1 Diabetes Centre, Department of Endocrinology, Odense University Hospital, DK-5000 Odense C; 2 Institute of Experimental Clinical Research, Aarhus University Hospital, DK-8000 Aarhus; 3 Department of Medical Physiology, University of Copenhagen, DK-2200 Copenhagen; and 4 Department of Clinical Biochemistry, Sønderborg Hospital, DK-6400 Sønderborg, Denmark
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
To
establish reference intervals for the pancreatic -cell response and
the counterregulatory hormone response to prolonged fasting, we studied
33 healthy subjects (16 males, 17 females) during a 72-h fast. Glucose,
insulin, C-peptide, and proinsulin levels decreased (P < 0.001), and the levels of counterregulatory factors increased
during the fast [P < 0.05; glucagon and free fatty
acids (FFA) with a linear increase and epinephrine, norepinephrine, and
cortisol with a clear underlying circadian rhythm]. Growth hormone
secretion increased from the first to third day of fasting (P < 0.05) but actually decreased from the second to
third day of fasting (P = 0.03). Males had higher
glucose and glucagon levels and lower FFA levels during the fast
(P < 0.05), whereas no effect of gender on
-cell
polypeptides was observed. A high body mass index resulted in higher
insulin and C-peptide levels during the fast (P < 0.05). In conclusion, we have provided reference intervals for
glucoregulatory factors during a 72-h fast. We observed a diminished
-cell response concomitant with an increased secretion of
counterregulatory hormones. These results should be of clinical and
scientific value in the investigation of hypoglycemic disorders.
catecholamines; glucagon; growth hormone; cortisol; free fatty acids
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
EPISODIC SYMPTOMS
suggestive of hypoglycemia are a common clinical presentation in
healthy-appearing adults seeking medical help. Often such patients
cannot be studied during a spontaneous episode of hypoglycemia, and, if
no obvious cause of hypoglycemia can be demonstrated, the supervised
72-h fast with continuous measurements of glucose, C-peptide,
proinsulin, and insulin has traditionally been the diagnostic test of
choice (36, 38). The 72-h fast is used to diagnose
insulin-mediated hypoglycemia, where fasting fails to suppress the
secretion of -cell polypeptides. The ability of this test to
diagnose abnormalities in the counterregulatory hormone response to
prolonged fasting has never been examined fully.
The diagnosis of hyperinsulinemic hypoglycemia has previously been
associated with some uncertainty, mainly because normal overnight
fasting serum levels of the -cell polypeptides have been used to
discriminate between normal subjects and subjects with hypoglycemia.
Recently, a diagnostic interpretation of data based on paired
observations of plasma glucose and insulin, C-peptide and proinsulin
obtained at the end of a 72-h fast has been suggested (36,
38). However, the diagnostic accuracy of these assays was
<100% in the plasma glucose range of 2.8 to 3.3 mmol/l (27, 33,
40), where some persons are known to experience hypoglycemic symptoms (35, 36). Furthermore, the normal subjects used
in these studies had been referred for evaluation of the possible presence of a hypoglycemic disorder and therefore did not represent an
unselected healthy population (27, 33, 40). Thus, to our
knowledge, well-defined reference intervals for insulin, C-peptide, and
proinsulin during a 72-h fast in healthy humans have never been
established. In addition to suppressed secretion of
-cell polypeptides, several counterregulatory factors are known to be activated in the prevention and correction of hypoglycemia (5, 8,
9, 21, 23). Increases in glucagon and catecholamines are
involved in the first line defence of hypoglycemia (7, 11, 12,
17, 30), whereas increases in growth hormone (GH) and cortisol
are supposed to play a role in the defense against prolonged
hypoglycemia (6, 7, 13, 14, 30). Furthermore, increases in
free fatty acids (FFA; mediated by epinephrine) have been reported to
play a role in the counterregulation to hypoglycemia (16, 17,
28). Increases in all of these factors have been demonstrated in
studies of prolonged fasting (4, 7, 19, 21, 23, 24, 28, 31,
34). However, reference intervals for the response of these
glucose counterregulatory factors to a 72-h fast have never been
published. Abnormalities in the secretion of these factors during
prolonged fasting could previously have gone undetected as the actual
cause of hypoglycemia, and this could explain why hormone deficiencies
are so seldomly reported as the cause of hypoglycemia. Therefore, such
reference intervals could be of large interest in cases where patients
under prolonged fasting actually experience Whipples triad but in whom
insulin-mediated hypoglycemia can be excluded.
The present study was designed to establish reference intervals for the
-cell polypeptides and the most important glucose counterregulatory
factors during a classic 72-h fast and to investigate the impact of
gender and body mass index (BMI) on these factors to improve the
diagnosis of hyperinsulinemic hypoglycemia and other hypoglycemic disorders.
![]() |
SUBJECTS AND METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Subjects.
Thirty-three healthy Caucasian volunteers with no previous history of
hypoglycemic symptoms were studied in the Department of Endocrinology
at Odense University Hospital, Denmark. To study the effects of gender
and BMI, an equal number of males and females and subjects with normal
and high BMI were included. A high BMI was á priori defined as
BMI 27 kg/m2 for males and BMI
25 kg/m2 for females. Age, BMI, and the proportion of subjects
with high BMI were similar in males and females (Table
1). Two to three weeks before the
prolonged fast, every subject underwent a routine medical examination,
including clinical history, medication, blood pressure,
electrocardiogram, and blood analysis including, hemoglobin, liver
enzymes, creatinine, and a lipid profile. An oral glucose tolerance
test according to the World Health Organization criteria was performed
to ensure a glucose tolerance in the normal range. None of the
participants was taking medications except oral contraceptives or
suffered from hyperlipidemia, hypertension, diabetes, ischemic heart
disease, or liver disease. In addition, all subjects had normal serum
creatinine precluding reduced renal clearance as a cause of high levels
of C-peptide or proinsulin. All participants underwent a 72-h fast
according to a standard protocol on patients admitted to our
department. The subjects were told to fast at home from 10:00 PM until
they presented at the department at 7:30 AM the next day. At 8:30 AM
the 72-h fast was started. A cannula was inserted in a forearm vein,
and blood samples were drawn every 3 h for determinations of
plasma levels of glucose and pancreatic glucagon and serum levels of
insulin, proinsulin, C-peptide, plasma, cortisol, and GH. Serum levels
of albumin and FFA were measured every 6 h. Serum epinephrine and
norepinephrine were measured in six volunteers only [3 males, 3 females, mean BMI 25.8 (19.9-36.5), mean age 25.5 (19.9-43.1)]. During the study, the participants were permitted
to drink water ad libitum and to walk around the hospital area while
under surveillance. One-half hour before sampling and between 10:00 PM
and 7:00 AM, subjects were required to rest. All participants completed
the 72-h fast without developing hypoglycemic symptoms or biochemical
hypoglycemia (plasma glucose <2.5 mmol/l). Compliance with the fast
was assessed by determination of FFA and hydration by determination of
albumin concentration. For unknown reasons, glucose, insulin,
C-peptide, and proinsulin levels increased just before termination of
the fast, and we have therefore decided that "the end of the fast"
refers to the values of these variables measured after 69 h of
fast, where nadir in almost every subjects was obtained.
|
Assays.
All blood samples were centrifuged immediately at 4°C and stored at
20°C until analysis. Care was taken to analyze samples from each
patient within the same run. All samples were analyzed in duplicate.
The only exception was GH, which due to a very low intra-assay
coefficient of variation (CV) was analyzed in single determinations
only. Plasma glucose was measured by the glucose dehydrogenase
oxidation method. Intra- and interassay CVs were lower than 2%. Serum
FFA was measured using an enzymatic colorimetric procedure with a
"NEFA C" kit (Wako Chemicals, Neuss, Germany). Serum insulin,
proinsulin, and C-peptide were measured by a noncompetitive time-resolved immunofluorometric assay (TR-IFMA; Wallac Oy, Turku, Finland) as described previously (41). Detection limits
for serum insulin, C-peptide, and proinsulin were 9, 18 and 2 pmol/l, respectively. Intra- and interassay CVs were 5 and 7% for insulin, 9 and 8% for proinsulin, and 10 and 8% for C-peptide. In the insulin assay, cross-reactivities with proinsulin, C-peptide, and 32-33 split proinsulin were all <0.4%. Serum GH was measured using a commercial noncompetitive TR-IFMA (Wallac Oy; see Ref.
18). Intra- and interassay CVs were <5%. Pancreatic
glucagon concentrations were measured in ethanol-extracted plasma
(final concentration 70% vol/vol) by an RIA using antiserum 4305, 125I-labeled porcine glucagon (a generous gift from Novo
Nordisk, Bagsvaerd, Denmark), and synthetic human glucagon (Peninsula, Merseyside, St. Helens, UK) as standards. The detection limit was ~1
pmol/l, the intra-assay CV was <5% at a level of 15 pmol/l (25), and the interassay CV was below 11%. Serum cortisol
was measured by RIA (Orion Diagnostics, Espoo, Finland). The detection limit was 5 nmol/l, and the intra-assay CV was <3%. Plasma
epinephrine and norepinephrine were measured by a radioenzymatic assay
with some modifications as described previously (29).
Intra-assay CVs for norepinephrine and epinephrine in samples
containing basal values were 6 and 8%, respectively. Corresponding
values of interassay CVs for norepinephrine and epinephrine were 7 and
11%, respectively.
Statistical analysis.
All statistical analyses were performed using SPSS/PC + 5.0. Nonparametric methods were used, as almost all variables were nonnormally distributed when tested by normal plot and the Shapiro-Wilk test. Mann-Whitney rank sum test was used to compare unpaired data,
whereas the Wilcoxon matched-pairs signed-ranks test was used in
evaluating differences between paired observations. Chi square test was
used to compare categorical data. The Sign test was used to test the
proportions of subjects having at least a 10% increase in each of the
-cell polypeptides from 69 to 72 h of fast. To reduce type 1 errors, 24-h mean concentrations of all variables for every day (1st
day, 2nd day, and 3rd day) were calculated. The 24-h mean levels (from
8:30 AM to 5:30 AM every day) were used to examine variables with a
circadian variation and effects of gender and BMI. Data are presented
as median and 5th and 95th percentiles defining 90% reference
intervals. Data are displayed graphically by plots of median values
with their reference intervals as error bars. This will of course hide
any variation in the shape of the curves for different individuals. Data for median curves that are not good indicators of the typical curve are discussed in the text. The impact of gender and BMI is
displayed as median curves in cases where a significant effect was
observed. The catecholamines are presented as means ± SE, but
data are tested by nonparametric tests. P values <0.05 were considered statistically significant.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Plasma glucose.
Plasma glucose decreased steadily from 5.2 (4.3-6.2) (median and
5th and 95th percentiles) mmol/l and reached a plateau after ~2 days
of fasting from which it remained constant at 3.7 (3.0-4.4) mmol/l
throughout the fast (P < 0.001; Fig.
1). During the entire fast, the median
level of plasma glucose was higher in males than in females (Fig. 1),
but differences in 24-h mean levels were only significant during the
last 48 h (day 2, P = 0.02 and day 3, P = 0.01). No significant differences in 24-h mean
levels of plasma glucose between subjects with normal or high BMI were
observed. The individual curves showed only minor variations compared
with the median curve. The lowest plasma glucose value measured was 2.7 mmol/l. The lowest individual values were observed between 18 and
72 h of fast (3.3-4.4 mmol/l for males, 2.7-4.6 mmol/l for females, 2.7-4.2 mmol/l for subjects with normal BMI, and 3.0-4.6 mmol/l for subjects with high BMI). Age correlated
positively with plasma glucose at the start (r = 0.41, P < 0.05) and at the end (r = 0.36, P < 0.05) of the fast but with none of the other variables measured.
|
-Cell polypeptides in subjects with plasma glucose
3.3 mmol/l.
In the period from 18 to 69 h of fast (excluding 72-h samples), we
observed 69 measurements of plasma glucose
3.3 mmol/l in 10 subjects
(30%). The corresponding values of insulin, C-peptide, and proinsulin
were
26, 249, and 10 pmol/l, respectively (Fig. 2). From 63 to 69 h of fast, the
corresponding
-cell polypeptides were all suppressed below the
recently published diagnostic criteria for hyperinsulinemia: insulin
18 pmol/l; C-peptide
200 pmol/l; and proinsulin
5 pmol/l
(38). From 18 to 60 h of fast, we observed 52 measurements of plasma glucose
3.3 mmol/l in nine subjects (27%; all
females). In 20 out of these 52 cases (38%; 6 out of 9 females), the
diagnostic criteria for hyperinsulinemia were exceeded by at least one
of the corresponding
-cell polypeptides. During the entire fast,
35% (6/17) of the healthy females had plasma glucose
3.3 mmol/l
concomitant with
-cell polypeptides, exceeding the diagnostic
criteria for hyperinsulinemia.
|
Serum insulin, proinsulin, and C-peptide.
For all three hormones, there was a marked decline during the first
30 h of fast, after which a more modest fall during the rest of
the fast was observed. The concentration of serum insulin decreased
from 33 (16-69) to 9 (<9-19) pmol/l at the end of the fast
(P < 0.001; Fig. 3). The
levels of serum proinsulin fell from 6 (3-17) to 3 (2-5)
pmol/l (P < 0.001; Fig. 3), and serum C-peptide
decreased from 421 () to 127 () pmol/l
(P < 0.001; Fig. 3). After comparison of 24-h mean
levels, no significant gender differences in circulating levels of
serum insulin, proinsulin, and C-peptide were observed, although the
median levels of C-peptide almost throughout the fast were higher in
males than in females (Fig. 4).
Twenty-four hour mean levels of C-peptide were significantly higher in
subjects with high BMI every day during the fast (P < 0.05). Insulin levels were also significantly higher in subjects with high BMI on the first day (P = 0.003) and the
second day (P = 0.03) and tended to be higher on the
third day (P = 0.06). Although the median values of
proinsulin at separate points of time were higher in subjects with high
BMI (Fig. 4), statistical significance comparing 24-h mean levels was
not reached on any day.
|
|
Plasma glucagon, serum cortisol, and serum GH.
Plasma glucagon increased gradually throughout the fast from 10 (5-17) to 23 (10-38) pmol/l after 72 h of fast
(P < 0.001; Fig. 5). The
24-h mean levels of plasma glucagon were higher in males on the first
day (P = 0.04), second day (P = 0.06),
and third day (P = 0.05). Glucagon median levels were
higher in subjects with normal BMI during the last half of the fast
(Fig. 5), but when comparing 24-h mean levels no significant difference
was observed on any day.
|
|
Plasma epinephrine and norepinephrine.
As for cortisol, both plasma epinephrine and norepinephrine
concentrations fluctuated in a circadian rhythm with increases at about
5:30-8:30 PM and decreases at about 2:30-5:30 AM every day
(Fig. 7). The 24-h mean levels of plasma
epinephrine increased significantly from 8.4 ± 3.7 (SE) ng/l the
first day to 15.2 ± 6.5 ng/l the second day (P = 0.005) and to 20.9 ± 6.7 ng/l the third day (P < 0.001). Likewise, the 24-h mean levels of norepinephrine rose
significantly from 115.6 ± 30.6 ng/l the first day to 130.6 ± 32.0 ng/l the second day (P = 0.03) and to
178.3 ± 37.6 ng/l the third day (P = 0.004).
|
Serum FFA and albumin.
Serum FFA increased gradually from 0.47 (0.20-0.81) mmol/l to a
final level of 1.46 (0.78-2.39) mmol/l after 72 h of fasting (P < 0.001; Fig. 8),
indicating that no food was consumed. FFA decreased markedly between
8:30 PM and 8:30 AM every night but returned to the expected increasing
levels in the mornings. Compared with the values at 8:30 PM every day,
69, 91, and 71% of the subjects experienced a decrease in FFA levels
at 2:30 AM on the first, second, and third day, respectively (all
P < 0.05). After correcting for changes in albumin
levels, 63, 88, and 67% of the subjects experienced a decrease in FFA
levels on the first, second, and third day, respectively, which was
significant on the second day only (P < 0.05). The
median levels of serum FFA were higher in females than in males during
the entire fast (Fig. 8), but, when comparing 24-h mean levels,
statistical significance was observed on the first and the third day
(P = 0.004 and P = 0.01). We observed no significant differences in the 24-h mean levels of FFA between subjects with normal or high BMI, although the median levels were higher in subjects with normal BMI during the last two-thirds of the
fast.
|
Increase in C-peptide, insulin, proinsulin, and glucose in the last
sample.
As seen in Fig. 3, most subjects experienced an increase in serum
C-peptide, insulin, and proinsulin from 69 to 72 h of fast. Demanding an increase of >10% (to ensure increases higher than the
CVs for these assays), the proportions of subjects with increasing levels of C-peptide, insulin, and proinsulin were 97, 87, and 81%,
respectively, and, for all three -cell polypeptides, this increase
was significant (P < 0.001). In contrast, plasma
glucose only increased >10% in 32% of the subjects, and this was not significant.
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The physiological mechanisms that prevent insulin-induced hypoglycemia and hypoglycemia during prolonged fasting have been studied and reviewed intensively (6, 7, 11, 13, 14, 16, 17, 28, 30, 31). Recently, a diagnostic interpretation of data obtained at the end of a 72-h fast has been suggested to distinguish between normal subjects and patients with hyperinsulinemic hypoglycemic disorders (27, 33, 36, 40). However, to our knowledge, well-defined reference values for most of the known glucose regulatory factors in healthy subjects have never been established. The present study was carried out to provide reference intervals for the most important factors contributing to glucose homeostasis during a classic 72-h fast and to study the impact of gender and BMI.
In the current study, plasma glucose, serum insulin, C-peptide, and
proinsulin all decreased as reported in other studies (4, 7, 27,
31, 33, 40), and, characteristically for all, a pronounced
initial decline during the first day of fast was followed by a more
modest fall during the rest of the fast, which is in agreement with a
previous study (31). The reference intervals of C-peptide,
insulin, and proinsulin narrowed markedly during the last 9 h of
the fast, and in most subjects serum levels of insulin and proinsulin
reached the detection limit of the assays. This implies that the
secretion of -cell polypeptides essentially ceased in most subjects.
In addition, we found glucose levels to be significantly lower in
females than in males during the last 48 h of the fast, also
reported previously (31, 40). We therefore expected higher
levels of
-cell polypeptides in males during the last part of the
fast (40). However, we found no effect of gender on the
levels of
-cell polypeptides, as reported earlier for insulin
(31).
-Cell polypeptides rapidly reached baseline
levels in both sexes, probably eliminating any difference in the last
part of the fast. Contrary to a recently published study
(40), we observed significantly higher levels of C-peptide and insulin throughout the fast in subjects with high BMI. Our data
further establish that the well-described insulin resistance in obese
individuals is preserved even after prolonged fasting.
The main purpose of a 72-h fast is to demonstrate endogenous
hyperinsulinism. In a series of studies, it was demonstrated that
diagnostic problems primarily exist in the plasma glucose range of
2.8-3.3 mmol/l (27, 33, 40), in which some subjects experience hypoglycemic symptoms and stop the fast (35,
36). In this glucose range, the diagnostic accuracy of insulin,
C-peptide, and proinsulin to discriminate between normal subjects and
patients with insulinoma was found to be <100% for all hormones. In
the period of 18-60 h of fast, as many as 35% of the healthy
females with plasma glucose 3.3 mmol/l in our study had
-cell
polypeptides exceeding the recently published diagnostic criteria for
hyperinsulinemia (36, 38), whereas from 63 to 69 h of
fast none out of the 10 subjects with plasma glucose
3.3 mmol/l
exceeded these diagnostic criteria. Our data suggest that, in females
who terminate the fast before 63 h have passed, the recently
published diagnostic criteria for hyperinsulinemic hypoglycemia should
be used with some caution (27, 33, 36, 38, 40). Among 178 patients with confirmed insulinoma, only one completed the 72-h fast
without fulfilling the diagnostic criteria for hyperinsulinemia
(37). This observation together with our data indicates
that the greatest problem must be not to overdiagnose hyperinsulinemic
hypoglycemia in females stopping the fast before 72 h had passed.
Contrary to the healthy subjects investigated in the studies from which the criteria for hyperinsulinemia are derived (27, 33,
40), the healthy subjects in our study were not referred for
evaluation of the possible presence of a hypoglycemic disorder, which
to some extent renders our data less biased. The reference intervals for insulin, C-peptide, and proinsulin could be useful not only in the
diagnosis of hyperinsulinemic hypoglycemia but also in the
investigation of other abnormalities in glucose homeostasis in which
the secretion of
-cell polypeptides could be of interest.
Glucagon and catecholamines play a primary role in the first line defense against hypoglycemia (5, 8, 11, 12, 17, 30). As expected, prolonged fasting induced a marked and gradual increase in the levels of pancreatic glucagon as reported earlier (4, 7, 31). Plasma glucagon levels were lower in females throughout the fast, and this finding contrasts to previous observations (31). This discrepancy must be due to methodological differences. In the study of Merimee and Fineberg (31), glucagon was measured using the 30K assay, in which "big plasma glucagon" is unidentified in contrast to the assay we used (42). Elevations of FFA within the physiological range have been shown to have a significant inhibitory effect on glucagon secretion in humans even during arginine-induced hypoglycemia (20), although the latter are more controversial (2, 22). Therefore, the higher FFA levels in women in our study can explain the lower but still increasing secretion of glucagon during fasting. Comparing the 24-h mean levels, we demonstrated a statistically significant increase of epinephrine and norepinephrine during the fast, as expected from earlier studies on the effect of fasting and on insulin-induced hypoglycemia (4, 7). The measurements of catecholamines were done to verify an increasing secretion and a circadian rhythm with nocturnal troughs. Because of the small number of subjects, reference intervals have not been calculated. In a previous fasting study, the circadian variation of norepinephrine and epinephrine was not demonstrated, probably because blood samples were taken only two times a day, at 8:30 AM and 8:30 PM (4). The blunted response of catecholamines at sleep is an important observation because it will significantly increase the risk of hypoglycemia in subjects in whom the glucagon response to hypoglycemia is absent. This phenomen is in fact believed to be responsible for the greater risk of hypoglycemic episodes at night in subjects with insulin-treated diabetes (26). In our study, the circadian variation and increase of both catecholamines were similar, suggesting a role for both hormones in the glucose counterregulation. Defects in catecholamine and glucagon release as a cause of hypoglycemia are very seldomly reported (9, 36), except in long-term diabetes. The presented reference intervals for glucagon are very broad, and therefore the applicability is limited to some extent. However, in cases of spontaneous hypoglycemia where hyperinsulinism can be excluded, a low glucagon response may point out an abnormality in the counterregulatory response of this hormone, which can then be further investigated by other means.
Cortisol and GH play a demonstrable role as counterregulatory hormones only in the defense against prolonged hypoglycemia (6, 13, 14, 30). In agreement with these studies and with observations during prolonged fasting (4, 7, 34), we found the 24-h mean values of cortisol to increase during the fast. In addition to this increase, there was an underlying circadian variation with peaks in the morning and nocturnal falls. As reported earlier, the cortisol levels did not increase significantly from the second to the third day of fasting (4), suggesting that the counterregulatory response of cortisol might not increase further beyond 72 h of fast. In contrast to the findings of Galvao et al. (19), high BMI did not diminish the day-to-night variation of cortisol levels in the present study. For cortisol, no effects of gender or BMI were found. In studies of the effect of prolonged fasting, significant increases in the 24-h mean levels of GH and in the amplitude and frequency of GH secretory bursts have been demonstrated (21, 23, 24). This is consistent with the GH data from our study, of which the GH levels were highly individual and highly episodic but with increasing 24-h mean levels from the first to the second day and from the first to the third day but actually decreasing from the second to the third day, exactly as reported in previous studies (4, 34). Apparently, the counterregulatory response of GH is attenuated with fasting beyond 2 days. In contrast to cortisol and catecholamines, the pattern of GH secretion was highly episodic, and, as stated before (21), it would be necessary to sample at least every 5 min to catch the pulses of this hormone. As reported earlier, we found no gender effect on the GH levels (4), whereas high BMI was accompanied by statistically significant lower levels of GH in the last day of the fast as previously reported (18, 19). Hypopituitarism and adrenal insufficiency are the common hormonal causes of hypoglycemia (10, 36). Especially in Addison's disease, other characteristic clinical features may be entirely absent. The applicability of the reference intervals for cortisol seems obvious, whereas that for GH seems to be of rather limited value in the light of the large individual variation.
FFA increased gradually during the entire fast as reported in other studies (7, 21, 23, 24, 31, 33), confirming that the subjects abided by the rules of the fast. However, we found unexpected reductions in FFA levels at 2:30 AM. These decrements were unexpected because single GH pulses have been reported to increase FFA levels typically after 2-3 h and typically after the nocturnal mean peak of GH (32). In the current study, we observed a nocturnal fall in catecholamine levels, and reduced catecholamine-mediated lipolysis may explain the concomitant nocturnal decline in FFA levels (16, 17). However, this hypothesis needs to be verified. The median levels of FFA were significantly higher in females on the first and the third day, in agreement with earlier reports (31), and there was no significant effect of high BMI on FFA levels. Patients with insulinomas have been reported to have significantly lower fasting levels of FFA than healthy subjects (33), but, from our data, it seems likely that this difference was obtained mainly because the FFA levels in healthy subjects after 72 h of fast were compared with levels of FFA in patients with insulinoma measured at the time they stopped the fast. Future studies may show whether the presented reference intervals for FFA can uncover diseases of glucose homeostasis.
Although the volunteers in our study were kept under continuous
surveillance during the entire fast, rested in their beds from 10:00 PM
to 8:30 AM the last day, and experienced a continuing increase in FFA
and glucagon levels, the levels of insulin, C-peptide, and proinsulin
increased significantly from 69 to 72 h of fasting. In another
study of prolonged fasting, this phenomenon was not observed
(4). This discrepancy could be explained by the fact that
their last samples were drawn 6 h before termination of the fast.
In studies of Service et al. (27, 33, 40), this was not
observed either. For insulin, this could be ascribed to the higher
detection limit of 35 pmol/l (22, 28, 31) vs. 9 pmol/l in
our study, whereas this is not the case with C-peptide. However, it is
well known that the anticipation of food intake stimulates gastric
secretion by stimulating vagal efferents (cephalic phase), and
accompanying increases in certain gastrointestinal hormones together
with stimulation of the vagal efferents to the pancreatic -cells may
explain this phenomenon (3). However, this warrants further investigations. Prolonged fasting in most cases is conducted to
establish the mechanisms for hypoglycemia. In this respect,
-cell
polypeptides are expected to be maximally suppressed at the end of a
72-h fast. However, our data suggest that the levels of
-cell
polypeptides at the terminal sampling time point should be interpreted
with some caution.
The strength of this study derives from the fact that the
presented reference intervals are derived from healthy subjects without
any previous hypoglycemic events, and, to our knowledge, this is the
largest study simultaneously measuring the pancreatic -cell response
and the counterregulatory hormone response to prolonged fasting.
However, the sample size is small with respect to the limits of the
reference intervals, and the generalization of these data to the
population at large should be done with caution. In the clinical
settings, the reference intervals should therefore be used with common
sense, and type 1 error risk could be reduced by sampling every 6 h instead of every 3 h. The relevance of measuring counterregulatory factors comes in question when hormone deficiencies are suspected. If blood is drawn for these analyses, the levels can be
determined afterward in subjects who terminate the fast before 72 h of fast and have plasma glucose in the hypoglycemic range (
3.3
mmol/l). Furthermore, it could be of notable scientific value to
compare the counterregulatory response of healthy subjects with the
response in patients with verified hypoglycemic diseases and subjects
with postprandial hypoglycemic symptoms, first-degree relatives of
diabetic subjects, etc.
In summary, the present study has provided reference intervals for the
-cell polypeptides and the most important glucose counterregulatory
factors during a classic 72-h fast. The secretion of
-cell
polypeptides essentially ceased in most subjects. Gender had no effect
on the secretion of
-cell polypeptides during prolonged fasting but
resulted in higher glucose levels, higher glucagon levels, and lower
FFA levels in males. High BMI resulted in higher insulin and C-peptide
levels and lower GH levels during the last day of fasting. These
results might be helpful in the future discrimination between different
hypoglycemic disorders, such as insulinoma, adult nesidioblastosis,
hormone deficiencies, and noninsulinoma pancreatogenous hypoglycemia
(39), and may be in yet unknown hypoglycemic disorders.
![]() |
FOOTNOTES |
---|
Address for reprint requests and other correspondence: K. Højlund, Dept. of Endocrinology, Odense University Hospital, Kloevervaenget 6, DK-5000 Odense C, Denmark (E-mail: k.hojlund{at}dadlnet.dk).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 5 July 2000; accepted in final form 8 September 2000.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
1.
Amiel, SA,
Maran A,
Powrie JK,
Umpleby AM,
and
Macdonald IA.
Gender differences in counterregulation to hypoglycaemia.
Diabetologia
36:
460-464,
1993[ISI][Medline].
2.
Andrews, SS,
Lopez-S A,
and
Blackard WG.
Effects of lipids on glucagon secretion in man.
Metabolism
24:
35-44,
1975[ISI][Medline].
3.
Becker, KL.
Principles and Practice of Endocrinology and Metabolism (2nd ed). Philadelphia, PA: Lippincott-Raven, 1995.
4.
Beer, SF,
Bircham PM,
Bloom SR,
Clark PM,
Hales CN,
Hughes CM,
Jones CT,
Marsh DR,
Raggatt PR,
and
Findlay AL.
The effect of a 72-h fast on plasma levels of pituitary, adrenal, thyroid, pancreatic and gastrointestinal hormones in healthy men and women.
J Endocrinol
120:
337-350,
1989[Abstract].
5.
Bolli, GB.
Importance of catecholamines in defense against insulin hypoglycemia in humans.
Adv Pharmacol
42:
627-630,
1998[Medline].
6.
Boyle, PJ,
and
Cryer PE.
Growth hormone, cortisol, or both are involved in defense against, but are not critical to recovery from, hypoglycemia.
Am J Physiol Endocrinol Metab
260:
E395-E402,
1991
7.
Boyle, PJ,
Shah SD,
and
Cryer PE.
Insulin, glucagon, and catecholamines in prevention of hypoglycemia during fasting.
Am J Physiol Endocrinol Metab
256:
E651-E661,
1989
8.
Cryer, PE.
Glucose counterregulation: prevention and correction of hypoglycemia in humans.
Am J Physiol Endocrinol Metab
264:
E149-E155,
1993
9.
Cryer, PE.
Role of growth hormone in glucose counterregulation.
Horm Res
46:
192-194,
1996[ISI][Medline].
10.
Cryer, PE.
Hypoglycemia: Pathophysiology, Diagnosis and Treatment. New York: Oxford Univ Press, 1997.
11.
Cryer, PE,
Tse TF,
Clutter WE,
and
Shah SD.
Roles of glucagon and epinephrine in hypoglycemic and nonhypoglycemic glucose counterregulation in humans.
Am J Physiol Endocrinol Metab
247:
E198-E205,
1984
12.
De Feo, P,
Perriello G,
Torlone E,
Fanelli C,
Ventura MM,
Santeusanio F,
Brunetti P,
Gerich JE,
and
Bolli GB.
Contribution of adrenergic mechanisms to glucose counterregulation in humans.
Am J Physiol Endocrinol Metab
261:
E725-E736,
1991
13.
De Feo, P,
Perriello G,
Torlone E,
Ventura MM,
Fanelli C,
Santeusanio F,
Brunetti P,
Gerich JE,
and
Bolli GB.
Contribution of cortisol to glucose counterregulation in humans.
Am J Physiol Endocrinol Metab
257:
E35-E42,
1989
14.
De Feo, P,
Perriello G,
Torlone E,
Ventura MM,
Santeusanio F,
Brunetti P,
Gerich JE,
and
Bolli GB.
Demonstration of a role for growth hormone in glucose counterregulation.
Am J Physiol Endocrinol Metab
256:
E835-E843,
1989
15.
Diamond, MP,
Jones T,
Caprio S,
Hallarman L,
Diamond MC,
Addabbo M,
Tamborlane WV,
and
Sherwin RS.
Gender influences counterregulatory hormone responses to hypoglycemia.
Metabolism
42:
1568-1572,
1993[ISI][Medline].
16.
Fanelli, C,
Calderone S,
Epifano L,
De Vincenzo A,
Modarelli F,
Pampanelli S,
Perriello G,
De Feo P,
Brunetti P,
Gerich JE,
and
Bolli GB.
Demonstration of a critical role for free fatty acids in mediating counterregulatory stimulation of gluconeogenesis and suppression of glucose utilization in humans.
J Clin Invest
92:
1617-1622,
1993[ISI][Medline].
17.
Fanelli, CG,
De Feo P,
Porcellati F,
Perriello G,
Torlone E,
Santeusanio F,
Brunetti P,
and
Bolli GB.
Adrenergic mechanisms contribute to the late phase of hypoglycemic glucose counterregulation in humans by stimulating lipolysis.
J Clin Invest
89:
2005-2013,
1992[ISI][Medline].
18.
Frystyk, J,
Vestbo E,
Skjaerbaek C,
Mogensen CE,
and
Orskov H.
Free insulin-like growth factors in human obesity.
Metabolism
44:
37-44,
1995[ISI][Medline].
19.
Galvao, TA,
Graves L,
Burke CW,
Fotherby K,
and
Fraser R.
Free cortisol in obesity; effect of fasting.
Acta Endocrinol Copenh
81:
321-329,
1976[Medline].
20.
Gerich, JE,
Langlois M,
Schneider V,
Karam JH,
and
Noacco C.
Effects of alterations of plasma free fatty acid levels on pancreatic glucagon secretion in man.
J Clin Invest
53:
1284-1289,
1974[ISI][Medline].
21.
Hartman, ML,
Veldhuis JD,
Johnson ML,
Lee MM,
Alberti KG,
Samojlik E,
and
Thorner MO.
Augmented growth hormone (GH) secretory burst frequency and amplitude mediate enhanced GH secretion during a two-day fast in normal men.
J Clin Endocrinol Metab
74:
757-765,
1992[Abstract].
22.
Hicks, HB,
Taylor CI,
Vij SK,
Pek S,
Knopf RF,
Floyd Jr JC,
and
Fajans SS.
Effect of changes in plasma levels of free fatty acids on plasma glucagon, insulin and growth hormone in man.
Metabolism
26:
1011-1023,
1977[ISI][Medline].
23.
Ho, KY,
Veldhuis JD,
Johnson ML,
Furlanetto R,
Evans WS,
Alberti KG,
and
Thorner MO.
Fasting enhances growth hormone secretion and amplifies the complex rhythms of growth hormone secretion in man.
J Clin Invest
81:
968-975,
1988[ISI][Medline].
24.
Ho, PJ,
Friberg RD,
and
Barkan AL.
Regulation of pulsatile growth hormone secretion by fasting in normal subjects and patients with acromegaly.
J Clin Endocrinol Metab
75:
812-819,
1992[Abstract].
25.
Holst, JJ.
Evidence that glicentin contains the entire sequence of glucagon.
Biochem J
187:
337-343,
1980[ISI][Medline].
26.
Jones, TW,
Porter P,
Sherwin RS,
Davis EA,
O'Leary P,
Frazer F,
Byrne G,
Stick S,
and
Tamborlane WV.
Decreased epinephrine responses to hypoglycemia during sleep.
N Engl J Med
338:
1657-1662,
1998
27.
Kao, PC,
Taylor RL,
and
Service FJ.
Proinsulin by immunochemiluminometric assay for the diagnosis of insulinoma.
J Clin Endocrinol Metab
78:
1048-1051,
1994[Abstract].
28.
Klein, S,
Holland OB,
and
Wolfe RR.
Importance of blood glucose concentration in regulating lipolysis during fasting in humans.
Am J Physiol Endocrinol Metab
258:
E32-E39,
1990
29.
Knudsen, JH,
Christensen NJ,
and
Bratholm P.
Lymphocyte norepinephrine and epinephrine, but not plasma catecholamines predict lymphocyte cAMP production.
Life Sci
59:
639-647,
1996[ISI][Medline].
30.
Lins, PE,
Adamson U,
Clausen N,
Hamberger B,
and
Efendic S.
The role of glucagon, catecholamines and cortisol in counterregulation of insulin-induced hypoglycemia in normal man.
Acta Med Scand
220:
39-46,
1986[ISI][Medline].
31.
Merimee, TJ,
and
Fineberg SE.
Homeostasis during fasting. II. Hormone substrate differences between men and women.
J Clin Endocrinol Metab
37:
698-702,
1973[ISI][Medline].
32.
Moller, N,
Jorgensen JO,
Moller J,
Orskov L,
Ovesen P,
Schmitz O,
Christiansen JS,
and
Orskov H.
Metabolic effects of growth hormone in humans.
Metabolism
44:
33-36,
1995[ISI][Medline].
33.
O'Brien, T,
O'Brien PC,
and
Service FJ.
Insulin surrogates in insulinoma.
J Clin Endocrinol Metab
77:
448-451,
1993[Abstract].
34.
Palmblad, J,
Levi L,
Burger A,
Melander A,
Westgren U,
von SH,
and
Skude G.
Effects of total energy withdrawal (fasting) on the levels of growth hormone, thyrotropin, cortisol, adrenaline, noradrenaline, T4, T3, and rT3 in healthy males.
Acta Med Scand
201:
15-22,
1977[ISI][Medline].
35.
Schwartz, NS,
Clutter WE,
Shah SD,
and
Cryer PE.
Glycemic thresholds for activation of glucose counterregulatory systems are higher than the threshold for symptoms.
J Clin Invest
79:
777-781,
1987[ISI][Medline].
36.
Service, FJ.
Hypoglycemic disorders.
N Engl J Med
332:
1144-1152,
1995
37.
Service, FJ.
Classification of hypoglycemic disorders.
Endocrinol Metab Clin North Am
28:
501-518,
1999[ISI][Medline].
38.
Service, FJ.
Diagnostic approach to adults with hypoglycemic disorders.
Endocrinol Metab Clin North Am
28:
519-532,
1999[ISI][Medline].
39.
Service, FJ,
Natt N,
Thompson GB,
Grant CS,
van Herden JA,
Andrews JC,
Lorenz E,
Terzic A,
and
Lloyd RV.
Noninsulinoma pancreatogenous hypoglycemia: a novel syndrome of hyperinsulinemic hypoglycemia in adults independent of mutations in Kir6.2 and SUR1 genes.
J Clin Endocrinol Metab
84:
1582-1589,
1999
40.
Service, FJ,
O'Brien PC,
McMahon MM,
and
Kao PC.
C-peptide during the prolonged fast in insulinoma.
J Clin Endocrinol Metab
76:
655-659,
1993[Abstract].
41.
Toivonen, E,
Hemmilä I,
Marniemi J,
Jørgensen PN,
Zeuthen J,
and
Lövgren T.
Two-site time-resolved immunofluorometric assay of human insulin.
Clin Chem
32:
637-640,
1986
42.
Von Schenck, H,
and
Nilsson OR.
Radioimmunoassay of extracted glucagon compared with three non-extraction assays.
Clin Chim Acta
109:
183-191,
1981[ISI][Medline].