Sensory Chronic Inflammatory Demyelinating Polyradiculoneuropathy: Neglected Immunotherapy-Responsive Sensory Neuropathy

In 1992, Oh et al.1 reported 10 cases of chronic sensory demyelinating neuropathy (CSDN) as a pure sensory neuropathy of chronic inflammatory demyelinating polyneuropathy (CIDP).

This entity was characterized by subacute or chronic progression, pure sensory neuropathy, high cerebrospinal fluid (CSF) protein level in most of cases, electrophysiological evidence of demyelination, demyelination in the nerve biopsy, and a good response to immunotherapy in the progressive phase. Over the past 30 years, there have been several papers on small number of cases reported as “sensory CIDP” as CIDP variant under the various different names and diagnostic criteria: sensory CIDP, pure sensory CIDP, sensory predominant CIDP, and chronic immune sensory polyradiculopathy (CISP).2, 3, 4, 5, 6, 7, 8, 9 A few questioned an existence of sensory CIDP as a variant of CIDP claiming that this is a transient clinical stage that precedes the appearance of weakness in CIDP.2, 8

To delineate the true nature of sensory CIDP, especially in term of immunotherapeutic response, we report the clinical, electrodiagnostic, histopathological, and immune therapeutic aspects of 56 cases of sensory CIDP collected for 30 years at the University of Alabama at Birmingham (UAB) Medical Centers. This study was approved by the UAB IRB (300007885). Informed consent on medical procedures was obtained from patient during the routine care of patients.

Sensory CIDP was diagnosed in this study when following two inclusion criteria and one exclusion criterion are met. The two inclusion criteria were 1) acquired, chronic progressive or relapsing symmetrical or asymmetrical sensory polyneuropathy solely based on clinical findings that had progressed for >2 months, and 2) definite electrophysiological and/or biopsy evidence of demyelinating neuropathy.10 The single exclusion criterion was anti-myelin associated glycoprotein neuropathy.11

Neuropathy was classified into four types: sensory neuropathy, painful neuropathy, ataxic neuropathy, and focal sensory neuropathy. Painful neuropathy required strong analgesics or narcotics for pain control. For ataxic neuropathy, patients had a gait abnormality due to ataxia as the chief complaint. For focal neuropathy, sensory neuropathy was present in one or multiple limbs asymmetrically. All other cases were classified as sensory neuropathy.

The electrophysiological criterion of demyelinating neuropathy was abnormal findings in a nerve conduction study (NCS) with an electrophysiological criterion of demyelination in any motor or sensory NCSs in at least two nerves following the UAB guidelines (Table 1).10

Table 1
Electrophysiological criteria of demyelination

• Click for larger image
• Click for full table
• Download as Excel file

Routine NCSs were performed using standard surface electrode placement as described previously.12 All patients underwent motor NCS in the median, ulnar, fibular, and tibial nerves; sensory NCS in the median, ulnar, and sural nerves; and mixed NCSs in the median and ulnar nerves. In 15 patients, the near-nerve needle (NNN) sensory NCS of the nerve was required to document demyelination in 13 in plantar nerve, 3 in ulnar nerve and 1 in median nerve.12, 13

A sural nerve biopsy was performed using the standard method behind the lateral malleolus, with the obtained samples investigated as the frozen, paraffin-embedded, and semi-thin sections in 36 patients.14 Inflammatory cells were assessed on the frozen and paraffin-embedded sections while demyelination was assessed on the semithin section as a routine. The teased nerve is required in a few cases. Demyelination was diagnosed based on many thinly myelinated fibers and/or onion-bulb formation on the semithin section or segmental demyelination in the teased nerve fibers.

In 43 patients, the CSF was obtained by lumbar puncture for diagnostic purposes. All patients received basic peripheral neuropathy work-up that included thyroid and rheumatology profiles including anti-Ro and anti La antibody tests, vitamin B12 and folic acid, hemoglobin A1C, sedimentation rate, and immunoelectrophoresis of serum and urine protein by immunofixation. All patients underwent common serum antibody tests for peripheral neuropathy available at the commercial laboratory at their initial visit including sufatide and GM1 antibodies. In none of our patients, IgG4 antibodies for nodal or paranodal protein were systematically checked.15 Magnetic resonance imaging (MRI) scan of spinal roots or plexus was not used as a routine procedure in our patients.

The degree of disability was assessed using the disability scoring system for peripheral neuropathy reported by UAB,16 in which the score is 0 is for normal and 20 for maximally disabled. The disability score is the total sum of scores for four different modalities: two motor functions (distal and proximal) and two sensory functions (pin-prick and touch, vibration and position).17 Improvement was defined when there was an objective improvement by at least 2 “disability scores” of 10 total sensory scores within one year of treatment.

The clinical and laboratory features of sensory CIDP were compared with those features of typical CIDP reported in 1996.17 From the UAB file, two CIDP cases were selected from the UAB file closest to one index of subacute inflammatory demyelinating neuropathy.17

“Chronic sensory demyelinating neuropathy (CSDN): CIDP presenting as pure sensory neuropathy” was used to represent this entity in the original paper in 1992.1 We adopted the term “sensory CIDP” purely based on the clinical manifestation and documentation of “demyelination” by NCS and/or histological studies. We believe that the use of “sensory CIDP” reflects better the commonly accepted view that this entity is a variant of CIDP.

We did not adopt the terminology of “sensory predominant CIDP” if motor nerve conduction slowing or motor conduction block are present as suggested by the latest recommendation.21 However, we reserved the terminology of “pure sensory CIDP” for sensory CIDP patients who had demyelination in the sensory nerve NCS only.19, 20 In our series, there were eight such patients who had normal motor nerve conduction but had demyelination in the NNN sensory nerve conduction only.

It is widely believed that CISP and CISP-plus represent one spectrum of a sensory variant of CIDP.4, 5, 22 In these cases, NCS showed either normal or “axonal.”12 Proximal root biopsy showed the loss of large fibers and onion-bulb formation were found in 3 of 15 CISP cases and in 11 of 44 CISP-plus and 22 CISP cases.4, 5 These 14 cases represent sensory CIDP under our diagnostic criteria even though NCS study showed no evidence of demyelination. However, the NNN sensory conduction study was done in any of the CISP or CISP plus cases. We believe that CISP represents a “proximal form” of sensory CIDP.

Some authors have questioned the existence of sensory CIDP as a separate variant of CIDP claiming that this is a transient clinical presentation in typical CIDP.7, 8 van Dijk et al.7 reported weakness within 0.9 and 6.3 years after onset of sensory symptoms in five out of seven CIDP. The mean duration of the disease before weakness developed was 3.1 years. Berger et al.8 reported weakness in 9–12 months after presentation in three of five sensory CIDP patients. In our study only 2 out of 56 patients developed weakness during the mean follow-up period of 6 years. The mean duration of symptoms in our patients was 3.9 years. These findings suggest that weakness develops mostly during the first four years of sensory CIDP.

In our electrodiagnostic criteria of demyelination, we have used the UAB criteria.10, 12 The UAB criteria are much stricter than the criteria by Joint Task for EFNS/PNS by using 50% decrease of compound muscle action potential (CMAP) as a criterion of conduction block.21, 22 The UAB criteria also include the demyelination criteria by the sensory nerve conduction to accommodate the diagnostic value of the NNN-sensory NCS in sensory CIDP.

In the present study, evidence of demyelination was documented by motor NCS alone in 70% of patients in the absence of motor weakness. It has been known that motor NCS is abnormal in many sensory neuropathies.12 There has not been a good correlation between neurological findings and nerve conduction abnormalities. Exception has been pure sensory neuronopathy pattern and pure motor neuropathy patterns.

The mean age at onset in our series was 55 years of age, which is 10 years older than for typical CIDP. Similar observation is also observed in an Italian study.6 In pediatric literatures, Łukawska et al.23 reported 1 case of sensory CIDP among 39 CIDP patients.

Our study found that the mean duration of symptoms in sensory CIDP was three times longer than that in typical CIDP. This suggests that patients perceived sensory symptoms as being less disabling than motor weakness, and explain why 20% of sensory CIDP patients did not need any immunotherapy and reflects the lower disability score in sensory CIDP than that for typical CIDP.

We have classified sensory CIDP into four clinical types. The most common type is sensory neuropathy type as expected followed by painful neuropathy type. In our 1992 paper, we found patients who needed narcotics or strong analgesic.1 Even in painful neuropathy type, the improvement with immunotherapy was as good for the other neuropathy types. A recent review found that the prevalence of pain was 46% in CIDP patients during their course of CIDP and that immune therapy may be adequate for the management of pain.24 The ataxic neuropathy type was seen in one case in our 1992 study and in five patients in the present series.1 The ataxic neuropathy type was also reported in five patients by Ohkoshi et al.25 and all of them responded to steroid sensory ataxia was main symptom in CISP and CISP-plus.3, 4 Those cases are different from sensory CIDP because the main pathological site is in the spinal root and the routine NCS is normal. The present series included one case of focal neuropathy type.18 So far two other cases with focal sensory CIDP were reported.9, 26

The most prominent laboratory abnormality was a high protein in the CSF which we found in 21 (49%)/47 patients. In 1992, we observed elevated CSF protein (>55 mg/dL) in four of seven tested patients. In Italian series, a high CSF protein was seen in 69% of cases.6 Thus, as in typical CIDP, the high CSF protein is the most prominent laboratory feature in sensory CIDP, representing an inflammatory process in the spinal roots. A peripheral antibody panel did not show any consistent pattern suggestive of sensory CIDP.

The most important diagnostic test in sensory CIDP is the extensive routine motor and sensory NCSs, since they can detect an electrophysiological evidence of demyelination in 70% of patients. Among the various demyelination markers, the NCV is the most common marker for demyelination as abnormality was observed in 57% of patients. Two caveats in the neuropathy work-ups at the UAB are that NCS was applied in one arm and both legs as a routine test and NNN-sensory NCS was performed in 15 patients. All of these NNN sensory NCSs produced evidence of demyelination. If we choose the NNN sensory NCS as the second diagnostic test, sural nerve biopsy would have been needed in two patients in whom the routine and NNN sensory NCS did not detect demyelination. Thus, we found the NNN sensory NCS to be a powerful tool for recognizing demyelinating neuropathy as previously reported by Odabasi and Oh.13 Thus, the NNN sensory NCS is essential as the second diagnostic test when a sural nerve biopsy is not available.

A sural nerve biopsy is the next important diagnostic test in sensory CIDP, and this detected demyelination in 29 (81%) of our 36 biopsied cases. In 10 patients in whom the routine NCS did not show demyelination, the definite diagnosis of sensory CIDP was made by the sural nerve biopsy. Three of these 10 patients did not undergo an NNN sensory NCS. This confirms the important diagnostic role of the sural nerve biopsy in sensory CIDP. Chin et al.2 presented eight patients with sensory CIDP patients in whom the sural nerve biopsy revealed demyelinating neuropathy by teased fiber preparation despite of axonal neuropathy being detected in the NCS. It is noteworthy that four of those patients who received IVIg therapy improved. Vallat et al.27 also were able to make the diagnosis of CIDP using nerve biopsy in eight patients with chronic idiopathic axonal polyneuropathy, five of whom responded well to immunotherapy and three had sensory CIDP.

Immunotheraohy was needed in 5 of the 10 patients reported in the 1992.1 A clinical improvement was achieved with steroid alone in three and with steroid/plasma exchange/azathioprine in two. Immunotherapy was required in 73% of the patients in the present series. In the rest of patients, the disease is either mild or patients’ symptoms responded to medications such as analgesic or narcotics, thus not requiring an immunotherapy. Improvement was achieved in 88% of immunotherapy treated patients, which is comparable with the response rate for typical CIDP in general.16

IVIg was the main mode of therapy in sensory CIDP patients and achieved an improvement in 77% of treated patients, usually with combined with other immunosuppressors. An improvement was also observed in seven patients with steroids and in two patients with azathioprin as the main immunosuppressors. These improvement rates with IVIg and steroids in our patients are comparable with those reported in Italian study.6

Our study showed that improvement rates with immunotherapy in the various clinical types ranged from 75% to-90%. This includes painful neuropathy type. Previous studies found good improvement in the ataxic neuropathy type.3, 4, 25 A review by Michaelides et al.24 found that treating CIDP reduced pain in 41 (89%) of 46 patients: 10 with steroid alone, 7 with IVIg alone, and 24 with combined therapy. Our study thus confirmed those findings that immunotherapy improved pain in CIDP patients. We found that an immunotherapy was also effective in the pure sensory neuropathy type.

As expected, sensory CIDP showed less typical clinical, laboratory, and therapeutic features of typical CIDP. The most striking differences are that the requirement for immunotherapy and recovering to normal are less common in sensory CIDP, indicating that typical CIDP is more disabling and thus needs better recovery than sensory CIDP. This is understandable because motor weakness is more disabling for the activities of daily living.

Doneddu et al.6 compared clinical features between 16 patients with sensory CIDP patients and 376 typical CIDP patients, and found continuation of sensory CIDP beyond 8 years, older age of onset, good response to immunotherapy, and lower disability score in sensory CIDP. These are exactly what were observed in the present study.

In the present study, we have not study MRI scan of the spinal roots and plexus as the routine study in peripheral neuropathy work-ups. We consider this to be a major limitation of study because we will not detect an enhancement of the roots or proximal segment of the nerve as observed in CISP or CISP-plus.3, 4

Since the first report of sensory CIDP in 1995, a subject of sensory CIDP as an immune responsive sensory neuropathy was discussed mostly in CIDP reviews but not in the general review papers on sensory neuropathy or neuronopathy.28, 29, 30 Because sensory CIDP responded to an immunotherapy in 88% of treated patients and that sensory CIDP could be be diagnosed by the routine motor NCS in 70% of patients and by the nerve biopsies in 18% of patients in the present study, it is time to recognize this neglected entity and treat the patients with sensory CIDP accordingly.

Four conclusions can be drawn on the basis of 56 patients with sensory CIDP in the present study: 1) sensory CIDP exists as a separate CIDP variant and is less disabling than typical CIDP; 2) the diagnosis of sensory CIDP is made in 70% of patients by the routine motor NCS, in other 18% of patients by sural nerve biopsy and in the last remaining 12% of patients by NNN sensory NCS; 3) sensory CIDP is responsive to an immunotherapy in 88% of treated patients; and 4) compared with typical CIDP, recovery to normal is less common in sensory CIDP.