| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Division of Internal Medicine, Karolinska Institute, Danderyd Hospital, S-182 88 Danderyd, Sweden
Address all correspondence and requests for reprints to: Dr. Lena Landstedt-Hallin, Division of Internal Medicine, Karolinska Institute, Danderyd Hospital, S-182 88 Danderyd, Sweden.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-cell responsiveness to hypoglycemia in these
patients. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Type 2 diabetes is a progressive disease and the use of more than one
glucose-lowering agent is often needed to attain glycemic goals (9).
Patients with secondary failure to oral treatment with sulfonylurea
(SU) derivatives are frequently treated with combined insulin and SU
therapy because the combination makes it possible to improve metabolic
control while using smaller amounts of exogenous insulin than with
insulin therapy alone (10). One reason for this is improved
insulinization of the liver by the higher portal insulin levels
achieved, leading to a reduction of hepatic glucose production (11).
However, it has been proposed that intraislet mechanisms regulate the
release of glucagon and that high intraislet insulin concentrations may
inhibit glucagon release in response to low blood glucose (12). SU
derivatives stimulate the ß-cell, not only during hyperglycemia but
also in normo- and hypoglycemic conditions, thus increasing the
intraislet insulin levels, which, in turn, may affect the -cell
responsiveness to hypoglycemia.
To test the hypothesis that the glucagon response during hypoglycemia is affected by SU in patients with type 2 diabetes, we performed a study on such patients, all of them treated with a combination of insulin and glibenclamide, the most commonly used SU. We induced hypoglycemia using the insulin-glucose clamp technique and studied the glucagon and epinephrine responses in the presence and absence of glibenclamide.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Thirteen patients (3 women and 10 men) with type 2 diabetes (mean age, 57 yr; range, 48–63), a mean body mass index of 28.2 kg/m2 (range, 23.1–33.3), and a known diabetes duration of 9 yr (range, 2–17) were studied on 2 occasions. All patients had been treated with a combination of insulin (10–48 U/day) and oral glibenclamide (10.5 mg/day) for at least 8 months (median, 18 months), and their mean hemoglobin A1c (HbA1c) was 7.6% (range, 6.6–8.5%; upper reference, 5.2%). None of the patients had signs of autonomic neuropathy. The study protocol was approved by the local ethics committee at the Karolinska Institute, and all patients signed informed consent forms.
Study design
Two weeks before the first experiment the patients stopped taking oral glibenclamide. They performed frequent self-monitoring of blood glucose and, when needed, subsequently increased their insulin dose to maintain glycemic control. After this glibenclamide wash-out period, the first hypoglycemic clamp (-GLIB) was performed. All patients then resumed combined glibenclamide and insulin therapy, and after 6 months they participated in the second hypoglycemic clamp (+GLIB). HbA1c did not differ significantly between the clamps (mean ± SD HbA1c at -GLIB, 7.6 ± 0.6%; at +GLIB, 7.5 ± 0.9%).
The patients were instructed to perform home-monitoring of blood glucose at least four times daily during the week before each test and to contact the investigator if they had hypoglycemic symptoms and/or blood glucose values below 3.5 mmol/L during that week. If hypoglycemia occurred within the preceding 72 h of the clamp, the experiment was postponed at least a week. The day before the experiment the patients injected small doses of pre-meal regular insulin, instead of their ordinary mixed insulin (30% regular and 70% NPH) to avoid hypoglycemia. During that day all patients had at least one telephone contact with the investigator, when they reported home-monitoring glucose values and received instructions concerning insulin dosage. The patients were also asked to eat a small snack in the evening before the clamp to avoid nocturnal hypoglycemia.
Clamp procedure
Nonstepwise hypoglycemic clamps were performed, aimed at reaching a stable hypoglycemia with a target nadir of 2.6 mmol/L (47 mg/dL) during the last hour of the experiments. The patients were admitted to the metabolic ward at 0700 h after an overnight fast and without taking their morning medication. On arrival on the morning of the second clamp (+GLIB), glibenclamide (7 mg) was given orally at least 1 h before the point defined as time zero. Three iv cannulas were inserted, one in each antecubital vein and one, retrogradely, in a dorsal vein of the nondominant hand. This hand was placed in a hot box where circulating warm air (55 C) heated the limb to arterialize venous blood (13). A constant infusion of regular insulin (0.144 U/kg BW per h) was started in the opposite arm. The time point when blood glucose reached approximately 6.0 mmol/L (108 mg/dL) was defined as time zero, and the experiment was continued for 150 min from that point. Blood was drawn every third minute from the cannula in the hot box for analysis of arterialized blood glucose, and the glucose level was adjusted with a variable infusion of 20% glucose, matching the rate of fall of blood glucose for the two clamps between individuals as well as between the two experiments for each individual. Samples for free plasma insulin, C peptide, epinephrine, and glucagon measurements were drawn from the antecubital vein at 0, 90, and 150 min, and additional samples for glucagon and epinephrine were drawn at 60 min. A serum sample for analysis of glibenclamide was taken at time zero.
Analytical methods
During the experiments blood glucose was analyzed every 3 min using a glucose oxidase method (YSI, Inc., Yellow Springs, OH). After precipitation of antibody-bound insulin using 25% polyethylene glycol, plasma free insulin was analyzed by RIA (Pharmacia Diagnostics AB, Uppsala, Sweden). A proteinase inhibitor (aprotinin) was added to the serum samples for C peptide determination, and analyses were made using an enzyme-linked immunosorbent assay (DAKO Corp., Carpenteria, CA). Plasma epinephrine was analyzed by high performance liquid chromatography with electrochemical detection (14); plasma glucagon was analyzed by RIA (Euro-Diagnostica AB, Malmo, Sweden). Glibenclamide was analyzed at the Hoechst-Marion-Roussel Laboratory by high performance liquid chromatography.
Statistical methods
All results are expressed as the mean ± SD unless otherwise noted. The area under the curve (AUC) was calculated using the trapezoidal method. Statistical analyses were performed using statistical JMP package, version 3.1.5 (SAS Institute, Inc., Cary, NC). After validation for normal distribution using Shapiro-Wilk’s test, Student’s paired t test was used for comparison between the two clamps. Parameters that were not normally distributed were tested with Wilcoxon’s signed rank test.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
None of the patients had hypoglycemia during the week before the two tests, and thus all clamps were performed as planned. Fasting blood glucose on arrival at -GLIB was 12.7 ± 2.1 mmol/L (mean ± SD) compared to 11.3 ± 2.9 mmol/L at +GLIB (P = 0.0653). The time from start of the insulin infusion to the time when blood glucose was 6 mmol/L, defined as time zero, was similar between -GLIB and +GLIB (83 ± 37 vs. 77 ± 42 min; P = 0.64), and this time was correlated to the fasting blood glucose level on arrival. Serum glibenclamide was below the detection limit of 20 ng/mL in all patients at the start of -GLIB, whereas the serum level varied between 51–328 ng/mL (median, 165) at the start of +GLIB.
From time zero, the arterialized venous blood glucose concentrations
were well matched and almost identical between patients and between the
two tests for each patient (Fig. 1
). The
mean blood glucose level during the last 60 min was 2.7 ± 0.2
mmol/L during -GLIB and 2.6 ± 0.2 mmol/L during +GLIB. The rate
of glucose infusion was similar during the first hour, before stable
hypoglycemia was reached, but thereafter the need was clearly higher
during +GLIB (Fig. 1). The median total amount of glucose infused was
4.1 g during -GLIB (range, 0–10.2) and 5.8 g during +GLIB
(range, 0–33.7), and this difference did not reach statistical
significance (P = 0.094). However, after hypoglycemia
had been induced, the need for infused glucose was higher when
glibenclamide was present (median, 3.7 vs. 1.3 g;
P = 0.034).
|
The peripheral plasma insulin concentrations remained constant
throughout the clamps, with no significant differences between -GLIB
and +GLIB (Fig. 2). C Peptide values,
reflecting endogenous insulin production and thus mirroring portal
insulin, were significantly higher at all time points during +GLIB
(Fig. 2). C Peptide secretion was suppressed during both clamps when
blood glucose was lowered, but at the end of +GLIB the mean C peptide
level was still almost as high as the mean C peptide level at the start
of -GLIB.
|
Epinephrine increased markedly from very low baseline values, and
neither peak values nor AUC differed significantly between -GLIB and
+GLIB (Fig. 3). The glucagon response was
blunted during the second clamp when glibenclamide was present,
resulting in significantly lower values of peak glucagon (212 ±
89 vs. 159 ± 43 ng/L for -GLIB and +GLIB,
respectively; P = 0.016), total AUC (409 ± 128
vs. 133 ± 101 ng/L·h; P = 0.016), as
well as incremental AUC (133 ± 101 vs. 64 ± 62
ng/L·h; P = 0.011; Fig. 3).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
When arterial blood glucose falls below 4.6 mmol/L, the earliest
physiological response in healthy subjects is the suppression of
insulin secretion (15). With further lowering of blood glucose,
counterregulatory hormones are activated. Among these hormones,
glucagon plays the most important role in the physiological defense
against hypoglycemia (7), as implied by the finding that epinephrine
cannot fully compensate for an absent glucagon response in normal
subjects (8). At least three different mechanisms have been suggested
to regulate glucagon secretion during hypoglycemia. Firstly, glucagon
may be released through direct stimulation of the -cells by low
glucose (16). However, such a mechanism could not be verified in a
study of purified -cells (17). Secondly, autonomic nervous input
stimulates the -cells via sympathetic and parasympathetic nerves as
well as via activation of the adrenal medulla (18). Thirdly, the
glucagon secretion may be regulated via an intraislet, paracrine
mechanism, where glucagon release is tonically inhibited by insulin
secretion from the ß-cells, so that suppression of endogenous insulin
secretion, e.g. by low glucose, leads to less inhibition of
the -cell, by which the release of glucagon is increased. This model
of intraislet interaction was proposed in 1983 (12, 19) after the
anatomy of the microvasculature of the islets of Langerhans had been
described by Bonner-Weir and Orci (20). The intraislet blood flow, from
ß- to -cells, was proposed to create an insulin-glucagon feedback
relationship in favor of insulin’s inhibiting effect on -cell
secretion (12).
A number of studies have shown that a high insulin level inhibits the release of glucagon both in vitro (21) and in vivo (22, 23, 24). The acute glucagon response to arginine was inhibited by insulin infusion in both healthy and diabetic individuals (22), and the magnitude of glucagon and epinephrine responses to mild hypoglycemia was suppressed by hyperinsulinemia in normal subjects (23). Similarly, we have shown that a 4-fold increase in plasma insulin levels reduced the glucagon response to hypoglycemia in healthy individuals (24). In contrast, one investigator reported that the glucagon response to insulin-induced hypoglycemia was independent of the level of both endogenous intraislet and exogenous arterial insulin concentrations in both healthy and diabetic subjects (25). However, the length of exposure to hyperinsulinemia seems to be important, as a brief exposure (0.5 h) to high insulin levels did not alter the counterregulatory response, whereas prolonged hyperinsulinemia (3.5 h) resulted in a selective blunting of the glucagon response to hypoglycemia (26).
When SU is present, stimulating insulin secretion from ß-cells, low blood glucose levels are not able to suppress insulin secretion as effectively as when SU is absent (27), i.e. there is a continued release of insulin despite hypoglycemia, as was also demonstrated in the present study. As systemic hyperinsulinemia, induced by exogenously given insulin, suppresses the glucagon response, it is most likely that intraislet hyperinsulinemia, induced by SU stimulation of the ß-cell, also results in a blunted glucagon response to hypoglycemia.
The effect of SU on the release of glucagon has been studied in both hypoglycemic and nonhypoglycemic conditions. In 1969, Samols and co-workers reported a significant decrease in glucagon when hypoglycemia was induced by a tolbutamide injection whereas glucagon increased after iv insulin in ducks (28), but others argued that this was a species-dependent result, as glucagon is a hormone essential to life in the fowl (29), and a species difference was later confirmed in a Japanese study (30). In studies of healthy volunteers, SU had no significant effect on glucagon release, either basally or that stimulated by arginine (31, 32). However, in patients with type 2 diabetes starting treatment with either SU or insulin, the glucagon response to an arginine infusion after 1–2 months of treatment was markedly reduced in patients treated with SU compared to that in the patients taking insulin (33). Another study showed that both fasting and postprandial glucagon levels were significantly lower after chronic SU therapy compared to glucagon levels after the patients had been withdrawn from the medication (34).
Pfeifer et al. infused tolbutamide in patients with type 2 diabetes and normal subjects and showed that arginine-stimulated glucagon release was suppressed in all subjects when glucose was clamped at the preinfusion level, but not if the blood glucose level was allowed to fall, concluding that SU suppression of glucagon secretion can be masked by falling glucose levels (35). Using a tolbutamide infusion vs. insulin is a method developed to compare portal and peripheral insulin delivery in humans (36), and with this method Peacey et al demonstrated that the glucagon response to hypoglycemia induced by infused tolbutamide was reduced compared to the response during an equivalent hypoglycemia induced by insulin infusion, yielding higher peripheral insulin levels (37). In a recent report from a Dutch group also studying healthy individuals, oral glibenclamide or placebo was given as a single dose before an insulin-induced stepwise hypoglycemic clamp, resulting in less suppressed C peptide levels and reduced glucagon secretion in the glibenclamide experiment compared to the placebo experiment (38).
An impaired glucagon response in the presence of SU derivatives may
also result from a direct effect of SU on the -cell. However,
in vivo, tolbutamide appeared to be a mild stimulant, rather
than a suppressant, of glucagon secretion in a study of patients with
type 1 diabetes and no C peptide response, i.e. in absence
of detectable insulin secretory capacity (39).
There is a considerable variation between individuals in the glucagon response to hypoglycemia in patients with type 2 diabetes, which is seen both in the present study as well as in reports from earlier investigators (40, 41, 42), where patients with type 2 diabetes were compared to nondiabetic controls. The main results from these studies are somewhat contradictory. Thus, impairment of the counterregulatory responses of one or more hormones have been reported by some (41, 43, 44), whereas others have found that patients with type 2 diabetes have counterregulatory responses similar to those of matched healthy controls (40, 42, 45, 46). However, the methods used to induce hypoglycemia as well as the hypoglycemic challenge differ between these studies. Furthermore, if SU affects the glucagon response to hypoglycemia, as in the present study, then this must be considered when interpreting the results from earlier studies.
In the present study the two experiments were performed in a fixed sequence, and it may be argued that the blunted glucagon response is the result of an order effect. We do not believe that this is the reason for our findings. Firstly, the defective glucagon response seen in patients with type 1 diabetes, a few years after onset of the disease (47), appears to be related to the loss of ß-cell function (48). In our patients there was no indication of a significant loss of endogenous insulin secretion within 6 months. In contrast, the C peptide levels were significantly higher during +GLIB. Furthermore, although type 2 diabetes is a progressive disease with respect to ß-cell function, a marked progress within 6 months is unlikely. In a prospective study, C peptide levels remained stable over an observation period of 4–6 yr in 49 type 2 patients, with age and disease duration similar to those of our patients (49). Secondly, to date no prospective studies showing a defective glucagon response to hypoglycemia in patients with type 2 diabetes have been published.
Two recent studies have reported that glycemic control may be of importance for the hormonal counterregulatory response during hypoglycemia in type 2 diabetes (46, 50). After a considerable improvement in glycemic control, from 11.3% to 8.1% in seven patients, all counterregulatory responses began at significantly lower plasma glucose levels, the epinephrine response was significantly lower, and no significant effect on the magnitude of the glucagon response was seen (50). Similarly, the level of HbA1c correlated to the glucose threshold for epinephrine in 11 patients with a wide range of glycemic control (46). In the present study, the mean HbA1c was stable between the 2 tests, but in individual patients there were changes in both directions, with a median change of -0.5% (range, -1.7 to +1.5%). No correlation was seen between the change in HbA1c and the change in glucagon response between the 2 tests.
Hypoglycemia is not an uncommon problem in type 2 diabetes (6, 51, 52), and severe hypoglycemia during SU treatment is associated with significant morbidity and mortality (53, 54). If the UKPDS results are implemented in the treatment of patients with type 2 diabetes, i.e. aiming for strict glycemic control, the risk for hypoglycemic events will increase considerably. Polypharmacy is often needed to reach and maintain good glycemic control (9), and combining SU and insulin has been established as a useful therapeutic approach for these patients during the last decade. However, our study demonstrates that when glibenclamide is present during hypoglycemia, the glucagon release is blunted, and the need for exogenous glucose is higher. We suggest that the impaired glucagon response is an effect of glibenclamide through stimulation of the ß-cell, resulting in high intraislet insulin levels in relation to the prevailing low glucose. We did not measure glucose kinetics, and we can therefore only speculate on the reasons for the higher glucose infusion needed in the presence of glibenclamide. The peripheral insulin levels in our patients (70–75 mU/L) were probably not high enough to completely suppress hepatic glucose production (55), which has to be taken into consideration when interpreting our results. We believe that the combination of a higher portal insulin level with an impaired glucagon response to hypoglycemia can explain why more glucose had to be infused when glibenclamide was present. A clinical correlate to our findings may be that hypoglycemia associated with SU is often protracted despite treatment with parenteral glucose (28, 53).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received February 12, 1999.
Revised May 19, 1999.
Accepted June 18, 1999.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-cell responses in type 1 diabetes. Diabetes. 47:995–1005.[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  I. Quesada, E. Tuduri, C. Ripoll, and A. Nadal Physiology of the pancreatic {alpha}-cell and glucagon secretion: role in glucose homeostasis and diabetes J. Endocrinol., October 1, 2008; 199(1): 5 - 19. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Gromada, I. Franklin, and C. B. Wollheim {alpha}-Cells of the Endocrine Pancreas: 35 Years of Research but the Enigma Remains Endocr. Rev., February 1, 2007; 28(1): 84 - 116. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. N. Zammitt and B. M. Frier Hypoglycemia in Type 2 Diabetes: Pathophysiology, frequency, and effects of different treatment modalities Diabetes Care, December 1, 2005; 28(12): 2948 - 2961. [Full Text] [PDF]
|
|
|
|
 |
|
 J. Gromada, X. Ma, M. Hoy, K. Bokvist, A. Salehi, P.-O. Berggren, and P. Rorsman ATP-Sensitive K+ Channel-Dependent Regulation of Glucagon Release and Electrical Activity by Glucose in Wild-Type and SUR1-/- Mouse {alpha}-Cells Diabetes, December 1, 2004; 53(suppl_3): S181 - S189. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Zhou, P. O. T. Tran, S. Yang, T. Zhang, E. LeRoy, E. Oseid, and R. P. Robertson Regulation of {alpha}-Cell Function by the {beta}-Cell During Hypoglycemia in Wistar Rats: the "Switch-off" Hypothesis Diabetes, June 1, 2004; 53(6): 1482 - 1487. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. M. Hope, P. O. T. Tran, H. Zhou, E. Oseid, E. Leroy, and R. P. Robertson Regulation of {alpha}-Cell Function by the {beta}-Cell in Isolated Human and Rat Islets Deprived of Glucose: the "Switch-off" Hypothesis Diabetes, June 1, 2004; 53(6): 1488 - 1495. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Ostergard, K. B. Degn, M.-A. Gall, R. D. Carr, J. D. Veldhuis, M. K. Thomsen, R. A. Rizza, and O. Schmitz The Insulin Secretagogues Glibenclamide and Repaglinide Do Not Influence Growth Hormone Secretion in Humans but Stimulate Glucagon Secretion during Profound Insulin Deficiency J. Clin. Endocrinol. Metab., January 1, 2004; 89(1): 297 - 302. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Cejvan, D. H. Coy, J. J. Holst, E. Cerasi, and S. Efendic Gliclazide Directly Inhibits Arginine-Induced Glucagon Release Diabetes, December 1, 2002; 51(90003): S381 - 384. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. McCrimmon, M. L. Evans, R. J. Jacob, X. Fan, Y. Zhu, G. I. Shulman, and R. S. Sherwin AICAR and phlorizin reverse the hypoglycemia-specific defect in glucagon secretion in the diabetic BB rat Am J Physiol Endocrinol Metab, November 1, 2002; 283(5): E1076 - E1083. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Banarer, V. P. McGregor, and P. E. Cryer Intraislet Hyperinsulinemia Prevents the Glucagon Response to Hypoglycemia Despite an Intact Autonomic Response Diabetes, April 1, 2002; 51(4): 958 - 965. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. W.M.T. ter Braak, A. M.M.F. Appelman, I. van der Tweel, D. W. Erkelens, and T. W. van Haeften The Sulfonylurea Glyburide Induces Impairment of Glucagon and Growth Hormone Responses During Mild Insulin-Induced Hypoglycemia Diabetes Care, January 1, 2002; 25(1): 107 - 112. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
