Clinical Unawareness of Hypoglycemia by Insulin-Dependent Diabetics
Summary After several years of insulin therapy, about 20 % of insulin-dependent diabetics have little or no perception of hypoglycaemia because of a loss of the adrenergic warning symptoms. This defect, poorly correlated with the presence of autonomic neuropathy, has been classically explained by a defect in the catecholamine secretion. We compared the hormonal counterregulation during hypoglycaemia induced by subcutaneous injection of insulin in 7 insulin-dependent diabetics with poor perception of hypoglycaemia and experiencing repeated episodes of severe hypoglycaemia (group A) and 7 insulin-treated diabetics with very good perception of hypoglycaemia and not experiencing severe hypoglycaemia (group B). Groups A and B were similar in terms of age, duration of diabetes, HbA1c level and degenerative complications. The glucagon levels were identical and non-reactive in the two groups. The basal levels and secretion peaks of adrenaline, noradrenaline, growth hormone and Cortisol were similar between the two groups, but there was a significant delay in secretion in group A with a blood glucose threshold of adrenergic secretion of between 3.1 ±0.5 and 1.6 ± 0.2 mmoles/l in group A and between 4.6 ± 0.3 and 3.2 ± 0.2 mmoles/1 in group B (P < 0.05). This delayed secretion could be explained by desensitisation of the hypothalamic glucostat and could be due to the frequency and/or severity of hypoglycaemic episodes. Key words Insulin Dependent Diabetes - Hypoglycaemia - Hormonal Counterregulation
Introduction The loss of perception of adrenergic symptoms during hypoglycaemia constitutes a major risk factor for severe hypoglycaemia in insulin-dependent diabetics, particularly during optimisation of insulin therapy (DCCT Research Group 1987). This defect, which generally occurs after several years of insulin therapy, was described for thefirsttime in 1928 by Maddock and Trimble (1928). It appears to affect about
Horm. metabol. Res. 22 (1990) 90-95 © Georg Thieme Verlag Stuttgart • New York
20% of diabetics treated by conventional insulin therapy (Goldgewicht, Slama, Papoz and Tchobroutsky 1983) and is even more common in diabetics receiving optimised insulin therapy {Arias, Renter, Zier, Navascuesand Pfeiffer 1985). The pathogenesis of this defect remains controversial. Initially attributed to autonomic neuropathy (Hoeldtke, Boden, Shuman and Owen 1982), it was subsequently explained by a defect in adrenergic secretion occurring even in the absence of any sympathetic dysautonomy (Kleinbaum and Shamoon 1983). However, along with several other authors (Sussman, Crout and Marble 1963), we have been surprised by the fact that adrenergic symptoms are not always absent in these patients who no longer perceive hypoglycaemia, but occur following a delay, after the first signs of neuroglucopenia in the form of intellectual slowing and drowsiness. It is this delay which eliminates the value of adrenergic manifestations as warning symptoms. We therefore tried to confirm the existence of a delay in adrenergic secretion on insulin-induced hypoglycaemic tests in these diabetic patients who no longer or only poorly perceive these symptoms, resulting in repeated episodes of severe hypoglycaemia. Patients and Method Patients We systematically questioned 151 insulin-treated diabetics followed in our department about the frequency and severity of hypoglycaemic episodes, their blood glucose threshold of perception, their adrenergic symptoms (hunger, hotflushes,cold feelings, moist hands, sweating, tremor, palpitations, piloerection) and their neuroglucopenic symptoms (bizarre feelings, disturbances of intellectual concentration, drowsiness, mood disorders, visual disturbances, speech disorders, gait disorders, abnormal movements, paraesthesiae) and about the nature of the first two symptoms usually experienced. This questioning enabled us to distinguish 4 groups of patients (see Table 1): - The 1 st group consisted of 26 patients (17%) with little or no perception of hypoglycaemia and presenting with severe unexplained episodes of hypoglycaemia (coma, convulsions, confusional syndrome, requiring injection of GLUCAGON or intravenous glucose). The first two symptoms experienced were adrenergic in only 19and4%of cases). - The 2nd group consisted of 45 patients (30%), described as "poorly" perceiving hypoglycaemia, either due to a lowering of the threshold of perception or to a defect in adrenergic symptoms which had become inconstant or delayed. Thefirsttwo symptoms were adrenergic in only 35 and 28 % of cases.
Received: 28 March 1989
Accepted: 9 June 1989
Downloaded by: University of Illinois. Copyrighted material.
A. Grimaldi1, F. Bosquet1,P. Davidoff1, J. P. Digy1, C. Sachon1, C. Landault2, F. Thervet1, F.Zoghbi2and J. C. Legrand2 1Service Diab6tologie and 2Service Biochimie, C. H. U. Pitie-Salpetriere, Paris, France
H o r m . metabol. Res. 22 (1990)
Unawareness of Hypoglycaemia
Table 1 Characteristics of 151 insulin-treated diabetics classified on the basis of their awareness of hypoglycaemia-group I, "inconsistent awareness" responsible for severe hypoglycaemic episodes- group II, "reduced awareness" - group III, "normal awareness" -group IV, "very good awareness" (mean values ± sem) (AS = adrenergic symptoms). Diabetes duration (years)
Insulin dose (Ul/kg)
Age (years)
B.M.I.
1 "A.Sympt.
2 ° A. Sympt.
17
22 ± 3
0.76 ± 0.04
46 ± 3
22.0 ± 0.4
19%
4%
30
15 ± 11
0.71 ± 0.03
42 ± 2
22.3 ± 0.4
35%
28%
32
16±1
0.70 ± 0.03
46±2
23.0 ± 0.4
75%
66%
21
11 ± 2
0.76 ± 0.07
41 ± 3
22.6 ± 0.5
90%
80%
P < 0.001
N.S.
N.S.
N.S.
%
Group I n = 26 Group II n = 45 Group III n = 48 Group IV n = 32
P < 0.001
Age (years)
Diabetes duration (years)
Peripheral Neuropathy
Autonomic Neuropathy
Retinopathy = R Proteinuria = P
HbA1C
(N = 4-5.6%)
GroupA(n = 7) 3 9/4 0?
•
44 (24-60)
17 (3-53)
3/7
2/7
-3 R. severe -3 R. modest -1P.
7.7 ± 0.5% (6.1-10.6%)
GroupB(n = 7) 2 9/5 c?
•
47 (24-66)
8 (3-15)
4/7
3/7
-4 R. severe -3P.
8.4 ±0.7% (5.4-11.6%)
N.S.
N.S.
N.S.
N.S.
Significativity
The 3rd group consisted of 48 patients (32%), described as "normally" perceiving hypoglycaemia, without any modification since the onset of their diabetes, with a threshold of perception > 40 mg/dl. The first two symptoms were adrenergic in 75 and 66 % of cases. The 4th group consisted of 32 patients (21 %), with "excessive" perception of hypoglycaemia, with a threshold of perception of about 80 mg/dl. The first two symptoms were adrenergic in 90 and 80% of cases. From these 151 diabetics, we selected 7 insulintreated diabetics belonging to the first group, no longer perceiving hypoglycaemia and developing episodes of severe hypoglycaemia (group A) and 7 insulin-treated diabetics with excessive perception of hypoglycaemia, belonging to the 4th group (group B). These patients were selected so that groups A and B did not differ in terms of age, duration of diabetes, degenerative complications (particularly neuropathy), or control of the diabetes evaluated by the HbAl c level (see Table 2).
Protocol The insulin-induced hypoglycaemia test was performed by subcutaneous injection of insulin in order to obtain a progressive decrease in blood glucose comparable to that induced by ambulant treatment of diabetic patients. All of the patients were admitted to hospital. On the evening prior to the test, they received their usual insulin therapy with a 25 % reduction of the dose. The absence of noc-
N.S.
N.S.
turnal hypoglycaemia was confirmed by monitoring the capillary blood glucose from the fingertip every two hours. In the morning, the patients remained fasted and received their usual dose of insulin retard and two or three times their usual dose of rapid insulin (ACTRAPID®), depending on the blood glucose level in the fingertip. A catheter was inserted into a forearm vein and the patients remained in the supine position throughout the test. After subcutaneous injection of ACTRAPID® insulin, the capillary blood glucose was monitored every 15 minutes and when it became less than or equal to 160 mg/dl, samples were collected every 30 minutes for assay of blood glucose, adrenaline, noradrenaline, Cortisol, growth hormone and glucagon. In parallel with these blood samples, the neuroglucopenic and adrenergic symptoms were recorded and the pulse and blood pressure were monitored. The test was suspended after at least two samples in the symptomatic hypoglycaemic phase by immediate administration of sugar either orally or by intravenous injection of 40 cc of 30 % glucose solution, in the presence of drowsiness.
Assay techniques The blood glucose was assayed by the glucose oxidase method. HbAlc, with a normal value of between 4 and 5.6 %, was determined by cation exchange resin chromatography using microcolumns (BIOREX). Adrenaline and noradrenaline were assayed by a radioenzymatic method, using the UPJOHN kit (Peuler and Johnson 1977). Cortisol, glucagon and GH were assayed by a radioimmunological method using the Immunotech France®, Biodata® and RIAGnost Behring® kits respectively.
Downloaded by: University of Illinois. Copyrighted material.
Table 2 Characteristics of 7 insulin dependent diabetics with very poor awareness of hypoglycaemia and subject to severe hypoglycaemic episodes (group A • ) and of 7 insulin-treated diabetics with very good awareness of hypoglycaemia and not subject to hypoglycaemic episodes (group B • ).
Horm. metabol. Res. 22(1990)
A. Grimaldi, F. Bosquet, P. Davidoff, J. P. Digy, C. Sachon et al.
Fig. 2 Plasma epinephrine levels during S/C insulin-induced hypoglycaemia in 7 IDD's with poor awareness of hypoglycaemia (group A • ) and in 7 insulin-treated diabetics with good awareness of hypoglycaemia (group B • ) (mean ± sem).
tion of hypoglycaemia, presented a spontaneous rise in blood glucose at the 120th minute (3.85 ± 0.42 mmoles/1) in contrast with the patients in group A, in whom the blood glucose The results were expressed in the form of their mean value ± SEM. The blood glucose thresholds for secretion of adrena- continued to decrease (1.45 ± 0.17 mmoles/1 - P < 0.05) (see line, noradrenaline and GH were defined arbitrarily by doubling the Fig. 1). basal levels followed by a rise. As the blood samples were collected every 30 minutes, we were unable to precisely determine this blood The blood glucose threshold of perception was glucose threshold. We estimated it by means of the two limit blood glu- significantly different: 2.2 ± 0.34 mmoles/1 in group A vs cose values, i. e. the blood glucose preceding and immediately follow- 3.8 ± 0.41 mmoles/1 in group B (p < 0.001). The first symping the rise in the particular hormone. The threshold of secretion of tom was adrenergic in 3 cases and neuroglucopenic in 4 cases Cortisol was defined by a level representing 1.5 times the basal level, also followed by a rise. Statistically, the quantitative variables were in group A compared with adrenergic in 5 cases and neuroglucompared by means of analysis of variance (F) and the qualitative vari- copenic in 2 cases in group B. During hypoglycaemia, the ables were compared by means of the Chi-squared test (X2). The limit number of adrenergic symptoms over the number of neurogluof statistical significance adopted was 5 %. copenic symptoms (A/N) was 1.0 ± 0.8 in group A compared with 2.1 ± 1 in group B(P < 0.001). Results The counterregulatory hormones (adrenaline, Comparable levels of hypoglycaemia were ob- noradrenaline, Cortisol, GH) demonstrated a significant rise tained in groups A and B, with a blood glucose at time 0 of during hypoglycaemia with similar basal and maximal levels 6.04 ± 0.21 mmoles/1 for group A and 6.2 ± 0.35 mmoles/1 in the two groups. Only glucagon, not different between the for group B and a blood glucose at the 90th minute of 2 ± 0.16 two groups, did not increase during hypoglycaemia. A delay in mmoles/1 for group A and 2.8 ± 0.31 mmoles/1 for group B adrenaline secretion was observed in group A with a signifi(N. S.). However, the patients in group B, with good percep- cant difference between the two groups at the 60th minute
Expression of the results and statistical methods
Downloaded by: University of Illinois. Copyrighted material.
Fig. 1 S/C insulin-induced hypoglycaemia in 7 diabetics with poor awareness of hypoglycaemia (group A • ) and 7 diabetics with very good awareness of hypoglycaemia (group B • ) .
Unawareness ofHypoglycaemia
Horm. metabol. Res. 22(1990)
Group A • (n-7)
Group B • (n = 7)
Statistical Results
Epinephrine Glycemic threshold (mmol/l) between Basal level (pg/ml) Peak level (pg/ml)
3.1 ± 0 . 5 and 1.6 ± 0 . 2 30 ± 1 0 320 ± 124
4.6 ± 0.3 and 3.2 ± 0.2 24 ± 6 283 ± 52
P < 0.05 N.S. N.S.
Norepinephrine Glycemic threshold (mmol/l) between Basal level (pg/ml) Peak level (pg/ml)
4.0 ± 0 . 5 and 2.0 ± 0 . 3 264 ± 57 468 ± 104
4.6 ± 0 . 3 and 3.2 ± 0 . 3 203 ± 37 444 ± 8 6
P < 0.05 N.S. N.S.
Cortisol Glycemic threshold (mmol/l) between Basal level (ug/dl) Peak level ((ug/dl)
1.9 ±0.15 and 1.7 ± 0 . 2 13 ± 1 . 5 20 ± 3
3.3 ± 0.3 abd 2.9 ± 0.3 13 ± 2 18 ± 3
P < 0.01 N.S. N.S.
GH Glycemic threshold (mmol/l) between Basal level (ng/ml) Peak level (ng/ml)
3.4 ± 0 . 8 and 1.9 ± 0 . 4 2 ±0.7 13 ± 1 . 7
3.8 ± 0.6 and 3.2 ± 0.6 0.9 ± 0.07 7 ± 1.9
P < 0.05 N.S. N.S.
63.9 ± 3.8 66 ± 9.8
67.1 ±11.7 74.4 ± 6.6
Glucagon Basal level (pg/ml) Peak level (pg/ml)
N.S. N.S.
Fig. 3 Plasma epinephrine levels as a function of blood glucose during subcutaneous insulin-induced hypoglycaemia in 7 IDD's with poor awareness of hypoglycaemia (group A 9 ) and 7 diabetics with good awareness of hypoglycaemia (group B • ) . The epinephrine values correspond to the sum of measured values and values extrapolated from the individual curves, for blood glucose values groups grouped around ± 0.2 mmol/l.
(44 ± 18 pg/ml in group A vs 182 ± 61 pg/ml in group B P < 0.05) (see Fig. 2). The other counterregulatory hormones were not significantly different at the various measuring times. On the other hand, the blood glucose thresholds of secretion, expressed by the two values for blood glucose preceding and immediately following the rise in hormone, were significantly different for adrenaline, noradrena-
line, Cortisol and growth hormone (see Table 3) with a delay in secretion in group A. When the values for adrenaline (measured values and values extrapolated from the individual curves) corresponding to a given blood glucose level (blood glucose ± 0.2 mmoles/1) are analysed separately for each group, an adrenaline/blood glucose curve can be plotted (see Fig. 3). This curve confirms the shift to the left in the patients in group A with poor perception of hypoglycaemia.
Downloaded by: University of Illinois. Copyrighted material.
Table 3 Basal levels, peak levels and glycaemic threshold of epinephrine, norepinephrine, Cortisol and growth hormone secretion during subcutaneous insulin-induced hypoglycaemia (the threshold of secretion is defined by the blood glucose levels immediately prior to and following the doubling of basal levels of epinephrine, norepinephrine, GH and the multiplication by 1,5 of Cortisol levels (mean values + sem).
Horm. metabol. Res. 22(1990)
A. Grimaldi, F. Bosquet, P. Davidoff, J. P. Digy, C. Sachon et al.
HbA1c was not correlated with the blood glucose threshold of perception (r = 0.27, N. S.) nor with the blood glucose threshold of adrenergic secretion (r = 0.18, N. S.).
to hypoglycaemia induced by repetition and/or severity of the hypoglycaemia itself:
We also separately studied the 5 patients with clinical signs of autonomic neuropathy (2 patients from group A and 3 patients from group B). Their basal adrenaline and noradrenaline levels were significantly lower than those of the other 9 diabetic patients (12.3 ±2.3 pg/ml vs 41.9 ±9.3 pg/ml for adrenaline and 157.4 ± 23.6 pg/ml vs 310 ± 48.8 pg/ml for noradrenaline, respectively - P < 0.001), but the thresholds of secretion and the secretory peaks were not significantly different.
- Goldfien, Moore, Zibelli, Havens, Boling and Thorn (1961) reported that repeated insulin hypoglycaemic coma induced during the treatment of schizophrenic patients resulted in a loss of the adrenergic reaction. - MacCall, Fixman, Fleming, Tornheim, Chid and Ruderman (1986) showed that hypoglycaemia, induced in the rat either by grafting an insulinoma or by insulin pump or by subcutaneous injection of IPZ insulin, induced an increase in the cerebral tolerance to hypoglycaemia after 4 days, related to an increase by about 30 % in the capacity of cerebral extraction of glucose.
We therefore found that diabetic patients with defective perception of adrenergic warning symptoms of hypoglycaemia, who develop recurrent episodes of severe hypoglycaemia, present a delay in adrenaline secretion. However, the scatter of the results (illustrated by the size of the standard deviation) should be stressed, reflecting the diversity of the alteration in hormonal counterregulation, ranging from a simple delay in secretion with a normal peak of secretion to an absolute defect of secretion, at least for the blood glucose levels measured. This delay in adrenaline secretion was accompanied by a similar delay in the secretion of all of the hypothalamo-hypophyseal counterregulatory hormones. This delay perfectly explains the delay in adrenergic symptoms during hypoglycaemia, which consequently lose their warning value.
- Gulan, Perlman, Sole, Albisser and Zinman (1988) compared the effect on hormonal counterregulation of the optimisation of insulin therapy by continuous intravenous or subcutaneous insulin infusion. While the HbAlc level was identical with the two modalities of insulin therapy, insulin therapy with an intravenous pump did not induce any alteration in hormonal counterregulation in contrast with subcutaneous treatment. This difference could be explained by the three times lower frequency of hypoglycaemia with the intravenous route than with the subcutaneous route. This hypothalamo-hypophyseal desensitisation by repeated hypoglycaemic episodes may also explain the absence or delay of adrenergic symptoms during hypoglycaemia induced by insulinomas, in which the neuroglucopenic manifestations are generally predominant (Assan 1987).
In contrast, we confirmed the rise in the blood The difference between the results and those of previous studies by Polonski, Bergenstal, Pons, Schneider, Jas- glucose threshold of adrenergic secretion following resection pan and Rubenstein (1982); Kleinbaumand Shamoon (1983); of an insulinoma in one of our patients. The threshold which White, Skor, Cryer, Levandoski, Bier and Santiago (1983) and was 1.15 mmoles/1 before the operation, rose to 2.58 Bolli, De Feo, De Cosmo, Perriello, Ventura, Massi Benedetti, mmoles/1 one month after resection of the insulinoma (GriSanteusanio, Gerich and Brunetti (1984), which essentially de- maldi, Bosquet, Sachon, Davidoff, Landault, Zoghbi and Duet monstrated a secretion defect, may be only apparent inasmuch 1989). as the hypoglycaemia induced by these authors was less intense than in our study and consequently did not allow the In conclusion, the absence of adrenergic warnthreshold and secretory peaks to be evaluated in every case. On ing symptoms during hypoglycaemia occurring in insulin-dethe other hand, our results are in line with those reported by pendent diabetics could be explained by a delay in adrenergic Hirsch and Shamoon (1987); Heller, MacDonald, Herbert and secretion with or without an absolute secretion defect. This Tattersall (1987) and Amiel, Tamborlane, Simonson and Sher- delay would be due to a "down resetting" of the hypothalamic win (1987), showing a reduction in the blood glucose threshold glucostat and could be induced by repeated and/or severe epiof adrenergic secretion in patients with poor perception of hy- sodes of hypoglycaemia. Attempts to correct the HbAlc level poglycaemia. at the expense of repeated hypoglycaemia may induce this type of desensitisation associated with a high risk of severe hyThe cause of this "downward resetting" of the poglycaemia. It therefore seems justified to determine, for hypothalamic glucostat remains unknown. In contrast with each diabetic patient, a maximal benefit/risk ratio, as proAmielet al. (1987) and HellereX al. (1987), we did not find any posed by Unger(l9S2). Hyperglycaemia of between 100 and 180 mg/dl would still be compatible with the objective of precorrelation between the threshold of adrenergic secretion and HbAl c. However, our patients were selected in order to obtain vention of diabetic microangiopathy while strict correction of a comparable level of blood glucose control. In reality, the cor- HbAlc is only required in particular situations such as diarelation reported by these authors between the threshold of betic pregnancy and microangiopathy incipiens. adrenergic secretion and HbAlc appears to be explained not by correction of the blood glucose, but by an increase in the Acknowledgements frequency of hypoglycaemia, which is usually accompanied by a reduction in HbA1 c. Several arguments plead in favour of We would like to thank Professor Guillemand and the hypothesis of hypothalamo-hypophyseal desensitisation Doctor Duet for the hormonal assays and Mrs. M. Leclaire for her excellent technical assistance.
Downloaded by: University of Illinois. Copyrighted material.
Discussion
ofHypoglycaemia
Horm. metabol. Res. 22 (1990)
Hirsch, B. R., H. Shamoon: Defective epinephrine and growth hormone responses in type I diabetes are stimulus specific. Diabetes 36:20-26(1987) Amiel, S. A., W. Tamborlane, D. C. Simonson, R. S. Sherwin:DefecHoeldtke, R. C, K. M. Boden, C. R. Shuman, O. E. Owen: Reduced tive glucose counterregulation after strict glycemic control of inepinephrine secretion and hypoglycemia unawareness in diabetic sulin-dependent diabetes mellitus. N. Engl. J. Med. 316: 1375autonomic neuropathy. Ann. Intern. Med. 96:459-462 (1982) 1383(1987) Arias, P., W. Kerner, H. Zier, I. Navascues, E. F. Pfeiffer:Incidence ofKleinbaum, J., H. Shamoon: Impaired counterregulation of hypoglycemia in insulin-dependent diabetes mellitus. Diabetes 32: 493hypoglycemic episodes in diabetic patients under continuous sub498 (1983) cutaneous insulin infusion and intensified conventional insulin MacCall, A.L..L.B. Fixman, N. Fleming, K. Tornheim, W. Chid, N. B. treatment: assessment by means of semiambulatory 24-hour conRuderman: Chronic hypoglycemia increases brain glucose transtinuous blood glucose monitoring. Diabetes Care 8: 134-140 port. Am. J. Physiol. 251:442-447 (1986) (1985) Maddock, S. J., H. C. Trimble: Prolonged insulin hypoglycemia Assan, R.:Tumeurs beta-langerhanciennes insulinosecretantes ou inwithout symptoms. Jour. Amer. Med. Assoc. 91:616-621 (1928) sulinomes. In: Tumeurs endocrines du pandreas, Bonfils, S., Peuler, J. D., G. A. Johnson: Simultaneous single isotope radioenzyMignon, M.. Douin Edit., Paris, p. 67 (1987) Bolli, G. B., P. De Feo, S. De Cosmo, G. Perriello, M. M. Ventura, M. matic assay of plasma norepinephrine and dopamine. Life Sci. 21: 625 (1977) Massi Benedetti, F. Santeusanio, J. E. Gerich, P. Brunetti: A reliaPolonski, K., R. Bergenstal, G. Pons, M. Schneider, J. Jaspan, A. ble and reproducible test for adequate glucose counterregulation Rubenstein: Relation of counterregulatory responses to hypoglyon type I diabetes mellitus. diabetes 33:732-737 (1984) cemia in type I diabetics. N. Engl. J. Med. 307:1106-1112 (1982) DCCT Research Group: Diabetes control and complications trial Sussman, K. E., J. R. Crout, A. Marble: Failure of warning in insulin(DCCT): results of feasibility study. Diabetes Care 10: 1-19 induced hypoglycemic reactions. Diabetes 12:38-45 (1963) (1987) Goldflen, A., R. Moore, S. Zibelli, L. L. Havens, L. Boling, G. W. Unger, R. //..-Meticulous control of diabetes: benefits, risks and precautions. Diabetes 31:479-483 (1982) Thorn: Plasma epinephrine and norepinephrine levels during inWhite, N.H.,D.A. Skor, E. C. Cryer, L. A. Levandoski, D. M. Bier, J. sulin-induced hypoglycemia in man. J. Clin. Endocrinol. Metab. V. Santiago: Identification of type I diabetic patients at increased 21:296-304(1961) risk for hypoglycemia during intensive therapy. N. Engl. J. Med. Goldgewicht, C, G. Slama, L. Papoz, G. Tchobroutsky: Hypogly308:485-491(1983) caemic reactions in 172 type 1 (insulin dependent) diabetic patients. Diabetologia 24:95-99 (1983) Grimaldi, A., F. Bosquet, C. Sachon, P. Davidoff, C. Landault, F. Zoghbi, M. Duet:Abaissement du seuil glycemique de reponse hy- Requests for reprints should be addressed to: pophysaire et surrenalienne a l'hypoglycemie lors d'un insuliA. Grimaldi nome. Diabete et Metabolisme 15:209 (1989) Gulan, M., K. Perlman, M. Sole, A. M. Albisser, B. Zinman:Counter- Service Diabetologie regulatory hormone responses preserved after long-term inC. H. U. Pitie-Salpetriere travenous insulin infusion compared to continuous subcutaneous 83 boulevard de l'hopital insulin infusion. Diabetes 37:526-531 (1988) F-75013 Paris (France) Heller, S. R., J. A. MacDonald, M. Herbert, R. B. Tattersall:Influence of sympathetic nervous system on hypoglycaemic warning symptoms. Lancet IIi: 359-363 (1987) References
Downloaded by: University of Illinois. Copyrighted material.
Unawareness
Summary After several years of insulin therapy, about 20 % of insulin-dependent diabetics have little or no perception of hypoglycaemia because of a loss of the adrenergic warning symptoms. This defect, poorly correlated with the presence of autonomic neuropathy, has been classically explained by a defect in the catecholamine secretion. We compared the hormonal counterregulation during hypoglycaemia induced by subcutaneous injection of insulin in 7 insulin-dependent diabetics with poor perception of hypoglycaemia and experiencing repeated episodes of severe hypoglycaemia (group A) and 7 insulin-treated diabetics with very good perception of hypoglycaemia and not experiencing severe hypoglycaemia (group B). Groups A and B were similar in terms of age, duration of diabetes, HbA1c level and degenerative complications. The glucagon levels were identical and non-reactive in the two groups. The basal levels and secretion peaks of adrenaline, noradrenaline, growth hormone and Cortisol were similar between the two groups, but there was a significant delay in secretion in group A with a blood glucose threshold of adrenergic secretion of between 3.1 ±0.5 and 1.6 ± 0.2 mmoles/l in group A and between 4.6 ± 0.3 and 3.2 ± 0.2 mmoles/1 in group B (P < 0.05). This delayed secretion could be explained by desensitisation of the hypothalamic glucostat and could be due to the frequency and/or severity of hypoglycaemic episodes. Key words Insulin Dependent Diabetes - Hypoglycaemia - Hormonal Counterregulation
Introduction The loss of perception of adrenergic symptoms during hypoglycaemia constitutes a major risk factor for severe hypoglycaemia in insulin-dependent diabetics, particularly during optimisation of insulin therapy (DCCT Research Group 1987). This defect, which generally occurs after several years of insulin therapy, was described for thefirsttime in 1928 by Maddock and Trimble (1928). It appears to affect about
Horm. metabol. Res. 22 (1990) 90-95 © Georg Thieme Verlag Stuttgart • New York
20% of diabetics treated by conventional insulin therapy (Goldgewicht, Slama, Papoz and Tchobroutsky 1983) and is even more common in diabetics receiving optimised insulin therapy {Arias, Renter, Zier, Navascuesand Pfeiffer 1985). The pathogenesis of this defect remains controversial. Initially attributed to autonomic neuropathy (Hoeldtke, Boden, Shuman and Owen 1982), it was subsequently explained by a defect in adrenergic secretion occurring even in the absence of any sympathetic dysautonomy (Kleinbaum and Shamoon 1983). However, along with several other authors (Sussman, Crout and Marble 1963), we have been surprised by the fact that adrenergic symptoms are not always absent in these patients who no longer perceive hypoglycaemia, but occur following a delay, after the first signs of neuroglucopenia in the form of intellectual slowing and drowsiness. It is this delay which eliminates the value of adrenergic manifestations as warning symptoms. We therefore tried to confirm the existence of a delay in adrenergic secretion on insulin-induced hypoglycaemic tests in these diabetic patients who no longer or only poorly perceive these symptoms, resulting in repeated episodes of severe hypoglycaemia. Patients and Method Patients We systematically questioned 151 insulin-treated diabetics followed in our department about the frequency and severity of hypoglycaemic episodes, their blood glucose threshold of perception, their adrenergic symptoms (hunger, hotflushes,cold feelings, moist hands, sweating, tremor, palpitations, piloerection) and their neuroglucopenic symptoms (bizarre feelings, disturbances of intellectual concentration, drowsiness, mood disorders, visual disturbances, speech disorders, gait disorders, abnormal movements, paraesthesiae) and about the nature of the first two symptoms usually experienced. This questioning enabled us to distinguish 4 groups of patients (see Table 1): - The 1 st group consisted of 26 patients (17%) with little or no perception of hypoglycaemia and presenting with severe unexplained episodes of hypoglycaemia (coma, convulsions, confusional syndrome, requiring injection of GLUCAGON or intravenous glucose). The first two symptoms experienced were adrenergic in only 19and4%of cases). - The 2nd group consisted of 45 patients (30%), described as "poorly" perceiving hypoglycaemia, either due to a lowering of the threshold of perception or to a defect in adrenergic symptoms which had become inconstant or delayed. Thefirsttwo symptoms were adrenergic in only 35 and 28 % of cases.
Received: 28 March 1989
Accepted: 9 June 1989
Downloaded by: University of Illinois. Copyrighted material.
A. Grimaldi1, F. Bosquet1,P. Davidoff1, J. P. Digy1, C. Sachon1, C. Landault2, F. Thervet1, F.Zoghbi2and J. C. Legrand2 1Service Diab6tologie and 2Service Biochimie, C. H. U. Pitie-Salpetriere, Paris, France
H o r m . metabol. Res. 22 (1990)
Unawareness of Hypoglycaemia
Table 1 Characteristics of 151 insulin-treated diabetics classified on the basis of their awareness of hypoglycaemia-group I, "inconsistent awareness" responsible for severe hypoglycaemic episodes- group II, "reduced awareness" - group III, "normal awareness" -group IV, "very good awareness" (mean values ± sem) (AS = adrenergic symptoms). Diabetes duration (years)
Insulin dose (Ul/kg)
Age (years)
B.M.I.
1 "A.Sympt.
2 ° A. Sympt.
17
22 ± 3
0.76 ± 0.04
46 ± 3
22.0 ± 0.4
19%
4%
30
15 ± 11
0.71 ± 0.03
42 ± 2
22.3 ± 0.4
35%
28%
32
16±1
0.70 ± 0.03
46±2
23.0 ± 0.4
75%
66%
21
11 ± 2
0.76 ± 0.07
41 ± 3
22.6 ± 0.5
90%
80%
P < 0.001
N.S.
N.S.
N.S.
%
Group I n = 26 Group II n = 45 Group III n = 48 Group IV n = 32
P < 0.001
Age (years)
Diabetes duration (years)
Peripheral Neuropathy
Autonomic Neuropathy
Retinopathy = R Proteinuria = P
HbA1C
(N = 4-5.6%)
GroupA(n = 7) 3 9/4 0?
•
44 (24-60)
17 (3-53)
3/7
2/7
-3 R. severe -3 R. modest -1P.
7.7 ± 0.5% (6.1-10.6%)
GroupB(n = 7) 2 9/5 c?
•
47 (24-66)
8 (3-15)
4/7
3/7
-4 R. severe -3P.
8.4 ±0.7% (5.4-11.6%)
N.S.
N.S.
N.S.
N.S.
Significativity
The 3rd group consisted of 48 patients (32%), described as "normally" perceiving hypoglycaemia, without any modification since the onset of their diabetes, with a threshold of perception > 40 mg/dl. The first two symptoms were adrenergic in 75 and 66 % of cases. The 4th group consisted of 32 patients (21 %), with "excessive" perception of hypoglycaemia, with a threshold of perception of about 80 mg/dl. The first two symptoms were adrenergic in 90 and 80% of cases. From these 151 diabetics, we selected 7 insulintreated diabetics belonging to the first group, no longer perceiving hypoglycaemia and developing episodes of severe hypoglycaemia (group A) and 7 insulin-treated diabetics with excessive perception of hypoglycaemia, belonging to the 4th group (group B). These patients were selected so that groups A and B did not differ in terms of age, duration of diabetes, degenerative complications (particularly neuropathy), or control of the diabetes evaluated by the HbAl c level (see Table 2).
Protocol The insulin-induced hypoglycaemia test was performed by subcutaneous injection of insulin in order to obtain a progressive decrease in blood glucose comparable to that induced by ambulant treatment of diabetic patients. All of the patients were admitted to hospital. On the evening prior to the test, they received their usual insulin therapy with a 25 % reduction of the dose. The absence of noc-
N.S.
N.S.
turnal hypoglycaemia was confirmed by monitoring the capillary blood glucose from the fingertip every two hours. In the morning, the patients remained fasted and received their usual dose of insulin retard and two or three times their usual dose of rapid insulin (ACTRAPID®), depending on the blood glucose level in the fingertip. A catheter was inserted into a forearm vein and the patients remained in the supine position throughout the test. After subcutaneous injection of ACTRAPID® insulin, the capillary blood glucose was monitored every 15 minutes and when it became less than or equal to 160 mg/dl, samples were collected every 30 minutes for assay of blood glucose, adrenaline, noradrenaline, Cortisol, growth hormone and glucagon. In parallel with these blood samples, the neuroglucopenic and adrenergic symptoms were recorded and the pulse and blood pressure were monitored. The test was suspended after at least two samples in the symptomatic hypoglycaemic phase by immediate administration of sugar either orally or by intravenous injection of 40 cc of 30 % glucose solution, in the presence of drowsiness.
Assay techniques The blood glucose was assayed by the glucose oxidase method. HbAlc, with a normal value of between 4 and 5.6 %, was determined by cation exchange resin chromatography using microcolumns (BIOREX). Adrenaline and noradrenaline were assayed by a radioenzymatic method, using the UPJOHN kit (Peuler and Johnson 1977). Cortisol, glucagon and GH were assayed by a radioimmunological method using the Immunotech France®, Biodata® and RIAGnost Behring® kits respectively.
Downloaded by: University of Illinois. Copyrighted material.
Table 2 Characteristics of 7 insulin dependent diabetics with very poor awareness of hypoglycaemia and subject to severe hypoglycaemic episodes (group A • ) and of 7 insulin-treated diabetics with very good awareness of hypoglycaemia and not subject to hypoglycaemic episodes (group B • ).
Horm. metabol. Res. 22(1990)
A. Grimaldi, F. Bosquet, P. Davidoff, J. P. Digy, C. Sachon et al.
Fig. 2 Plasma epinephrine levels during S/C insulin-induced hypoglycaemia in 7 IDD's with poor awareness of hypoglycaemia (group A • ) and in 7 insulin-treated diabetics with good awareness of hypoglycaemia (group B • ) (mean ± sem).
tion of hypoglycaemia, presented a spontaneous rise in blood glucose at the 120th minute (3.85 ± 0.42 mmoles/1) in contrast with the patients in group A, in whom the blood glucose The results were expressed in the form of their mean value ± SEM. The blood glucose thresholds for secretion of adrena- continued to decrease (1.45 ± 0.17 mmoles/1 - P < 0.05) (see line, noradrenaline and GH were defined arbitrarily by doubling the Fig. 1). basal levels followed by a rise. As the blood samples were collected every 30 minutes, we were unable to precisely determine this blood The blood glucose threshold of perception was glucose threshold. We estimated it by means of the two limit blood glu- significantly different: 2.2 ± 0.34 mmoles/1 in group A vs cose values, i. e. the blood glucose preceding and immediately follow- 3.8 ± 0.41 mmoles/1 in group B (p < 0.001). The first symping the rise in the particular hormone. The threshold of secretion of tom was adrenergic in 3 cases and neuroglucopenic in 4 cases Cortisol was defined by a level representing 1.5 times the basal level, also followed by a rise. Statistically, the quantitative variables were in group A compared with adrenergic in 5 cases and neuroglucompared by means of analysis of variance (F) and the qualitative vari- copenic in 2 cases in group B. During hypoglycaemia, the ables were compared by means of the Chi-squared test (X2). The limit number of adrenergic symptoms over the number of neurogluof statistical significance adopted was 5 %. copenic symptoms (A/N) was 1.0 ± 0.8 in group A compared with 2.1 ± 1 in group B(P < 0.001). Results The counterregulatory hormones (adrenaline, Comparable levels of hypoglycaemia were ob- noradrenaline, Cortisol, GH) demonstrated a significant rise tained in groups A and B, with a blood glucose at time 0 of during hypoglycaemia with similar basal and maximal levels 6.04 ± 0.21 mmoles/1 for group A and 6.2 ± 0.35 mmoles/1 in the two groups. Only glucagon, not different between the for group B and a blood glucose at the 90th minute of 2 ± 0.16 two groups, did not increase during hypoglycaemia. A delay in mmoles/1 for group A and 2.8 ± 0.31 mmoles/1 for group B adrenaline secretion was observed in group A with a signifi(N. S.). However, the patients in group B, with good percep- cant difference between the two groups at the 60th minute
Expression of the results and statistical methods
Downloaded by: University of Illinois. Copyrighted material.
Fig. 1 S/C insulin-induced hypoglycaemia in 7 diabetics with poor awareness of hypoglycaemia (group A • ) and 7 diabetics with very good awareness of hypoglycaemia (group B • ) .
Unawareness ofHypoglycaemia
Horm. metabol. Res. 22(1990)
Group A • (n-7)
Group B • (n = 7)
Statistical Results
Epinephrine Glycemic threshold (mmol/l) between Basal level (pg/ml) Peak level (pg/ml)
3.1 ± 0 . 5 and 1.6 ± 0 . 2 30 ± 1 0 320 ± 124
4.6 ± 0.3 and 3.2 ± 0.2 24 ± 6 283 ± 52
P < 0.05 N.S. N.S.
Norepinephrine Glycemic threshold (mmol/l) between Basal level (pg/ml) Peak level (pg/ml)
4.0 ± 0 . 5 and 2.0 ± 0 . 3 264 ± 57 468 ± 104
4.6 ± 0 . 3 and 3.2 ± 0 . 3 203 ± 37 444 ± 8 6
P < 0.05 N.S. N.S.
Cortisol Glycemic threshold (mmol/l) between Basal level (ug/dl) Peak level ((ug/dl)
1.9 ±0.15 and 1.7 ± 0 . 2 13 ± 1 . 5 20 ± 3
3.3 ± 0.3 abd 2.9 ± 0.3 13 ± 2 18 ± 3
P < 0.01 N.S. N.S.
GH Glycemic threshold (mmol/l) between Basal level (ng/ml) Peak level (ng/ml)
3.4 ± 0 . 8 and 1.9 ± 0 . 4 2 ±0.7 13 ± 1 . 7
3.8 ± 0.6 and 3.2 ± 0.6 0.9 ± 0.07 7 ± 1.9
P < 0.05 N.S. N.S.
63.9 ± 3.8 66 ± 9.8
67.1 ±11.7 74.4 ± 6.6
Glucagon Basal level (pg/ml) Peak level (pg/ml)
N.S. N.S.
Fig. 3 Plasma epinephrine levels as a function of blood glucose during subcutaneous insulin-induced hypoglycaemia in 7 IDD's with poor awareness of hypoglycaemia (group A 9 ) and 7 diabetics with good awareness of hypoglycaemia (group B • ) . The epinephrine values correspond to the sum of measured values and values extrapolated from the individual curves, for blood glucose values groups grouped around ± 0.2 mmol/l.
(44 ± 18 pg/ml in group A vs 182 ± 61 pg/ml in group B P < 0.05) (see Fig. 2). The other counterregulatory hormones were not significantly different at the various measuring times. On the other hand, the blood glucose thresholds of secretion, expressed by the two values for blood glucose preceding and immediately following the rise in hormone, were significantly different for adrenaline, noradrena-
line, Cortisol and growth hormone (see Table 3) with a delay in secretion in group A. When the values for adrenaline (measured values and values extrapolated from the individual curves) corresponding to a given blood glucose level (blood glucose ± 0.2 mmoles/1) are analysed separately for each group, an adrenaline/blood glucose curve can be plotted (see Fig. 3). This curve confirms the shift to the left in the patients in group A with poor perception of hypoglycaemia.
Downloaded by: University of Illinois. Copyrighted material.
Table 3 Basal levels, peak levels and glycaemic threshold of epinephrine, norepinephrine, Cortisol and growth hormone secretion during subcutaneous insulin-induced hypoglycaemia (the threshold of secretion is defined by the blood glucose levels immediately prior to and following the doubling of basal levels of epinephrine, norepinephrine, GH and the multiplication by 1,5 of Cortisol levels (mean values + sem).
Horm. metabol. Res. 22(1990)
A. Grimaldi, F. Bosquet, P. Davidoff, J. P. Digy, C. Sachon et al.
HbA1c was not correlated with the blood glucose threshold of perception (r = 0.27, N. S.) nor with the blood glucose threshold of adrenergic secretion (r = 0.18, N. S.).
to hypoglycaemia induced by repetition and/or severity of the hypoglycaemia itself:
We also separately studied the 5 patients with clinical signs of autonomic neuropathy (2 patients from group A and 3 patients from group B). Their basal adrenaline and noradrenaline levels were significantly lower than those of the other 9 diabetic patients (12.3 ±2.3 pg/ml vs 41.9 ±9.3 pg/ml for adrenaline and 157.4 ± 23.6 pg/ml vs 310 ± 48.8 pg/ml for noradrenaline, respectively - P < 0.001), but the thresholds of secretion and the secretory peaks were not significantly different.
- Goldfien, Moore, Zibelli, Havens, Boling and Thorn (1961) reported that repeated insulin hypoglycaemic coma induced during the treatment of schizophrenic patients resulted in a loss of the adrenergic reaction. - MacCall, Fixman, Fleming, Tornheim, Chid and Ruderman (1986) showed that hypoglycaemia, induced in the rat either by grafting an insulinoma or by insulin pump or by subcutaneous injection of IPZ insulin, induced an increase in the cerebral tolerance to hypoglycaemia after 4 days, related to an increase by about 30 % in the capacity of cerebral extraction of glucose.
We therefore found that diabetic patients with defective perception of adrenergic warning symptoms of hypoglycaemia, who develop recurrent episodes of severe hypoglycaemia, present a delay in adrenaline secretion. However, the scatter of the results (illustrated by the size of the standard deviation) should be stressed, reflecting the diversity of the alteration in hormonal counterregulation, ranging from a simple delay in secretion with a normal peak of secretion to an absolute defect of secretion, at least for the blood glucose levels measured. This delay in adrenaline secretion was accompanied by a similar delay in the secretion of all of the hypothalamo-hypophyseal counterregulatory hormones. This delay perfectly explains the delay in adrenergic symptoms during hypoglycaemia, which consequently lose their warning value.
- Gulan, Perlman, Sole, Albisser and Zinman (1988) compared the effect on hormonal counterregulation of the optimisation of insulin therapy by continuous intravenous or subcutaneous insulin infusion. While the HbAlc level was identical with the two modalities of insulin therapy, insulin therapy with an intravenous pump did not induce any alteration in hormonal counterregulation in contrast with subcutaneous treatment. This difference could be explained by the three times lower frequency of hypoglycaemia with the intravenous route than with the subcutaneous route. This hypothalamo-hypophyseal desensitisation by repeated hypoglycaemic episodes may also explain the absence or delay of adrenergic symptoms during hypoglycaemia induced by insulinomas, in which the neuroglucopenic manifestations are generally predominant (Assan 1987).
In contrast, we confirmed the rise in the blood The difference between the results and those of previous studies by Polonski, Bergenstal, Pons, Schneider, Jas- glucose threshold of adrenergic secretion following resection pan and Rubenstein (1982); Kleinbaumand Shamoon (1983); of an insulinoma in one of our patients. The threshold which White, Skor, Cryer, Levandoski, Bier and Santiago (1983) and was 1.15 mmoles/1 before the operation, rose to 2.58 Bolli, De Feo, De Cosmo, Perriello, Ventura, Massi Benedetti, mmoles/1 one month after resection of the insulinoma (GriSanteusanio, Gerich and Brunetti (1984), which essentially de- maldi, Bosquet, Sachon, Davidoff, Landault, Zoghbi and Duet monstrated a secretion defect, may be only apparent inasmuch 1989). as the hypoglycaemia induced by these authors was less intense than in our study and consequently did not allow the In conclusion, the absence of adrenergic warnthreshold and secretory peaks to be evaluated in every case. On ing symptoms during hypoglycaemia occurring in insulin-dethe other hand, our results are in line with those reported by pendent diabetics could be explained by a delay in adrenergic Hirsch and Shamoon (1987); Heller, MacDonald, Herbert and secretion with or without an absolute secretion defect. This Tattersall (1987) and Amiel, Tamborlane, Simonson and Sher- delay would be due to a "down resetting" of the hypothalamic win (1987), showing a reduction in the blood glucose threshold glucostat and could be induced by repeated and/or severe epiof adrenergic secretion in patients with poor perception of hy- sodes of hypoglycaemia. Attempts to correct the HbAlc level poglycaemia. at the expense of repeated hypoglycaemia may induce this type of desensitisation associated with a high risk of severe hyThe cause of this "downward resetting" of the poglycaemia. It therefore seems justified to determine, for hypothalamic glucostat remains unknown. In contrast with each diabetic patient, a maximal benefit/risk ratio, as proAmielet al. (1987) and HellereX al. (1987), we did not find any posed by Unger(l9S2). Hyperglycaemia of between 100 and 180 mg/dl would still be compatible with the objective of precorrelation between the threshold of adrenergic secretion and HbAl c. However, our patients were selected in order to obtain vention of diabetic microangiopathy while strict correction of a comparable level of blood glucose control. In reality, the cor- HbAlc is only required in particular situations such as diarelation reported by these authors between the threshold of betic pregnancy and microangiopathy incipiens. adrenergic secretion and HbAlc appears to be explained not by correction of the blood glucose, but by an increase in the Acknowledgements frequency of hypoglycaemia, which is usually accompanied by a reduction in HbA1 c. Several arguments plead in favour of We would like to thank Professor Guillemand and the hypothesis of hypothalamo-hypophyseal desensitisation Doctor Duet for the hormonal assays and Mrs. M. Leclaire for her excellent technical assistance.
Downloaded by: University of Illinois. Copyrighted material.
Discussion
ofHypoglycaemia
Horm. metabol. Res. 22 (1990)
Hirsch, B. R., H. Shamoon: Defective epinephrine and growth hormone responses in type I diabetes are stimulus specific. Diabetes 36:20-26(1987) Amiel, S. A., W. Tamborlane, D. C. Simonson, R. S. Sherwin:DefecHoeldtke, R. C, K. M. Boden, C. R. Shuman, O. E. Owen: Reduced tive glucose counterregulation after strict glycemic control of inepinephrine secretion and hypoglycemia unawareness in diabetic sulin-dependent diabetes mellitus. N. Engl. J. Med. 316: 1375autonomic neuropathy. Ann. Intern. Med. 96:459-462 (1982) 1383(1987) Arias, P., W. Kerner, H. Zier, I. Navascues, E. F. Pfeiffer:Incidence ofKleinbaum, J., H. Shamoon: Impaired counterregulation of hypoglycemia in insulin-dependent diabetes mellitus. Diabetes 32: 493hypoglycemic episodes in diabetic patients under continuous sub498 (1983) cutaneous insulin infusion and intensified conventional insulin MacCall, A.L..L.B. Fixman, N. Fleming, K. Tornheim, W. Chid, N. B. treatment: assessment by means of semiambulatory 24-hour conRuderman: Chronic hypoglycemia increases brain glucose transtinuous blood glucose monitoring. Diabetes Care 8: 134-140 port. Am. J. Physiol. 251:442-447 (1986) (1985) Maddock, S. J., H. C. Trimble: Prolonged insulin hypoglycemia Assan, R.:Tumeurs beta-langerhanciennes insulinosecretantes ou inwithout symptoms. Jour. Amer. Med. Assoc. 91:616-621 (1928) sulinomes. In: Tumeurs endocrines du pandreas, Bonfils, S., Peuler, J. D., G. A. Johnson: Simultaneous single isotope radioenzyMignon, M.. Douin Edit., Paris, p. 67 (1987) Bolli, G. B., P. De Feo, S. De Cosmo, G. Perriello, M. M. Ventura, M. matic assay of plasma norepinephrine and dopamine. Life Sci. 21: 625 (1977) Massi Benedetti, F. Santeusanio, J. E. Gerich, P. Brunetti: A reliaPolonski, K., R. Bergenstal, G. Pons, M. Schneider, J. Jaspan, A. ble and reproducible test for adequate glucose counterregulation Rubenstein: Relation of counterregulatory responses to hypoglyon type I diabetes mellitus. diabetes 33:732-737 (1984) cemia in type I diabetics. N. Engl. J. Med. 307:1106-1112 (1982) DCCT Research Group: Diabetes control and complications trial Sussman, K. E., J. R. Crout, A. Marble: Failure of warning in insulin(DCCT): results of feasibility study. Diabetes Care 10: 1-19 induced hypoglycemic reactions. Diabetes 12:38-45 (1963) (1987) Goldflen, A., R. Moore, S. Zibelli, L. L. Havens, L. Boling, G. W. Unger, R. //..-Meticulous control of diabetes: benefits, risks and precautions. Diabetes 31:479-483 (1982) Thorn: Plasma epinephrine and norepinephrine levels during inWhite, N.H.,D.A. Skor, E. C. Cryer, L. A. Levandoski, D. M. Bier, J. sulin-induced hypoglycemia in man. J. Clin. Endocrinol. Metab. V. Santiago: Identification of type I diabetic patients at increased 21:296-304(1961) risk for hypoglycemia during intensive therapy. N. Engl. J. Med. Goldgewicht, C, G. Slama, L. Papoz, G. Tchobroutsky: Hypogly308:485-491(1983) caemic reactions in 172 type 1 (insulin dependent) diabetic patients. Diabetologia 24:95-99 (1983) Grimaldi, A., F. Bosquet, C. Sachon, P. Davidoff, C. Landault, F. Zoghbi, M. Duet:Abaissement du seuil glycemique de reponse hy- Requests for reprints should be addressed to: pophysaire et surrenalienne a l'hypoglycemie lors d'un insuliA. Grimaldi nome. Diabete et Metabolisme 15:209 (1989) Gulan, M., K. Perlman, M. Sole, A. M. Albisser, B. Zinman:Counter- Service Diabetologie regulatory hormone responses preserved after long-term inC. H. U. Pitie-Salpetriere travenous insulin infusion compared to continuous subcutaneous 83 boulevard de l'hopital insulin infusion. Diabetes 37:526-531 (1988) F-75013 Paris (France) Heller, S. R., J. A. MacDonald, M. Herbert, R. B. Tattersall:Influence of sympathetic nervous system on hypoglycaemic warning symptoms. Lancet IIi: 359-363 (1987) References
Downloaded by: University of Illinois. Copyrighted material.
Unawareness
