Article Text

Download PDFPDF
Chronic inflammatory demyelinating polyradiculoneuropathy: classification and treatment options
  1. James R Overell,
  2. Hugh J Willison
  1. Department of Neurology, Institute of Neurological Sciences, Southern General Hospital, Glasgow, UK
  1. Correspondence to:
 Dr J Overell, Department of Neurology, Institute of Neurological Sciences, Southern General Hospital, 1345 Govan Road, Glasgow G51 4TF, UK

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) is an uncommon and probably underdiagnosed disorder that often results in appreciable disability. Its importance lies in its generally good response to treatment. Although it is often contextualised by neurologists as “a chronic form of Guillain-Barré syndrome” (GBS), this oversimplifies the relationship between the two disorders: there is more to the distinction than one taking longer than the other to present and develop. Because CIDP is treatable, and because treatment is complex, largely non-evidence based, and potentially toxic, it is an important disorder to understand. The clinical spectrum is broad, but clinical subclassification is worthwhile, because it informs treatment decisions—certain subtypes respond better to certain treatment options, and some subtypes don’t respond well, if at all. We will discuss the clinical presentation and suggest a simple and practical classification system as a step towards rationalising the many different immunomodulatory strategies available.


When should the clinician consider CIDP? There are three main scenarios:

  • The typical clinical picture is of a symmetrical, progressive, mixed sensory and motor neuropathy (but largely motor and often proximal), which results in significant disability (typically problems with walking) and areflexia.

  • Sometimes CIDP is diagnosed when “GBS” develops over a longer period than expected (clinical deterioration should proceed for less than four weeks in GBS and more than eight weeks in CIDP, the middle ground being classified as “SIDP”—subacute IDP). The neurophysiological criteria for CIDP have been defined and modified by numerous authors and are discussed in detail elsewhere,1,2 but essentially require a combination of motor conduction block, reduced motor and sensory conduction velocities with temporal dispersion, prolonged distal motor latencies, and prolonged F wave latencies.

  • A third “route in” to the diagnosis is finding demyelinating neurophysiology in a patient with a paralytic syndrome that hadn’t been clearly classified through its clinical features.

Although this is the reality of clinical practice, attempts to classify “CIDP” clinically start with the population eventually diagnosed as having CIDP, and aim to categorise the phenotypes within that population.3,4 Other classification schemes include categorisation according to whether the neuropathy is associated with another disease (for example, myeloma), a serum protein (for example, a monoclonal IgG or IgM band), or a particular antibody reactivity (for example, a specific anti-ganglioside antibody). The advantages of clinical classification systems are their practicality, their inclusiveness, and the fact that clinical phenotypes can usefully predict clinical course and response to treatment. The balance between oversimplicity and unwieldly complexity is as important in CIDP as in many other areas of neurology. One simple classification scheme proposed by Saperstein et al divides patients into four clinical categories, leading to four different dysimmune neuropathy syndrome diagnoses:3

  1. Symmetrical proximal and distal weakness—that is, CIDP;

  2. Symmetrical distal weakness or sensory loss—that is, distal acquired demyelinating symmetrical (DADS) neuropathy;

  3. Asymmetrical weakness—that is, multifocal motor neuropathy (MMN);

  4. Asymmetrical weakness and sensory loss ie multifocal acquired demyelinating sensory and motor (MADSAM) neuropathy.

Clinical and laboratory features that distinguish each of these syndromes from each other are detailed in table 1. The common symmetrical presentation of proximal and distal weakness and sensory loss (what most people would call CIDP) is still called CIDP, and it is responsive to both steroids and intravenous immunoglobulin (IVIg). Large fibre sensory dominant cases with or without distal motor features (and with prolonged distal motor latencies on neurophysiology) are commonly seen in association with a monoclonal IgM band and with serum antibodies against myelin associated glycoprotein (MAG). These are classified clinically as “DADS neuropathy” (distal acquired demyelinating symmetrical neuropathy).5 Such neuropathies respond poorly to immunomodulatory therapy (but that doesn’t mean it shouldn’t ever be tried). Asymmetrical weakness, normally beginning in the upper limbs, and with multifocal conduction block on motor nerve conduction studies leads to a diagnosis of multifocal motor neuropathy (MMN). This is often associated with serum anti-GM1 antibodies, and responds specifically to IVIg, but not steroids. Finally multifocal acquired demyelinating sensory and motor neuropathy (MADSAM), which many neurologists refer to as Lewis-Sumner syndrome, is similarly asymmetrical and upper limb dominant, but mixed sensorimotor. It responds to both steroids and IVIg.


A comparison of the clinical and investigation findings in the different clinical types of chronic inflammatory demyelinating polyradiculoneuropathy (adapted from Saperstein et al3)

A potential difficulty with this scheme lies in its oversimplicity as there are a number of neuropathy syndromes that fall within the CIDP spectrum that do not fit neatly into this classification. Purely axonal variants, purely sensory variants, and purely motor symmetrical variants are described and are not easily incorporated. Some authors classify sensory ataxic neuropathies (often associated with antibodies to disialylated gangliosides) alongside CIDP.4 Many draw distinctions between chronic worsening symptoms, and subacute relapsing symptoms. Others emphasise the importance of systemic disease association (see below) or the presence of a monoclonal band, and classify such neuropathies separately. Unfortunately no classification scheme is without its problems, leading to confusingly different approaches across centres.

Many systemic illnesses can be associated with inflammatory demyelinating polyneuropathies. These are listed in table 2 and should be considered in all patients, screened for where possible, and actively sought in suggestive clinical contexts. An algorithm based on a clinical classification approach should include such a process, as illustrated in figure 1. Patients with paraproteins should be referred to a haematologist for assessment of any haematological malignancy, and in patients with monoclonal gammopathy of undetermined significance (MGUS), yearly electrophoresis should be performed with a close check kept on the level of the monoclonal protein (>30g/l is highly suggestive of malignancy). Nerve biopsy is rarely indicated, unless vasculitis is strongly suspected with an asymmetrical sensorimotor presentation; it is seldom helpful for clinical management.


Systemic disorders associated with CIDP

  Figure 1
Figure 1

A diagnostic and initial treatment trial scheme for chronic inflammatory demyelinating polyradiculoneuropathy (CIDP). DADS, distal acquired demyelinating symmetric neuropathy; MMN, multifocal motor neuropathy; MADSAM, multifocal acquired demyelinating sensory and motor neuropathy; IVIg, intravenous immunoglobulin.


There is a vast array of treatment options for chronic dysimmune neuropathies. These therapies are directed at blocking immune processes to arrest inflammation and demyelination, thereby improving conduction block and creating an environment conducive to neural repair. Many therapies have been summarised in a series of excellent Cochrane systematic reviews. Steroid responsiveness is well documented and was confirmed by the only randomised trial of prednisolone.6 The Cochrane review concurred with the generally held view that steroids are effective, while drawing attention to the lack of good quality randomised data.7 Plasma exchange has been shown to be effective in two good quality trials, summarised in another Cochrane review.8 Yet another Cochrane review has confirmed the treatment effect of IVIg, which is generally seen within two weeks.9

The question for most clinicians is which of these therapies to choose in any individual patient, and in what order. Small, short term comparative trials demonstrate equality of IVIg and plasma exchange,10 and equality of IVIg and prednisolone.11 Combinations of plasma exchange, IVIg, and prednisolone have not been systematically studied. Of course prednisolone is associated with significant adverse effects both in the elderly, and in the young CIDP population who may need treatment for many years. Plasma exchange is inconvenient for patients, especially if monthly or even more frequent exchanges are required. IVIg is limited in availability and is costly.

Starting with intravenous immunoglobulin

In patients with significant disability requiring therapy we generally opt for IVIg in the first instance. An advantage of this approach is that a clear outcome—positive or negative—is readily evident within two weeks. The central tenet of starting therapy in CIDP is to be objective about the outcome; many people feel a “lift” from steroids and from IVIg, but are they actually stronger? Objectivity is improved by using impairment and disability scales, and useful suggestions have been published recently.13 Subjective improvements in fatigue or endurance are harder to assess, but equally significant for the patient. The neuropathy monitoring sheets that we use are shown in the online chart (see

We aim to start IVIg at 2.0g/kg and assess the effect objectively at two weeks. Thereafter, if a decision to continue IVIg is made, the dose and dose frequency can be adjusted to the lowest possible that achieves and maintains adequate clinical improvement. Demonstrating efficacy with IVIg does not necessarily commit the physician or patient to continued use, but may simply serve the useful purpose of confirming the presence of a treatment responsive neuropathy.


We start prednisolone as a daily dose (about 1mg/kg/day) and hold this initial dose for a month. We then taper the daily dose by 10 mg per month until the patient is taking 40 mg per day. At this point we switch to alternate day dosing, reducing the dose on alternate days by 10 mg every two weeks to 40 mg alternate days. After that we slow the reduction to 10 mg alternate days per month to 20 mg alternate days, then 5 mg alternate days per month until as low a dose as possible has been achieved. We have concerns about using alternate day prednisolone as initial therapy, though there is a clear rationale for alternate day dosing in the longer term in the hope that it may minimise adverse effects.12 Gastric and osteoporosis prophylaxis should be considered before therapy is started. Patient held steroid dosing sheets are very helpful in the explanation of alternate day and changing regimes (see online appendix A at


An oral immunosuppressant can be used in combination with IVIg, prednisolone, or plasma exchange. The risks and benefits have to be balanced in each individual. Generally, patients that respond poorly to the primary therapies, or who relapse early, are considered for long term (1–2 years) immunosuppression. The different immunosuppressant or immunomodulatory therapies that are used in CIDP are summarised in table 3. Unfortunately, none is supported by good randomised evidence, and all are potentially toxic. A Cochrane review has summarised the current position by the all too familiar conclusion that there is inadequate evidence to support any particular strategy.14 We have had apparently good results in individual patients with azathioprine (see online appendix B), methotrexate (see online appendix C), and cyclosporine (see online appendix D), and deliberate about the order in which they should be tried in individual patients. Azathioprine is usually well tolerated and is familiar, but is slower to take effect than cyclosporin or methotrexate. Methotrexate (given orally once per week) is widely used by rheumatologists but is less familiar to neurologists; it is shortly to be studied in a Europe-wide randomised trial, with the dose of IVIg or prednisolone required in treated versus non-treated patients being used as the outcome measure (R Hughes, personal communication). The merits and risks of all immunosuppressive drugs need to be discussed with patients, and we use detailed information sheets and patient held monitoring forms (for blood results and dosage instructions) that simplify the process rather than complicate it (see online appendices B to D).


Immunosuppressant drugs used in CIDP spectrum disorders: what they do, and the evidence for each

In patients with significant disability requiring therapy we generally opt for IVIg in the first instance

It is important for both the physician and patient to “know when to stop” if there is no evidence of improvement, to avoid subjecting the patient to toxic long term therapies

Multifocal motor neuropathy

It is reported that CIDP with only motor involvement and MMN can be worsened by steroids, and that steroids should be avoided in such patients.15 On the other hand IVIg is often dramatically and rapidly effective. A Cochrane review in the MMN subgroup has concluded that the limited evidence available shows that IVIg has a beneficial effect on strength and suggests trends in improvement in disability.16 Thus repeated IVIg courses remain the mainstay of treatment for MMN. Although there are reports from small series that cyclophosphamide is effective, the relatively low level of disability in MMN, the very high toxicity and cumulative lifetime dose limits of cyclophosphamide, and the fact that most MMN patients are IVIg responsive mean that cyclophosphamide is best avoided in all but the severest situations.


Neuropathies associated with a paraprotein are usually treated along the same lines as described for “idiopathic” CIDP. The best treatment for CIDP associated with an IgG or IgA monoclonal band is unknown—anecdotally patients respond less well to therapy than idiopathic CIDP, which may in large part be due to the common axonal involvement. Gorson reported improvement in eight of 20 patients with IgG MGUS and CIDP with IVIg: demyelinating features on neurophysiology were associated with a response.17 IgG or IgA paraprotein associated neuropathies may respond better to plasmapheresis than IgM paraprotein associated neuropathies.18 A Cochrane review on IgG and IgA associated neuropathies is in progress.19 Generally patients with IgM monoclonal proteins with or without anti-MAG activity respond poorly to treatment with steroids, IVIg, and plasma exchange, and many patients end up receiving no long term treatment after completing several failed trials of therapy. In patients in whom a trial of therapy seems justified on clinical disability grounds, we currently use IVIg in the first instance—a recent Cochrane review in this subgroup concluded that IVIg is associated with short term benefits.20 Current hopes rest on the use of newer agents including Rituximab (a monoclonal anti-CD20 antibody that specifically targets B cells) which has been reported to improve such patients in small series,21,22 and which is currently being tested in a randomised trial.

A common debate between neurologists and haematologists is one or other party suggesting that a paraproteinaemia or a paraprotein associated neuropathy is of insufficient severity to be treated. These cases need careful thought, based on the clinical features and the patient’s therapeutic enthusiasm and expectations. Haematologists may correctly argue that the B cell disorder (for example, a low level MGUS or chronic lymphatic leukaemia) does not warrant treatment in its own right; however this should be overturned in the presence of a severe and worsening neuropathy. B cell directed therapy (such as chemotherapy or Rituximab, administered by the haematologist), either in addition to or in place of standard neuropathy treatment with IVIG or steroids, may be the most appropriate approach in these circumstances.

Managing borderline cases or treatment failures

The spectrum of different presentations and treatment options that we have summarised reflects a diverse disease and only partially successful strategies to treat it. Not all patients with a particular CIDP subtype will respond in the same way, and lack of response to one particular treatment strategy (for example, IVIg) does not mean that others (for example, prednisolone or plasmapheresis) will fail. This may reflect differences in the underlying disease in responders to different treatments that have not yet been fully characterised. Decision making is further confused by the fact that many clinical cases are borderline (for example, soft demyelinating features in a patient with a clinical axonopathy) or mixed (for example, axonopathy in a diabetic with some evidence of conduction block and a mildly raised CSF protein). The undoubted placebo effect of all of the treatment options also needs to be considered.

All of these issues make it particularly important that throughout the management process the clinician remains aware of his or her confidence in the diagnosis, the severity of the patient’s disability, and the objective evidence of improvement after each active therapy. If diagnostic confidence is strong, a broad range of treatment trials is indicated. Generally in mixed and borderline cases whose clinical severity warrants a treatment trial we use IVIg in the first instance, with as objective an assessment of improvement as possible at two weeks. If the patient is IVIg unresponsive then our enthusiasm for prednisolone or plasmapheresis is lower, but we might still consider prednisolone depending on the context. It is important for both the physician and patient to “know when to stop” if there is no evidence of improvement, to avoid subjecting the patient to toxic long term therapies.


  • The quality of life of patients with CIDP and related variants can improve markedly on appropriate treatment, making its recognition vital.

  • However, it is heterogeneous in both its clinical presentation and response to therapy, and this heterogeneity is enhanced by its association with numerous other diseases and syndromes. A simple clinical classification system is a useful starting point, but must be combined with a full immunological and haematological investigational screen, and a thorough search for associated diseases.

  • Whichever of the currently available treatment strategies is adopted, objective assessment of the effect of therapy will inform correct subsequent treatment decisions.

  • The future lies in the recognition of more specific immunomodulatory therapies that will hopefully lessen the requirements for multiple repeated therapies and attendances, and prove less toxic than currently available immunosuppressants.


The authors thank Professor Richard Hughes and Dr Mary Reilly for advice on neuropathy monitoring sheets. The immunosuppressant patient advice sheets published by the Arthritis and Rheumatism Council of Great Britain provided the basis for our immunosuppressant monitoring sheets.

This article was reviewed by Dr Jane Pritchard, London, UK


Supplementary materials

  • Files in this Data Supplement:


Other content recommended for you