ACTA

Acta Diabetol 28: 162-168, 1991

DIABETOLOGICA 9 Springer-Verlag 1991

Effect of sulphonylurea on glucose-stimulated insulin secretion in healthy and non-insulin dependent diabetic subjects: a dose-response study L. C. Group 1, K. Ratheiser 2, L. Luzi 3, A. Melander 4, D. C. Simonson 2, A. Petrides 3, R. C. Bonadonna 3, E. Widen 1, and R. A. DeFronzo 3 1 Fourth Department of Medicine, Helsinki University Hospital, Helsinki, Finland 2 Yale University School of Medicine, New Haven, Connecticut 3 Diabetes Division, Department of Medicine, University of Texas Health Science Center, Unioninkatu 38, SF-00170 and Audie L. Murphy VA Hospital, San Antonio, Texas, USA 4 Malm6 General Hospital, Department of Clinical Pharmacology, University of Lund, Malm6, Sweden

Abstract. The effect of a rapid-acting sulphonylurea, glipizide, on the dose-response relationship between the /3-cell response (insulin and C-peptide secretion) and the ambient plasma glucose concentration was examined in 12 healthy and 6 non-insulin-dependent diabetic subjects. The subjects participated in two sets of experiments which were performed in random order: (A) four hyperglycaemic clamp studies, during which the plasma glucose concentration was raised for 120 min by 1 (only in healthy subjects), 3, 7, and 17 mmol/1; and (B) the same four hyperglycaemic clamp studies preceded by ingestion of 5 mg glipizide. All subjects participated in a further study, in which glipizide was ingested and the plasma glucose concentration was maintained at the basal level. In control subjects in the absence of glipizide, the firstphase plasma insulin response (0-10 min) increased progressively with increasing plasma glucose concentration up to 10 mmol/1, above which it tended to plateau. Glipizide augmented the first-phase insulin response without changing the slope of the regression line relating plasma insulin to glucose concentrations. The second-phase plasma insulin response (20-120 min) increased linearly with increasing hyperglycaemia (r = 0.997). Glipizide alone increased the plasma insulin response by 180 pmol/1. A similar increase in plasma insulin response following glipizide was observed at each hyperglycaemic step, indicating that glipizide did not affect the sensitivity of the/3-cell to glucose. First-phase insulin secretion was reduced in the type 2 (non-insulin-dependent) diabetic patients, and was not influenced by glipizide. The dose-response curve relating second-phase insulin secretion to the ambient plasma glucose concentration was significantly (P < 0.001) flatter in the diabetic patients than in the control subjects. Glipizide alone increased the plasma insulin response by 60 pmol/1 without changing the slope of the dose-response curve. It is concluded that, in both type 2 diabetic patients and healthy subjects: (A) sulphonylurea augments glucose-stimulated second-phase insulin secretion without changing the sensitivity of the/3-cell to glucose; Offprint requests to: L. C. Groop

(B) first-phase insulin secretion is reduced in non-insulindependent diabetic patients with fasting hyperglycaemia and is not influenced by sulphonylurea.

Key words: Insulin - C-peptide - Glucose - Glipizide Non-insulin-dependent diabetes mellitus

Introduction Sulphonylurea drugs lower blood glucose primarily by stimulating insulin secretion from the pancreatic beta cells [1-8]. During acute exposure, sulphonylureas have been shown to enhance the insulin response to both oral and i.v. glucose [2-7]. In contrast, several studies have failed to show any increase in insulin secretion during chronic sulphonylurea treatment [9-13]. However, in these studies the diabetic subjects were rechallenged with sulphonylurea at a time when their plasma glucose concentrations were significantly lower than in the initial study. When plasma glucose concentrations were raised to the pretreatment level, the insulin response to glucose and other stimuli was significantly enhanced by sulphonylurea therapy [14-16]. It has been suggested that sulphonylureas restore the impaired sensitivity of the /3-cell to glucose [15, 17], and in particular, the reduced first-phase insulin secretion [18, 19] in patients with noninsulin-dependent diabetes mellitus. However, in a recent study we were unable to demonstrate any potentiating effect of either glibenclamide or glipizide on the insulin response to a hyperglycaemic clamp [20]. These discrepant results could be explained if the level of hyperglycaemia had already maximally stimulated insulin secretion, thereby masking a possible potentiating effect of the sulphonylurea. To examine this question in more detail, we studied the effect of glipizide on the dose-response relationship between the plasma glucose concentration and the insulin/ C-peptide response to i.v. glucose in type 2 (non-insulindependent) diabetic patients and non-diabetic subjects.

163

L. C. Groop et al.: Insulinotropic effects of sulphonylurea

Subjects and methods Twelve healthy volunteers (8 males and 4 females; mean age = 28 -I-5 years, BMI = 23.6 +__0.2 kg/m 2) and 6 non-insulin-dependent subjects (4 males, 2 females; mean age=57_5 years; BMI=26.8_ 0.7 kg/m 2) participated in the study. Three of the diabetic patients were controlled with diet alone, and three patients with sulphonylurea. The sulphonylurea-treated patients had been off therapy for 1 month. None of the subjects was taking any medication during the studies. Throughout the study, all subjects were on a weight-maintenance diet containing at least 200 g carbohydrate/day. Before participating, each subject received an explanation of the purpose, nature and potential risks of the study and gave informed written voluntary consent. The study protocol was approved by the local ethical committees.

Experimental protocol The non-diabetic subjects participated in nine and the diabetic subjects in seven experiments performed in random order at 1-2 week intervals: (A) 4 (3 for diabetic patients) hyperglycaemic clamp studies during which the plasma glucose concentration was raised by 1 (non-diabetic subjects only), 3, 7 and 17 mmol/1 for 120 rain; (B) 4 (3) identical hyperglycaemic clamps with the ingestion of 5 mg glipizide (Glucotrol, Roerig Pfizer, New York, NY, USA) 30 min prior to study; (C) plasma glucose clamped at the basal glucose level with prior intake of 5 mg glipizide. All studies were performed at 8 a.m. following an overnight fast of 10-12 h. On the morning of the test, a catheter was inserted into an antecubital vein for infusion of glucose. A second catheter was inserted into a wrist vein for blood sampling, and the hand was placed in a heated box (70 ~ to ensure arterialization of venous blood.

Hyperglycaemic clamp studies without glipizide. Six basal samples were obtained for determination of plasma glucose, insulin and C-peptide, and the plasma glucose concentration was acutely raised by (1), 3, 7 and 17 mmol/1, respectively, and then maintained at the desired hyperglycaemic level for 120 rain with a variable infusion of 20% glucose [21]. During the first 10 min of the hyperglycaemic clamp (first-phase insulin secretion) plasma glucose, insulin and C-peptide concentrations were measured every 2 min. From 10120 rain (second-phase insulin secretion) plasma glucose, insulin and C-peptide concentrations were measured at 5-15 rnin intervals.

Hyperglycaemic clamp with glipizide. Three basal samples were taken for determination of plasma glucose, insulin and C-peptide, 5 mg glipizide was ingested ( - 30 min), and the plasma concentrations of glucose, insulin and C-peptide were measured at -30, - 1 5 and 0 min. At 0 min the plasma glucose concentration was acutely raised by (1), 3, 7 and 17 mmol/1, respectively, and maintained at this level for 120 rain. Plasma glucose, insulin and C-peptide concentrations were measured on samples taken as above. Euglycaemic clamp study with glipizide. Three basal samples were taken for determination of plasma glucose, insulin and C-peptide concentrations, and 5 mg glipizide was ingested. Thereafter, plasma glucose was determined every 5 min and variable glucose infusion was adjusted to maintain euglycaemia. Samples for determination of plasma insulin and C-peptide were obtained at 10 min intervals and those for determination of serum glipizide at 10-15 min intervals. The coefficient of variation in plasma glucose concentrations during the different studies averaged 3.3 +0.4%.

Analytical methods' Plasma glucose concentration was determined by the glucose oxidase method on a Beckman Glucose Analyzer II (Fullerton, Calif., USA). Serum concentrations of glipizide were determined by high-

performance liquid chromatography [22]. The concentrations of insulin and C-peptide in plasma were determined by radioimmunoassay [23, 24]. The sensitivity of the insulin assay was 5 pmol/1 and the interassay coefficient of variation 5%. The sensitivity of the C-peptide assay was 0.02 nmol/1 and the interassay coefficient of variation 6%.

Calculations and statistics The first-phase insulin/C-peptide response was calculated as the weighted mean of the plasma concentrations at 0, 2, 4, 6 and 10 min. The second-phase insulin/C-peptide response was calculated as the weighted mean of the values at 10 rain intervals between 20 and 120 min. The mean glucose infusion rate (M-value) from 60 to 120 min corrected for urinary glucose losses was used as a measure of glucose uptake in the body. This calculation assumes that hepatic glucose production is completely suppressed. To obtain a measure of insulin sensitivity, the M-value was divided by the mean insulin concentration (I) during the same time period. All data are presented as the mean_+ SEM. The significance of differences from baseline was tested by analysis of variance for repeated measurements. Testing for differences between groups was carried out using the Mann-Whitney rank sum test. Correlations were tested by linear regression analysis.

Results

Effect of graded hyperglycaemia (without glipizide) on insulin and C-peptide secretion The plasma glucose, insulin and C-peptide concentrations during graded hyperglycaemia in control and diabetic subjects are s h o w n in Figs. 1 and 2.

Healthy subjects. The basal p l a s m a glucose c o n c e n t r a t i o n was similar in all four experiments, 4.7 + 0.1 mmol/1, and increased in a dose-dependent m a n n e r to 5.8 _+0.1 mmol/1, 7.5 _+0.1 retool/l, 11.4_+ 0.2 mmol/1 and 21.1 _+0.3 mmol/1, respectively during the four hyperglycaemic clamp studies. The basal plasma insulin and C-peptide concentrations were similar in all four experiments, with m e a n s o f 50 _+ 7 pmol/1 and 0.21 _+0.03 nmol/l, respectively. The plasma C-peptide c o n c e n t r a t i o n increased in parallel with the increase in plasma insulin c o n c e n t r a t i o n (Figs. 1 - 4 ) . The first-phase insulin responses increased linearly with increasing plsma glucose concentrations up to 10 mmol/l with a plateau a b o v e this (Fig. 3). The secondphase plasma insulin response rose progressively and linearly ( r = 0 . 9 9 7 ; P < 0 , 0 0 1 ) t h r o u g h o u t the range o f plasma glucose concentrations examined (Fig. 3). The second-phase C-peptide response curve tended to flatten at the higher glucose concentrations (Fig. 4). Type 2 (non-insulin-dependent) diabetic subjects. The basal plasma glucose concentrations during the three experiments were 9 . 5 _ 1.8 mmol/1, 9.6 _+ 1.7 mmol/1 and 9 . 3 _ 1.3 mmol/1, respectively, and rose to 1 2 . 3 + 1.2 mmol/1, 17.1 __+1.4 mmol/1 a n d 2 6 . 7 + 1.0 mmol/1, respectively, during the clamp. The basal p l a s m a insulin ( 1 3 8 + 3 8 , 148 _+29 and 127_+ 36 pmol/1) and C-peptide ( 0 . 6 4 _ 0.12, 0.62 _+0.11 and 0.62 _+0.15 nmol/1) were n o t significantly

164

L.C. Groop et al.: Insulinotropic effects of sulphonylurea

20l o 1,5

E "~ 1.0

~0.5

-60

-30

0

30 60 Time (min)

90

120

- 3 '0

0'

i -30 '~

1 0

30 Time (min)

6o

'2

go

1 0

I 90

-] 120

2.5

700 r

2.0

600t 5001

1,5

4001 1.0

3001 O

200t

0.5 100t O00t -60

900

t I 30 60 Time (min)

8001 o~ 700~

Fig. 2. Mean plasma C-peptide concentrations in 12 healthy subjects during + 1 (solid circles), + 3 (open circles), + 7 (solid triangles) and + 17 (open triangles) mmol/1hyperglycaemicclamp studies with (bottom) and without (top) prior ingestion of 5 mg glipizide. The line with crosses represents the responses to ingestion of 5 mg glipizide (arrow) while maintaining basal euglycaemia

600~ 500 F 4OO}

300} 200~ 100~ 0

~i ..... -30

I 0

3i0

6i0

I 90

I 120

Time (min)

Fig, 1. Mean plasma insulin concentrations in 12 healthy subjects during + 1 (solid circles), + 3 (open circles), + 7 (solid triangles) and + 17 (open triangles) mmol/1 hyperglycaemicclamps with (bottom) and without (top) intake of 5 mg glipizide. The line with crosses represents the response to glipizide (arrow) while maintaining basal euglycaemia

different in the three experiments. First-phase insulin secretion was impaired in the diabetic subjects (Figs. 3, 5). Second-phase insulin and C-peptide concentration responses were markedly attenuated compared with healthy controls (Figs. 3-6) with significantly different slopes (P < 0.001 vs controls). Effect of glipizide on insulin and C-peptide secretion Healthy subjects. The fasting plasma glucose concentration before glipizide was similar in the five experiments, 4.8_+0.2 mmol/1 and was maintained at the basal level

during the 30-min period following glipizide ingestion. During the four hyperglycaemic clamp steps the plasma glucose concentration increased to 5.8 4- 0.1, 7.3 4- 0.2, 11.2 4- 0.3, 20.7 + 0.4 mmol/1. The plasma glipizide concentration had increased significantly by 35 min after glipizide ingestion ( P < 0.05 vs basal) and reached peak values of 448 4- 65 nmol/1 during the last hour of the basal (euglycaemic) clamp. As we have previously reported [25], the peak plasma glipizide concentrations decreased in a dose-dependent manner with increasing plasma glucose concentration, 430 4-56, 327_+58, 197_+82, and 185_+50nmol/1, respectively. However, the mean plasma glipizide concentration still remained above the level required for maximal effect, i.e. about 150 nmol/1 [26]. The basal plasma insulin (58 4-_7 pmol/1) and C-peptide (0.23 4-0.02 nmol/1) concentrations before the intake of glipizide were not significantly different in the five experiments. The plasma insulin and C-peptide concentrations began to rise 30 min following glipizide. During the hyperglycaemic clamp studies performed with glipizide there was a dose-dependent increase in both the plasma insulin and C-peptide concentrations (Figs. 1-4), which paralleled that observed during the hyperglyca-

L. C. Groop et al.: Insulinotropic effects of sulphonylurea

165 Table 1. Glucose disposal rates and insulin sensitivity indices (M/I)

during graded hyperglycaemia in 12 healthy subjects and in 6 type 2 (non-insulin-dependent) diabetic patients (values are mean +-SEM)

1000 1 900

Glucose infusion rate (gmol/kg per min)

M/I ( x 100)

No drug Glipizide

No drug Giipizide

1.9_+0.1 4.4+0.6**

2.0___0.1 1.6___0.4

+ 3 mmol/1 Controls Type 2 diabetic subjects

3.4+0.2 5.9+0.5** 1.6+_0.2 2.0_+0.3

2.0_+0.2 1.3+_0.3 1.5___0.41.2+0.5

+ 7 mmol/1 Controls Type 2 diabetic subjects

7.8-t-1.0 10.3+_1.0"* 2.7-1-0.3 3.2_+0.4

2.1__+0.2 2.2+_0.4 1.2+_0.4 1.3+_0.4

+ 17 mmol/1 Controls Type2 diabetic subjects

20.8+_2.0 23.1+_2.3" 5.0+_0.9 5.4+0.8

2.7+_0.3 1.8+0.4 1.3_+0.4 1.1+-0.4

800 I

+ 1 mmol/1 Controls

400 I

I

200 F

1~176 L 400

o 300 := 200 IO0

0

I

5

I

I

I

I

10 15 20 25 Plasma Glucose (mmol/I)

*P<0.01 vs no drug; ** P<0.001

Fig. 3. Dose-response curves relating the first (bottom) and second

(top) phases of plasma insulin responses to the ambient plasma glucose concentration in 12 healthy (circles) and 6 non-insulindependent diabetic (triangles) subjects. Open symbols denote values without and solid symbols values with prior intake of glipizide. (Values are mean + SEM)

2.0

~ 1.5 c

~

0

1,0

0.5 I

I

I

i

I

~

1

--~ 1.0

"~ 0.5

9

0

I

5

10 15 Plasma Glucose

I

I

I

20

25

30

(mmol/I)

Fig. 4. Dose-response curves relating the first (bottom) and second

(top) phases of plasma C-peptide responses to the ambient plasma glucose concentration in 12 healthy (circles) and 6 non-insulindependent) diabetic (triangles) subjects. Open symbols denote values without and solid symbols values with prior intake of glipizide. (Values are mean _ SEM)

emic clamp studies without glipizide. During each of the four hyperglycaemic clamp steps performed with glipizide, the second-phase plasma insulin response was about 180 pmol/1 greater than the corresponding hyperglycaemic step Without glipizide while the second-phase plasma C-peptide response was about 0.25 nmol/1 greater (Figs. 2, 4). These increments in plasma insulin and C-peptide responses were nearly identical to the increments observed when glipizide was ingested and the plasma glucose concentration was maintained at the basal euglycaemic level.

Type 2 (non insulin-dependent) diabetic subjects. The fasting plasma glucose concentration before glipizide was not significantly different in the four experiments, 9.8 + 2.0, 9.6_+ 1.7, 9.4+_ 1.9, and 9.9+ 1.9 mmol/1, respectively. While the plasma glucose concentration was maintained at the basal level in the first experiment (9.6___0.6 mmol/1), it rose to 12.6+_0.8, 16.2_+0.8 and 26.5_+1.0mmol/1, respectively, during the hyperglycaemic clamps. The basal plasma insulin (132__35, 121 +-29, 127+-30, 125_+ 33 pmol/1) and C-peptide (0.62 ___0.18, 0.61 +- 0.20, 0.54__+ 0.21, 0.62_+0.24 nmol/1) did not differ significantly between the experiments. In keeping with the findings in healthy controls, there was a glucose-dependent decrease in plasma glipizide concentrations in the diabetic subjects during the 2nd h of the clamp (450 +- 84, 325 +- 64, 255 ___95, and 161 ___80 nmol/l). The first-phase insulin concentration increase was not restored by glipizide (Figs. 3, 5). During each clamp, the second-phase insulin and C-peptide levels were about 50 pmol/1 and 0.20 nmol/1, respectively, higher than during the corresponding study without glipizide. However, the dose-response curves were still significantly different

166

L.C. Groop et al.: Insulinotropic effects of sulphonylurea 1000

2.0

900 600 700 ~1.0

"~ 600 D. = 50O '~ 400

~0.5

3OO

I

2OO

t

-60

-30

I

0

100 I

-60

I

-30

t

0

I

,.-I

30 60 Time (rain)

I

I

I

90

120

2.0

r

90

I

30 60 Time (min)

120

3

10001

1.5 1.0

~'0.5 O 1

5001-

1 I

-60

I

-60

I

-30 ,~

I

O

I

/

30 60 Time (min)

I

90

120

Fig. 5. Mean plasma insulin concentrations in 6 non-insulin-dependent diabetic subjectsduring + 3 (open circles), + 7 (solid triangles) and + 17 (open triangles) mmol/1hyperglycaemicclamp studies with (bottom) and without (top) prior ingestion of 5 mg glipizide. The line with crosses represents the responses to ingestion of 5 mg glipizide (arrow) while maintaining basal glycaemia

from those observed in the healthy controls (P<0.001). To obtain an estimate of insulin sensitivity, we calculated the ratio of glucose infusion rate (M-value) and secondphase insulin concentration (I) with and without glipizide in control and non-insulin-dependent diabetic subjects (Table 1). Although insulin sensitivity was significantly impaired in diabetic compared with control subjects at all glucose concentrations (P < 0.001), acute administration of glipizide did not significantly influence insulin sensitivity.

Discussion

In the present study we have examined the separate effects of, as well as the interaction between, hyperglycaemia and sulphonylurea on insulin secretion in healthy and type 2 (non-insulin-dependent) diabetic subjects. Each hyperglycaemic clamp was performed on a separate day so that the prestimulus glucose concentration was similar in all studies. In healthy subjects, the first-phase insulin concentration rose in a dose-dependent manner with increasing

[

-30

I

0

t

30 Time (min)

I

I

I

60

90

120

Fig. 6. Mean plasma C-peptide concentrations in 6 non-insulindependent diabetic subjects during + 3 (open circles), +7 (solid triangles) and + 17 (open triangles) mmol/1 hyperglycaemic clamp studies with (bottom) and without (top) prior ingestion of 5 mg glipizide. The line with crosses represents the responses to ingestion of 5 mg glipizide (arrow) while maintaining basal glycaemia

plasma glucose concentrations up to 9-10 mmol/1. With further increases in the plasma glucose concentration, the first-phase insulin secretion tended to plateau (Fig. 5), suggesting a curvilinear relationship between the firstphase insulin response and the ambient plasma glucose concentration. These data are in keeping with two previous reports using somewhat different experimental designs [5, 27]. In addition, van Haeften et al. [28] demonstrated that the half-maximally stimulating blood glucose concentration (ED 5o) for first-phase insulin secretion was significantly lower than the EDs0 for second-phase insulin secretion (8.4 compared with 14.3 mmol/1). Taken together, the data indicate that first-phase insulin secretion is linearly related to the plasma glucose concentration within the physiological range but saturates at higher glucose concentrations. As expected, first-phase insulin secretion in response to glucose was reduced in the diabetic subjects. Although glipizide enhanced the stimulatory effect of glucose on diabetic first-phase insulin secretion in healthy subjects, the sulphonylurea did not influence first-phase insulin secretion in type 2 patients. In contrast to these results, a previous study indicated that another sulphonylurea, gliclazide, restored the first-phase C-peptide but not the first-phase insulin response to glucose [18]. It is of note that glipizide stimulated both insulin and C-peptide secretion during the euglycaemic control experiment in type 2 diabetic patients (Fig. 5). During the hyperglycaemic clamp this effect is superimposed on glucose-stim-

167

L. C. Groop et al.: Insulinotropic effects of sulphonylurea ulated insulin secretion and might be interpreted as an increase in first-phase insulin secretion. In order to determine the specific effect of sulphonylurea on glucose-stimulated first-phase insulin secretion, the stimulatory effect of sulphonylurea alone must be subtracted. Our results, however, do not rule out the possibility that sulphonylureas influence first-phase insulin secretion in patients with less severe degrees of hyperglycaemia. Hosker et al. [18] chose diet-treated patients with a fasting plasma glucose of 6.0 mmol/1, whereas the fasting plasma glucose in our patients was about 9 mmol/1. On the other hand, one of our patients had a fasting plasma glucose concentration below 7 mmol/1. Despite this, no sulphonylurea effect on glucose-stimulated first-phase insulin secretion was observed. Similarly, in newly diagnosed type 2 (noninsulin-dependent) diabetic subjects with fasting plasma glucose concentrations below 8 retool/l, glipizide enhances the early insulin response to a test meal [19]. Given these results, it seems likely, that sulphonylureas must be initiated early in the disease if they are to influence first-phase insulin secretion. In both type 2 diabetic and heathy subjects, the second-phase plasma insulin and C-peptide concentration showed a strong linear correlation with the increase in the plasma glucose concentration. While glipizide moved the dose-response curve to the left (enhanced #-cell responsiveness to glucose), the drug did not influence /%cell sensitivity to glucose. Similar conclusions were drawn by Shapiro et al. [26], who showed that the predominant mechanism of chronic glyburide therapy is to increase the responsiveness of the/~-cell to glucose. This effect may, however, be masked by the fall in fasting plasma glucose concentration during therapy. In type 2 (non-insulin-dependent) diabetic patients, the stimulatory effect o f sulphonylurea on glucose-stimulated second phase insulin secretion was only about 30% of that observed in the healthy controls. Although the patients were characterized by manifest fasting hyperglycaemia, at present they are the major target group for sulphonylurea therapy. Given the limited efficacy of sulphonylurea in this group of subjects, the data emphasize the need to start sulphonylurea therapy earlier in the disease. It should, however, be kept in mind that these are acute experiments measuring the effect of sulphonylurea on glucose-stimulated insulin secretion. It is still possible that, during chronic therapy, sulphonylureas exert other effects on non-glucose-stimulated insulin secretion, e.g. meal-stimulated insulin secretion. The decrease in plasma glipizide concentrations accompanying increasing hyperglycaemia deserves some comments. When we first reported this finding in normal subjects, we proposed that the impaired absorption was a consequence of delayed gastric emptying [29]. Hyperglycaemia has subsequently been shown to cause a dosedependent reduction in gastric emptying in patients with insulin-dependent diabetes [25]. The present data provide further evidence that the same mechanisms are operative in patients with type 2 diabetes mellitus, i.e. hyperglycaemia causes a dose-dependent impairment in sulphonylurea absorption. It could therefore be argued that we underestimated the effect o f glipizide on glucose-stimu-

lated insutin/C-peptide secretion at the higher glucose concentrations. We think this is unlikely, since intravenous glyburide results in maximal enhancement of glucose-stimulated insulin secretion at steady-state plasma drug concentrations of 120-150 nmol/1 [30]. The glipizide concentrations measured at the highest hyperglycaemic levels in both normal and diabetic subjects were above these values. We think it is justifiable to generalize data from glyburide, since identical plasma concentrations of glipizide and glyburide seem to exert the same effect on glucose-stimulated insulin secretion [20]. Acute glipizide administration significantly enhanced the rate of glucose disposal in both healthy and diabetic subjects. However, sulphonyiurea did not influence the M/I value, i.e. insulin sensitivity. Similar conclusions were drawn by Hosker et al. [18], who were unable to demonstrate any effect o f gliclazide on insulin sensitivity assessed during a hyperglycaemic clamp. In conclusion, sulphonylureas stimulate insulin secretion in both normal and type 2 (non-insulin-dependent) diabetic subjects by enhancing responsiveness of the t-cell to glucose without having any effect on t-cell sensitivity to glucose. This effect appears to be independent of the ambient plasma glucose concentration.

Acknowledgements. We thank the nurses and technicians of the Clinical Research Center at the Yale New Haven Hospital for their support with the studies and all the subjects who faithfully completed the extensive study protocol. The studies were supported in part by Clinical Research Center grant RR 125, NIH grant AM-24092 and by a grant from Pfizer Roerig, New York, NY, USA. L.G. was the recipient of a Fogarty International Fellowship (F05-TWO3451). L.L. was the recipient of a grant from Istituto Scientifico San Raffaele Hospital, Milan, Italy.

References 1. Loubatieres A, The hypoglycemic sulfonamides: history and development of the problem from 1942 to 1945. Ann NY Acad Sci 71:4 11, 1957 2. Brunzell JD, Robertson RP, Lerner RL, Hazzard WR, Ensinck JW, Bierman EL, Porte D Jr, Relationship between fasting plasma glucose levels and insulin secretion during intravenous glucose tolerance tests. J Clin Endocrinol Metab 42: 222-229, 1979 3. Chen M, Porte D Jr, The effect of rate and dose of glucose infusion on acute insulin response in man. J Clin Endocrinol Metab 42: 1168-1175, 1976 4. Joost HG, Beckman J, On the relation of glucose and insulin secretion in the fasting state. Metabolism 29: 23-27, 1980 5. Ferner RE, Ashworth L, Tronier B, Alberti KGMM, Effects of short-term hyperglycemia on insulin secretion in normal humans. Am J Physiol 250:E655-E661, 1986 6. Halter JB, Graf RJ, Porte D Jr, Potentiation of insulin secretory responses by plasma glucose levels in man: evidence that hyperglycemia in diabetes compensates for impaired glucose potentiation. J Clin Endocrinol Metab 48: 946-954, 1979 7. Ward WK, Beard JC, Porte D Jr, Clinical aspects of islet B-cell function in non-insulin dependent diabetes mellitus. Diabetes/ Metabol Rev 2: 297-313, 1986 8. Grodsky GM, Bennett LL, Smith D, Nemecheck K, Tolbutamide after 10 years. Excerpta Medica, New York, p 11, 1967 9. Duckworth WC, Solomon SS, Kitabchi AE, Effect of chronic sulfonylurea on plasma insulin and proinsulin levels. J Clin Endocrinol Metab 35: 585-591, 1972

168 10. Reaven G, Dray J, Effect of chlorpropamide on serum glucose and immunoreactive insulin concentrations in patients with maturity onset diabetes mellitus. Diabetes 16:487-492, 1967 11. Chu P-C, Conway M J, Krouse HA, Goodner CJ, The pattern of response of plasma insulin to meals and fasting during chlorpropamide therapy. Ann Intern Med 68: 757-768, 1968 12. Tsalikian E, Dunphy TW, Bohannon NV, Lorenzi M, Gerich JE, Forsham PH, Kane JP, Karam JH, The effect of chronic oral antidiabetic therapy on insulin and glucagon responses to a meal. Diabetes 26: 314-321, 1977 13. Simonson DC, Ferrannini E, Bevilacqua S, Smith D, Barrett G, Carlson R, DeFronzo RA, Mechanisms of improvement in glucose metabolism after chronic glyburide therapy. Diabetes 33:838-845, 1984 14. Hosker JP, Burnett MA, Davies EG, Harris EA, Turner RC, Sulphonylurea therapy doubles B-cell response to glucose in type 2 diabetic patients. Diabetologia 28:809-814, t985 15. Judzewitsch RG, Pfeifer MA, Best JD, Beard JC, Halter JB, Porte D Jr, Chronic chlorpropamide therapy of noninsulin-dependent diabetes augments basal and stimulated insulin secretion by increasing islet sensitivity to glucose. J Clin Endocrinol Metab 55: 321-328, 1982 16. Groop L, Groop P-H, Stenman S, Saloranta C, Trtterman K-J, Fyhrquist F, Melander A, Comparison of pharmacokinetics, metabolic effects and mechanisms of action of glyburide and glipizide during long-term treatment. Diabetes Care 10: 671678, 1986 17. Pfeiffer MA, Halter JB, Beard JC, Porte D Jr, Differential effects of tolbutamide on first and second phase insulin secretion in noninsulin-dependent diabetes mellitus. J Clin Endocrinol Metab 53:1256-1262, 1981 18. Hosker JP, Rudenski AS, Burnett MA, Matthews DR, Turner RC, Similar reduction of first- and second phase B-cell responses at three different glucose levels in type II diabetes and effect of gliclazide therapy. Metabolism 38: 767-772, 1989 19. Bitzrn PO, Melander A, Scherstrn B, Wgthlin-Boll E, The influence of glipizide on early insulin release and glucose disposal before and after diet regulation in diabetic patients with different degrees of hyperglycemia. Eur J Clin Pharmacol 35:31-37, 1988

L. C. Groop et al.: Insulinotropic effects of sulphonylurea 20. Groop L, Luzi L, Melander A, Groop P-H, Ratheiser K, Simonson DC, DeFronzo RA, Different effects of glyburide and glipizide on insulin secretion and hepatic glucose production in normal and NIDDM subjects. Diabetes 36:1320-1328, 1987 21. DeFronzo RA, Tobin JD, Andres R, Glucose clamp technique: a method for quantifying insulin secretion and insulin resistance. Am J Physiol 237: E214-E223, 1979 22. Wfihlin-Boll E, Melander A, High performance liquid chromatographic determination of glipizide and some other sulfonylurea drugs in serum. J Chromatogr 164:541-546, 1979 23. Hales CN, Randle P J, Immunoassay of insulin with insulin antibody precipitate. Biochem J 88: 137-146, 1963 24. Heding LG, Radioimmunological determination of human C-peptide in serum. Diabetologia 11: 541-548, 1975 25. Fraser RJ, Horowitz M, Maddox AF, Harding PE, Chatterton BE, Dent J, Hyperglycemia slows gastric emptying in type 1 (insulin-dependent) diabetes mellitus. Diabetologia 33: 675-680, 1990 26. Shapiro ET, Cauter E van, Tillil H, Given RD, Hirsch L, Beebe C, Rubenstein AH, Polonsky KS, Glyburide enhances the responsiveness of the/?-cell to glucose but does not correct the abnormal patterns of insulin secretion in noninsulin-dependent diabetes mellitus. J Clin Endocrinol Metab 69: 571- 576, 1989 27. Rudenski AS, Hosker JP, Burnett MA, Matthews DR, Turner RC, The beta cell glucose stimulus-response curve in normal humans assessed by insulin and C-peptide secretion rates, Metabolism 37: 526-534, 1988 28. Haeften TW van, Boonstra E, Veneman TF, Gerich JE, Veen EA van der, Dose-response characteristics for glucose-stimulated insulin release in man and assessment of influence on arginine-stimulated insulin release. Metabolism 39:1292-1299, 1990 29. Groop L, Luzi L DeFronzo RA, Melander A, Hyperglycemia and absorption of sulfonylurea drugs. Lancet II: 129-130, 1989 30. Groop LC, Barzilai N, Ratheiser K, Luzi L, Wgthlin-Boll E, Melander A, DeFronzo RA, Dose-dependent effects of glibenclamide on insulin secretion and glucose uptake in man. Diabetes Care 14: 724-727, 1991