0021-972x/96/503,00/0
Journal
of Clinical
Endocrinology
and Metabolism
Copyright
0 1996 by The Endocrine
Society
Abnormalities
Oscillations
Dependent
Reduction*
Vol. 81, No. 6
Printed in U.S.A.
of Insulin
Pulsatility
and Glucose
during
Meals in Obese NoninsulinDiabetic
Patients:
Effects of Weight
BARRY GUMBINER,
EVE VAN CAUTER,
WILLIAM
DITZLER,
KAY GRIVER,
KENNETH
S. POLONSKY,
F. BELTZ,
TIMOTHY
M.
AND ROBERT R. HENRY
Departments
of Medicine, Monroe Community
Hospital and University of Rochester (B.G.), Rochester,
New York 14620; University of California-San
Diego and Veterans Administration
Medical
Center
School
(W.F.B., T.M.D., K.G., R.R.H.), La Jolla, C a l’fI ornia 92161; and University of Chicago, Pritzker
of Medicine
(E.V.C., K.S.P.), Chicago, Illinois 60637
ABSTRACT
Twenty-seven obese patients, including 8 with normal glucose tolerance, 10 with subclinical NIDDM,
and 9 with overt noninsulindependent diabetes mellitus (NIDDM),
were studied before and after
prolonged weight loss to assess the effects of the underlying defects
of diabetes per se from those of obesity and chronic hyperglycemia on
the regulation
of pulsatile
insulin
secretion.
Serial measurements
of
insulin
secretion
and plasma
glucose were obtained
during
3 standardized
mixed
meals consumed
over 12 h. Insulin
secretion
rates
were calculated by deconvoluting plasma C peptide levels using a
mathematical model for C peptide clearance and kinetic parameters
derived individually in each subject. Absolute (nadir to peak) and
relative
(fold increase
above nadir) amplitudes
of each insulin
secretory pulse and glucose oscillation
were calculated.
Compared
to the obese controls,
the subclinical
and overt NIDDM
patients
manifested
the following
abnormal
responses:
1) decreased
relative
amplitudes
of insulin
pulses, 2) reduced frequency
of glucose
oscillations,
3) increased
absolute
amplitudes
of glucose oscillations,
4) decreased
temporal
concomitance
between
peaks of insulin
pulses
and glucose oscillations,
5) reduced
correlation
between
the relative
amplitudes
of glucose
oscillations
concomitant
with insulin
pulses,
and 6) temporal
disorganization
of the insulin
pulse profiles.
These
defects were more severe in the overt NIDDM
patients,
and weight
loss only partially
reversed
these abnormalities
in both NIDDM
groups.
These findings
indicate
that p-cell responsiveness
is reduced,
and
the regulation
of insulin
secretion
is abnormal
under
physiological
conditions
in all patients
with NIDDM,
including
those without
clinical manifestations
of the disease.
These abnormalities
are not completely
normalized
with weight
loss, even in patients
who achieve
metabolic
control comparable
to that in obese controls.
The results are
consistent
with the presence
of an inherent
p-cell defect that contributes
to secretory
derangements
in subclinical
NIDDM
patients.
This abnormality
precedes
frank
hyperglycemia
and may ultimately
contribute
to the development
of overt NIDDM.
(J Clin Endocrinol
Metub 81:2061-2068,1996)
R
ECENT STUDIES have described the pulsatile nature of
meal-related
insulin secretion (l-4). The periodicity
of
insulin pulses range from 90-150 min and occur independent
of the high frequency insulin pulses that recur every lo-15
min (5, 6). In patients with noninsulin-dependent
diabetes
mellitus (NIDDM),
insulin pulses are decreased in amplitude
and disorganized
in their temporal
pattern (3,4). In contrast,
insulin
pulse amplitudes
are increased
in obese glucosetolerant patients, but regulation
of insulin pulsatility
is otherwise normal.
To what degree defects in pulsatile
insulin
secretion in
NIDDM
are a result of the chronic hyperglycemia
is unReceived
September
1, 1995. Revision
received
December
29, 1995.
Accepted
January
16, 1996.
Address all correspondence
and requests for reprints
to: Barry Gumbiner, M.D., Monroe
Community
Hospital,
435 East Henrietta
Road,
Rochester,
New York 14620.
*This work was supported
in part by grants from the Medibase
Weight
Management
Program,
the Medical
Research
Service
of the
Veterans Affairs Medical Center, the NIH (RR-00827
and RR-00044 from
the General Clinical Research Branch; Division
of Research Diabetes and
Research
Training
Center;
DK-20595,
5-T32-DK-07494-04,
DK-38949,
DK-31842,
and DK-41814),
the Whitaker
Foundation,
and a Rochester
Area Pepper Center Jr. Faculty Award
(AG-10463;
to B.G.).
2061
known.
To determine
this, in viuo insulin
secretion pulses
during mixed meals were compared
in obese patients with
varying degrees of glucose intolerance.
To further explore
this issue, patients were restudied
after weight loss therapy
to determine
whether
defects in pulsatile
insulin secretion
persist despite reduced glycemia and an improved
metabolic
milieu.
Subjects
and
Methods
Subjects
Studies were performed
in 27 obese subjects. The data in Table 1 and
other data on these subjects have been previously
reported
(7). Eight of
the subjects were classified
as controls
based on normal
glucose tolerance by oral glucose tolerance
test (OGTT)
criteria (8) and a negative
family history
for NIDDM.
Ten subjects had normal
to nondiagnostic
fasting glucose levels ((7.8 mmol/L),
but were either diabetic by OGTT
criteria (n = 9) or had a nondiagnostic
OGTT (i.e. a test result between
criteria
for impaired
glucose tolerance
and diabetes mellitus
in which
O-2 h glucose levels were 7.8-11.1 mmol/L;
n = 1). Six of these subjects,
including
the subject with a nondiagnostic
OGTT, had positive
family
histories
for NIDDM.
This group
was classified
as impaired
glucose
tolerance
or NIDDM
without
fasting hyperglycemia.
For the purposes
of simplification,
this group is identified
throughout
this report as subclinical NIDDM.
The remaining
9 subjects were classified
as NIDDM
with fasting hyperglycemia
on the basis of a fasting plasma glucose level
2062
GUMBINER
JCE & M . 1996
Volt31 . No 6
ET AL.
more than 7.8 mmol on 2 occasions.
For the purposes
of simplification,
these subjects
are identified
as overt NIDDM
throughout
this report.
Three of the overt NIDDM
subjects
had positive
family
histories
for
NIDDM.
The metabolic
and clinical
characteristics
of the study groups
are detailed
in Table 1. All subjects
were otherwise
healthy.
Those
subjects receiving
medications
(e.g. oral hypoglycemic
agents) had them
discontinued
2 weeks before the studies.
Experimental
protocol
Subjects were admitted
to the San Diego V.A. Medical
Center Special
Diagnostic
and Treatment
Unit. The subjects
consumed
a standardized
isocaloric
meal plan f-30 Cal/kg)
composed
of 55% carbohydrate,
30%
fat, and 15% protein
for at least 24 h before the studies.
All procedures
were initiated
after a 14-h overnight
fast.
Preweigl~t
loss studies. Insulin
secretion
rates for pulse analysis
were
determined
from measurements
of each individual’s
C peptide
kinetics
and C peptide
responses
to mixed meals. C peptide
kinetics
were determined
as previously
described
in detail (9, lo), and previous
studies
of these subjects
indicate
that C peptide
kinetics
were not significantly
altered by weight
loss (7).
On a separate day, a 12-h meal profile study was conducted.
The meal
plan was isocaloric
f-30
Cal/kg),
with 20% of the calories
served at
breakfast,
40% at lunch, and 40% at dinner.
Each meal was a standardized solid mixed meal composed
of 55% carbohydrate,
30% fat, and 15%
protein
and was consumed
within
a 40-min period beginning
immediately after blood samples
obtained
at 0900, 1200, and 1700 h. Plasma C
peptide
and glucose were sampled
from an indwelling
venous
catheter
every 15 min during
the entire 12-h study period
(0900-2100
h).
Weight reduction and postweight
loss studies. After the above studies,
all
subjects
consumed
a weight-reducing
formula
diet of 600 Cal/day,
as
previously
described
(7). Dieting
continued
for at least 6 weeks to attain
a minimum
of 10% weight
loss, a reduction
that results in significant
improvement
in virtually
all metabolic
parameters
(11). After completion of the diet phase, subjects
were refed solid mixed meals for at least
2 weeks until isocaloric
intake was attained.
The C peptide
kinetic study
and the 12-h meal profile
were then repeated.
Assays
Plasma glucose
levels were determined
at the bedside
by an automated glucose
oxidase
method
(no. 23A, Yellow
Springs
Instruments,
Yellow Springs,
OH). Serum insulin
was assayed
by a double antibody
technique
(12). C peptide
samples were collected
in tubes containing
100
PL ethylenediamine
tetraacetate
(1.5 mg/mL)
and 150 PL Trasylol
(500
kallikrein
inhibiting
units; F.B.A. Pharmaceutical,
New York, NY), centrifuged
to separate
the plasma,
stored at -20 C, and later assayed by
a nonequilibrium
ethanol
precipitation
RIA (13). The lower
limit of
sensitivity
of the C peptide
assay was 20 pmol/L,
and the intraassay
coefficient
of variation
was 4%. Each individual’s
pre- and postweight
loss samples
were assayed
in the same batch.
Data
analysis
Insulin secretion rates. Individual
C peptide kinetic parameters
[fractional
rate constants
(K,, K,, and K,), volume
of distribution,
and MCRI were
derived
by compartmental
analysis
of the individual
decay curves,
as
previously
described
(9, 10) and reported
(7). Insulin
secretion
rates
during
the meal studies
were calculated
by deconvolution
of plasma C
peptide
levels using the rate constants
derived
for each individual.
Pulse analysis.
Pulses of insulin
secretion
and oscillations
in glucose
concentration
were considered
true physiological
events if the percent
incremental
rise from the preceding
nadir to the peak and the percent
decremental
fall from the peak to the subsequent
nadir both met threshold criteria
(2,3,14).
The criteria chosen were based on the measurement
error (i.e. intraassay
coefficient
of variation),
expected
pulse frequency,
and theoretical
rate of occurrence
of false positive
and false negative
pulses (14). The threshold
chosen for glucose
peaks was twice the maximum assay error (2 X 3% = 6%). The threshold
for identification
of
pulses of insulin
secretion
was set at 3 times the intraassay
coefficient
INSULIN
PULSATILITY
AFTER WEIGHT
Temporal concomitance between peaks of glucose oscillations and
insulin pulses. The percentage of glucose oscillations concomitant with insulin pulses was similar among the groups both
before and after weight loss.
Stntisticul analysis. Data are expressed
as the mean 5 SEM and were
compared
among groups by ANOVA.
To identify
specific groups differences, the Newman-Keuls
test was used, and a statistical
difference
was identified
if P < 0.05. Pearson correlation
coefficients
were determined to analyze
the relationship
between
the relative
amplitudes
of
insulin
pulses concomitant
with glucose
oscillations.
Differences
between correlation
coefficients
were tested for statistical significance
after
Fisher’s
r to z transformation
(15). Analyses
were performed
using
CLINFO
and BMDl’
statistical
software
on the University
of Rochester
and University
of California-San
Diego General Clinical Research Center
computer
systems.
secretory pulses
2063
in relative amplitudes, but the pattern among the groups
remained the same. The obese controls had the highest relative amplitudes, and weight loss had the greatest impact on
this group. In contrast, the overt NIDDM patients had the
lowest relative amplitudes, and the effect of weight loss did
not reach statistical significance in this group (P = 0.08).
of variation
of C peptide (3 x 4% = 12%) to exclude additional
noise that
could be contributed
by the deconvolution
process.
Pulse analysis was facilitated
by the computer
program
ULTRA
(14),
which generates a “clean” series of data points to identify
true pulses and
eliminates
peaks that do not meet threshold
criteria.
The absolute
amplitude of each pulse was calculated
as the difference
between
the peak
and the preceding
nadir value. Amplitudes
were also calculated
in terms
of their relative
(i.e. fold) increase over the preceding
nadir. The 12-h
profiles
obtained
during the meal studies were analyzed
for temporal
concomitance
between
pulses of insulin secretion and glucose. Peaks of
insulin secretion and glucose oscillations
that occurred
within 15 min of
each other were considered
to be concomitant
events.
Insulin
LOSS
Glucose oscillations
(Table 3)
Frequency. Both the subclinical and overt diabetic subjects
had significantly
fewer meal-related glucose oscillations
than the obese controls (P < 0.05). Only the subclinical diabetic patients had increased numbers of glucose oscillations
after weight loss (P < 0.05).
Absolute amplitude. As expected, the absolute amplitudes of
the glucose oscillations were significantly increased in both
diabetic groups (P < 0.05). Weight loss was associatedwith
a decreasein the amplitudes of all diabetic subjects(P < 0.05),
but only the subclinical diabetic patients achieved a level
comparable to that in the obese controls.
Results
(Table 2)
Frequency.The number of pulses during the 12-h meal profile
was similar among all groups. However, inspection of the
data indicates a potential trend of a weight loss effect in the
overt NIDDM group. Thus, although the ANOVA was not
significant, separate testing of the overt group indicated that
there was a decreasein pulse number after weight loss (P <
0.05).
Relative amplitude. The relative amplitudes of the glucose
oscillations were similar among all groups, and treatment
had no significant effect.
Absolutepulseamplitude.The subclinical diabetic patients had
significantly higher pulse amplitudes (P < 0.05), whereas the
overt NIDDM patients had significantly lower pulse amplitudes than the other two groups (P < 0.05). After weight loss,
the mean pulse amplitude in the subclinical NIDDM patients
remained unchanged. In contrast, modest changes in overt
NIDDM patients and obese controls resulted in postweight
losspulse amplitudes that were comparable between the two
groups. However, overall there was no consistent effect by
weight loss per seon absolute pulse amplitude.
Temporalconcomitancebetweenpeaksof insulin pulsesand glucoseoscillations.The percentage of insulin pulses concomitant
with glucose oscillations was significantly lower in subclinical and overt NIDDM patients (both P < 0.05). Weight loss
in the subclinical diabetic subjects was associated with an
increase in the percentage of concomitant pulses that was no
longer statistically different from that in the obesecontrols.
Relative pulseamplitude.Differences in relative amplitudes of
secretory pulses were found among the groups and in response to treatment (P < 0.05). The obese controls had the
largest relative amplitudes, and the overt NIDDM subjects
had the lowest. Weight loss was associatedwith an increase
Obesecontrols. Regression analyses of glucose oscillations
concomitant with insulin pulses indicated a significant association between the relative amplitudes of these parameters in the obese subjects (r = 0.73; P < 0.05) that was unchanged by weight loss (r = 0.75; P < 0.05).
TABLE
2. Insulin
pulses
Frequency
Obese (n = 8)
Subclinical
NIDDM
(n = 10)
Overt
NIDDM
(n = 91
MPS,
aP
bP
"P
dP
eP
Twelve-hour
Regressionanalyses of glucose oscillations and insulin
pulses (Fig. 1)
meal
Absolute
(no./MPS)
~omol/mz
~I~
Pre
Post
Pre
9.9 + 0.6
9.8 -t 0.6
10.9 ? 0.6
9.6 z 0.6
9.6 I! 0.5
8.7 + 0.6b
148 lr 14
217 + lFd
92 ? 13”r’
profile
< 0.05, by ANOVA.
< 0.05, pre vs. post.
< 0.05, vs. obese.
< 0.05, us. overt NIDDM.
< 0.05, vs. subclinical
NIDDM.
study;
pre,
preweight
loss;
post,
amplitude
. minY
Relative
Post
postweight
125 ? 13
196 f 17cad
101 + 8
loss.
amplitude
(fold
above nadir)”
increase
Concomitance
plucose
with
(%
insulin)
Pre
Post
Pre
Post
1.76 2 0.20
0.97 2 0.11”
0.78 2 0.06”
3.85 2 0.596
1.60 I! 0.16’,”
1.02 2 0.11”
77 k 5
85 k 6
73 k 12
83 k 5
76 2 7
70 k 5
GUMBINER
TABLE
3.
ET AL.
JCE & M . 1996
Vol81.
No 6
Glucose pulses
Frequency (no.iMPSP
Obese (n = 81
Subclinical NIDDM (n = 10)
Overt NIDDM (n = 9)
Relative amplitude (fold
increase above nadir)
Pre
Post
Absolute amplitude (mmol/L)”
Pre
Post
6.5 2 0.5
3.8 " 0.46
3.4 2 0.36
6.0 + 0.7
5.3 " 0.6"
4.0 'I 0.46
Pre
MPS, Twelve-hour meal profile study; pre, preweight
a P < 0.05,by ANOVA.
bP < 0.05vs. obese.
cP < 0.05 pre vs post.
d P < 0.05 vs overt NIDDM.
e P < 0.05 vs subclinical NIDDM.
Post
1.8 + 0.2
1.7 + 0.2
3.0 + 0.5b.d
4.8 k 0.46,'
0.34
0.46
0.32
2.1 + 0.3
3.3 If: 0.6"'
loss; post, postweight
+ 0.04
k 0.06
I? 0.03
0.39
0.38
0.46
k 0.04
+ 0.05
2 0.08
Post
Overt
r = 0.75
Post
3
53 k 6
42 " 5
34 2 56
NIDDM
Pre r = 0.20
Pre r = 0.51
20
53 -c 5
35 k 5"
25 -+ 5b
loss.
Subclinical NIDDM
Pre r = 0.73
Concomitance (%
insulin with glucose)”
Pre
Post
Post
r = 0.58
r = 0.59
I
/
0 Post
l
a
15
$
I
g
s
*
lo
5
2
0
7
0.C1
0.2
0.4
0.6
0.8
1.0
1.2
0.0
0.2
0.4
0.6
0.8
1.0
:
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Relative Glucose Oscillations
FIG. 1. Regression analyses between relative amplitudes of insulin pulses concomitant
and overt NIDDM patients pre- (-1 and post- (- - -) weight loss.
SubclinicalNIDDM. A significant, albeit weaker, correlation
between the relative amplitudes of glucose oscillations concomitant with insulin pulses was present in subclinical diabetic patients (r = 0.51; P < 0.05). However, the correlation
was significantly weaker than that observed in the obese
controls (P < 0.05). Weight loss did not significantly change
these observations (r = 0.59).
Overt NIDDM. The correlation between the relative amplitudes of glucose oscillations concomitant with insulin pulses
in the untreated overt NIDDM group (r = 0.28) was not
significant. After weight loss, a significant correlation was
observed (r = 0.59; P < 0.051,but was significantly lessthan
that in the obesecontrols (P < 0.05).
Individual meal profiles (Figs. 2-4)
Examples of individual meal profiles illustrate the differences in glucose and insulin responses among the three
groups.
Obesecontrols. The two control profiles depicted are typical
of the responsesof obese nondiabetic patients before and
after weight loss.All were similar in pulse amplitude, timing,
and concomitance. Meals were consistently demarcated by a
prominent pulse immediately following food intake. Other
with glucose oscillations
in obese, subclinical
NIDDM
than an increase in relative insulin pulse amplitudes, weight
loss had little effect on the profiles.
Subclinical NIDDM. A sluggish glycemic response, a low
concomitancy rate between glucose and insulin pulses, and
an indistinct dinner response were characteristic of subclinical NIDDM patients. Improvement in these parameters after
weight loss varied among the individuals in this group.
Overt NIDDM. As previously described (3), insulin pulsesin
these patients were disorganized. Meals were demarcated by
a large glucose oscillation, but not a prominent insulin pulse.
The decreasein insulin pulse amplitudes and concomitancy
demonstrated by these two individuals are typical of all of
the overt NIDDM subjects. In most of these patients, these
abnormal patterns did not improve with weight loss.
Discussion
One of the major pathophysiological mechanismscausing
hyperglycemia in NIDDM is impaired insulin secretion (1,3,
6, 16). The results of the current study indicate that defects
in P-cell responsiveness and regulation of insulin secretion
are evident under physiological conditions even in patients
with subclinical derangements in metabolic control. These
defects were found to be more severe in the overt NIDDM
INSULIN
PULSATILITY
AFTER
WEIGHT
2065
LOSS
0
B
0900
Clock Time
L
1300
D
1700
2100
Clock Time
2. Individual
meal profiles
of two obese control subjects
(subject
1, A and B; subject 2, C and D), pre- (solid line) and postweight loss. The data depicted
are a “clean” series generated
by ULTRA
(see Materials and Methods). Breakfast
(B) was given at
(L) at 1300 h, and dinner(D)
at 1700 h (see Results for description).
Arrows identify true pulses and oscillations
(downward arrows
loss; upward arrows are postweight
loss). Solid arrows indicate
that the glucose oscillation
in the upperpanel is concomitant
with
insulin
pulse identified
by solid arrows in the lower panel. Open arrows indicate
that there is no concomitancy
associated
with
oscillation.
FIG.
patients, and reducing glycemia with weight loss only partially reversed these abnormalities. Although these are crosssectional data, as many patients with modest subclinical
hyperglycemia progress to overt NIDDM, the findings of the
current study support the contention that P-cell dysfunction
progressesas glucose tolerance deteriorates (17). However,
the defects were not completely reversible, even in the subclinical diabetic subjects, providing support for the contention that the aberrances in p-cell responsivenessand regulation of insulin secretion are inherently characteristic of
NIDDM (1, 16).
When measured in absolute terms, the total amount of
insulin secretedper pulse was decreasedin overt NIDDM. In
contrast, absolute pulse amplitudes were increased in the
subclinical NIDDM subjects. Thus, unlike that in overt
NIDDM, the increase in absolute pulse amplitudes in subclinical NIDDM patients is consistent with the 24-h hypersecretion of insulin (7) and the commonly observed hyperinsulinemia associated with less severe disease (18-21).
Whether the quantity of insulin secreted by the subclinical
NIDDM patients was appropriate for their mild degree of
hyperglycemia could not be directly addressedin the current
study. However, in studies of patients with impaired glucose
tolerance, when glycemia is matched to normal subjects,
insulin secretion is reduced (1). Therefore, it is likely that
(shaded
line)
0900 h, lunch
are preweight
the respective
the pulse or
patients with subclinical NIDDM also secrete less insulin
relative to their degree of glucose intolerance.
Additional insight into whether insulin secretion by the
subclinical NIDDM patients is appropriate for their modest
degree of postprandial hyperglycemia is provided by the
analysis of insulin pulses in relative terms. This analysis
normalizes data for differences in baseline or nadir rates of
secretion, thus providing an index of pulsatile insulin secretory responsesby which individuals and groups can be compared. The results indicate that pulsatile insulin secretion
increased almost 2-fold above its rate of nadir secretion in
nondiabetic obesepatients. In contrast, the responsesin the
subclinical and overt diabetic patients were significantly
lower. Therefore, P-cell pulsatile responsiveness to mealrelated secretagoguesis decreased not only in patients with
overt NIDDM, but in subclinical NIDDM patients without
fasting hyperglycemia as well.
The meal profiles reveal an interesting pattern of glucose
responsesnot previously highlighted by studies performed
under physiological conditions (2-4). Both diabetic groups
had a significant reduction in the number of postmeal glucoseoscillations. As expected, the oscillations were of greater
amplitude, when measured in absolute terms, in the diabetic
patients. However, when measured in relative terms, there
was no difference among the groups.
GUMBINER
2066
‘GSF
.t?!!E
$ CT-
co E
!s$$
YE
24
-
ET
JCE & M . 1996
Vol81.No6
AL.
600
500
400
300
200
100
0
B
0900
L
1300
D
1700
2100
Clock Time
B
0900
L
1300
D
1700
2100
Clock Time
3. Individual
meal profiles
of two subclinical
diabetic
patients
(subject
1, A and B; subject 2, C and D), pre- (solid line) and post- (shaded
line) weight loss. The data depicted
are a “clean”
series generated
by ULTRA
(see Materials
and Methods).
Breakfast
(B) was given at 0900
h, lunch (L) at 1300 h, and dinner
(D) at 1700 h (see Results for description).
Arrows
identify
true pulses and oscillations
(downward
arrows
are preweight
loss; zqwaral
arrows
are postweight
loss). Solid arrows
indicate
that the glucose oscillation
in the upper panel is concomitant
with the respective
insulin
pulse identified
by solid arrows
in the lower panel. Open arrows
indicate
that there is no concomitancy
associated
with the pulse or oscillation.
FIG.
The significance of these glucose oscillations has only recently begun to be appreciated. Postprandial increases in
glucose serve to amplify insulin secretion (2,3). In addition,
a tight link exists between the oscillatory pattern of glucose
and subsequent patterns of insulin secretion (22). This entrainment of insulin pulses by glucose oscillations is a manifestation of a glucose-insulin feedback loop. When glucose
oscillations are manipulated by a glucose infusion, entrainment hasbeen found to be disrupted in both mild and severe
glucose intolerance (17). The results of the current study
extend these findings by demonstrating that the relationship
between insulin pulses and glucose oscillations in glucoseintolerant patients are disturbed in the physiological setting
of mixed meals. Not only is the percent concomitancy decreased,but the correlations between the relative amplitudes
of insulin pulses concomitant with glucose oscillations are
also reduced. As these abnormalities are more severe in the
overt NIDDM group, this suggests that the feedback loop
between glucose and insulin becomes more disrupted as
metabolic control deteriorates.
The responseto weight lossdiffered substantially among
the groups. Relative amplitudes of insulin pulses increased
in all groups after weight loss. As the relative amplitudes of
glucose oscillations were unchanged after treatment, this
indicates that p-cell responsiveness to meal-related secreta-
gogues improved. However, these improvements were
much lessdramatic in the diabetic groups. The overt NIDDM
patients were unable to achieve responsescomparable to the
levels observed in the obese controls before weight loss. In
subclinical NIDDM patients, although the posttreatment
weight and body mass index were comparable to the pretreatment weight and body massindex of the obesecontrols,
they remained mildly hyperglycemic. In this metabolic milieu, almost none of the mean insulin and glucose parameters
achieved levels equal to or exceeding the control values
before treatment. Whether treatment that results in normal
body weight and glycemic control could completely reverse
these defects requires further evaluation.
The individual pulse profiles illustrate the abnormalities in insulin and glucose responses. Insulin pulses of the
obese subjects were organized and clearly demarcate
meals. In contrast, the profiles of the overt NIDDM subjects were markedly disorganized, and it is unclear when
meals were consumed. The profiles of subclinical diabetic
patients also illustrate that abnormal temporal organization occurs even with mild disease. After weight loss, the
pattern of insulin pulsatility is still disorganized in the
subclinical NIDDM patients, implicating factors other
than hyperglycemia as the cause of abnormalities of insulin secretion regulation.
INSULIN
PULSATILITY
AFTER WEIGHT
2067
LOSS
C
262422-
t
t
20.
18161412-
600
B
1
L
1300
D
1700
Clock Time
600,
L
1300
D
1700
Clock Time
4. Individual
meal profiles of two overt diabetic
patients
(subject
1, A and B; subject 2, C and D), pre- (solid line) and postweight loss. The data depicted
are a “clean” series generated
by ULTRA
(see Materials
and Methods).
Breakfast
(B) was given at
(L) at 1300 h, and dinner(D)
at 1700 h (seeResults
for description).Arrows
identify
true pulses and oscillations
(downward
arrows
loss; upward
arrows
are postweight
loss). Solid arrows indicate
that the glucose oscillation
in the upper panel is concomitant
with
insulin
pulse identified
by solid arrozus in the lower panel. Open arrows
indicate
that there is no concomitancy
associated
with
oscillation.
FIG.
Numerous studies suggest that the contribution of impaired insulin secretion VS.insulin resistance to the pathophysiology of NIDDM is heterogeneous. It has been speculated that insulin resistance contributes more to the
pathophysiology of NIDDM patients with modest fasting
hyperglycemia, but significant postglucose load hyperglycemia with impaired insulin secretion is of greater significance in the pathophysiology of NIDDM patients with frank
fasting hyperglycemia (23). The results of the current study
would not support this postulate. Both categories of NIDDM
patients manifest serious insulin secretory defects, and neither is completely reversed by treatment, even though previous studies indicate that both insulin sensitivity and secretion improve after weight loss (11).
In conclusion, p-cell responsiveness is reduced, and the
regulation of pulsatile secretion is disrupted in patients with
NIDDM under physiological conditions. This occurs regardless of whether the disease is manifested clinically, but is
more severe in patients with frank hyperglycemia. Weight
loss and improvement in glycemia only partially reverse
these abnormalities. By this scenario, weight loss reverses
only defects that are acquired and exacerbated by other processes,such as obesity and glucose toxicity (24). The high
incidence of a positive family history of NIDDM in the pa-
(shaded line)
0900 h, lunch
are preweight
the respective
the pulse or
tients studied (VS.no family history in any controls) would
also argue in favor of defects that predispose these individuals to NIDDM. Therefore, these findings are consistent with
the perspective that an inherent pancreatic defect plays a
pivotal role in the underlying pathophysiology and progression of NIDDM.
Acknowledgments
We thank Paul Rue and William
Pugh for performing
numerous
C
peptide and insulin assays, the laboratory
and nursing staff of the Special
Diagnostic
and Treatment
Unit at the San Diego V.A. Medical Center for
their dedicated
work, and the volunteers
for their diligence
in carrying
out this study.
References
1. Polonsky
KS 1995 The B-cell in diabetes:
from molecular
genetics
to clinical
research.
Diabetes.
44:705-717.
2. Polonsky
KS, Given
BD, Cauter
EV. 1988 Twenty-four-hour
profiles
and
pulsatile
patterns
of insulin
secretion
in normal
and obese subjects.
J Clin
Invest. 81:442-448.
3. Polonsky
KS, Given
BD, Hirsch
LJ, et al. 1988 Abnormal
patterns
of insulin
secretion
in non-insulin-dependent
diabetes
mellitus.
N Engl J Med. 318:12311239.
4. Calles-Escandon
J, Jaspan J, Robbins
DC. 1989 Postprandial
oscillatory
patterns of blood
glucose
and insulin
in NIDDM.
Abnormal
diurnal
insulin
GUMBINER
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
secretion
patterns
and glucose
homeostasis
independent
of obesity.
Diabetes
Care. 12:709-714.
Lang DA, Matthews
DR, Phil D, Pete J, Turner
RC. 1979 Cyclic
oscillations
of basal plasma
glucose
and insulin
concentrations
in human beings. N Engl
J Med. 301:1023-1028.
Simon
C, Brandenberger
G, Follenius
M. 1987 Ultradian
oscillations
of
plasma
glucose,
insulin,
and C-peptide
in man during
continuous
enteral
nutrition.
J Clin Endocrinol
Metab.
64:669-674.
Polonsky
KS, Gumbiner
B, Ostrega
D, Griver
K, Tager H, Henry
RR. 1994
Alterations
in immunoreactive
proinsulin
and insulin
clearance
induced
by
weight loss in NIDDM.
Diabetes.
43:871-877.
National
Diabetes
Data Group.
1979 Classification
and diagnosis
of diabetes
mellitus
and other categories
of glucose
intolerance.
Diabetes.
28:1039-1057.
Eaton RF, Allen RC, Schade DS, Erickson
KM, Standefer
J. 1980 l’rehepatic
insulin
production
in man: kinetic
analysis
using peripheral
connecting
peptide behavior.
J Clin Endocrinol
Metab.
51:520-528.
Polonsky
KS, Licinio-Paixao
J, Given
BD, et al. 1986 Use of biosynthetic
human
C-peptide
in the measurement
of insulin
secretion
rates in normal
volunteers
and type I diabetic
patients.
J Clin Invest. 77~98-105.
Henry RR, Gumbiner
B. 1991 Benefits
and limitations
of very-low-calorie
diet
therapy
in obese NIDDM.
Diabetes
Care. 14802-823.
Morgan
CR, Lazarow
A. 1963 Immunoassay
of insulin:
two antibody
systemsplasma
insulin
levels of normal
and subdiabetic
rats. Diabetes.
12:115-126.
Faber OK, Binder
C, Markussen
J, et al. 1978 Characterization
of seven
C-peptide
antiserum.
Diabetes.
27:170-177.
Van Cauter E 1988 Estimating
false-positive
and false-negative
errors in analyses of hormonal
pulsatility.
Am J Physiol.
254:E786-E794.
ET AL.
JCE & M . 1996
Vol81
. No 6
15. Dawson-Saunders
B, Trapp RG. 1990 Correlation
and regression.
In: Basic and
clinical
biostatistics.
East Norwalk:
Appleton
and Lange;
161-185.
16. Porte Jr D. 1990 Banting
Lecture.
Beta-cells
in type II diabetes
mellitus.
Diabetes. 40:166-180.
17. O’Meara
NM, Sturis J, Van Cauter E, Polonsky
KS. 1993 Lack of control
by
glucose of &radian
insulin
secretory
oscillations
in impaired
glucose tolerance
and in non-insulin-dependent
diabetes
mellitus.
J Clin Invest. 92:262-271.
18. Bogardus
C, Lillioja
S, Howard
BV, Reaven
G, Mott D. 1984 Relationships
between
insulin
secretion,
insulin
action,
and fasting plasma glucose
concentration
in nondiabetic
and noninsulin-dependent
diabetic
subjects. J Clin Invest. 741238-1246.
19. Seltzer
HS, Allen
EW, Herron
AL, Brennan
MT. 1967 Insulin
secretion
in
response
to glycemic
stimulus:
relation
of delayed
initial
release to carbohydrate intolerance
in mild diabetes
mellitus.
J Clin Invest. 46:323-335.
20. Perley MJ, Kipnis
DM. 1967 Plasma insulin
response
to oral and intravenous
glucose:
studies
in normal
and diabetic
subjects.
J Clin Invest. 46:1954-1962.
21. DeFronzo
RA 1988 The triumvirate:
p-cell, muscle,
liver: a collusion
responsible for NIDDM.
Diabetes.
37:667-687.
22. Sturis
J, Van Cauter
E, Blackman
JD, Polonsky
KS. 1991 Entrainment
of
pulsatile
insulin
secretion
by oscillatory
glucose
infusion.
J Clin Invest. 87:
439-445.
23. O’Rahilly
S, Hattersley
A, Vaag A, Gray H. 1994 Insulin
resistance
as the
major cause of impaired
glucose
tolerance:
a self-fulfilling
prophesy?
Lancet.
344585-589.
24. Unger
RH, Gmndy
S. 1985 Hyperglycaemia
as an inducer
as well as a consequence
of impaired
islet cell function
and insulin
resistance:
implications
for
the management
of diabetes.
Diabetologia.
28:119-121.