
state ketone body turnover rates in man. Am J
Physiol240:E253-E262, 1981.
12. Miles JM, Haymond MW, Rizza RA, and Gerich JE:
Determination of 14C radioactivity in ketone bodies:
A new, simplified method and its validation. Lipid
Res 21:646-650, 1980.
13. Miles JM, Schwenk WF, McClean KL, and Hay-
mond MW: A dual isotope technique for determina-
tion of in vivo ketone body kinetics. Am 1 Physiol
251:E185-E191, 1986.
14. Bougn6res P, and Ferre P: Study on ketone body
kinetics in children by a combined perfusion of
and 'H3 tracers. Am J Physiol253:E496-E502, 1987.
15. Steele R: Influences of glucose loading and of in-
jected insulin on hepatic glucose output. Ann NY
Acad Sci 82:420-430, 1959.
16. Balasse EO, F6ry F, and Neef MA: Changes induced
by exercise in rates of turnover and oxidation of
ketone bodies in fasting man. J Appl Physiol44:5-11,
1978.
17. Landau BR: A potential pitfall in the commonly used
isotopic techniques for measurement of ketone pro-
duction rates. In Diabetes 2985. Serrano-Rios M and
Lef6bvre PJ, Eds. Elsevier Science Publishers, Am-
sterdam, 1986, pp 340-342.
18. Des Rosiers C, Fink G, Dalose T, David F, Garneau
M, Montgomery JA, Mamer OA, Dalose P, Landau
BR, and Brunengraber H: Evidence for pseudoke-
togenesis in vivo. ASM Conference on experimental
BALASSE AND FERY
and theoretical analysis of metabolic processes, Big-
fork, Montana, 1987.
19. Nosadini R, Avogaro A, Sacca L, Vigorito C, de
Kreutzenberg S, Cobelli C, Toffolo G, Trevisan R,
Tessari P, Tiengo A, and Crepaldi G: Ketone body
metabolism in normal and diabetic human skeletal
muscle. Am J Physiol 249:E131-E136, 1985.
20. Miles JM, Haymond MW, Nissen SL, and Gerich JE:
Effects of free fatty acid availability, glucagon excess
and insulin deficiency on ketone body production in
postabsorptive man. J Clin Invest 71:1554-1561,1983.
21. Nosadini A, Avogaro A, Trevisan R, Duner E,
Marescotti C, Ion E, Cobelli C, and Toffolo G:
Acetoacetate and 3-hydroxybutyrate kinetics in
obese and insulin-dependent diabetic humans. Am J
Physiol248:R611-R620, 1985.
22. Reichard GA Jr, Owen OE, Haff AC, Paul P, and
Bortz WM: Ketone body production and oxidation in
fasting obese humans. J Clin Invest 53:508-515,1974.
23. F6ry F, and Balasse EO: Response of ketone body
metabolism to exercise during transition from post-
absorptive to fasted state. Am J Physiol 250:E495-
E501, 1986.
24. Bjorntorp P, and Schersten T: Effect of p-hydroxy-
butyrate on lipid mobilization. Am ] Physiol212:683-
687, 1967.
25. Balasse E, and Ooms HA: Changes in the concentra-
tion of glucose, free fatty acids, insulin and ketone
bodies in the blood during sodium p-hydroxybu-
tyrate infusions in man. Diabetologia 4:133-135,1968.
26. F6ry F, and Balasse EO: Differential effects of so-
dium acetoacetate and acetoacetic acid infusions on
alanine and glutamine metabolism in man. J Clin
Invest 66:323-331, 1980.
27. Binkiewicz A, Sodeghi-Nejad A, Hochman H, Lori-
dan L, and Senior B: An effect of ketones on the
concentration of glucose and of free fatty acids in
man, independent of the release of insulin. J Pediatr
84:226-231, 1974.
28. Madison LL, Mebane D, Unger RH, and Lochner A:
The hypoglycemic action of ketones. 11. Evidence for
a stimulatory feedback of ketones on the pancreatic
beta cells. J Clin Invest 43:408-415, 1964.
29. Balasse EO, Ooms HA, and Lambilliotte JP: Evi-
dence for a stimulatory effect of ketone bodies on
insulin secretion in man. Horm Metab Res 2:371-372,
1970.
30. Malaisse WJ, and Malaisse-Lagae F: Stimulation of
insulin secretion by non-carbohydrate metabolites. ]
Lab Clin Med 72:438-448, 1968.
31. Balasse EO: Importance of ketone bodies in endoge-
nous fat transport. Clin Nutr 5:73-80, 1986.
32. Balasse EO, and Neef MA: Inhibition of ketogenesis
by ketone bodies in fasting humans. Metabolism
24~999-1007,1975.
33. Miles JM, Haymond MW, and Gerich JE: Suppres-
sion of glucose production and stimulation of insulin
secretion by physiological concentrations of ketone
bodies in man. J Clin Endocrinol Metab 52:34-37,
1981.
34. Metzger P, Franken P, and Balasse EO: Permissive
role of glucose on the insulinotropic effect of ketone
bodies in vivo. Horm Metab Res 5:313-315, 1973.
35. Green A, and Newsholme EA: Sensitivity of glucose
269
uptake and lipolysis of white adipocytes of the rat to
insulin and effects of some metabolites. Biochem J
180:365-370, 1979.
36. Keller U, Lustenberger M, and Stauffacher W: Effect
of insulin on ketone body clearance studied by a
ketone body "clamp" technique in normal man.
Diabetologia 31:24-29, 1988.
37. Issekutz B Jr, Bortz WM, Miller HI, and Paul P:
Turnover rate of plasma FFA in humans and in
dogs. Metabolism 16:lOOl-1009, 1967.
38. Owen OE, and Reichard GA Jr: Human forearm
metabolism during progressive starvation. J Clin
Invest 50:1536-1545, 1971.
39. F6ry F, and Balasse EO: Ketone body production
and disposal in diabetic ketosis. A comparison with
fasting ketosis. Diabetes 34:326-332, 1985.
40. Campbell J, and Best CH: Physiologic aspects of
ketosis. Metabolism 5:95-113, 1956.
41. Dietze G, Wicklmayr M, and Mehnert H: On the key
role of ketogenesis for the regulation of glucose
homeostasis during fasting: Intrahepatic control,
ketone levels and peripheral pyruvate oxidation. In
Biochemical and Clinical Aspects of Ketone Body Metabo-
lism. Soling HD, and Seufert CD, Eds. G. Thieme,
Stuttgart, 1978, pp 213-225.
42. Wahren J, Hagenfeldt L, and Felig P: Splanchnic and
leg exchange of glucose, amino acids and free fatty
acids during exercise in diabetes mellitus. ] Clin
Invest 55:1303-1314, 1975.
43. Owen OE, Block BSB, Pate1 M, Boden G, Mc-
Donough M, Kreulen T, Shuman CR, and Reichard
GA Jr: Human splanchnic metabolism during dia-
betic ketoacidosis. Metabolism 26:381-398, 1977.
44. Miles JM, Rizza RA, Haymond MW, and Gerich JE:
Effects of acute insulin deficiency on glucose and
ketone body turnover in man. Diabetes 29:926-930,
1980.
45. Hall SEH, Wastney ME, Bolton TM, Braaten JT, and
Berman M: Ketone body kinetics in humans: The
effects of insulin-dependent diabetes, obesity and
starvation. J Lipid Res 25:1184-1194, 1984.
46. Keller U, Schnell H, Sonnenberg GE, Gerber PPG,
and Stauffacher W: Role of glucagon in enhancing
ketone body production in ketotic diabetic man.
Diabetes 32:387-391, 1983.
47. Sherwin RS, Hendler RG, and Felig P: Effect of
diabetes mellitus and insulin on the turnover and
metabolic response to ketones in man. Diabetes
25:776- 784, 1976.
48. Bondy PK, Bloom WL, Whitner US, and Farrar BW:
Studies of the role of the liver in human carbohy-
drate metabolism by the venous catheter technique.
II. Patients with diabetic ketosis before and after the
administration of insulin. J Clin Invest 28:1216-1221,
1949.
49. Foster DW, and McGarry JD: The metabolic de-
rangements and treatment of diabetic ketoacidosis.
N Engl J Med 309:150-159, 1983.
50. Hagenfeldt L, and Wahren J: Human forearm mus-
cle metabolism during exercise. 111. Uptake, release
and oxidation of p-hydroxybutyrate and observa-
tions on the p-hydroxybutyratelacetoacetate ratio.
Scand J Clin Lab Invest 21:314-320, 1968.
51. Hagenfeldt L, and Wahren J: Human forearm mus-
270
cle metabolism during exercise. IV. Substrate utili-
zation in prolonged fasting. Scand J Clin Lab Znvest
27299-306, 1971.
52. Wahren J, Hagenfeldt L, and Felig P: Splanchnic and
leg exchange of glucose, amino acids and free fatty
acids during exercise in diabetes mellitus. J Clin
Znvest 55:1303-1314, 1975.
53. F6ry F, and Balasse EO: Ketone body turnover
during and after exercise in overnight-fasted and
starved humans. Am J Physiol 245:E318-E325,
1983.
54. Havel RJ, Naimark A, and Borchgrevinck CF: Turn-
over rate and oxidation of free fatty acids of blood
plasma in man during exercise: studies during con-
tinuous infusion of paImitate-I-14C. J Clin Znvest
42:1054-1063, 1963.
55. Young DR, Pelligra R, Shapira J, Adachi PR, and
Skrettingland K: Glucose oxidation and replacement
during prolonged exercise in man. J Appl Physiol
23:734-741, 1967.
56. Smith EE, Guyton AC, Manning RD, and White RJ:
Integrated mechanisms of cardiovascular response
and control during exercise in the normal human.
Prog Curdiovusc Dis 18:421-443, 1976.
57. Bjorkman 0, and Eriksson LS: Splanchnic glucose
metabolism during leg exercise in 60-hour-fasted
human subjects. Am J Physiol 245:E443-E448,
1983.
58. Wahren J, Sat0 Y, Ostman J, Hagenfeldt L, and Felig
P: Turnover and splanchnic metabolism of free fatty
acids and ketones in insulin-dependent diabetics at
rest and in response to exercise. J Clin Znvest 73:
1367-1376, 1984.
59. Fenselau A, and Wallis K: 3-oxoacid coenzyme A-
transferase in normal and diabetic rat muscle. Bio-
chem J 158:509-512, 1976.
60. Felig P, and Wahren J: Fuel homeostasis in exercise.
N Engl J Med 293:1078-1084, 1975.
61. Courtice FC, and Douglas CG: The effects of pro-
longed muscular exercise on the metabolism. Proc
ROY SOC 119Br391-439, 1936.
62. Hagenfeldt L, and Wahren J: Turnover of free fatty
acids during recovery from exercise. J Appl Physiol
39:247-250, 1975.
63. Rennie MJ, Jennett S, and Johnson RH: The meta-
bolic effects of strenuous exercise: a comparison
between untrained subjects and racing cyclists.
Quart E x p Physiol 59:201-212, 1974.
64. Maehlum S, Felig P, and Wahren J: Splanchnic
glucose and muscle glycogen metabolism after glu-
cose feeding during postexercise recovery. Am J
Physiol 235:E255-E260, 1978.
65. Maizels EZ, Ruderman NB, Goodman MN, and Lau
D: Effect of acetoacetate on glucose metabolism in
KETONE BODY PRODUCTION AND DISPOSAL
the soleus and extensor digitorum longus muscles of
the rat. Biochem J 162:557-568, 1977.
66. Berger M, Hagg SA, Goodman-MN, and Ruderman
NB: Glucose metabolism in perfused skeletal mus-
cle. Effects of starvation, diabetes, fatty acids,
acetoacetate, insulin and exercise onglucose uptake
and disposition. Biochem J 158:191-202, 1986.
67. Zinman B, and Vranic M: Diabetes and exercise. Med
Clin North Am 69:145-157, 1985.
68. Gottstein U, Miiller W, Berghoff W, Gartner H, and
Held K: Ziir Utilisation von nicht-veresterten Fet-
tsauren und Ketonkorpen in gehirn des Menschen.
Klin Wschr 49:406-411, 1971.
69. Wicklmayr M, and Dietze G: Effect of continuously
increasing concentrations of plasma ketone bodies
on the uptake and oxidation of glucose by muscle in
man. Eur J Clin Znvest 8:415-421, 1978.
70. Wicklmayr M, and Dietze G: Effect of metraprotere-
no1 on ketone body metabolism on the forearm in
healthy and diabetic subjects. Horm Metub Res
11~1-6, 1979.
71. Lyngsae J, Clausen JP, Trap-Jensen J, Sestoft L,
Schaffalitzky De Muckadell 0, Holst JJ, Nielsen SL,
and Rehfeld JF: Exchange of metabolites in the leg of
exercising juvenile diabetic subjects. Clin Sci Mol
Med 55:73-80, 1978.
72. Rennie MJ, Park DM, and Sulaiman WR: Uptake
and release of hormones and metabolites by tissues
in exercising leg in man. Am J Physiol 231:967-973,
1976.
73. Havel RJ, Kane JP, Balasse EO, Segel N, and Basso
LV: Splanchnic metabolism of free fatty acids and
production of triglycerides of very low density lipo-
proteins in normotriglyceridemic and hypertri-
glyceridemic humans. J Clin Znvest 49:2017-2035,
1970.
74. Sestoft L, Trap-Jensen J, Lyngsae J, Clausen JP,
Holst JJ, Nielsen SL, Rehfeld JF, and Schaffalitzky
De Muckadell 0:Regulation of gluconeogenesis and
ketogenesis during rest and exercise in diabetic
subjects and normal men. Clin Sci Mol Med 53:411-
418, 1977.
75. Wolfe BM, Havel JR, Marliss-EB, Kane JP, Seymour
J, and Ahuja SP: Effects of a 3-day fast and of ethanol
on splanchnic metabolism of FFA, amino acids, and
carbohydrates in healthy young men. J Clin Znvest
57329-340, 1976.
76. Garber AJ, Menzel PH, Boden G, and Owen OE:
Hepatic ketogenesis and gluconeogenesis in hu-
mans. J Clin Znvest 54:981-989, 1974.
77. F6ry F, De Maertelaer V, and Balasse EO: Mecha-
nisms of the hyperketonaemic effect of prolonged
exercise in insulin-deprived type 1 (insulin-depen-
dent) diabetic patients. Diubetologiu 30:298-304, 1987.