Like Guillain–Barré syndrome (GBS), chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) is a syndrome with both classic and variant phenotypes. In the case of GBS, the classic phenotype is referred to as acute inflammatory demyelinating polyradiculoneuropathy (AIDP), a disorder characterized by areflexia, generalized and usually symmetric weakness, and sensory involvement. Likewise, the classic phenotype of CIDP, typically referred to simply as CIDP, shares many of the characteristic clinical and electrodiagnostic (EDX) features as GBS.

As the cause of CIDP is unknown and there is no universally agreed upon diagnostic gold-standard for CIDP (15 CIDP and 16 EDX-proposed diagnostic criteria paradigms to date), the classification of the chronic acquired demyelinating neuropathies remain in flux.^1,2 This chapter will address the majority of the chronic, acquired, predominantly demyelinating, and presumably inflammatory neuropathy syndromes. POEMS, with its associated CIDP-like neuropathy phenotype, will be the notable exception, discussed separately in Chapter 19.^3,4

Current classification schemes often consider any neuropathy that fulfills the EDX criteria for CIDP to be categorized as CIDP or a CIDP variant. Multifocal motor neuropathy (MMN) is the notable exception, now being considered as a separate entity by virtually all neuromuscular experts.⁵ We however, consider the CIDP spectrum to include only those disorders which have common phenotypic, EDX, cerebrospinal fluid, and therapeutic responsiveness features with the classic syndrome.^2,6,7 Included in this category are multifocal acquired demyelinating and sensory motor neuropathy (MADSAM) (a.k.a. Lewis Sumner syndrome) as well as some pure motor and pure sensory variants. Conversely, we will consider other chronic, acquired, and predominantly demyelinating neuropathy syndromes as separate entities if they differ significantly, particularly in consideration of their phenotype, natural history, and their therapeutic responsiveness profiles. We do so even if they fulfill EDX criteria for CIDP. Within this latter category, we will discuss distal acquired demyelinating sensorimotor (DADS) neuropathy, MMN, chronic immune sensory polyradiculopathy (CISP), and multifocal acquired motor axonopathy (MAMA) and attempt to distinguish them based on clinical, EDX, natural history, and response to treatment data (Table 14-1).

The first report of apparent CIDP, referred to as recurrent polyneuritis, is credited to Eichorst in 1890.⁸ In the mid-1950s, animal models of both acute and chronic experimental allergic neuritis provided scientific support for an autoimmune pathophysiology. This concept was further cemented by a seminal report by Austin in 1958 describing steroid responsiveness in patients with relapsing polyneuritis.⁹ Nonetheless, “chronic relapsing polyneuritis” was often considered to be a form of GBS in early publications until the mid 1970s.^10,11 Arguably, two developments promoted widespread neurological awareness of CIDP. The first was the availability and widespread utilization of nerve conduction studies, techniques that allowed for the noninvasive recognition of the demyelinating features that characterize this syndrome, demyelination of peripheral nerves having been previously demonstrable only by pathological means. The second seminal moment in the history of CIDP occurred with the 1975 publication by Peter Dyck, considered by many to be the patriarch of peripheral neuropathy in the United States.¹²

CLINICAL FEATURES

The prevalence of CIDP has been reported to range from 0.8 to 8.9/10⁵ patients depending on the population studied. Although this may suggest different susceptibility between different geographical locations or ethnicities, this data may be heavily biased based on the diagnostic criteria utilized in different studies.¹³ CIDP usually presents in adults (peak incidence at about 30–60 years of age) but it can manifest at any age including infants and children.^{11,12,14–30} The relapsing form tends to present earlier, usually in the twenties.^12,14 There is a slightly increased male prevalence, up to two-thirds of cases in some series.^12,31–33 Like GBS, CIDP may begin or relapse in association with an antecedent event that may include infection (10–30%), vaccination, surgery, trauma, or pregnancy.^{12,13,15,24,25,34,35} CIDP may account for 10–33% of initially undiagnosed peripheral neuropathies in some series.^14,36,37 It is possible however, that these statistics represent a biased perspective considering that CIDP patients with their attendant morbidity are more likely to be referred to academic neuropathy clinics than are the far more abundant, indolent, length-dependent, sensory predominant, and frequently idiopathic axonopathies of the elderly.

The natural history of CIDP needs to be considered both with and without the influence of treatment. Up to 18% of CIDP patients will evolve acutely in such a manner to be initially confused with GBS.^32,38–40 Approximately 12% will evolve subacutely over a 4–8 week period, also confounding the distinction of CIDP from GBS in some cases.^41–44 These subacute cases may have a monophasic course with recovery reminiscent of GBS or have a progressive or relapsing course thereby justifying a CIDP diagnosis. In our minds subacute inflammatory demyelinating polyradiculoneuropathy represents a “holding” diagnosis until such time as categorization of GBS or CIDP can be made.

Regarding the natural history of treatment of naïve CIDP patients, the seminal paper by Peter Dyck and colleagues in 1975 describe four disease trajectories: (1) chronic monophasic (15%), (2) chronic relapsing (fluctuations of weakness or improvement over weeks or months) (34%), (3) stepwise progressive (34%), and (4) steady progressive (15%).¹² With treatment, a consensus group of experts, using a CIDP disease activity status scale, determined that a minority of patients (11%) achieved a status of “cure,” defined by a stable examination off treatment for more than 5 years.³¹ Half of these patients had normal examinations and the remainder had apparent mild and presumably minor clinical findings. Twenty percent of their 106 patients were considered in remission, similar to the cure group with the exception that they had not been medication-free for 5 years. Forty-four percent of patients in this series were classified as having stable active disease in which ongoing treatment was required. Another 7% improved in response to treatment which had recently been administered, precluding classification elsewhere. Finally, 18% were labeled as unstable active disease. A third of this cohort had not received treatment and two-thirds or 11% of the entire group were identified as unstable and active with treatment unresponsiveness.

Classic CIDP is a subacute to chronic, motor predominant disorder that presents with a non–length-dependent and symmetric pattern of weakness.^{2,6,12–14,32,36,45,46} It is estimated that approximately half of CIDP patients present in this manner although these statistics may be biased by the inclusion of patients with DADS and CISP variants within the CIDP denominator.³³ Although distal limb muscles may be more severely affected early, significant and at times dominant involvement of proximal limb muscles is characteristic of the syndrome. This is estimated to occur in approximately 75–90% of patients, 90% of which will have a symmetric pattern.^{9,11,13–16,31,47} In addition, approximately 80–94% of CIDP patients, like GBS, will have sensory symptoms and usually signs that may be the presenting or relapsing manifestation but that are usually rapidly overshadowed by the morbidity of their weakness.^13,31,47 Although the sensory symptoms are usually most pronounced in the distal extremities, they can be frequently identified as being nonlength dependent by affecting the hands before, at the same time, or soon after involvement of the feet. This is notably different from most length-dependent axonal neuropathies. Loss of large-fiber sensory modalities is typically more pronounced than their small-fiber counterparts but loss of all sensory modalities is not rare. Accordingly, dysesthesias occur in 15–50% of affected individuals.^12,15,48 When back pain occurs early in the syndrome in a manner similar to GBS, it is presumed to represent nerve root inflammation.⁴⁹ In addition, CIDP patients along with most acquired demyelinating polyneuropathies, have generalized areflexia (70%) or hyporeflexia in 97% of cases. Pathophysiologically, this is a presumptive effect of desynchronous impulse transmission associated with the temporal dispersion so frequently identified electrodiagnostically.^13,31

Involvement of cranial nerves, phrenic nerves, and other nerves innervating intercostal and other ventilatory muscles, and the autonomic nervous system may occur in CIDP but are far less prevalent than in GBS.¹³ Bifacial weakness although typically mild, occurs more commonly in our experience than the 15% reported figure.^{12,14,15,31,47} Ophthalmoplegia, vestibulocochlear symptoms, and bulbar weakness are less common manifestations.^{12,14–16,20,50,51} A rare presentation is neck extensor weakness leading to the dropped head syndrome.⁵² Papilledema may be seen in rare patients with CIDP, but its presence should heighten consideration toward POEMS syndrome (polyneuropathy, organomegaly, endocrinopathy, monoclonal gammopathy, and skin changes).^12,53 Symptoms of dysautonomia are uncommon and typically mild when present, affecting distal postganglionic axons and producing mild, cholinergic, predominantly sudomotor dysfunction.^{12,15,54–56} Ventilatory failure in CIDP also occurs much less frequently than in GBS, but can occur perhaps more commonly in POEMS patients and is one source of mortality in this disease.^33,56–58 We have had one experience with a patient with typical CIDP, exquisitely treatment responsive for years, who eventually became refractory to multiple different therapies, eventually becoming “locked-in” and ventilator dependent. Postural tremor may occur in CIDP. Rarely it may be the presenting and initially dominant symptom. At times, patients seem to be ataxic in a manner disproportionate to the degree of their weakness or sensory loss, reminiscent of Miller Fisher syndrome.

Pure motor forms of CIDP are estimated to occur in 5–15% of CIDP patients, pure sensory forms in 15–35%, and notable asymmetries in 30% of patients.^32,33 These statistics have to be interpreted with caution however as different series may include patients that may represent DADS, MMN, or MADSAM neuropathies. Pure motor forms are fairly easy to identify in most cases as the characteristic EDX features of acquired demyelination are the norm. A steroid responsive, pure motor axonal form of CIDP, analogous to the acute motor axonal neuropathy (AMAN) variant of GBS, has been described.^57,59

Pure sensory variants of CIDP are more difficult to define as they may or may not have concomitant demyelinating features on motor conduction studies which are arguably, the most reliable means by which to classify them as CIDP variants.^{15,32,60–62} Some of these patients will develop significant weakness and evolve into typical CIDP.⁶³ Many, however will remain as pure sensory syndromes. Whether most if not all of these should be more appropriately classified as DADS neuropathy or CISP is a rhetoric question for which there is no consensus answer. Patients that have been described as sensory CIDP seem to fit into one of three categories.³³ Approximately, half will have EDX features of demyelination despite absence of any clinical evidence of motor involvement, similar if not identical to the DADS phenotype at presentation.^33,64 The second group, constitutes approximately a quarter of the pure sensory phenotype.¹⁵ These patients have sensory signs and symptoms, coupled with abnormal sensory nerve action potential (SNAP) amplitudes, but normal motor conduction studies. These patients have been classified as CIDP, largely as a result of characteristic nerve histopathology. It is our suspicion that this group is the one more likely to have an alternative diagnosis, for example, sensory neuronopathy. The third group, representing the remaining 25% of the pure sensory CIDP phenotype, manifests as a pure sensory clinical presentation often associated with sensory ataxia, coupled with normal routine conduction studies.¹⁵ In these cases, elevated cerebral spinal fluid (CSF) protein levels and abnormal somatosensory conductions in the face of normal SNAPs implicate a nerve root localization. As such, these patients are similar if not identical to the described CISP syndrome.^60,65–68

As many as 3% of patients with CIDP develop evidence of central nervous system (CNS) demyelination clinically, electrophysiologically (evoked potential studies), or by magnetic resonance imaging (MRI).^69–75 Attacks of CNS demyelination can precede or follow the onset of CIDP. Like multiple sclerosis (MS), asymptomatic lesions may be detected by MRI and may represent a CIDP variant or coexisting MS.

DIAGNOSIS AND DIFFERENTIAL DIAGNOSIS

As previously mentioned, there are 15 published diagnostic criteria for CIDP that place variable emphasis on the importance of clinical, EDX, cerebrospinal fluid, nerve root imaging, nerve biopsy features, and response to treatment data.^2,75 Of these, the criteria proposed by the European Federation of Neurological Societies (EFNS) in conjunction with the Peripheral Nerve Society (PNS) in 2006 are viewed as having the optimal balance between diagnostic sensitivity and specificity with a 97% and 92% positive and negative predictive value, respectively (Table 14-2).^2,32,45 In summary, these criteria emphasize the importance of the clinical presentation and findings. They consider the results of CSF analysis, MRI, nerve biopsy, and response to treatment to be supportive considerations (Fig. 14-1). In other words, a diagnosis of definite CIDP can be made without any supportive tests in a patient with symmetric proximal and distal weakness, sensory signs and symptoms, and generalized hypo- or areflexia if their illness progresses or relapses over a period of more than 2 months in the setting of typical EDX findings. The probable and possible diagnostic categories are also determined by a combination of clinical and EDX criteria. When available, supportive data allows escalation of the diagnostic criteria. Our beliefs parallel the ENFS/PNS criteria with the following exceptions. We are very cautious about making a CIDP diagnosis in pure sensory cases, particularly in those with a DADS phenotype, particularly if there is an IgM monoclonal protein (MCP), with or without myelin-associated glycoprotein (MAG) autoantibodies.

Accurate diagnosis is important for research purposes, both in consideration of clinical trial inclusion and in order to be able to eventually illuminate the cause(s) of this syndrome. We believe however, that in clinical practice, the most important value of accurate diagnosis is to decide who to treat and what to treat them with. Avoiding expensive and potentially harmful treatment without likely benefit while at the same time ensuring that potentially treatment responsive individuals are identified are two self-evident benefits of accurate diagnosis. In order to achieve that goal we place the greatest emphasis on the patient’s phenotype and the supporting EDX data. We follow the lead of the EFNS/PNS in considering a characteristic CSF pattern, found in approximately 90% of patients, to be helpful but not diagnostically mandatory. For a number of reasons, we do not routinely recommend nerve biopsy as a means to diagnose CIDP, particularly when the diagnosis is well established by clinical and EDX means. We are more apt to perform nerve biopsies in atypical cases with the primary goal of excluding alternative diagnosis.

Regarding the application of EDX to CIDP, we believe that its role is to support the diagnosis in a patient with a compatible phenotype, not to establish it in the absence of a compatible clinical picture. Our position is supported by the knowledge that an EDX pattern compatible with an acquired demyelinating neuropathy is characteristic of a number of disparate phenotypes whose natural history and response to treatment may vary considerably from classic CIDP, for example, POEMS syndrome and DADS neuropathy.^{3,4,7,54,66,81} Even hereditary demyelinating neuropathies may at times have EDX patterns compatible with CIDP.⁷⁷ We do not adhere to the concept that a diagnosis of CIDP can be established by electrodiagnosis alone.

The differential diagnosis of CIDP begins in its distinction from GBS. As previously mentioned, up to 18% of individuals who will eventually develop a relapsing or progressing course justifying a CIDP diagnosis will have an initial evolution rapid enough to suspect GBS.^33,38–40 In these individuals, it would be pragmatic to initiate treatment with either intravenous immunoglobulin (IVIg) or plasma exchange (PLEX) and reserve any consideration of corticosteroids until the trajectory of the illness justifies a CIDP diagnosis.

CIDP is typically considered a primary diagnosis. In approximately a quarter of individuals, a concomitant systemic disease may be identifiable^{2,14,34,45,78–82} with lymphoma and diabetes being the most prevalent in one series. (Table 14-3)³³ Not included in this table are CIDP phenotypes which have been described in association with drugs such as cyclosporine, tacrolimus, and tumor necrosis alpha blockers.^{79,80,93–97} It is likely that the neuropathy in such cases is caused by the altered immune status of the patient, not to a direct toxic effect of the drug. Associations between CIDP and other disorders may represent two different diseases sharing a common pathophysiological mechanism or in the case of diabetes, a common disorder with a similar phenotype occurring coincidentally. Alternatively, and probably less commonly, CIDP might be caused by consequences of the primary systemic disease. The last consideration is the coexistence of two unrelated conditions, the second disorder being identified as a byproduct of the diagnostic scrutiny provided by the evaluation of the first.

In the early part of this century, there was considerable debate as to whether there was an increased incidence of CIDP in diabetic patients, or whether these were two disorders that might in some cases have overlapping clinical, EDX, or CSF features.^98–103 The practical consideration of this question revolves around the consideration of immunomodulating treatment in diabetics with “CIDP-like” features. Confounding this decision is the realization that nerve biopsy specimens of patients with diabetic radiculoplexopathy may have inflammatory features and may respond to immunomodulating treatment, without invoking any consideration of CIDP phenotype a demyelinating EDX signature.^104,105

Arguably, the most confusing consideration relevant to conditions associated or perhaps causally related to CIDP is the coexistence of an MCP. That a relationship exists is supported by a reported incidence of MCPs in CIDP as 22–30% ^{32,33,106,107} in comparison to 10% in polyneuropathies in general which in turn, is 6–10 fold greater than in an age-matched population.¹⁰⁶ Confusion arises as the neuropathies associated with MCPs may share phenotypic and EDX features with CIDP, or they may not. For example, the neuropathy associated with IgM MGUS commonly fulfills EDX criteria for CIDP, even if the phenotype (DADS) is disparate in the majority of cases. In addition, patients with IgM MCPs differ in most cases from CIDP without MCPs or those associated with IgG/IgA MCPs as they have (1) a definable autoantibody in many cases (anti-MAG), (2) a differing natural history, (3) a distinctive histopathology, and (4) a different response to treatment.^{14,106,108,109}

Conversely, although IgA/IgG-associated neuropathies frequently have an axonal EDX profile, they may manifest with both the clinical and EDX features of CIDP.¹⁰⁹ They tend to manifest more motor involvement, faster progression, less of a demyelinating EDX profile, and better treatment responsiveness than do their IgM counterparts. For these reasons, like others,^{2,14,45,102,109} we do not consider the presence of an MCP, particularly if IgG or IgA, to preclude a CIDP diagnosis as have others.^12,15

When a neuropathy coexists with an MCP, it is usually an MCP of unknown significance (MGUS). It is important to remain aware that an MCP may be representative of an underlying systemic disease which needs to be considered not only at the time that the MCP is recognized but subsequently as MGUS may evolve into lymphoma, Waldenstrom macroglobulinemia, chronic lymphocytic leukemia, multiple myeloma, POEMS syndrome, amyloidosis, heavy chain disease, or cryoglobulinemia.¹⁰⁹

Most diagnostic criteria for CIDP require exclusion of other neuropathies with definable etiologies that may have overlapping features. This list has to be to some extent flexible, given the uncertainties of CIDP diagnostic boundaries as discussed above. According to the EFNS criteria, hereditary neuropathies including demyelinating CMT genotypes and Refsum disease should be considered and excluded when relevant, to the extent possible.⁴⁵ In addition, these criteria consider MMN and DADS neuropathy associated with (but not without) demonstrable MAG autoantibodies as separate disorders. We find this latter argument somewhat specious as the ability to detect MAG autoantibodies varies considerably depending on the technique used and as on the whole, DADS patients with and without MAG autoantibodies are indistinguishable.¹¹⁰ It is our personal bias that amyloidosis should warrant special consideration in individuals with “CIDP-like” phenotypes with axonal EDX signatures, particularly in association with symptoms of extraneural or autonomic involvement.⁹⁰

LABORATORY FEATURES

Most patients (80–95%) have an elevated CSF protein (>45 mg/dL) with a mean of 135 mg/dL and levels that may exceed 1,200 mg/dL.^{10–12,14,16,48} Very high CSF protein levels associated with CIDP should trigger consideration of POEMS syndrome or blockade of the spinal canal from hypertrophic nerve roots. Similar to GBS, the CSF cell count is usually normal, although up to 10% of patients have greater than five lymphocytes/mm³. Elevated CSF cell counts should lead to the consideration of HIV infection, sarcoidosis, Lyme disease, and lymphomatous or leukemic infiltration of nerve roots. Oligoclonal bands may be demonstrated in the CSF in approximately 65% of patients.^111,112

Blood testing in CIDP is done largely in consideration of alternative diagnoses, associated diseases, or secondary causes. As previously discussed, a monoclonal gammopathy (IgA, IgG, or IgM) is present in up to 30% of patients with CIDP.^{14,47,48,63,102,107,113} A small number of patients have GM1, P₂, and particularly P₀ (20%) autoantibodies.^45,114–117 Antitubulin antibodies were reported in one study of CIDP¹¹⁷ but was not seen in others.^118,119 More recently, autoantibodies directed at novel nodal and paranodal myelin and axonal antigens have been reported in small numbers of CIDP patients (see below).^127–129 Many would consider the detection of MAG autoantibodies helpful, particularly in those patients with a DADS phenotype. In our estimation however, detection of these antibodies provide little if any practical value beyond the detection of an IgM MCP in most cases.¹⁰⁹ We believe that autoantibody testing currently has a limited role in the evaluation of a CIDP suspect.

Otherwise, as always, we feel that the evaluation of a CIDP patient should be determined in consideration of the clinical context of the individual case. In general, we follow the recommendations of the EFNS and obtain a complete blood count, urinalysis, C reactive protein or sedimentation rate, glycosylated hemoglobin, creatinine, transaminases, serology for Lyme disease, hepatitis B and C, and antinuclear antibodies in addition to a serum and in some cases urine immunofixation.⁴⁵ In cases in which an MCP is detected, evaluation for POEMS syndrome, lymphoma, myeloma, and amyloid are undertaken as warranted.

Hypertrophy and enhancement of the nerve roots and peripheral nerves may be appreciated with MRI which has been proposed as supportive criteria for diagnosis.^45,120–122 Rarely, a myelopathy can develop secondary to the markedly enlarged nerve roots compressing the spinal cord.¹²³

ELECTRODIAGNOSTIC FEATURES

Multiple nerves should be evaluated on NCS because of the multifocal nature of the disease process; some nerves can have normal conduction studies, while other nerves are abnormal. Various electrophysiological criteria for demyelination have been devised.^{1,2,14,45,78,124,125} Again, our position is that EDX assessment is an integral component of the CIDP diagnosis, but that it does not determine the diagnosis independent of clinical assessment. Less than two-thirds of patients with CIDP fulfill electrophysiological criteria for demyelination, regardless of the criteria utilized.^1,33

Motor Conduction Studies

Motor conduction studies include assessment of compound muscle action potential (CMAP) amplitudes, distal latencies, conduction velocities (CVs), F wave latencies, and waveform morphological changes such as temporal dispersion or conduction block. These are the most useful EDX tools in the evaluation of a patient with suspected CIDP.^1,2,14,45 As previously mentioned, there are at minimum 16 EDX criteria for CIDP that have been proposed. For purposes of simplicity, we will summarize those EDX criteria proposed by the EFNS/PNS which require demonstration of one or more of the following characteristic demyelinating abnormalities in at least two motor nerves (Table 14-2)⁴⁵:

1. 
   

   Prolongation of motor distal latencies by >50% of the upper limits of normal (ULN).

   
   
2. 
   

   Reduction in CV to <30% of the lower limits of normal (LLN).

   
   
3. 
   

   Conduction block as defined by CMAP reduction of >50%.

   
   
4. 
   

   Temporal dispersion as defined by >30% increase in CMAP duration in proximal compared to distal CMAP or distal CMAP duration exceeding 9 ms.

   
   
5. 
   

   20% increase in F wave latency if CMAP >80% LLN, >50% increase in F wave latency if CMAP <80% LLN, or absence of F waves in motor nerves with CMAP >20% LLN.

We would add the following caveats to the interpretation of these criteria. We are cautious about including conduction block at common compression entrapment sites, conduction block of the tibial nerve, and absent F waves of the peroneal nerve as fulfilling the diagnostic criteria for CIDP. Although of uncertain utility, phrenic nerve conduction abnormalities in CIDP patients occur and do so with greater frequency than in MMN.¹²⁶ In addition, we would be willing to identify a conduction block with a reduction of CMAP amplitude of less than 50% if the waveform change occurs abruptly over a short segment of a nerve in a location where neither entrapment or compression are likely.

As patients’ strength and function improve, repeat NCS may demonstrate evidence of improvement with increases in CMAP amplitudes and CVs along with reduction in the magnitude of conduction block.^16,127–133 In keeping with this, conduction block in motor nerves is an uncommon findings in patients with DADS neuropathy in keeping with the frequently pure sensory (clinical) nature of their phenotype, at least at onset.^106,109 Clinical improvement is primarily the result of resolving conduction block, although some may be attributed to improved ion channel function without remyelination, collateral sprouting, or regeneration of axons.

Sensory Conduction Studies

Most patients with CIDP have SNAPs that are reduced in amplitude or absent in both the upper and the lower extremities.^{10,14–16,61,107,133,134} When present, SNAPs may demonstrate conduction slowing either as prolonged distal latencies and/or slow CVs. However, this “slowing” is usually not as severe as that demonstrated in motor nerves. In a study of 18 patients using the near-nerve technique, sensory conduction slowing was only moderately slow in proportion to the degree of amplitude loss.¹³⁵ A helpful feature when present is the identification of median, ulnar, or radial SNAPs abnormalities when the sural or superficial peroneal SNAPs are normal. This pattern of “sural” sparing suggests a non–length-dependent process (most axonal neuropathies are length dependent). When sensory EDX abnormalities are worse in the arms than in the legs, one needs to consider a predominantly demyelinating neuropathy or sensory ganglionopathy.

Needle Electromyography

Insertional and spontaneous activities are often normal on needle electromyography (EMG). However, fibrillation potentials are not rare as there is often some element of secondary axonal loss. Occasionally, myokymic discharges may be seen related to ephaptic transmission between demyelinated nerve fibers. The earliest and perhaps only abnormality one might see on EMG is reduced recruitment (fast-firing) motor unit action potentials (MUAPs) that otherwise appear morphologically normal.

Evoked Potential and Autonomic Studies

Rare patients have abnormal visual-evoked, brainstem auditory-evoked, and somatosensory-evoked potentials suggestive of superimposed central demyelination.^72,73 Another application of SSEPs, like H reflexes, is their utility in identifying pathology of the dorsal roots. SSEP and H reflex abnormalities occurring in the context of normal SNAPs referable to the same dermatome(s) imply pathology of the dorsal root such as believed to occur in CISP (see below).^60,65–68 Although most patients are not symptomatic, evidence of subclinical autonomic neuropathy is not uncommon.^54–56 Blood pressure and heart rate responses to tilt testing followed by the 30/15 heart rate ratio in response to deep breathing are the most frequently abnormal autonomic studies.

HISTOPATHOLOGY

Nerve biopsies in CIDP should be interpreted with the realization that a sensory nerve is being examined in a disorder that typically has a motor predominant phenotype. As a result, neither the extent nor type of abnormality identified may be fully representative of the entire disease process. Nerve biopsy is a useful, but not requisite diagnostic tool, for CIDP, arguably of greater utility in excluding other disorders than in proving the existence of CIDP.^136,137 Biopsies are particularly useful when vasculitis, lymphomatous infiltration, amyloidosis, or sarcoidosis are considered.

Nerve biopsies may reveal segmental demyelination and remyelination which may not be evident due to the multifocal nature of the process (Fig. 14-2).^{12,14,70,136,138,139} Chronic demyelination and remyelination result in proliferation of surrounding Schwann cell processes known as “onion bulbs.” These are the basis of hypertrophic nerves and are seen CIDP although they are not as prominent as in demyelinating forms of Charcot–Marie–Tooth disease (see Fig. 3-31). Myelinated fibers are usually reduced in number. Fibers examined in semithin sections demonstrate myelin thickness that is disproportionately thin in relationship to axon diameter indicate remyelination (Fig. 14-2C). Teased nerve fiber analysis demonstrates segmental demyelination and/or remyelination in 23–46%, axonal degeneration in 21–42%, mixed demyelinating and axonal features in 12.5%, and normal findings in 18–43.5% of nerve biopsies from studied CIDP patients.^12,14 (see Fig. 3-30).

Endoneurial and perineurial edema may also be appreciated on biopsy. Inflammatory cell infiltrate may be evident in the epineurium, perineurium, or endoneurium. It is often perivascular when detectable but often quite subtle or absent in nerve biopsy specimens (Fig. 14-2A).^12,14 Inflammatory cells are better appreciated with immunostaining for lymphocytes (Fig. 14-2B).^140,141 This inflammatory component comprises of macrophages, CD3+-activated T cells (mainly CD8+ but also CD4+ cells lymphocytes), and dendritic cells.^141,142 Of note, a similar frequency of inflammatory cell infiltrate within nerves is seen in a variety of neuropathies, raising questions concerning the pathogenic role of these cells.¹⁴⁰ The matrix metalloproteinases MMP-2 and MMP-9 (gelatinase A and B) are overexpressed in the peripheral nerves in patients with CIDP.¹⁴³ These enzymes are secreted by T cells and are capable of digesting basement membrane proteins, thereby facilitating the infiltration of inflammatory cells into peripheral nerves.

On electron microscopy (EM), macrophages may be noted to penetrate the basement membrane with displacement of the Schwann cell cytoplasm, lyse superficial myelin lamellae, penetrate along intraperiod lines, and engulf the disrupted myelin by endocytosis. By doing so, they disrupt the nodes of Ranvier, and by doing so, presumptively saltatory conduction.¹⁴⁴ Subsequently, Schwann cells are recruited to remyelinate the demyelinated internodes. The demyelinated axons diminish in diameter as much as 50% but later regain some of their diameter following remyelination.

PATHOGENESIS

There are numerous aspects of both the anatomy and physiology of peripheral nerve that are relevant to both normal impulse transmission, and to the mechanisms of abnormal nerve transmission associated with the acquired, immune-mediated, and demyelinating polyneuropathies. These variables include axonal diameter which varies under normal circumstance between the node of Ranvier and the internode as well as the intervening paranode and juxtaparanode (Fig. 14-3). In addition, normal and abnormal nerve impulse transmission are related to many factors pertaining to the surrounding myelin sheath, and to the type, location, and density of the ion channels upon which action potential generation is ultimately dependent.

There are many lines of evidence that would identify CIDP as an autoimmune disorder although disease mechanisms remain poorly understood.¹⁴⁵ In particular, the antigen(s) to which the immune attack is targeted and specific roles of the humoral and cellular system played in the pathogenesis of CIDP remain in large part unknown. Anatomically, the pathological changes associated with CIDP are demonstrable at the root, plexus, and nerve level.¹⁴⁶ Autoantibodies against glycolipids such as GM1 and GD1a are logical choices as responsible pathogenic agents in CIDP as these antigenic targets are located on the nodal axolemma of motor nerves. GM1 and GD1a autoantibodies are capable of interfering with ion channel function although perhaps not at an order of magnitude sufficient to disrupt impulse transmission. In addition, these gangliosides are found as well as on the cell surface of Campylobacter jejuni, the precipitating agent in many cases of the AMAN variant of GBS with which these autoantibodies are so closely associated. Furthermore, GM1 and GD1 knockout mice demonstrate that these gangliosides are essential for the integrity of node/internode junctions. Paranodal loops in these animal models do not attach to the axolemma, Na channels are disrupted, and K channels are mislocated to the paranode.¹⁴⁵ Nonetheless, ganglioside autoantibodies have not been routinely identifiable in CIDP patients.¹²¹

As previously mentioned, some CIDP patients will have antibodies directed against myelin P₀, which along with PMP-22 are transmembrane components of compact myelin. These antigens are probably more relevant to hereditary than acquired demyelinating neuropathies. Current opinion does not strongly support either a pathogenic or diagnostic role for these autoantibodies in CIDP. Recently, a small number of GBS and CIDP patients have been shown to have autoantibodies targeting novel antigens residing in nodal and paranodal regions many of which are responsible for the positioning and anchoring of ion channels in strategic locations along the axolemma.¹⁴⁵ Specifically, Na channels are anchored to spectrin of the axonal cytoskeleton via ankyrin-G and to gliomedin of the Schwann cell microvilli via Nr-CAM and neurofascin-186. The juxtaparanodal type 1 K channels are anchored by contactin-associated protein-2 (Caspr-2) and transient axonal glycoprotein-1 (TAG-1), the latter being an adhesion molecule that resides on the juxtaparanodal axolemma and apposing Schwann cell membranes (Fig. 14-3). Of apparent physiological importance is the separation of clustered Na channels in nodal regions from the clustered potassium channels in the juxtaparanodal regions which is dependent upon interactions between axonal Caspr and contactin which are connected to neurofascin-155 of paranodal myelin loop. Autoantibodies directed against contactin, neurofascin, Nr-CAM, and gliomedin have been described in small numbers of CIDP patients.^144,147,148

Failure of regulatory T-cell mechanism is thought to underlie persistent or recurrent disease, differentiating CIDP from the acute inflammatory demyelinating polyneuropathy form of GBS.^45,116 CD8+ T-cell–mediated autoimmunity has been demonstrated.¹⁴² The rapid improvement that occasionally follows PLEX or IVIg and the demonstration of immunoglobulin and complement on peripheral nerve tissues suggest a role of the humoral arm of the immune system as well.^16,149 Theoretically, autoantibodies may impair ion channel function and by doing so produce an EDX pattern suggesting rapidly reversible axon loss similar to the AMAN variant of GBS. Ion channel dysfunction may also theoretically result in EDX evidence of conduction block as well.¹⁴⁵

Physiologically, conduction block is most frequently attributed to paranodal and internodal demyelination which impairs axon potential propagation.^145,146,150 Demyelination of a nerve segment produces an increased transverse capacitance and reduced resistance in the area. This causes a leakage of current and increases the time required for the longitudinal current to reach the next node of Ranvier. If the current leakage is too excessive, current may be insufficient to depolarize the next node of Ranvier, necessary for continued impulse transmission. It is this conduction block not just the slowing of velocity, which is responsible for motor weakness. Weakness may also occur as a result of axon loss, the mechanism of which in CIDP is poorly understood.

TREATMENT

Corticosteroids, PLEX, and IVIg have been demonstrated to be beneficial in randomized controlled trials in CIDP patients, both in adults and children (Table 14-4).^{7,33,128,151–156} In one study, 66%, 59%, and 62% of individuals responded to IVIg, corticosteroids, and PLEX respectively.³³ This same therapeutic equivalence between IVIg and PLEX was demonstrated in one prospective trial.¹²⁹ Another clinical trial comparing IVIg versus prednisolone for 6 weeks demonstrated no short-term differences in efficacy.¹⁵¹ It is generally held that failure to respond to one of these three treatments does not predict failure with the other two.¹³ In one study, half of the patients who failed corticosteroids responded to IVIg.¹⁵⁷ Although CIDP should be considered a chronic disease, it is also a treatable disease with one study describing 40% of patients treated largely with IVIg and/or corticosteroids becoming independent of activities of daily living achieving a modified Rankin scale of two or less.³³ Despite treatment responsiveness, it is also important to point out that not all CIDP patients require treatment, and as always, clinical judgment of potential risk versus benefit is essential.³³ Our decision to treat is determined primarily by the neuropathy phenotype and by the extent that the patient’s comfort and function are affected by the neuropathy. Like others, we do not consider the coexistence of an MCP, particularly if IgA/IgG as a reason to withhold treatment in a patient with a typical generalized and symmetric pattern of weakness.¹⁴

Corticosteroids

Anecdotal cases and small series have long shown that corticosteroids can be beneficial in both adults and children,^{9,10,14,16,33,158} further supported in a randomized control trial of oral prednisone in patients with CIDP.¹⁵⁹ Despite this limited evidence-based support, their use is supported by a Cochrane review in view of their long-term use and general acceptance of their benefit.¹⁵⁵ When we treat CIDP with corticosteroids, prednisone is initiated at a dose of 1–1.5 mg/kg (up to 100 mg) per day for 2–4 weeks, with transition to alternate day treatment as outlined in Chapter 4 (e.g., either 100 mg QOD or 60 mg alternating with 50 mg every other day).^14,36 Patients remain on relatively high doses of prednisone until their strength is normalized or there is a clear plateau in clinical improvement. When weaning begins, the trajectory it takes is dependent on the contextual features of the individual case. Typically, we begin the taper after a month or two of the induction dose, particularly if there are either signs of improvement or the development of unwanted side effects. We slowly taper the prednisone by 5 mg every 2–3 weeks until the dose is 20 mg every other day. At that point, dosing is reduced no faster than 2.5 mg every week or 5 mg every 2 weeks. Using this method of treatment, the time of initial improvement ranges from several days to 5 months (mean 1.9 months), the time to maximum improvement averages 6.6 months, with significant improvement in strength and function appreciated in 95% of treated patients after 1 year.¹⁴

In addition to traditional oral prednisone regimens, pulsed steroids either with intravenous methylprednisolone or with oral dexamethasone have been attempted with the hope of more rapid therapeutic onset and the potential for fewer side effects.^{157,160–162} One study using pulsed dexamethasone at a dose of 40 mg daily for 4 consecutive days for a total of 6 cycles (if required) demonstrated faster improvement, longer remissions, fewer relapses, and fewer side effects in comparison to a more conventional oral prednisolone regimen.¹⁵⁷ Another study using a very similar design was unable to demonstrate any advantage over a similar pulsed oral dexamethasone regimen in comparison to a more traditional oral prednisolone.¹⁶² Further information regarding the use of corticosteroids including mechanism(s) of action and adverse effects can be found in Chapter 4.

Plasma Exchange

PLEX, although likely to be equally effective, is less likely to be used as a CIDP treatment in comparison to corticosteroids or IVIg.^156,163 In all probability, this is based on considerations of availability and convenience as well as safety concerns. It remains an effective option in patients who are refractory or have contraindications to IVIg and/or corticosteroids. Efficacy of PLEX was demonstrated in prospective, randomized, double-blinded, placebo-controlled trials using sham PLEX.^128,131 However, the response to PLEX is transient, usually lasting only a few weeks, and therefore repeated courses of PLEX must be given intermittently or other, augmentative immunomodulating treatments used. The characteristic regimen is to exchange approximately 200–250 mL/kg body weight five to six times over a 2-week period. This identifies another notable benefit of IVIg which is the ability to treat on consecutive days, potentially reducing length-of-stay in a potentially hospitalized population. PLEX is usually used in combination with prednisone in patients with severe generalized weakness as the combination may provide a more rapid response than prednisone alone. In addition, protracted maintenance treatment with PLEX is often inconvenient, impractical and not endorsed by the evidence-based guideline from the American Academy of Neurology.¹⁶³ In those patients who require maintenance exchanges, the strategy once again is to reduce the number and frequency of exchanges while maintaining an optimal, achievable functional and comfort level. Like IVIg, a therapeutic/diagnostic trial of PLEX may be useful in patients in whom the CIDP diagnosis is uncertain or in whom corticosteroids are optimally avoided, for example, those with diabetes, HIV, or those in whom a diagnosis of GBS is considered.⁹⁸ Once again, the relatively rapid response to PLEX allows an earlier determination of whether or not such patients could have an immune-responsive neuropathy. Further detail regarding PLEX and potential adverse effects may be found in Chapter 4.

Intravenous Immunoglobulin

IVIg has been demonstrated to be effective in CIDP treatment, as demonstrated in controlled clinical trials and supported by a Cochrane meta-analysis.^{131,154,157,164–166} Three of four trials comparing IVIg against placebo and another three comparing IVIg against either steroids or PLEX-demonstrated benefit.¹⁶⁷ Efficacy is estimated to be as high as 82% in one study.³³ Another observer-blinded, prospective, randomized trial found no clear difference in efficacy between IVIg and PLEX.¹²⁹ Unlike PLEX, it is endorsed by the American Academy of Neurology for both the long-term as well as short-term treatment of CIDP.¹⁶⁸ In further comparison to corticosteroids, the INCAT group demonstrated essentially no significant difference in the short-term between IVIg and prednisolone during a 6-week period.¹⁵¹ In addition, retrospective studies have demonstrated that approximately 25% of CIDP patients achieve remission and 65% clinical stability with long-term IVIg treatment.¹⁶⁹

There is a strong consensus that IVIg is the treatment of choice in CIDP for a number of reasons in addition to its efficacy.⁷ It is generally accepted that the long-term side effect profiles favor IVIg over corticosteroids. In addition, there are occasional individuals who achieve a durable remission, estimated at 55% after 24 weeks in one study, potentially obviating the need for protracted steroid treatment.¹⁷⁰ Finally, initial use of IVIg as opposed to steroids eliminates the problem by the desire to initiate rapid treatment in a patient where there is diagnostic uncertainty between CIDP and GBS.

As the effects of IVIg are often transient, repeated courses of IVIg are necessary. The treatment regimen needs to be individualized. We initiate IVIg at a dose of 2 g/kg body weight over 2–5 days monthly for at least 3 months. Subsequent dosing is dependent on clinical response with a regimen of 1 g/kg every 3 weeks demonstrated to be effective in one study.¹⁷⁰ As always, the strategy is to ascertain the minimally effective dose, typically 1–2 g/kg every 3–8 weeks based on individual response.¹⁷⁰ In patients who become refractory to IVIg, courses of PLEX may restore IVIG responsiveness.¹⁷¹

IVIg is generally well tolerated. Although it has been suggested that individuals with IgA deficiency may be at risk for anaphylaxis from IVIg,¹⁷² we agree with others that this reaction is rare and that the risk may be overstated.^167,173,174 Other adverse effects are addressed in Chapter 4.

Other Treatment Options

Interferon-β, etanercept, rituximab, cyclophosphamide, azathioprine, mycophenolate mofetil, fludarabine, alemtuzumab, methotrexate nataluzimab, and cyclosporine have all been used in the treatment of CIDP.¹⁵² None have demonstrated benefit according to a Cochrane review with the caveat that the studies may have been too small to identify a modest benefit.¹⁷⁵ In addition, as reviewed in Chapter 4 and mentioned earlier in this chapter, tumor necrosis factor alpha blockers have been associated with CIDP, prompting caution in their use in this disorder. In essentially all cases, treatments other than IVIg, PLEX, and corticosteroids are used only as secondary agents, only when an acceptable response cannot be obtained by the three primary treatments. Again details pertaining to the mechanisms of action, administration, and adverse effects of these agents can be found in Chapter 4.

Small anecdotal reports suggest a beneficial effect of azathioprine at doses of 100–300 mg/d.^{15,16,176–179} A prospective, randomized, but nonblinded 9-month study of 27 patients with CIDP failed to demonstrate a benefit of azathioprine (2 mg/kg/d) when added to prednisone.¹⁸¹ The dose of azathioprine may have been too small however (we go up to 3 mg/kg/d) and the duration of this study may have been too short for a beneficial effect to become evident. A beneficial response from azathioprine may require a longer than 9 month exposure to adequate doses.