Spring 2000

Residual effects
following GBS

What is the role of diabetes mellitus in patients with CIDP?

GBS and campylobacter jejuni infection

Disability after "recovery" from GBS

Search our Newsletters


What is the Role of Diabetes Mellitus in Patients with CIDP?
Kenneth C. Gorson, M.D. Neurology Service St. Elizabeth’s Medical Center Associate Professor of Neurology Tufts University School of Medicine

Diabetes mellitus is the most common cause of peripheral neuropathy in the United States. Most patients have chronic, slowly progressive symptoms that include numbness, tingling and neuropathic pain, usually limited to the feet and legs, but some patients develop a disabling condition associated with generalized limb weakness, severe sensory loss, imbalance, and absent deep tendon reflexes that simulates chronic inflammatory demyelinating polyneuropathy (CIDP). The situation is further confused because some patients with diabetic polyneuropathy have demyelinating features on EMG studies, an elevated spinal fluid protein concentration, and respond to immune treatments (e.g., prednisone. plasma exchange, or IVIG); their polyneuropathy therefore emulates CIDP. To begin to clarify the role of diabetes in patients with a CIDP-like syndrome, we evaluated 14 patients with diabetes mellitus who also fulfilled clinical and EMG research criteria for midi and compared them to 60 patients with CIDP and no other associated illness (idiopathic CIDP).

All patients were seen at St. Elizabeth's Medical Center in Boston over a period of 8 years. In the diabetic group, there were 10 men and 4 women with type 11 diabetes mellitus for an average duration of 9 years. Seven were treated with oral medications for diabetes, 4 required insulin, and 3 were treated with dietary modification only.

The patients with diabetes and CIDP were older (average age of 67 years vs. 49 years in those with idiopathic CIDP) and more often complained of imbalance (100% of diabetic-clop patients vs. 58% with idiopathic CIDP), but the severity and distribution of weakness, sensory features, gait disturbance and functional disability were similar to patients with idiopathic CIDP. The spinal fluid protein was elevated in a similar proportion of patients (100% of diabetic CIDP patients vs. 84% with idiopathic CIDP), and the average amount of the protein elevation was identical in both groups (128 mg/ dl). EMG studies showed that the frequency and severity of demyelinating features also were similar between the groups. For example, 79% of diabetic-CIDP patients had at least l nerve showing evidence of abnormal conduction of electrical impulses (conduction block), compared to 83% in patients with idiopathic CIDP. However, nerve conduction studies also showed a much lower average number of healthy nerve fibers in several motor and sensory nerves (peroneal, ulnar, sural) in the diabetic group; this observation was confirmed by finding more severe nerve fiber loss (termed axonal loss) with examination of sensory nerve biopsy material.

The rate of response to standard immune therapies was similar in both groups

  Diabetes and CIDP Idiopathic CIDP P Value
Treatment response*
N= 58
3/5 (60%)
14/34 (41%)
2/10 (20%)
22/49 (45%)
Plasma Exchange
4/9 (44%)
22/39 (56%)

Legend: * independent of treatment order, including patients who received multiple treatment modalities due to relapses or lack of response. P-value >0.05 indicates that the difference between the groups is not statistically significant.

However, the degree of improvement was less favorable in patients with CIDP and diabetes, with a lower average strength score and greater functional disability at the time of the last evaluation.

Our experience indicates that patients with diabetes mellitus and a CIDP-like syndrome have symptoms, neurological findings, and functional disability that are indistinguishable from those with idiopathic CIDP. The older age of patients with diabetes and CIDP probably reflects the age distribution of patients with type 11 diabetes. Therefore, in patients who meet clinical criteria for midi a history of diabetes does not appreciably alter the clinical features of the illness.

The EMG and nerve biopsy findings suggest that in patients with diabetes and midi diabetes seems to contribute an element of axonal nerve fiber damage, in contrast to the typical demyelinating features observed with CIDP. This conclusion is not surprising when one considers the frequency of axonal loss that has been linked to most forms of diabetic neuropathy. Furthermore, the prominent axonal loss in the diabetic CIDP group may explain their less favorable outcome after therapy.

A relationship between diabetes and CIDP has been suggested in several previous studies. Krendel and colleagues reported 6 patients with insulin-dependent diabetes and a demyelinating neuropathy. Half had widespread, symmetrical weakness and an elevated CSF protein level, all had conduction block in one or more motor nerves, two-thirds had features of demyelination on nerve biopsy and all improved with various combinations of immune treatment. Stewart and coworkers reported 7 patients with CIDP and diabetes and stressed the difficulties in distinguishing between diabetic polyneuropathy and CIDP: 1) there may be substantial overlap of the clinical features; 2) the spinal fluid protein level often is elevated in patients with diabetes; 3) EMG studies frequently show demyelinating changes in patients with typical diabetic polyneuropathy; and 4) nerve biopsies show varying degrees of axonal and demyelinating changes in both conditions. These investigators concluded that more rapid worsening than expected from diabetic neuropathy and substantial improvement following immune treatment provided the strongest evidence that their diabetic patients most likely had CIDP. The uncertainty at this time is whether patients such as these have coincident diabetes and midi or have a form of demyelinating neuropathy attributable solely to diabetes. This is a separate issue from the additive effects of diabetic axonal damage to conventional CIDP.

Printer Friendly Version