Characterization of the late posthypoglycemic insulin resistance in insulin-dependent diabetes mellitus

Characterization

of the Late Posthypoglycemic Insulin Insulin-Dependent Diabetes Mellitus

Jesper Fowelin,

Stig Attvall,

Henning von Schenck,

Resistance

in

Ulf Smith, and Ibe Lager

The insulin effect (6.5 to 7.5 hours) following hypoglycemia was studied with the euglycemic clamp technique in eight patients with insulin-dependent diabeteses mellitus (IDDM). The results were compared with a control study with the same insulin infusion, but where hypoglycemia was prevented by a glucose infusion. Glucose production (Ra) and utilization (Rd) were evaluated with D-(3-3H) glucose infusion. Hypoglycemia (glucose nadir, 1.5 + 0.1 mmol/L) caused a marked increase in cortisol and growth hormone, whereas the release of adrenaline and, in particular, glucagon was low. The plasma free insulin levels were similar in the studies, including during the clamp periods. The glucose infusion rates (GIR) were significantly lower after the hypoglycemia as compared with the control study (control, 2.4 ? 0.3; hypoglycemia, 1.5 + 0.3 mg/kg x min; P < .05). Thus, hypoglycemia induces prolonged insulin resistance. The posthypoglycemic insulin resistance during a moderate hyperinsulinemic (-30 mu/L) clamp was mainly due to a decreased insulin effect on glucose utilization (control, 2.9 + 0.2; hypoglycemia, 2.2 * 0.2 mg/kg x min; P < .02), whereas the insulin effect on glucose production was not significantly different after hypoglycemia. 0 1996 by W-B. s8ffffd8fS Compny.

H

YPOGLYCEMIA is followed by a state of insulin resistance both in healthy subjects’,’ and in patients with insulin-dependent diabetes mellitus (IDDM).3,4 However, the clinical importance of insulin resistance in causing a rebound hyperglycemia is still under debate. Failure of nocturnal hypoglycemia to induce fasting hyperglycemia in patients with IDDM has recently been reported.jv6 However, in the study by Periello et al,’ asymptomatic nocturnal hypoglycemia induced elevated fasting and, in particular, postprandial glucose levels in a group of IDDM patients with very good metabolic control. These discrepant results may be explained by the different metabolic control of the patients studied. We have recently found that the degree of insulin resistance induced by hypoglycemia is related to the initial insulin effect, ie, the inhibitory effect induced by the counterregulatory hormones is exerted on the insulin-stimulated part of the glucose disposal.’ The insulin-antagonistic effect of a hypoglycemia can, in principle, influence glucose production and/or utilization. The extent to which these effects are exerted would primarily influence the fasting or the postprandial glucose concentrations, respectively.9,‘0 The present study was performed to evaluate the influence of hypoglycemia on the insulin effect late (6.5 to 7.5 hours) following hypoglycemia in insulindependent diabetics. PATIENTS AND METHODS

Patients Eight patients with IDDM characterized according to the criteria of the National Diabetes Data group” were studied. The postprandial plasma C-peptide levels were below 0.2 nmol/L in all patients.

From the Department of Medicine II, Sahlgren’s Hospital, University of Gothenburg, Gothenburg, Sweden; and the Department of Clinical Chemistry, University of Linkiiping. Linkijping, Sweden. Supported by grants from the Swedish Medical Research Council (project B-3506). the Gateborg Diabetes Association, and the Giiteborg Medical Society. Address reprint requests to fbe Lager, MD, Department of Medicine 11, Sahlgrens Hospital, S-413 45 Gothenburg, Sweden, B 1990 by W.B. Saunders Company. 00260495/90/390g-OOOgos803.00/0 822

The patients were on regular treatment with a four-dose insulin regimen (Insulin Actrapid Human and Protaphan Human, NOVO, Denmark or Insulin Velosulin Human and Insulatard Human, Nordisk, Denmark). None of the patients had evidence of autonomic neuropathy demonstrable with the immediate heart-rate response to standing” or heart-rate variation to deep breathing.13 Their metabolic control was stable, and the insulin doses remained unchanged between the studies. The characteristics of the patients are shown in Table 1. Informed consent was obtained from all patients and the study was approved by the Ethical Committee of the University of Gothenburg. Infusions All infusions were made through a catheter in a cubital vein. Arterialized blood samples were drawn from a dorsal hand vein in the contralateral arm. Insulin (Actrapid Human) was infused at a concentration of 40 mU/mL in isotonic saline with 4 mg/mL albumin added to prevent glass adhesion. Glucose (200 g/L; Baxter A/S, Halden, Norway) was infused at variable rates in the same catheter as insulin. Potassium chloride (0.1 moi/L) was infused at a rate of 7 mmol/h during the first 120 minutes, at a rate of 5 mmol/h when the insulin infusion rate was decreased, and at a rate of 7 mmol/h during the euglycemic clamp period. Rates of glucose appearance (Ra) and disappearance (Rd) were determined by the infusion of D-(~-~H) glucose (New England Nuclear, Boston, MA) dissolved in saline. A primed infusion of 25 pCi was administered followed by a constant infusion of 15 &i/h. Study

Design

The design of the study is shown in Fig 1. Intermediate-acting insulin was not administered to patients for 24 hours before the study. During the night before the study, blood glucose levels were regulated with a slow, continuous insulin infusion. On the morning of the study, at 8:00 AM, patients were admitted to the laboratory in the postabsorptive state and insulin was infused at a rate of 2.5 mu/kg xmin during 40 minutes. The insulin infusion was then stopped until 120 minutes. Glucose was infused for 10 minutes to restore euglycemia; thereafter an insulin infusion was administered at a rate of 0.1 mu/kg x min for 4 hours. Glucose levels during these 4 hours were kept between 5.0 and 6.0 mmol/L, with infusion of glucose (200 g/L) if necessary. Six hours after induction of hypoglycemia, a euglycemic clamp (insulin 0.5 mu/kg x min) was performed for 2 hours. Infusion of D-(~-~H) glucose was started 2 hours before the clamp. Each patient was studied twice in a randomized order in conjunction with a hypoglycemia and in a control study where hypoglycemia

Merabolism, Vol 39, No 8 (August), 1990: pp 822-826

POSTHYPOGLYCEMIC INSULIN RESISTANCE IN IDDM

823

Table 1. Patient Characteristics Daily Height

Weight

BMI

Diabetes

Insulin

Duration

DOS0

HbA,c Control

Hypoglycemia

(Vr)

NJ/kg)

study

study

22.1

14

0.89

8.2

9.0

23.7

13

0.77

8.5

9.3

87

24.1

7

0.83

6.8

7.1

169

68

23.8

19

0.9 1

7.2

6.9

31

176

72

23.2

6

0.69

8.3

6.7

35

178

56

1717

28

0.91

5.7

5.7

37

181

70

21.4

14

0.69

9.0

8.9

sex

Age (vr)

(cm)

Ike)

M

20

169

63

M

23

173

71

M

25

190

F

26

M F M

(kg/d

F

37

164

68

25.3

8

0.59

7.7

7.7

Mean + SE

29 * 7

175 * 8

69 + 9

22.6 f 2.3

14 * 7

0.79 * 0.12

7.7 * 1.1

7.7 + 1.3

NOTE.

Upper

limmit

of normalrangefor HbA,c is 5.4%.

Abbreviation: BMI, bodY mass index.

by glucose infusion. During induction of hypoglycemia and during the euglycemic clamps, the blood glucose levels were measured every 5 minutes with a Reflectometer (Reflomat, Clinicon International, Mannheim, FRG; correlation with the glucose dehydrogenase technique described below, r = .99, n = 532, P < .OOl). Glucose was determined every 10 minutes throughout the rest of the study. was prevented

Analytical Procedures Blood samples for the determinations of glucose, free insulin, adrenaline, glucagon, cortisol, and growth hormone levels were drawn as shown in the Results. During the isotopic equilibration period, blood samples were taken at 240 minutes, ie, before infusion of D-(3-%) glucose, after 340 minutes, and every 10 minutes thereafter for chemical determination of glucose and measurement of specific activity. Blood samples for the determination of C-peptide were drawn before and 1 hour following a standardized meal. All samples were kept on ice, rapidly centrifuged, and the plasma stored at - 20°c’ until analyzed. The glucose levels shown in the Results were determined with a glucose-6-dehydrogenase method (Beckman Instruments, Fullerton, CA). Measurements of the specific activity of glucose were performed on plasma samples deproteinized with Ba (OH), and ZnSo, and evaporated at 40°C. The samples were counted in a Packard Liquid Scintillation counter for 20 minutes. The free insulin and the C-peptide levels were determined on the supernatant of polyethyleneglycol-treated samplesI with radioimmunoassay techniques (Phadeseph Instrument, Pharmacia, Uppsala, Sweden, and NOVO, respectively). Glucagon was assayed with antiserum E7.” This antiserum, which recognizes the carboxyterminal region of glucagon, has been characterized with antiserum 30K 0

1

2

3

,

Insulin bdJ~k&niri’)~

5

6

7

sh

Clamp

E-GlUCOSe

091

0.5

3-3H-Glu~are KCI Glucose

Fig 1. Study design. Hypoglycemia wao induced with an insulin infusion and the insulin affect wa6 ev6luated 8.6 to 7.6 houra later with 6 euglycemic cbmp. Euglycemia wa6 maintained before the clampa. Compari6on6 were made to e control rtudy with the 6eme in6ulin infusion. but without hypoglycemia. Hypoglycemia, -; control, . . . - . . .

as reference.lb The adrenaline levels were determined with a liquid chromatographic method, cortisol with a fluorimeteric method, and growth hormone with a double-antibody technique. Glycosylated hemoglobin levels (HbA,,) were determined with a high-performance liquid chromatographic technique.” The insulin and glucagon levels during the euglycemic clamps shown in the Results are the mean values of three samples collected with an interval of 30 minutes during the second hour of each clamp period, and the adrenaline, cortisol, and growth hormone levels the mean values of two samples collected with an interval of 60 minutes during the same periods. The glucose infusion rates (GIR), glucose production (Ra), and utilization (Rd) in the Results are the mean values during the second hour of each clamp period. The rates of glucose production and utilization were calculated with the non-steady-state equation of de Bodo et al.” This method assumes that rapid changes in the specific activity and concentrations of glucose do not occur uniformly within the entire glucose pool. To compensate for nonuniform mixing, the non-steady-state term of the equation was multiplied by a correction factor (pool fraction) of 0.65.‘9,20During the glucose infusions, Ra was calculated by subtracting the rate of infusion of exogenous glucose from the tracer determined total rate of glucose production. Statistical Analysis Data are shown as means t SEM. Significances of differences were evaluated with Student’s t test for paired data. RESULTS

Hypoglycemic Period Blood glucose levels were similar in the two studies before induction of hypoglycemia (control, 7.0 r 0.5; hypoglycemia, 7.1 + 0.4 mmol/L). The increased insulin infusion produced hypoglycemia where the lowest glucose levels were seen 10 minutes after the insulin infusion was stopped (1.5 k 0.1 mmol/L). In the control study, glucose levels were maintained above 5.0 mmol/L (Fig 2). Following the hypoglycemia and before the clamps, glucose levels were maintained at 5.0 to 6.0 mmol/L. The free insulin levels were similar before the hypoglycemia (control, 12 + 2; hypoglycemia, 17 + 3 mu/L), as well as during and following the increased insulin infusion (Fig 2). The basal adrenaline and glucagon levels were similar in the two studies. A marked increase in adrenaline was seen during the hypoglycemia, whereas glucagon only increased slightly (Fig 2). These hormones did not increase in the

FOWELIN ET AL

824

l

P’O.05

Fig 3. GIR and plasma free insulin levels before and during the euglycemic clamps. lP c.06.

tively). The free insulin levels before the clamps were stable and similar in the two studies (control, 10 + 2; hypoglycemia, 11 f 2 mu/L). During the clamps, free insulin levels, as well as the levels of counterregulatory hormones, were similar in the two studies (Fig 3, Table 2). GIR the hour before the clamps were similar in the two studies (control, 0.7 * 0.4; hypoglycemia, 0.4 2 0.2 mg/ kg + min). However, GIR during the second hour of the clamp were significantly lower after hypoglycemia (control, 2.4 k 0.3; hypoglycemia, 1.5 f 0.3 mg/kg . min, P < .OS). Thus, a state of insulin resistance had been induced by hypoglycemia 7 hours previously (Fig 3). Glucose production (Ra) rates were somewhat higher in the control study before the clamps. However, the difference was not statistically different (Fig 4). Ra decreased during the clamps and similar rates were found during the second hour of the clamps. The decrease in Ra was not statistically different in the two studies. Glucose utilization (Rd) rates were equal before the clamps and increased during the clamps in the control study. However, following hypoglycemia, no increase was seen (Fig 4). Glucose utilization rates were, therefore, higher during the clamps in the control study (control, 2.9 + 0.2; hypoglycemia, 2.2 f 0.2 mg/kg - min, P < .02).

Time(min) Fig 2. Blood glucose and plasma levels of free insulin. adrenaline, glucagon, growth hormone, and cortiaol during the studies. Control, k-k, hypoglycemia. 0. * - *O.

control study. The basal cortisol levels were similar in the two studies, but increased during hypoglycemia; the highest levels were seen 40 minutes after glucose nadir (764 + 90 nmol/L) (Fig 2). The basal growth hormone levels were also similar in the studies, but increased markedly during hypogly cemia. A moderate increase was also seen in the control study (Fig 2).

DISCUSSION

The present study shows that a late posthypoglycemic insulin resistance occurs in IDDM. The relative insulin sensitivity in the liver is higher than that for glucose utilization.” This must be taken into consideration when the insulin-antagonistic effect of hypoglycemia on Ra and Rd is evaluated. During the low insulin infusion rate before the clamps, glucose production was higher, albeit not statistically significant, following hypoglycemia. Glucose utilization was then similar in the two studies. During the clamps, with moderately high plasma insulin levels, the insulin-antagonis-

Euglycemic Period

Glucose levels were similar in both studies before the clamps and were 5.0 + 0.1 and 5.6 * 0.3 mmol/L in the control and hypoglycemic study, respectively, at the beginning of the clamp periods. During the euglycemic clamps, the glucose levels were 4.7 k 0.1 mmol/L in the control study and 4.6 5 0.2 mmol/L in the hypoglycemic study (coefficients of variation 4.9% + 0.9% and 7.8% f 1.5%, respecTable 2. Counterreguletory

Hormones

Before (at 6 hours) and During (7 to 8 hours) the Euglycemic Clamps

Control study

Hypoglycemic Study

Before Glucagon (ng/L) Adrenaline (nmol/L) Cortiaol (nmol/L) Growth hormone (mu/L) NOTE. Results are means + SEM.

41 +2 0.12

zk 2

365 * 60 2.0 f 0.7

Before

Clamp 35 * 2 0.12

* 0.01

236 f 39 3.8 + 1.7

43 * 3 0.22

f 0.07

382 f 41 4.6 + 3.1

Clamp 37 * 3 0.18

i 0.04

210 f 5 7.3 * 2.9

825

POSTHYPOGLYCEMIC INSULIN RESISTANCE IN IDDM

I

Fig 4. Glucose production (Ra) and utilization during the euglycemic clamps. lP 1.02.

(Rd) before and

tic effect of hypoglycemia was instead principally seen on glucose utilization. However, the preclamp data concerning Ra and Rd must be carefully interpreted, as the tritiated glucose infusion began only 2 hours before the clamps. Taken together, these results show that the late posthypoglycemic insulin resistance influences the insulin effect both on Ra and Rd. However, at moderately high free insulin levels, the effect on Rd is predominant. Similar results have recently been found in healthy subjects where the insulinantagonistic effect of hypoglycemia was more pronounced on glucose utilization than on glucose production at plasma insulin levels of approximately 30 mU/L.8 Fasting hyperglycemia following a nocturnal hypoglycemia has not been found or has only been small in patients with IDDM.‘.’ This can be explained by the deficient glucose counterregulation regularly found in these patients22~23and also by the high insulin sensitivity of the liver, as the plasma insulin levels were greater than 11 mU/L when the measures were performed.’ The postprandial glucose levels are mainly determined by glucose disposal rates, and plasma insulin levels are then higher. Glucose production rates are, therefore, less important for the glucose levels in these situations. This can explain the more pronounced postprandial hyperglycemia following hypoglycemia found by Periello et al’ and in a recent study in our laboratory.24 Similar free insulin levels as those in the present study during the clamps are often found postprandially in patients with IDDM.24-26 The mechanism for the late posthypoglycemic insulin resistance found in the present study is probably the in-

creased release of growth hormone and cortisol during the hypoglycemia.* Growth hormone has insulin-antagonistic effects both on glucose production and utilization. However, the latter effect is predominant during the moderately high plasma insulin levels used during the clamps in the present effect of cortisol also study.27-29 The insulin-antagonistic seems to influence both Ra and Rd.3o The results of the present study, with a lower insulin effect following hypoglycemia, principally on glucose utilization, during moderately high free insulin levels are in agreement with the results discussed above and can probably explain the elevated postprandial glucose levels found in the studies by Periello et al’ and Fowelin et a1?4 The attenuated adrenaline release and the virtually obliterated glucagon response during the hypoglycemia in the present study are expected in a group of IDDM patients with a mean diabetes duration of 14 years.22*2313’ The growth hormone levels also increased in the control study despite euglycemia. No corresponding increase was seen in adrenaline, glucagon, and cortisol levels. Increased release of growth hormone in diabetic patients during hyperinsulinemia and normal blood glucose levels have previously been described by Sharp et a1.32.33The mechanisms for this are unclear, but the hyperinsulinemia per se seems to be the most probable explanation. However, the growth hormone levels were approximately three times higher during the hypoglycemia than in the control study. Thus, if the insulin effect after hypoglycemia is compared with a control study without insulin infusion, the magnitude of the posthypoglycemic insulin resistance induced by growth hormone may be overestimated. In conclusion, the present study shows that a late (6.5 to 7.5 hours) posthypoglycemic insulin resistance occurs in patients with IDDM. This resistance is under the currently employed experimental conditions with moderately high insulin levels mainly due to a decreased insulin effect on glucose utilization. These results can explain why the most pronounced effect of a hypoglycemia on glucose control is seen postprandially. ACKNOWLEDGMENT

The skillful assistance by Renate Schmidt, RN, is gratefully acknowledged. Excellent secretarial help was given by Kristina Ullgren and Helena Carlsten.

REFERENCES

1. Attvall S, Eriksson BM, Fowelin J, et al: Early posthypoglycemic insulin resistance in man is mainly an effect of fi-adrenergic stimulation. J Clin Invest 80:437-442, 1987 2. Clore JN, Brennan JR, Gebhart SP, et al: Prolonged insulin resistance following insulin-induced hypoglycaemia. Diabetologia 30:851-858, 1987 3. Attvall S, Fowelin J, von Schenck H, et al: Insulin resistance in type 1 (insulin-dependent) diabetes following hypoglycaemiaEvidence for the importance of &adrenergic stimulation. Diabetologia 30:691-697, 1987 4. Kollind M, Adamson U, Lins P-E, et al: Importance of growth hormone for blood glucose regulation following insulin-induced nocturnal hypoglycemia in insulin-dependent diabetes mellitus. Acta Med Stand 223:159-164, 1988 5. Tordjman KM, Havlin CE, Levandovski LA, et al: Failure of

nocturnal hypoglycemia to cause fasting hyperglycemia in patients with insulin-dependent diabetes mellitus. N Engl J Med 3 17:15521559,1987 6. Havlin CE, Cryer PE: Nocturnal hypoglycemia does not commonly result in major morning hyperglycemia in patients with diabetes mellitus. Diabetes Care 10:141-147, 1987 7. Periello G, de Feo P, Calcinaro F, et al: Nocturnal hypoglycemia exaggerates fasting and postbreakfast hyperglycemia in CSIItested IDDM. N Engl J Med 319:1233-1239, 1988 8. Fowelin J, Attvall S, von Schenck H, et al: Combined effect of growth hormone and cortisol on late posthypoglycemic insulin resistance in humans. Diabetes 38:1357-1364, 1989 9. Campbell PJ, Mandarin0 LJ, Gerich JE: Quantification of the relative impairment in actions of insulin on hepatic glucose produc-

FOWELIN ET AL

tion and peripheral glucose uptake in non-insulin-dependent diabetes mellitus. Metabolism 37:15-21, 1988 10. Ferrannini E, Simonson DC, Katz LD, et al: The disposal of oral glucose load in patients with non-insulin-dependent diabetes. Metabolism 37:79-85, 1988 11. National Diabetes Data Group: Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 28:1039-1057, 1979 12. Ewing DJ, Campbell IW, Murray A, et al: Immediate heart-rate response to standing: Simple test for autonomic neuropathy in diabetes. Br Med J 1:145-147, 1978 13. Mackay JD, Page MMcB, Cambridge J, et al: Diabetic autonomic neuropathy in diabetes. Diabetologia 18:471-478, 1980 14. Arnqvist H, Olsson PO, von Schenck H: Free and total insulin as determined after precipitation with polyethylene-glycol: Analytical characteristics and effects of sample handling and storage. Clin Chem 33193-96, 1987 15. von Schenck H: Production and characterization of an antiserum against pancreatic glucagon. Clin Chim Acta g&455-463, 1977 16. von Schenck H, Nilsson OR: Radioimmunoassay of extracted glucagon compared with three non-extraction assays. Clin Chim Acta 109:183-191, 1981 17. Holme I: Determination of HbAlc with cation exchange (mono-S) method. Academic Dissertation, 1980, University of Gothenburg, Sweden 18. de Bodo RC, Steele R, Altszuler N, et al: On the hormonal regulation of carbohydrate metabolism; studies with Cl4 glucose. Recent Prog Horm Res 19:445-488, 1963 19. Cowan JS, Hetenyi G Jr: Glucoregulatory responses in normal and diabetic dogs recorded by a new tracer method. Metabolism 20:360-372, 1971 20. Radziuk J, Norwich KH, Vranic M: Experimental validation of measurements of glucose turnover in nonsteady state. Am J Physiol234:E84-E93,1978 2 1. Rizza R, Mandarin0 L, Gerich J: Dose-response characteristics for the effects of insulin on production and utilization of glucose in man. Am J Physiol240:E630-E639, 1981 22. Bolli GB, Tsalikian E, Haymond MV, et al: Defective glucose

counterregulation after subcutaneous insulin in noninsulin-dependent diabetes mellitus. J Clin Invest 73:1532- 1541, 1984 23. Lager I, Von Schenck H, Smith U: Improved but not normalized glucose counter-regulation during glucagon infusion in type-l (insulin-dependent) diabetes. Diabetologia 26:337-342, 1984 24. Fowelin J, Attvall S, von Schenck H, et al: Postprandial hyperglycemia following a morning hypoglycemia in patients with type-l diabetes mellitus. Diabetes Med 7:156-161, 1990 25. Schriffin A, Parikhs S: Accommodating planned exercise in type-l diabetic patients on intensive treatment. Diabetes Care 8:337-342, 1985 26. Trovati M, Anfossi G, Vitali S, et al: Postprandial exercise in type 1 diabetic patients on multiple daily insulin injection regimen. DiabetesCare 11:107-110, 1988 27. Bratusch-Marrain PR, Smith D, DeFronzo RA: The effect of growth hormone on glucose metabolism and insulin secretion in man. J Clin Endccrinol Metab 55:973-982, 1982 28. Rizza RA, Mandarin0 LJ, Gerich JE: Effects of growth hormone and insulin action in man. Mechanisms of insulin resistance, impaired suppression of glucose production, and impaired stimulation of glucose utilization. Diabetes 3 1:663-669, 1982 29. Moller M, Butler PC, Antsiferov MA, et al: Effects of growth hormone on insulin sensitivity and forearm metabolism in normal man. Diabetologia 32:105-l 10, 1989 30. Rizza RA, Mandarin0 LJ, Gerich JE: Cortisol-induced insulin resistance in man: Impaired suppression of glucose production and stimulation of glucose utilization due to a postreceptor defect of insulin action. J Clin Endocrinol Metab 54: 13I- 138, 1982 31. Bolli G, de Feo P, Compagnucci P, et al: Abnormal glucose counterregulation in insulin-dependent diabetes mellitus. Interaction of anti-insulin antibodies and impaired glucagon and epinephrine secretion. Diabetes 32:134-141, 1983 32. Sharp PS, Foley K, Vitelli F, et al: Growth hormone response to hyperinsulinemia in insulin-dependent diabetics. Comparison of patients with and without retinopathy. Diabetes Med 1:55-58, 1984 33. Sharp PS, Viswanathan M, Maneschi F, et al: Changes in plasma growth hormone in diabetic and nondiabetic subjects during the glucose clamp. Metabolism 36:71-75, 1987