- Email: [email protected]
Severe insulin resistance in a patient with type 1 diabetes and stiff-man syndrome treated with insulin lispro
Diabetes Research and Clinical Practice 41 (1998) 197 – 202
Case report
Severe insulin resistance in a patient with type 1 diabetes and stiff-man syndrome treated with insulin lispro ˚ ke Lernmark, Alan Chait Irl B. Hirsch *, David D’Alessio, Lily Eng, Connie Davis, A UWMC, 1959 NE Pacific Street, Box 356176, Seattle, WA 98195, USA Accepted 26 May 1998
Abstract We describe a patient with type 1 diabetes with recurrent diabetic ketoacidosis and severe insulin resistance. Extensive evaluation of the etiology of the insulin resistance did not reveal an etiology, and well over 1000 U of daily insulin did not prevent the ketoacidosis. Her blood glucose and insulin requirements were improved with glucocorticoids and octreotide, but the effects of both of these agents were short-lived. She was given a trial of insulin lispro with immediate and dramatic effects, lowering her HbA1c from 14.6 to 5.1% in 7 months with a decrease in insulin requirements of 1600–100 U per day. Besides her diabetes, she had a history of pain and stiffness affecting numerous muscle groups, and hospitalization was required for pain control. The diagnosis of stiff-man syndrome (SMS) was confirmed with high titers of glutamic acid decarboxylase 65 antibodies in both serum and cerebral spinal fluid. In summary, we describe the first patient with type 1 diabetes, SMS, and severe insulin resistance. Although the etiology of the insulin resistance is unknown, due to the efficacious response to insulin lispro, hydrocortisone, and perhaps octreotide, we propose an immune-mediated etiology. Although rare, this syndrome needs to be considered as an etiology of insulin resistance. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Glucocorticoid; Stiff man syndrome; Insulin lispro; Insulin resistance; Octreotide
1. Case presentation The patient is a 33-year-old Caucasian woman who presented to the University of Washington in * Corresponding author. Tel.: + 1 206 5484882; fax: +1 206 5484976.
March of 1994. She had a 2-year history of type 1 diabetes complicated by numerous episodes of diabetic ketoacidosis (DKA) requiring hospitalization. During the previous 9 months, a 9-kg weight loss accompanied symptoms of uncontrolled hyperglycemia. Her insulin dose was 60 U per day of human insulin, split into two injections
0168-8227/98/$ - see front matter © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0168-8227(98)00072-2
I.B. Hirsch et al. / Diabetes Research and Clinical Practice 41 (1998) 197–202
198
of NPH and regular insulin administered prior to breakfast and supper. Her only other medication was thyroxine. Physical examination revealed an emaciated woman, with a weight of 53.6 kg (body mass index=19.2 kg/m2). Her supine blood pressure and pulse were 110/70 and 52, respectively, and there were no postural changes. The rest of the exam was unremarkable and there were no neurological abnormalities. Initial laboratory assessment revealed: HbA1c, 12.2% (nl, 4.0–6.0%) and mild DKA. She was treated with intravenous fluids, insulin, and potassium and during the hospitalization her subcutaneous insulin dose was increased. Although there was concern the patient was not administering her insulin, there was no previous psychiatric history. One week after hospitalization the patient was receiving 90 U/day of insulin. However, a random plasma glucose was 32.3 mmol/l with a carbon dioxide of 15 mEq/l and an anion gap of 17. Serum C-peptide was
0.07 nmol/l and total serum insulin was 606 pmol/l. During the next 3 months she was admitted to the University of Washington Medical Center on three occasions for the treatment of DKA. She usually required between 3 and 8 U/h of intravenous insulin, and continuous increases in her subcutaneous insulin dose were required (see Table 1). The nursing staff injected all subcutaneous insulin during the hospitalizations. By late May the patient was receiving 920 U of insulin per day, administered as twice daily ultralente and thrice daily regular insulin. Because of the apparent insulin resistance, further evaluation was undertaken. Plasma IGF1 and glucagon were normal, as was a 24-h urinary free cortisol. Insulin receptor antibodies, insulin antibodies, and antinuclear antibodies were negative. Sedimentation rate was 8. Antimicrosomal antibodies were positive at 1:40, supporting a diagnosis of Hashimoto’s thyroiditis.
Table 1 Patient characteristics and medication doses Date
HbA1c (%) (nl 4.0–6.0%)
BMI (kg/m2)
Insulin (units/day)
Prednisone (mg/day)
Octreotide (mg/day)
3/30/94 4/25/94 5/11/94 6/21/94 7/20/94 9/13/94 11/14/94 12/30/94 2/24/95 4/24/95 6/23/95 8/28/95 11/13/95 12/19/95 1/23/96 2/28/96 3/26/96 5/1/96 6/24/96 3/17/98
12.2
19.0 20.1 20.9 20.0 21.0 22.5 22.0 22.7 22.3 22.5 21.0 20.3 20.0 23.4 22.7 21.4 21.9 21.9 21.9 24.8
60 150 240 1500 250 350 1500 150 500 450 800 1200 1600 300b 300b 250c 200c 150c 100c 175c
20 10 20 10 10 10 10 10 10 10 10 7.5 20 20 20 10
200 200 300 150a 300a 700a 1000a 550 550 550 300 100 0 0
a
13.5 13.9 10.5 14.3
14.6 12.6 8.5 7.0 7.1 5.1 5.2 6.7
Continuous subcutaneous octreotide via external pump. Insulin lispro via insulin infusion pump with lispro and regular insulin before meals. c Insulin lispro alone via insulin infusion pump. b
I.B. Hirsch et al. / Diabetes Research and Clinical Practice 41 (1998) 197–202
Fig. 1. Response to hydrocortisone in a patient with severe insulin resistance. Blood glucose was maintained with an intravenous insulin infusion at a rate of 7–8 U/h. Two hours after administering the hydrocortisone, the intravenous insulin infusion was discontinued. HC, hydrocortisone.
To test the absorption of insulin from the subcutaneous space, 200 U of human regular insulin from a new vial was administered by one of her physicians (IBH), serial blood samples were drawn, and aliquots of plasma assayed by radioimmunoassay directly (total insulin) or after precipitation of immunoglobulins with 30% polyethylene glycol (free insulin): Time Glucose Total Free insulin (30% (h) (mmol/l) insulin polyethylene glycol) (pmol/l) (pmol/l) 0 1 2
26.7 23.3 21.9
1830 3540 4230
1920 3370 4140
These results suggested that the patient’s absorption of insulin was adequate and that there was no antibody interfering with the circulating insulin. In late June of 1994, the patient presented with another episode of DKA. Based on the hypothesis that her profound insulin resistance was mediated by an immune process, 100 mg of intravenous hydrocortisone was administered. Almost immedi-
199
ately after receiving the steroid, she had a dramatic drop in her plasma glucose (Fig. 1). Twenty-four hours later she received a second dose with a similar response. Despite a normal diet, no further insulin was required for 18 h. Following discharge the patient was maintained on 20 mg/day of prednisone and insulin requirements dramatically decreased (Table 1). Subsequently, the effect of glucocorticoid on her blood glucose was tested with a double-blind injection of hydrocortisone and placebo on two different weeks. Again, there was a dramatic glucose-lowering effect from the intravenous hydrocortisone but no effect of placebo. Although the patient’s level of glycemia was improved for several weeks with glucocorticoid treatment, this effect gradually diminished and her blood glucose rose. Also during the summer of 1994 she developed several episodes of intermittent abdominal pain which required evaluation. The pain was intermittent, with no clear antecedent provocation. It was thought that her pain was from pancreatitis, although her work-up was unremarkable except for one pancreatic amylase level which was slightly elevated on one occasion. By the fall of 1994, she had also developed lower extremity pain and weakness. Neurological examination revealed normal sensation in the lower extremities to light touch, but 3–4/5 weakness of the left iliopsoas, quadriceps, anterior tibialis, extensor hallucis longus, and 4/5 strength of her left ankle inversion and eversion compared to 5/5 strength of all of these muscle groups on the right. Her right Achilles reflex was 1+ on the right, 0 on the left. Reflexes of the upper extremities were brisk bilaterally. MRI of the brain and cervical spine were normal. A sural nerve biopsy revealed mild axonopathy, and a muscle biopsy showed mild denervation and reinnervation. Treatment with nortriptyline and oxycodone were ineffective. By March of 1995, her blood glucose levels were again unmanageable despite ongoing treatment with prednisone. She had several further admissions for DKA, and was requiring greater than 600 U of insulin per day, including 25 U of regular insulin every 2 h while awake to prevent metabolic decompensation. Because of the possibility that her unresponsiveness to insulin was the result of abnor-
200
I.B. Hirsch et al. / Diabetes Research and Clinical Practice 41 (1998) 197–202
mal counter-regulatory hormone secretion a trial of octreotide was initiated. She was started on a dose of 50 mg before each meal, and blood glucose levels dropped dramatically after each dose. Over time, the effect of octreotide also waned and higher doses were required to affect her glucose concentrations. She eventually required the octreotide via a continuous subcutaneous pump at 800 mg/day, and her blood glucose levels improved and her weight increased for several months (Table 1). An octreotide scan to search for an occult neuroendocrine tumor was negative. By November of 1995 the patient required weekly intravenous saline and insulin to prevent ketoacidosis, despite ongoing treatment with octreotide and prednisone. She was taking 1600 U of insulin per day, although her insulin requirements were two-fold higher without octreotide. Her HbA1c was 14.6% and she was started on sertraline for depression. It was decided to administer a trial of insulin lispro (Humalog®) provided by Eli Lilly, (Indianapolis, IN), and approved for compassionate use by the Food and Drug Administration. On November 16 she was admitted to the University of Washington Clinical Research Center. She had a quick and dramatic response to the insulin lispro. It became apparent that her blood glucose levels were affected little by ultralente insulin, and her response to regular insulin was also diminished. By comparison, the insulin lispro decreased her blood glucose quickly. She was administered the insulin lispro via continuous subcutaneous infusion, and discontinued other insulins. Her insulin requirements and HbA1c dramatically decreased, and her weight increased. She continued with meticulous glycemic control for the next 28 months (Table 1). Although her blood glucose control improved, her abdominal pain and lower extremity weakness continued. In addition, she developed intense pain in her upper and lower extremities. She also complained of rigidity and spasms of her proximal limb muscles making ambulation difficult. Glutamic acid decarboxylase (GAD) 65 antibodies were measured [1] and were found to be positive with an end-point titer of 1:5000. Cerebral spinal fluid GAD65 antibodies revealed an end-point titer of 1:500, and a dramatic response with im-
provement of her pain and stiffness with 10 mg of oral diazepam confirmed the diagnosis of stiff man syndrome (SMS) [2]. Baclofen was added to her regimen. EMG was read as normal, but diazepam and blaclofen administered prior to the study likely confounded the results.
2. Discussion We describe the first case of SMS and type 1 diabetes with severe insulin resistance. The etiology of the insulin resistance is unknown. She does not have the type A or B syndromes of insulin resistance, and there was no clinical evidence of lipodystrophy. Endocrine causes for her insulin resistance were ruled out, and her normal absorption of injected insulin ruled out the subcutaneous insulin degradation syndrome [3]. Insulin resistance to a mild or moderate degree is seen frequently in a number of physiological and nonphysiological conditions. These disorders would include type 2 diabetes, obesity, pregnancy, puberty, sepsis, and uremia. Severe insulin resistance is defined as distinct syndromes with extreme hyperinsulinemia and severe target-cell resistance to insulin [4]. In adults, these would include classic type A syndrome of insulin resistance with acanthosis nigricans, variants of the type A syndrome with muscle cramps, pseudoacromegaly, or lipodystrophy, or the type B syndrome with autoantibodies to the insulin receptor. Insulin resistance is also present in various endocrinopathies, such as acromegaly, Cushing’s syndrome, thyrotoxicosis, glucagonoma, and pheochromocytoma. None of these conditions explained the insulin resistance in our patient. SMS is characterized by painful spasms and marked rigidity of the skeletal musculature, predominantly affecting the muscles of the back, abdomen, lower extremities, and neck [5]. The clinical presentation usually occurs in adulthood and is followed by a slowly progressing and often fluctuating symptomatic course [6]. The majority of patients with SMS have high levels of glutamic acid decarboxylase (GAD) antibodies in their serum [7,8]. Our patient had all of the classic features of SMS except for the characteristic
I.B. Hirsch et al. / Diabetes Research and Clinical Practice 41 (1998) 197–202
EMG findings. Although this could be an atypical case of SMS, it is more likely the EMG was influenced by the diazepam and baclofen she was receiving at the time. GAD is the rate-limiting step for the synthesis of gamma-aminobutyric acid (GABA), the principle neurotransmitter controlling motor neuron activity. Approximately 30% of SMS subjects develop type 1 diabetes, whereas at least 40% have at least one of a number of autoimmune disorders [2,7]. GAD autoantigen exists in at least two forms, GAD65 and GAD67. Both antibodies are present in SMS, whereas GAD67 antibody does not occur with type 1 diabetes [9]. In addition, the GAD65 antibody is 100 – 500-fold higher in SMS than in type 1 diabetes [9]. Besides the paradox between subcutaneous and intravenous insulin requirements, our patient had an impressive response to insulin lispro, an insulin analogue which rapidly dissociates following subcutaneous injection resulting in peak serum levels in about 1 h and a duration of action of about 4 h [10]. Another patient with type 1 diabetes and insulin resistance but without SMS has recently been described to have a beneficial response to insulin lispro [11]. The etiology of this observation was not reported. Although we cannot determine the etiology of her insulin resistance, it is possible that she has antibody or other protein interference with the relatively slowly absorbed subcutaneous hexameric insulin. Based on her blood glucose response, she appeared to have a small effect from the regular insulin and no effect from the ultralente insulin. That the more rapidly absorbed insulin lispro, and the intravenous insulin, retain effectiveness suggests anti-insulin activity in the subcutaneous space. Although we could not measure insulin antibodies, she did have a quick and dramatic response to the hydrocortisone, raising the possibility of an immunological etiology. Unfortunately, the effectiveness of the glucocorticoid was short-lived. The effectiveness of the octreotide to transiently improve the severe insulin resistance in our patient is not readily explainable but may provide some insight into the underlying mechanism. Theoretically, octreotide could be helpful in the treatment of diabetes which is characterized by
201
paradoxical hypersomatotropinemia and hyperglucagonemia [12]. Somatostatin has been shown to suppress growth hormone and glucagon levels and to reduce hyperglycemia in type 1 diabetes [13]. However, our patient had normal peripheral levels of growth hormone and glucagon. On the other hand, a low-dose octreotide infusion for 4 days in subjects with type 1 diabetes has been shown to increase insulin sensitivity and glucose disposal and enhance suppression of hepatic glucose production [14]. Thus, octreotide may have improved this patient’s insulin sensitivity through metabolic effects, a different mechanism than the glucocorticoid. Alternatively, perhaps the effectiveness of the octreotide in this patient can also be attributed to immune regulation. Specific receptors for somatostatin have been identified on more than 95% of normal mitogen-activated human peripheral lymphocytes [15]. Octreotide could modulate the immune response by altering modification of immunoglobin secretion by plasma cells, changing lymphocyte proliferation, cytotoxicity, and changes in cytokine production; release of mediators by basophils, recruitment of eosinophils; and changes in macrophage function [16]. Whether or not these in vitro actions of octreotide are involved in this unusual syndrome of insulin resistance will require further study. Insulin resistance is not generally observed with SMS. Neuromuscular disorders, in general, are not associated with insulin resistance. The only other syndrome associated with glucose intolerance and insulin resistance is myotonic dystrophy [17]. The treatment of SMS is palliative in nature, usually consisting of diazepam, baclofen, clonazepam, sodium valproate, and physical therapy [5]. At this time we cannot comment if altering her immune response to GAD would improve her insulin resistance. Finally, it would have been ideal to perform glucose clamp studies to quantitate her insulin resistance; we are confident there was no possibility of malingering or patient subterfuge. In the hospital both nurses and physicians were administering her insulin. Severe hyperinsulinemia was documented on many occasions, and the initial response to insulin lispro was witnessed by all of
202
I.B. Hirsch et al. / Diabetes Research and Clinical Practice 41 (1998) 197–202
her physicians. She has now continued to do well with excellent blood glucose control and no further DKA for 2 years since starting the insulin lispro. In conclusion, we describe a patient with SMS and type 1 diabetes with insulin resistance responsive to insulin lispro. The etiology of the insulin resistance is unclear. Although rare, this form of unstable diabetes needs to be considered in patients with frequent DKA.
Acknowledgements We thank Terri Daniels and Eric P. Sullivan for their expert technical assistance. Supported in part by NIH grants DK62190, DK42654, and the Juvenile Diabetes Foundation International.
References ˚ . Lernmark, Ra[1] A. Falorni, E. Ortqvist, B. Persson, A dioimmunoassays for glutamic acid decarboxylase (GAD65) and GAD65 autoantibodies using 35S or 3H recombinant human ligands, J. Immunol. Methods 186 (1995) 89 – 99. [2] E.E. Gordon, D.M. Janusko, L. Kaufman, A critical survey of stiff-man syndrome, Am. J. Med. 42 (1967) 582 – 599. [3] G.R. Friedenberg, N. White, S. Cataland, et al., Diabetes response to intravenous but not subcutaneous effectivness of aprotinin, New Engl. J. Med. 305 (1981) 363. [4] D.E. Moller, J.S. Flier, Syndromes of insulin resistance: mechanisms, syndromes, and implications, Nw Engl. J. Med. 325 (1991) 935–948.
.
[5] T.R. Lorish, G. Thorsteinsson, F.M. Howard, Stiff-man syndrome updated, Mayo Clin. Proc. 64 (1989) 629 – 636. [6] R.B. Layzer, Stiff-man syndrome-an autoimmune disease?, New Engl. J. Med. 318 (1988) 1060 – 1062. [7] M. Solimena, F. Folli, R. Aparisi, et al., Autoantibodies to GABA-ergic neurons and pacreatic beta cells in stiffman syndrome, New Engl. J. Med. 322 (1990) 1555 – 1560. ˚. [8] L. Li, W.A. Hagopian, H.R. Brashear, T. Daniels, A Lernmark, Identification of autoantibody epitopes of glutamic acid decarboxylase in Stiff-man Syndrome patients, J. Immunol. 152 (1994) 930 – 934. [9] K. Daw, N. Ujihara, M. Atkinson, A.C. Powers, Glutamic acid decarboxylase autoantibodies in stiff-man syndrome and insulin-dependent diabtetes exhibit similarities and differences in epitope recognition, J. Immunol. 156 (1996) 818 – 825. [10] D.C. Howey, R.R. Bowsher, R.L. Brunelle, J.R. Woodworth, [Lys(B28),Pro(B29)]-human insulin: a rapidly absorbed analogue of human insulin, Diabetes 43 (1994) 396 – 402. [11] H.R. Henrichs, H. Unger, M.E. Trautmann, A. Pfu¨tzner, Severe insulin resistance treated with insulin lispro (Letter), Lancet 348 (1996) 1248. [12] H. Orskov, A. Flyvbjerg, J. Frystyk, et al., Octreotide and diabetes: theoretical and experimental aspects, Metabolism 41 (9 (Suppl 2) (1992) 66 – 71. [13] J.E. Gerich, Role of somatostatin and its analogue in the pathogenesis and treatment of diabetes mellitus, Metabolism 39 (1990) 52 – 54. [14] L. Orskov, N. Moller, J.F. Bak, et al., Effects of the somatostatin analog, octreotide, on glucose metabolism and insulin sensitivity in insuin-dependent diabetes, Metabolism 45 (1996) 211 – 217. [15] K. Hiruma, T. Koike, H. Nakamura, Somatostatin receptors on human lymphocytes and leukaemia cells, Immunology 71 (1990) 480 – 485. [16] P.M. Van Hagen, E.P. Krenning, D.J. Kwekkeboom, et al., Somatostatin in the immune and haematopoietic system; a review, Eur. J. Clin. Invest. 241 (1994) 91 – 99. [17] R.T. Moxley, W.J. Kingston, R.C. Griggs, Lack of rapid enhancement of insulin action after oral glucose tolerance challenge in myotonic dystrophy, Diabetes 36 (1987) 693 – 701.
Case report
Severe insulin resistance in a patient with type 1 diabetes and stiff-man syndrome treated with insulin lispro ˚ ke Lernmark, Alan Chait Irl B. Hirsch *, David D’Alessio, Lily Eng, Connie Davis, A UWMC, 1959 NE Pacific Street, Box 356176, Seattle, WA 98195, USA Accepted 26 May 1998
Abstract We describe a patient with type 1 diabetes with recurrent diabetic ketoacidosis and severe insulin resistance. Extensive evaluation of the etiology of the insulin resistance did not reveal an etiology, and well over 1000 U of daily insulin did not prevent the ketoacidosis. Her blood glucose and insulin requirements were improved with glucocorticoids and octreotide, but the effects of both of these agents were short-lived. She was given a trial of insulin lispro with immediate and dramatic effects, lowering her HbA1c from 14.6 to 5.1% in 7 months with a decrease in insulin requirements of 1600–100 U per day. Besides her diabetes, she had a history of pain and stiffness affecting numerous muscle groups, and hospitalization was required for pain control. The diagnosis of stiff-man syndrome (SMS) was confirmed with high titers of glutamic acid decarboxylase 65 antibodies in both serum and cerebral spinal fluid. In summary, we describe the first patient with type 1 diabetes, SMS, and severe insulin resistance. Although the etiology of the insulin resistance is unknown, due to the efficacious response to insulin lispro, hydrocortisone, and perhaps octreotide, we propose an immune-mediated etiology. Although rare, this syndrome needs to be considered as an etiology of insulin resistance. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Glucocorticoid; Stiff man syndrome; Insulin lispro; Insulin resistance; Octreotide
1. Case presentation The patient is a 33-year-old Caucasian woman who presented to the University of Washington in * Corresponding author. Tel.: + 1 206 5484882; fax: +1 206 5484976.
March of 1994. She had a 2-year history of type 1 diabetes complicated by numerous episodes of diabetic ketoacidosis (DKA) requiring hospitalization. During the previous 9 months, a 9-kg weight loss accompanied symptoms of uncontrolled hyperglycemia. Her insulin dose was 60 U per day of human insulin, split into two injections
0168-8227/98/$ - see front matter © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0168-8227(98)00072-2
I.B. Hirsch et al. / Diabetes Research and Clinical Practice 41 (1998) 197–202
198
of NPH and regular insulin administered prior to breakfast and supper. Her only other medication was thyroxine. Physical examination revealed an emaciated woman, with a weight of 53.6 kg (body mass index=19.2 kg/m2). Her supine blood pressure and pulse were 110/70 and 52, respectively, and there were no postural changes. The rest of the exam was unremarkable and there were no neurological abnormalities. Initial laboratory assessment revealed: HbA1c, 12.2% (nl, 4.0–6.0%) and mild DKA. She was treated with intravenous fluids, insulin, and potassium and during the hospitalization her subcutaneous insulin dose was increased. Although there was concern the patient was not administering her insulin, there was no previous psychiatric history. One week after hospitalization the patient was receiving 90 U/day of insulin. However, a random plasma glucose was 32.3 mmol/l with a carbon dioxide of 15 mEq/l and an anion gap of 17. Serum C-peptide was
0.07 nmol/l and total serum insulin was 606 pmol/l. During the next 3 months she was admitted to the University of Washington Medical Center on three occasions for the treatment of DKA. She usually required between 3 and 8 U/h of intravenous insulin, and continuous increases in her subcutaneous insulin dose were required (see Table 1). The nursing staff injected all subcutaneous insulin during the hospitalizations. By late May the patient was receiving 920 U of insulin per day, administered as twice daily ultralente and thrice daily regular insulin. Because of the apparent insulin resistance, further evaluation was undertaken. Plasma IGF1 and glucagon were normal, as was a 24-h urinary free cortisol. Insulin receptor antibodies, insulin antibodies, and antinuclear antibodies were negative. Sedimentation rate was 8. Antimicrosomal antibodies were positive at 1:40, supporting a diagnosis of Hashimoto’s thyroiditis.
Table 1 Patient characteristics and medication doses Date
HbA1c (%) (nl 4.0–6.0%)
BMI (kg/m2)
Insulin (units/day)
Prednisone (mg/day)
Octreotide (mg/day)
3/30/94 4/25/94 5/11/94 6/21/94 7/20/94 9/13/94 11/14/94 12/30/94 2/24/95 4/24/95 6/23/95 8/28/95 11/13/95 12/19/95 1/23/96 2/28/96 3/26/96 5/1/96 6/24/96 3/17/98
12.2
19.0 20.1 20.9 20.0 21.0 22.5 22.0 22.7 22.3 22.5 21.0 20.3 20.0 23.4 22.7 21.4 21.9 21.9 21.9 24.8
60 150 240 1500 250 350 1500 150 500 450 800 1200 1600 300b 300b 250c 200c 150c 100c 175c
20 10 20 10 10 10 10 10 10 10 10 7.5 20 20 20 10
200 200 300 150a 300a 700a 1000a 550 550 550 300 100 0 0
a
13.5 13.9 10.5 14.3
14.6 12.6 8.5 7.0 7.1 5.1 5.2 6.7
Continuous subcutaneous octreotide via external pump. Insulin lispro via insulin infusion pump with lispro and regular insulin before meals. c Insulin lispro alone via insulin infusion pump. b
I.B. Hirsch et al. / Diabetes Research and Clinical Practice 41 (1998) 197–202
Fig. 1. Response to hydrocortisone in a patient with severe insulin resistance. Blood glucose was maintained with an intravenous insulin infusion at a rate of 7–8 U/h. Two hours after administering the hydrocortisone, the intravenous insulin infusion was discontinued. HC, hydrocortisone.
To test the absorption of insulin from the subcutaneous space, 200 U of human regular insulin from a new vial was administered by one of her physicians (IBH), serial blood samples were drawn, and aliquots of plasma assayed by radioimmunoassay directly (total insulin) or after precipitation of immunoglobulins with 30% polyethylene glycol (free insulin): Time Glucose Total Free insulin (30% (h) (mmol/l) insulin polyethylene glycol) (pmol/l) (pmol/l) 0 1 2
26.7 23.3 21.9
1830 3540 4230
1920 3370 4140
These results suggested that the patient’s absorption of insulin was adequate and that there was no antibody interfering with the circulating insulin. In late June of 1994, the patient presented with another episode of DKA. Based on the hypothesis that her profound insulin resistance was mediated by an immune process, 100 mg of intravenous hydrocortisone was administered. Almost immedi-
199
ately after receiving the steroid, she had a dramatic drop in her plasma glucose (Fig. 1). Twenty-four hours later she received a second dose with a similar response. Despite a normal diet, no further insulin was required for 18 h. Following discharge the patient was maintained on 20 mg/day of prednisone and insulin requirements dramatically decreased (Table 1). Subsequently, the effect of glucocorticoid on her blood glucose was tested with a double-blind injection of hydrocortisone and placebo on two different weeks. Again, there was a dramatic glucose-lowering effect from the intravenous hydrocortisone but no effect of placebo. Although the patient’s level of glycemia was improved for several weeks with glucocorticoid treatment, this effect gradually diminished and her blood glucose rose. Also during the summer of 1994 she developed several episodes of intermittent abdominal pain which required evaluation. The pain was intermittent, with no clear antecedent provocation. It was thought that her pain was from pancreatitis, although her work-up was unremarkable except for one pancreatic amylase level which was slightly elevated on one occasion. By the fall of 1994, she had also developed lower extremity pain and weakness. Neurological examination revealed normal sensation in the lower extremities to light touch, but 3–4/5 weakness of the left iliopsoas, quadriceps, anterior tibialis, extensor hallucis longus, and 4/5 strength of her left ankle inversion and eversion compared to 5/5 strength of all of these muscle groups on the right. Her right Achilles reflex was 1+ on the right, 0 on the left. Reflexes of the upper extremities were brisk bilaterally. MRI of the brain and cervical spine were normal. A sural nerve biopsy revealed mild axonopathy, and a muscle biopsy showed mild denervation and reinnervation. Treatment with nortriptyline and oxycodone were ineffective. By March of 1995, her blood glucose levels were again unmanageable despite ongoing treatment with prednisone. She had several further admissions for DKA, and was requiring greater than 600 U of insulin per day, including 25 U of regular insulin every 2 h while awake to prevent metabolic decompensation. Because of the possibility that her unresponsiveness to insulin was the result of abnor-
200
I.B. Hirsch et al. / Diabetes Research and Clinical Practice 41 (1998) 197–202
mal counter-regulatory hormone secretion a trial of octreotide was initiated. She was started on a dose of 50 mg before each meal, and blood glucose levels dropped dramatically after each dose. Over time, the effect of octreotide also waned and higher doses were required to affect her glucose concentrations. She eventually required the octreotide via a continuous subcutaneous pump at 800 mg/day, and her blood glucose levels improved and her weight increased for several months (Table 1). An octreotide scan to search for an occult neuroendocrine tumor was negative. By November of 1995 the patient required weekly intravenous saline and insulin to prevent ketoacidosis, despite ongoing treatment with octreotide and prednisone. She was taking 1600 U of insulin per day, although her insulin requirements were two-fold higher without octreotide. Her HbA1c was 14.6% and she was started on sertraline for depression. It was decided to administer a trial of insulin lispro (Humalog®) provided by Eli Lilly, (Indianapolis, IN), and approved for compassionate use by the Food and Drug Administration. On November 16 she was admitted to the University of Washington Clinical Research Center. She had a quick and dramatic response to the insulin lispro. It became apparent that her blood glucose levels were affected little by ultralente insulin, and her response to regular insulin was also diminished. By comparison, the insulin lispro decreased her blood glucose quickly. She was administered the insulin lispro via continuous subcutaneous infusion, and discontinued other insulins. Her insulin requirements and HbA1c dramatically decreased, and her weight increased. She continued with meticulous glycemic control for the next 28 months (Table 1). Although her blood glucose control improved, her abdominal pain and lower extremity weakness continued. In addition, she developed intense pain in her upper and lower extremities. She also complained of rigidity and spasms of her proximal limb muscles making ambulation difficult. Glutamic acid decarboxylase (GAD) 65 antibodies were measured [1] and were found to be positive with an end-point titer of 1:5000. Cerebral spinal fluid GAD65 antibodies revealed an end-point titer of 1:500, and a dramatic response with im-
provement of her pain and stiffness with 10 mg of oral diazepam confirmed the diagnosis of stiff man syndrome (SMS) [2]. Baclofen was added to her regimen. EMG was read as normal, but diazepam and blaclofen administered prior to the study likely confounded the results.
2. Discussion We describe the first case of SMS and type 1 diabetes with severe insulin resistance. The etiology of the insulin resistance is unknown. She does not have the type A or B syndromes of insulin resistance, and there was no clinical evidence of lipodystrophy. Endocrine causes for her insulin resistance were ruled out, and her normal absorption of injected insulin ruled out the subcutaneous insulin degradation syndrome [3]. Insulin resistance to a mild or moderate degree is seen frequently in a number of physiological and nonphysiological conditions. These disorders would include type 2 diabetes, obesity, pregnancy, puberty, sepsis, and uremia. Severe insulin resistance is defined as distinct syndromes with extreme hyperinsulinemia and severe target-cell resistance to insulin [4]. In adults, these would include classic type A syndrome of insulin resistance with acanthosis nigricans, variants of the type A syndrome with muscle cramps, pseudoacromegaly, or lipodystrophy, or the type B syndrome with autoantibodies to the insulin receptor. Insulin resistance is also present in various endocrinopathies, such as acromegaly, Cushing’s syndrome, thyrotoxicosis, glucagonoma, and pheochromocytoma. None of these conditions explained the insulin resistance in our patient. SMS is characterized by painful spasms and marked rigidity of the skeletal musculature, predominantly affecting the muscles of the back, abdomen, lower extremities, and neck [5]. The clinical presentation usually occurs in adulthood and is followed by a slowly progressing and often fluctuating symptomatic course [6]. The majority of patients with SMS have high levels of glutamic acid decarboxylase (GAD) antibodies in their serum [7,8]. Our patient had all of the classic features of SMS except for the characteristic
I.B. Hirsch et al. / Diabetes Research and Clinical Practice 41 (1998) 197–202
EMG findings. Although this could be an atypical case of SMS, it is more likely the EMG was influenced by the diazepam and baclofen she was receiving at the time. GAD is the rate-limiting step for the synthesis of gamma-aminobutyric acid (GABA), the principle neurotransmitter controlling motor neuron activity. Approximately 30% of SMS subjects develop type 1 diabetes, whereas at least 40% have at least one of a number of autoimmune disorders [2,7]. GAD autoantigen exists in at least two forms, GAD65 and GAD67. Both antibodies are present in SMS, whereas GAD67 antibody does not occur with type 1 diabetes [9]. In addition, the GAD65 antibody is 100 – 500-fold higher in SMS than in type 1 diabetes [9]. Besides the paradox between subcutaneous and intravenous insulin requirements, our patient had an impressive response to insulin lispro, an insulin analogue which rapidly dissociates following subcutaneous injection resulting in peak serum levels in about 1 h and a duration of action of about 4 h [10]. Another patient with type 1 diabetes and insulin resistance but without SMS has recently been described to have a beneficial response to insulin lispro [11]. The etiology of this observation was not reported. Although we cannot determine the etiology of her insulin resistance, it is possible that she has antibody or other protein interference with the relatively slowly absorbed subcutaneous hexameric insulin. Based on her blood glucose response, she appeared to have a small effect from the regular insulin and no effect from the ultralente insulin. That the more rapidly absorbed insulin lispro, and the intravenous insulin, retain effectiveness suggests anti-insulin activity in the subcutaneous space. Although we could not measure insulin antibodies, she did have a quick and dramatic response to the hydrocortisone, raising the possibility of an immunological etiology. Unfortunately, the effectiveness of the glucocorticoid was short-lived. The effectiveness of the octreotide to transiently improve the severe insulin resistance in our patient is not readily explainable but may provide some insight into the underlying mechanism. Theoretically, octreotide could be helpful in the treatment of diabetes which is characterized by
201
paradoxical hypersomatotropinemia and hyperglucagonemia [12]. Somatostatin has been shown to suppress growth hormone and glucagon levels and to reduce hyperglycemia in type 1 diabetes [13]. However, our patient had normal peripheral levels of growth hormone and glucagon. On the other hand, a low-dose octreotide infusion for 4 days in subjects with type 1 diabetes has been shown to increase insulin sensitivity and glucose disposal and enhance suppression of hepatic glucose production [14]. Thus, octreotide may have improved this patient’s insulin sensitivity through metabolic effects, a different mechanism than the glucocorticoid. Alternatively, perhaps the effectiveness of the octreotide in this patient can also be attributed to immune regulation. Specific receptors for somatostatin have been identified on more than 95% of normal mitogen-activated human peripheral lymphocytes [15]. Octreotide could modulate the immune response by altering modification of immunoglobin secretion by plasma cells, changing lymphocyte proliferation, cytotoxicity, and changes in cytokine production; release of mediators by basophils, recruitment of eosinophils; and changes in macrophage function [16]. Whether or not these in vitro actions of octreotide are involved in this unusual syndrome of insulin resistance will require further study. Insulin resistance is not generally observed with SMS. Neuromuscular disorders, in general, are not associated with insulin resistance. The only other syndrome associated with glucose intolerance and insulin resistance is myotonic dystrophy [17]. The treatment of SMS is palliative in nature, usually consisting of diazepam, baclofen, clonazepam, sodium valproate, and physical therapy [5]. At this time we cannot comment if altering her immune response to GAD would improve her insulin resistance. Finally, it would have been ideal to perform glucose clamp studies to quantitate her insulin resistance; we are confident there was no possibility of malingering or patient subterfuge. In the hospital both nurses and physicians were administering her insulin. Severe hyperinsulinemia was documented on many occasions, and the initial response to insulin lispro was witnessed by all of
202
I.B. Hirsch et al. / Diabetes Research and Clinical Practice 41 (1998) 197–202
her physicians. She has now continued to do well with excellent blood glucose control and no further DKA for 2 years since starting the insulin lispro. In conclusion, we describe a patient with SMS and type 1 diabetes with insulin resistance responsive to insulin lispro. The etiology of the insulin resistance is unclear. Although rare, this form of unstable diabetes needs to be considered in patients with frequent DKA.
Acknowledgements We thank Terri Daniels and Eric P. Sullivan for their expert technical assistance. Supported in part by NIH grants DK62190, DK42654, and the Juvenile Diabetes Foundation International.
References ˚ . Lernmark, Ra[1] A. Falorni, E. Ortqvist, B. Persson, A dioimmunoassays for glutamic acid decarboxylase (GAD65) and GAD65 autoantibodies using 35S or 3H recombinant human ligands, J. Immunol. Methods 186 (1995) 89 – 99. [2] E.E. Gordon, D.M. Janusko, L. Kaufman, A critical survey of stiff-man syndrome, Am. J. Med. 42 (1967) 582 – 599. [3] G.R. Friedenberg, N. White, S. Cataland, et al., Diabetes response to intravenous but not subcutaneous effectivness of aprotinin, New Engl. J. Med. 305 (1981) 363. [4] D.E. Moller, J.S. Flier, Syndromes of insulin resistance: mechanisms, syndromes, and implications, Nw Engl. J. Med. 325 (1991) 935–948.
.
[5] T.R. Lorish, G. Thorsteinsson, F.M. Howard, Stiff-man syndrome updated, Mayo Clin. Proc. 64 (1989) 629 – 636. [6] R.B. Layzer, Stiff-man syndrome-an autoimmune disease?, New Engl. J. Med. 318 (1988) 1060 – 1062. [7] M. Solimena, F. Folli, R. Aparisi, et al., Autoantibodies to GABA-ergic neurons and pacreatic beta cells in stiffman syndrome, New Engl. J. Med. 322 (1990) 1555 – 1560. ˚. [8] L. Li, W.A. Hagopian, H.R. Brashear, T. Daniels, A Lernmark, Identification of autoantibody epitopes of glutamic acid decarboxylase in Stiff-man Syndrome patients, J. Immunol. 152 (1994) 930 – 934. [9] K. Daw, N. Ujihara, M. Atkinson, A.C. Powers, Glutamic acid decarboxylase autoantibodies in stiff-man syndrome and insulin-dependent diabtetes exhibit similarities and differences in epitope recognition, J. Immunol. 156 (1996) 818 – 825. [10] D.C. Howey, R.R. Bowsher, R.L. Brunelle, J.R. Woodworth, [Lys(B28),Pro(B29)]-human insulin: a rapidly absorbed analogue of human insulin, Diabetes 43 (1994) 396 – 402. [11] H.R. Henrichs, H. Unger, M.E. Trautmann, A. Pfu¨tzner, Severe insulin resistance treated with insulin lispro (Letter), Lancet 348 (1996) 1248. [12] H. Orskov, A. Flyvbjerg, J. Frystyk, et al., Octreotide and diabetes: theoretical and experimental aspects, Metabolism 41 (9 (Suppl 2) (1992) 66 – 71. [13] J.E. Gerich, Role of somatostatin and its analogue in the pathogenesis and treatment of diabetes mellitus, Metabolism 39 (1990) 52 – 54. [14] L. Orskov, N. Moller, J.F. Bak, et al., Effects of the somatostatin analog, octreotide, on glucose metabolism and insulin sensitivity in insuin-dependent diabetes, Metabolism 45 (1996) 211 – 217. [15] K. Hiruma, T. Koike, H. Nakamura, Somatostatin receptors on human lymphocytes and leukaemia cells, Immunology 71 (1990) 480 – 485. [16] P.M. Van Hagen, E.P. Krenning, D.J. Kwekkeboom, et al., Somatostatin in the immune and haematopoietic system; a review, Eur. J. Clin. Invest. 241 (1994) 91 – 99. [17] R.T. Moxley, W.J. Kingston, R.C. Griggs, Lack of rapid enhancement of insulin action after oral glucose tolerance challenge in myotonic dystrophy, Diabetes 36 (1987) 693 – 701.