Chronic inflammatory demyelinating polyneuropathy

Patient selection.

Clinical, laboratory, and electrophysiologic information as well as nerve biopsies were reviewed retrospectively for 67 consecutive patients with a diagnosis of CIDP who were seen by four neuromuscular specialists at St. Elizabeth's Medical Center from January 1, 1991, through January 1, 1995. Clinical details were collected during this period in a uniform manner. The diagnosis of CIDP was established by (1) the presence of a progressive or relapsing sensorimotor, or only motor or sensory, dysfunction of more than one limb of a peripheral nerve nature, developing over at least 2 months associated with hypo- or areflexia and (2) EMG features with use of the research criteria established by Cornblath et al. and the Ad Hoc Subcommittee of the American Academy of Neurology AIDS Task Force. ^[13] Patients with associated systemic conditions in addition to CIDP ("secondary" CIDP) and those with CIDP-MGUS were analyzed separately.

Laboratory studies.

These included complete blood count, routine chemistries, liver and renal function tests, total protein, B₁₂ and folate levels, erythrocyte sedimentation rate, ANA, thyroid function tests, RPR, and serum glucose. Other tests, such as Lyme serology (27 patients), hepatitis B (17 patients), SSA, SSB, or HIV titers (7 patients), rheumatoid factor (18 patients), angiotensin converting enzyme levels (16 patients), and cryoglobulins (14 patients) were performed in some patients.

Immunoelectrophoresis or immunofixation electrophoresis with quantitation of serum immunoglobulins and kappa and lambda light chains was obtained in 60 patients with a high-resolution agarose-gel technique according to the procedure described by Keren et al. ^[14] Fifteen had a monoclonal gammopathy (CIDP-MGUS) and were evaluated for a lymphoproliferative disorder with a skeletal bone survey and in most cases with a bone marrow biopsy, urine protein electrophoresis, and chest and abdominal CTs. Antibodies directed against myelin-associated glycoprotein (MAG) (50 patients) and GM1 (49 patients) were measured with ELISA (Athena Diagnostics, Worcester, MA). Spinal fluid was obtained from 52 patients.

Electrophysiologic studies.

Nerve conduction studies were performed in a standard fashion according to the methods of Albers and Kelly. ^[6] The median, ulnar, peroneal, and tibial motor nerves and the median, ulnar, and sural sensory nerves were examined in most but not all patients, and enough nerves were studied to fulfill diagnostic criteria for an acquired demyelinating polyneuropathy. ^[13] At least three of four AAN criteria for the EMG diagnosis of CIDP were satisfied in all patients: two nerves with slowing of conduction velocities <70% of the lower limit of normal; two nerves with prolonged distal latencies >150% of the upper limit of normal; two nerves with absent or prolonged F waves (minimum 10 trials) >150% of the upper limit of normal; and at least one nerve with partial conduction block with >20% drop in peak-to-peak amplitude between proximal and distal sites. Needle electrode examination included sampling of at least one distal and proximal muscle in the arm and leg, with use of semiquantitative assessment of motor unit potentials and abnormal spontaneous activity.

Nerve biopsy.

Eighteen patients had sural nerve biopsies. One segment was sectioned at 1 micro m and stained with toluidine blue. A second segment was embedded in paraffin, cut at 8 micro m and stained with hematoxylin and eosin, and with Congo red. Teased fibers were examined in three patients.

Functional assessment.

Strength at the time of the maximum deficit and following recovery after treatment was assessed in a proximal and distal muscle of the arm and leg (deltoid, finger extensors, iliopsoas, tibialis anterior) with the Medical Research Council (MRC) scale of 0 to 5. An average MRC score was determined from the four muscle groups. The modified Rankin disability scale was used to determine functional impairment at the time of the maximal neurologic deficit and following treatment: 0 = asymptomatic; 1 = nondisabling symptoms that do not interfere with daily activities; 2 = slight disability, unable to carry out all activities but still able to look after themselves; 3 = moderate disability, requiring assistance with some activities but able to walk without assistance; 4 = moderately severe disability, unable to walk without assistance and unable to attend to own bodily needs without assistance; 5 = severe disability, totally dependent, requiring constant nursing care and attention.

Clinical course and treatment.

Strength, sensory testing, and Rankin evaluations were performed at the initial evaluation and following each course of treatment. In patients who failed to improve after the initial therapy, an evaluation was recorded after the second, third, or fourth therapy. A treatment response was defined as improvement of at least one grade of the Rankin score for at least 2 months.

We followed most patients regularly during the course of their illness. For those who were followed elsewhere, information was obtained by telephone follow-up with the patient and local physician. The median duration of follow-up was 3 years (range, 1 to 24 years).

All but two patients who had mild disease were treated with either plasma exchange (PE, 250 ml/kg during five treatments every other day), intravenous gammaglobulin (IVIG, 0.4 g/kg/d for 5 days), or prednisone (60 mg/d for 2 months). The choice of initial therapy was left to the individual physician. The response was assessed after 2 months. Those who failed to improve with the initial therapy received an alternative second, third, and rarely, a fourth therapy.

Statistical methods.

Patients with CIDP-I were compared with patients with CIDP-MGUS with regard to clinical features, EMG abnormalities, and response to treatment. Chi square or Fisher's exact test was used to compare categorical responses, and the Mann-Whitney independent rank sum test or the Kruskal-Wallis one-way analysis of variance (ANOVA) was used to compare ordinal and continuous variables. The comparisons were average MRC score, sensory signs, Rankin score, proportion of patients responding to treatment, and clinical course. For these comparisons, a p value of 0.01 was used as the criterion for statistical significance.

Initial clinical features (67 patients).

There were 38 men and 29 women, with a mean age at the time of initial examination of 54 years (range, 11 to 90 years). The average duration of symptoms at the time we first saw these patients was 28 months (range, 2 months to 20 years). Fourteen (21%) recalled an antecedent illness in the weeks preceding the onset of symptoms: eight had an upper respiratory syndrome or flu-like illness, one each had a gastrointestinal syndrome, herpes zoster infection, leg cellulitis, cervical laminectomy, coronary bypass surgery, or influenza vaccination.

The main clinical features (Table 1) were limb weakness, sensory symptoms, and pain. Limb weakness was present in 52 patients (78%)-limited to the legs in 33%, restricted to the hands or arms in 9%, and generalized in 36%. Twenty-four (36%) had falls. Sensory symptoms were present initially in 52 (78%); two-thirds had foot numbness as the initial sensory symptom, and 55% complained of numb hands. Pain was prominent in 28 patients (42%); it took the form of burning dysesthesia in 14, radicular pain in 5, aching or cramping muscle pain in 5, and lancinating pains in four.

There was a predilection for distal leg weakness (mean MRC, 3.2), followed by proximal leg (mean MRC, 3.6), and hand weakness (mean MRC, 3.7). Pinprick sensation was absent or diminished in the feet in 48 and in the arms in thirty-seven. Fourteen had leg ataxia and four had upper extremity dysmetria. There was generalized areflexia in 42 patients (63%); reflex loss was limited to the legs in 9 (13%) and to the arms in three (4%). Four had only absent ankle reflexes (6%).

As the illness progressed, 19 became bedbound or wheelchair bound, 14 required a walker, and 9 used a cane. Nine developed a postural tremor. Seven (10%) had facial weakness, 2 (3%) had ophthalmoparesis with diplopia, 4 (6%) had dysphagia from oropharyngeal weakness, and 6 (9%) had respiratory failure requiring ventilator support during the course of the illness. The mean Rankin disability score for the entire group at the time of maximal disability was 3.0 (range, 1 to 5).

CIDP variant presentations.

Several patients had unusual features early in their illnesses while still corresponding to the accepted clinical and EMG criteria for CIDP (see Table 1). Seven (10%) had a pure or predominantly motor syndrome, superficially resembling motor neuron disease, and eight (12%) had a predominantly sensory syndrome with severe sensory loss that involved vibration, touch, and joint position sensations, imbalance, and a disabling sensory ataxia with normal or virtually normal strength. Six (9%) had a symmetric, predominantly upper limb syndrome with sensorimotor deficits in the distribution of specific nerve territories (bilateral median, ulnar, or radial), simulating a mononeuritis multiplex pattern. Three (4%) had a paraparetic syndrome with severe leg weakness and sensory loss with little or no involvement of the upper limbs.

Eleven (16%) presented with acute generalized weakness that was indistinguishable from Guillain-Barre syndrome; three had weakness that progressed beyond 4 weeks, and eight improved after initial treatment (plasma exchange or IVIG) but then had a chronic relapsing course (three to eight relapses). Relapses usually occurred within 2 months of the initial presentation (thereby qualifying as a treatment failure-six of eight patients), progressed rapidly, were similar to the initial symptoms, and usually responded to reinstitution of plasma exchange or IVIG.

CIDP-MGUS group.

Fifteen patients had CIDP-MGUS-12 IgM (9 kappa, 3 lambda) and 3 IgG (1 kappa, 2 lambda). In one, the IgM monoclonal protein became a persistent polyclonal IgM elevation following treatment. The urine protein electrophoresis showed a monoclonal IgG in one and kappa light chain in another and was normal in twenty. A skeletal survey and chest and abdominal CTs were normal in all patients tested. One patient had a mildly dysplastic bone marrow, and another subsequently developed a plasmacytoid lymphocytic lymphoma. Six of the 12 with an IgM paraprotein had elevated anti-MAG antibodies; 2 had low anti-MAG titers (1:1,600) associated with SGPG antibodies (1:3,200, 1:26,600). A polyclonal gammopathy was found in 16 (24%)-IgM in 5, IgG in 4, IgA in 5, IgA and IgM in 1, and IgA and IgG in one. ^[15]

CIDP associated with other diseases.

Fourteen patients (21%) had one or more associated systemic illnesses, independent of MGUS, that may have contributed to the neuropathy; diabetes (six patients) and hypothyroidism (six patients) were the most common. In addition, two patients had gout, asthma, and pernicious anemia, and one patient each had rheumatoid arthritis, mixed connective tissue disorder, micronodular cirrhosis, end-stage renal disease, Addison's disease, subhepatic abscess, or autoimmune hemolytic anemia. One patient had T-cell lymphoma (unassociated with MGUS), and one had clinical evidence and evidence on brain MRI of multiple sclerosis. None had serologic or clinical evidence of HIV infection, hepatitis, sarcoidosis, or Lyme disease.

Cerebrospinal fluid.

CSF was obtained in 52 patients. The protein concentration was elevated (>45 mg/dL) in 43 patients (mean, 131 mg/dL; range, 21 to 1,236 mg/dL). The fluid was acellular in most, but 14 patients had three to eight white blood cells and one patient had 58 lymphocytes; the lumbar puncture was not repeated, but there was no clinical evidence of HIV or other systemic illness in this patient.

Electrophysiologic studies.

All patients fulfilled at least three of four criteria for demyelination, but there was a wide range of electrophysiologic abnormalities. A total of 231 motor nerves were examined. Distal latencies were prolonged in 104 (45%), conduction velocity was slowed in 139 (60%), and F waves were prolonged or absent in 143 (62%). Conduction block was demonstrated in at least one nerve in 49 patients (73%). Twenty-one patients (31%) had a pure demyelinating polyneuropathy, whereas the remainder demonstrated a mixed axonal and demyelinating neuropathy with low compound muscle action potential amplitudes or fibrillation potentials on needle electrode examination. The tibial motor potential was absent in 36% of those tested, and the peroneal potential was unexcitable in 13%.

Nerve biopsy.

Eleven of 18 patients with sural nerve biopsies showed varying degrees of axonal loss, ranging from a mild decrease in the number of myelinated nerve fibers to virtually complete loss of all axons ("end stage" nerve). Seven biopsies had thinning of myelin sheaths on semithin sections suggestive of demyelination. One specimen had prominent onion-bulb formation, and three (18%) had endoneurial inflammatory changes with mononuclear cell infiltration. Teased-fiber studies showed segmental demyelination in two and were normal in another. Six specimens (35%) showed only axonal loss, three (18%) had a profile of demyelination or remyelination, five (29%) had mixed demyelinating and axonal changes, and three (18%) were normal.

Comparison of CIDP-I and CIDP-MGUS patients.

Treatment and outcome in the MGUS group.

The neuropathy associated with MGUS was progressive in 10 and relapsing in five. All seven patients treated initially with PE responded, whereas only two of six improved with IVIG, and one of two with steroids (p = 0.02). Two could not tolerate the initial therapy. Of the five who did not improve, three responded to plasma exchange when offered as an alternative treatment. Five patients responded but relapsed after 2 months, with four benefitting from retreatment. Two patients who relapsed with PE responded after cyclophosphamide was combined with PE. In all, 13 patients (87%) improved after therapy with PE, IVIG, steroids, or cyclophosphamide. At the most recent followup (median, 3.1 years), 10 patients (67%) had minimal impairment (Rankin score 1 or 2) and 5 (33%) had moderate disability (Rankin score 3).

Treatment and outcome in the idiopathic group.

(Table 4 and Table 5) There was no difference in the mean age, sex, duration of symptoms, strength (average MRC score), baseline Rankin disability score, or mean motor nerve conduction studies among the 44 patients in the three treatment groups.

Comparison of patients with CIDP-MGUS and idiopathic CIDP.

Patients with CIDP-MGUS made up 25% of our cohort. Although there were distinguishing features between the CIDP-MGUS and CIDP-I groups, we found more in the way of similarities. In contrast to recent studies, ^[1-3,7,9,23] the age at presentation, duration of symptoms at the time of diagnosis, and severity of neurologic impairment (Rankin disability score) were similar in the groups. In particular, we did not find that the CIDP-MGUS patients had a slower, more indolent course compared with CIDP-I patients ^[1,9]; the mean duration of symptoms at the time of evaluation was longer in CIDP-MGUS patients, but because of the wide range, this was not significant. As appreciated from previous reports, ^[1,9] patients with CIDP-I had greater weakness at the time of initial evaluation, whereas CIDP-MGUS patients were more likely to have imbalance, leg ataxia, and vibratory sensation loss in the hands; tremor was only marginally more frequent in the MGUS group. Sensory conduction studies were also more often absent in the upper extremities in the CIDP-MGUS patients. Despite these group differences, there was substantial overlap of clinical and electrodiagnostic features, and we found it difficult to reliably distinguish between CIDP-MGUS and CIDP-I in an individual patient.

Because most of our CIDP-MGUS patients had an IgM gammopathy, we are unable to comment about the differences between IgG/IgA and IgM groups. Several previous studies ^[23,38-40] have suggested that patients with IgM paraproteinemic neuropathy have more severe sensory loss and ataxia and can be differentiated clinically and electrophysiologically from those with IgG or IgA paraproteinemic neuropathies. The greater sensory impairment in the CIDP-MGUS group is presumably explained by a high proportion of IgM patients in our cohort.

More CIDP-MGUS patients had a favorable response to PE compared with CIDP-I patients, but the trend was not significant (p = 0.02). Others have observed improvement in patients with paraproteinemic neuropathies following PE. ^[41-44] Our retrospective data suggested that the main therapies for CIDP-I are just as effective for patients with CIDP-MGUS.

Treatment in patients with CIDP-I.

Only 39% of our CIDP-I patients responded initially to one of the three modalities of treatment. Prior studies ^{[1-4,11,12,22]} have suggested that most CIDP patients improve with treatment, with response rates ranging from 40 to 95%. Most of these reported overall response rates to several different treatment regimens administered sequentially or in combination, which may in part explain why our results after initial therapy are somewhat poorer. The relatively long duration of illness when we first treated these patients is likely to have been an additional factor. Nonetheless, the response rate in our PE patients is identical to the prospective trial of Dyck et al. ^[45] from the Mayo Clinic (5 of 15 patients improved). Others ^[11,46,47] observed similar response rates. In another series from the Mayo Clinic, ^[4] only 39% of patients responded to steroids, similar to our 27%.

When patients who failed to improve or relapsed within 2 months were treated with an alternative second or third therapy, one-third improved with each treatment modality, increasing the overall response rate to 66%. Simmons et al. ^[1] and van der Meche and van Doorn ^[48] found that almost one-third of their initial nonresponders improved with a second or third treatment modality. Others have reported success rates of two-thirds in smaller series. ^[49] We therefore emphasize that fewer than one-half of patients with CIDP-I respond to the first treatment, but the majority will improve if alternative therapies are used.

PE and IVIG are equally effective for CIDP, ^[50] but there has not been a comparison with steroids. We found that the response rates were similar among all three modalities when used as an initial, second, or third-line treatment. However, the degree of neurologic impairment (Rankin score) improved to a greater extent with PE compared with steroids. There were no differences in the Rankin score between PE and IVIG or between IVIG and steroids. We therefore have favored PE or IVIG as initial therapy for CIDP, followed by the alternative therapy in those who do not respond. This conclusion is based on data reviewed retrospectively and should be interpreted accordingly. Almost one-half of patients responding to therapy relapsed after 2 months, but all improved with retreatment.

In summary, we found CIDP to be clinically more heterogeneous than previously appreciated. The majority of CIDP patients had axonal loss on EMG, suggesting that electrodiagnostic studies must be extensive to establish the presence of demyelination. Patients with CIDP associated with MGUS were only subtly different from those with CIDP-I, and most CIDP-MGUS patients responded to the conventional therapies. Patients with CIDP-I had similar response rates to PE, IVIG, and steroids, and most patients who fail to improve with initial therapy will respond to an alternative treatment modality.