Diabetic Amyotrophy is Immune-Mediated

abstract & commentary

Source: Klekar P, et al. Distinctive pathologic findings in proximal diabetic neuropathy (diabetic amyotrophy). Neurology 2000;5:83-88.

Fifteen patients, eight men and seven women, ages 49-79 years, with proximal diabetic neuropathy (PDN, diabetic amyotrophy) of 5-52 weeks duration, underwent nerve and muscle biopsy. PDN was characterized by progressive, proximal, painful, asymmetric weakness. Diagnosis required the absence of complicating illness, including peripheral vascular disease, structural spinal cord or pelvic disease, previous radiation, or malignant infiltration of lumbar plexus. None had evidence of polyneuropathy and all demonstrated active denervation on electrodiagnostic studies. Two diabetic patients without PDN and five nondiabetics, two each with chronic inflammatory demyelinating polyneuropathy and muscular dystrophy, and one with inclusion body myositis, served as biopsy controls.

Four PDN patients, biopsied within six weeks (n = 3) or during flare-ups of symptoms (n = 1), demonstrated transmural polymorphonuclear infiltration of epineurial postcapillary venules, without fibrinoid necrosis or thrombosis. Biopsies in the other patients, performed at 3-12 months after onset, were absent of these features. Six patients, however, showed T lymphocyte perivascular mononuclear collections, without vasculitis but involving small epineurial vessels. Neurogenic atrophy was seen in all muscle biopsies with endomysial lymphocytic infiltration in only one patient. No inflammation or mural infiltration was observed in any of the controls. PDN is an immune-mediated vasculitis, and not an ischemic microangiopathic neuropathy. The inciting antigen remains unknown.


Using indirect immunofluorescence, circulating autoantibodies have been well reported in diabetic patients. Among 154 diabetic patients, 12% demonstrated anti-GM1 antibody, predominantly in those with distal symmetric polyneuropathy with demyelinating features (Contemp Intern Med 1994;16:41-55). Antiphospholipid antibodies, with their tendency to thrombotic complications and neural insults, were seen in 88% (n = 18) of this group compared to 2% of the general population and 32% of diabetics without neurologic complications (Diabetes Care 1995;18:1225-1232). In another report, antiglutamic acid decarboxylase (GAD) autoantibodies were present in 55% and 21% of type I and type 2 diabetics, respectively, supporting an autoimmune role in diabetes pathogenesis (Diabetes Res Clin Pract 2000;49:33-40). Circulating antiendothelial (antipericyte) autoantibodies may be active in diabetic retinopathy (Retina 1999;19:390-400). Anti-autonomic nervous system antibodies, including autoantibodies to adrenal medulla, vagus nerve, and sympathetic ganglion cells, are only rarely observed in symptomatic diabetic autonomic neuropathy (Diab Med 1997;14:461-465). Greater insulin requirement in the early years of clinical diabetes appears to correlate with multiple autoantibody positivity (anti GAD, anti IA-2 protein, anti-insulin, anti-islet cell) and may reflect more aggressive islet-cell destruction (J Clin Endocrinol Metab 1999;84:1534-1539). Prevention of autoimmune diabetes and its complications may involve protecting susceptible cells (islet cells, venules) from autoimmune attack by promoting immune tolerance in the host (Drugs 1997;53:943-956). —michael rubin

Diabetics demonstrate:

a. antiglutamic acid decarboxylase (GAD) autoantibodies are in 10% and 21% of type I and type 2 diabetics, respectively.

b. multiple autoantibody positivity including anti GAD, anti IA-2 protein, anti-insulin, and anti-islet cell autoantibodies.

c. anti-GM1 antibodies, predominantly in those with proximal asymmetric neuropathy.

d. antiphospholipid antibodies, in up to 28%.

e. None of the above