Treatments for chronic inflammatory demyelinating polyradiculoneuropathy (CIDP): an overview of systematic reviews

  • Protocol
  • Overview



This is the protocol for a review and there is no abstract. The objectives are as follows:

To summarise the evidence from Cochrane and non-Cochrane systematic reviews of all treatments for CIDP. If possible we will compare the effectiveness of agents with similar rationale for use in CIDP.


Description of the condition

Chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) is a chronic progressive or relapsing and remitting disease, usually causing weakness and loss of sensation in the limbs (Vallat 2010). It is caused by inflammation of the peripheral nervous system resulting in damage to the myelin sheaths which normally insulate nerve fibres. This produces a state of 'demyelination' of the nerves, which slows and may block nerve conduction. Although demyelination is the predominant pathology, the inflammation also damages the central conducting core of the nerve fibre, called the axon, resulting in axonal degeneration. The parts of the peripheral nervous system most affected are the motor and sensory spinal nerve roots and the peripheral nerves. The cranial nerves controlling eye movements and the facial, swallowing and speech muscles may also be affected. The nerves supplying the breathing muscles and the autonomic nerves which control the bladder, bowel and circulation are usually spared.

There is no single diagnostic test for CIDP. The diagnosis is made from the clinical picture, nerve conduction tests (which show evidence of demyelination), and the exclusion of other causes. Sometimes supportive tests, such as lumbar puncture and cerebrospinal fluid analysis, magnetic resonance imaging of spinal nerve roots, nerve biopsy, and therapeutic trials of immunomodulating agents, are used to help in diagnosis. Many different sets of diagnostic criteria have been proposed. Very strict research criteria require a characteristic nerve biopsy for a definite diagnosis (Ad hoc 1991). Less strict but widely accepted recent criteria rely on clinical history, examination, nerve conduction evidence and exclusion of other causes for a definite diagnosis (Van den Bergh 2010). In addition to weakness and sensory symptoms, CIDP may cause neuropathic pain and fatigue. Neuropathic pain differs from ordinary pain in that it arises spontaneously and is due to dysfunction of the nerves and not tissue injury.

The prevalence of CIDP has been between 1 and 9 per 100,000 population in different studies, with most reporting a prevalence of 2 to 3 per 100,000 (Mahdi-Rogers 2010b). It was between 1.4 and 4.7 times more common in men than women in eight population-based studies in which the sex ratio was reported. The average age of onset in four population-based studies in which the figure was given was 48 to 58 years (Mahdi-Rogers 2010b). It is uncommon in children but becomes more common with advancing age, reaching a peak prevalence in the eighth decade. In a population based study, 9 of 62 (14.5%) patients had a progressive, 44 (71%) a relapsing and remitting, and 9 (14.5%) a monophasic disease course (Mahdi-Rogers 2010b). According to Vallat 2010, 7 to 50% of patients have a monophasic or progressive course, and 20 to 35% a relapsing remitting course in different series.

CIDP is important because it can be severely disabling, although its severity is variable. On the prevalence day in the population based study (Mahdi-Rogers 2010b), 28 (68.2%) patients could walk independently, 10 (24.4%) required unilateral and three (7.3%) bilateral support to walk 10 metres. No patient needed a wheelchair. At nadir, 31 (75.6%) patients had disability in their upper limbs, 17 (41.5%) could walk independently, 11 (26.8%) needed unilateral support, 6 needed bilateral support and 7 patients used a wheelchair. Most commonly the disease causes progressive weakness leading to the need for aids to walk followed by improvement with treatment. However, treatment may need to be repeated or prolonged. In the population based study mentioned 64 (76.2%) of 84 patients had required treatment and 51 (79.7%) improved with at least one of the main line treatments, corticosteroids, intravenous immunoglobulin or plasma exchange. Less commonly the disease is so mild that treatment is not necessary. CIDP is also important because its treatment is expensive, especially when intravenous immunoglobulin (IVIg) is used. According to Blackhouse 2010 the cost per quality-adjusted life year (QALY) of using IVIg rather than corticosteroids was 696,598 USD.

The cause of CIDP is not known. Active disease sites in the spinal nerve roots and nerve trunks show inflammation and stripping of the myelin sheaths from the axons by macrophages which can be seen with the microscope in nerve biopsies or at post mortem. The inflammation is probably due to an autoimmune reaction. There is debate whether this is due to antibodies, to T cells directed against the Schwann cell or myelin, or both. There is some evidence that there is impairment of the regulatory T and B cells which normally control autoimmune responses (Vallat 2010).

Description of the interventions

Treatments aimed at the underlying disease

The most common first-line treatments for CIDP are corticosteroids, given as tablets every day or as intravenous infusions every four weeks, or immunoglobulin, usually given as intravenous infusions over two to five days for the first dose and then over one day every two to eight weeks for follow-up courses. Corticosteroids are anti-inflammatory drugs used in many types of inflammatory conditions such as asthma and arthritis. They are quite cheap but their long-term use risks serious side-effects including high blood pressure, diabetes mellitus, obesity, thinning of the bones, and cataracts (Mehndiratta 2002). Immunoglobulin is extracted from the plasma of several thousand blood donors and then purified to reduce the risk of transmitting virus and other infections. It is very expensive and often causes minor short-term side-effects such as headache and muscle aching; serious side-effects such as stroke, severe skin reactions and kidney failure do occur but are rare (Eftimov 2009).

Plasma exchange is an alternative first-line treatment. It involves drawing blood from the patient, separating the plasma from the cellular fraction and replacing the plasma with a plasma substitute. This is done with an automated machine which is able to replace the whole plasma volume in a few hours. A course of about five treatments on consecutive or alternate days is commonly used to initiate treatment, but regimes vary, and single exchanges every few weeks are sometimes used for long-term treatment. Plasma exchange is usually safe, although there are rare side-effects, such as bleeding and injuries arising from inserting large tubes into veins. Its main drawback is its inconvenience (Mehndiratta 2004).

When first-line treatments are inadequate, neurologists often prescribe immunosuppressant agents and in the past immunomodulatory agents have also been tried. The most commonly used immunosuppressants have been azathioprine, cyclophosphamide, ciclosporin and methotrexate. These can be given orally, although cyclophosphamide can also be given by intravenous infusion and methotrexate by intramuscular injection. They all reduce the white cell count and carry an increased risk of infection. Individual agents can have idiosyncratic side-effects. For instance, amongst others, azathioprine can cause hypersensitivity reactions, cyclophosphamide hair loss, ciclosporin kidney failure and methotrexate liver dysfunction. More recently-developed immunosuppressant drugs affect relatively specific components of the immune system and are beginning to be used. These include rituximab, which deletes most of the circulating B cells, and fingolimod, which prevents T cells from leaving the lymph nodes. In addition to immunosuppressant agents, an immunomodulatory agent, interferon beta, has been tested in randomised controlled trials for CIDP. Interferon beta is a naturally occurring molecule called a chemokine which can be manufactured commercially and became the first licensed treatment for multiple sclerosis. It is given by subcutaneous or intramuscular injection on one or several days a week. The potential risks of each treatment have to be balanced against the possible benefits (Mahdi-Rogers 2010a).

Symptomatic treatments

In addition to considering treatments aimed at the underlying disease process, this review will also provide an overview of reviews of treatments for symptoms of CIDP including those for neuropathic pain and fatigue. Trials and Cochrane reviews of such treatments generically for peripheral neuropathy do exist, but this review will focus on those reviews and trials which deal specifically with CIDP. Drugs which have been tested in trials for neuropathic pain include the tricyclic antidepressant amitriptyline (Saarto 2007) and anti-epileptic drugs such as gabapentin and pregabalin. These drugs are all given daily as tablets to try to reduce the pain. There are no trials known to us of drugs for fatigue caused by CIDP but there is a Cochrane review about exercise for fatigue in peripheral neuropathy (White 2004).

How the intervention might work

Corticosteroids, IVIg and broad spectrum immunosuppressive agents are expected to treat CIDP by inhibiting multiple components of the presumed inflammatory autoimmune response. The precise ways in which they do this in CIDP are incompletely understood, in part because the mechanisms causing nerve damage in CIDP are themselves not understood (Vallat 2010). The ways in which immunosuppressant drugs work vary from drug to drug, and will have been described or referenced in the constituent reviews. The most obvious way in which plasma exchange might work is by removing antibodies. However, it also removes other soluble plasma substances, including the small molecules called chemokines which might affect T cell function. Similarly, the most obvious action of rituximab, which engages the CD20 molecule on B cells, is to reduce the circulating B cell population, but this may in turn interfere with antigen presentation and so affect T cell function (Maloney 2002). Interferon beta has multiple actions on the immune system which tend to down-regulate harmful immune responses (Kieseier 2011). Exercise might help fatigue in CIDP by a combination of improving aerobic fitness, strengthening relevant muscle groups and having positive psychological effects.

Why it is important to do this overview

Treatment of the underlying disease in CIDP is covered by four Cochrane reviews on corticosteroids (Mehndiratta 2002), IVIg (Eftimov 2009), plasma exchange (Mehndiratta 2004) and other immunosuppressive and immunoregulatory drugs (Mahdi-Rogers 2010a). The treatment of CIDP is or might potentially be included in other Cochrane and non-Cochrane systematic reviews. There are several completed Cochrane reviews and some review protocols for treatment of neuropathic pain, for instance (Saarto 2007). There is one Cochrane review protocol of treatment for fatigue in peripheral neuropathy (White 2009), and a review of exercise for peripheral neuropathy (White 2004). Drawing together the reviews in one place will make their combined information more accessible to people with CIDP, healthcare professionals and researchers. An overview will allow some indirect comparisons of treatments which have not yet been directly compared in trials. It will be likely to draw attention to, and be a stimulus for, more research.


To summarise the evidence from Cochrane and non-Cochrane systematic reviews of all treatments for CIDP. If possible we will compare the effectiveness of agents with similar rationale for use in CIDP.


Criteria for considering reviews for inclusion

Types of reviews

In accordance with the advice on overviews of reviews in the Cochrane Handbook for Systematic Reviews of Interventions (Becker 2011), we will consider all Cochrane and non-Cochrane systematic reviews of randomised controlled trials for all treatments for CIDP. We will not include non-systematic reviews. Systematic reviews will be defined according to the Cochrane Handbook as those having:

  1. pre-defined objectives

  2. pre-defined criteria for eligibility of evidence

  3. an objective systematic search for evidence applying predetermined inclusion and exclusion criteria, and

  4. explicit and systematic methods for synthesising evidence which attempt to reduce bias

If we found a randomised controlled trial of another therapy for which a plausible rationale or empirical basis existed which had not been included in a Cochrane systematic review, we would note the existence of the trial in our overview, and draw the attention of the Cochrane Neuromuscular Disease Group to the existence of such a trial so that it could be incorporated into an existing or new review. We would not include it in our overview until the appropriate systematic intervention review was complete.

Types of participants

We will include reviews of all forms of CIDP approximating to the definite, probable and possible diagnostic criteria in the European Federation of Neurological Societies/Peripheral Nerve Society (EFNS/PNS) guideline (Van den Bergh 2010). Different reviews and different trials are likely to have used slightly different criteria and these differences will be noted. There is some evidence that patients who have CIDP associated with paraproteins have a different natural history and response to treatment than those without paraproteins. We will include those with IgG or IgA paraproteins, provided that they fulfil the clinical and electrophysiological criteria of the EFNS/PNS guideline for CIDP. Such patients have been considered in a separate Cochrane review (Allen 2007). We will exclude those with IgM paraproteins because the neuropathy of most such patients has a different pathogenesis caused by antibodies to myelin-associated glycoprotein, and responds differently to treatment compared with CIDP (Lunn 2012). We will exclude patients who have a CIDP-like illness associated with malignancy.

Some randomised controlled trials and reviews have included a minority of participants who have IgM paraproteins or malignancy. We will try to obtain data which exclude the results from such participants. If this is not possible and less than 5% of participants in a trial have either of these exclusion criteria, we will include the trial. If there are more than 5% of participants we will exclude the trial from our primary analysis but include it in a sensitivity analysis.

Some reviews may have included trials involving participants with other forms of peripheral nerve disease than CIDP. We will try to obtain the results of participants with CIDP alone. If this is not possible, and more than 95% of participants fulfil the inclusion criteria for this overview, we will include the trial. If there are less than 95% of participants with CIDP we will exclude the trial from our primary analysis but include it in a sensitivity analysis.

Types of interventions

We will include all interventions for CIDP, whether pharmacological or physical. We will consider treatments both of the underlying disease process and of symptoms which occur in CIDP such as pain and fatigue.

Types of outcomes

In the narrative part of our overview we will report the outcomes reported in the individual Cochrane systematic reviews. We will give priority to change in disability after 12 months as the primary efficacy outcome. Change after six months will be accepted as an alternative. Although of less interest we will also report short-term outcomes after periods of at least four weeks but less than six months, since we are aware that many trials have only reported outcomes at these times. Scales used must have been validated as having good reproducibility, face validity and correlation with other scales measuring the same attribute. Recent studies have used the Inflammatory Neuropathy Cause and Treatment (INCAT) scale (Hughes 2001), Overall Disability Sum Score (Merkies 2002), and Overall Neuropathy Limitations Scale (ONLS) (Graham 2006). Earlier studies have used the Modified Rankin Scale (Bamford 1989) and Guillain-Barré Syndrome Scale (Hughes 1978).

We also wish to report as secondary efficacy outcomes change in impairment scores and quality of life. Scales used must have been validated as having good reproducibility, face validity and correlation with other scales measuring the same attribute. We anticipate that most studies will have included the MRC sum score (Kleyweg 1991) and a sensory sum score (Merkies 2000).

We will report serious adverse events, being those which require or prolong hospitalisation or are fatal at any time during treatment or within a biologically plausible time after cessation. They will have been collected for all trials and attempts made to report them in the relevant included systematic reviews. We will also report adverse events which, although not serious, would be likely to influence the decision to prescribe the treatment being tested. Examples are diabetes mellitus (relevant for corticosteroids), skin rash (IVIg) and abnormal liver function (methotrexate).

Search methods for identification of reviews

We will search the Cochrane Database of Systematic Reviews (CDSR) and the Database of Abstracts of Reviews of Effects (DARE) (current issues in The Cochrane Library), MEDLINE (January 1966 to current), EMBASE (January 1980 to current) and CINAHL Plus (January 1937 to current issue) for systematic reviews of CIDP. We will supplement the searches for RCTs in the existing Cochrane systematic reviews by searching for RCTs of any treatment of CIDP, including those with CIDP and fatigue or pain, in the Cochrane Neuromuscular Disease Group Specialized Register, the Cochrane Central Register of Controlled Trials (CENTRAL, current issue in The Cochrane Library), MEDLINE (January 1966 to current), EMBASE (January 1980 to current) and CINAHL Plus (January 1937 to current).

The search strategy for MEDLINE is shown in Appendix 1.

Data collection and analysis

Selection of reviews

Two overview authors will select systematic reviews for inclusion and will resolve disagreements by reference to a third author.

Data extraction and management

Two overview authors will independently collect data from published systematic reviews with a data collection form designed specifically to include all the data needed. We will contact the review authors or extract data from the relevant trials ourselves if information is lacking.

Assessment of methodological quality of included reviews

Two overview authors, who were not authors of the original included reviews, will independently assess the methodological quality of each review included in the overview. For this purpose we will use the assessment of multiple systematic reviews (AMSTAR) tool developed by Shea 2007 and shown to have acceptable inter-rater agreement, construct validity and feasibility (Shea 2009). Two overview authors will also independently assess the quality of the evidence in the included reviews with the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. This has become the method preferred by the Cochrane Collaboration for evaluating the quality of evidence, and by many guideline bodies for assessing the strength of recommendations (GRADE 2008; Guyatt 2008). In both cases, disagreements will be resolved by discussion, if necessary with a third author.

Data synthesis

We anticipate that the principal method for presenting data from the constituent reviews will be a narrative review. We will report the evidence for each intervention from each review and its strength estimated using the GRADE approach. Where more than one eligible review of a particular intervention is found and the conclusions agree we will report this; where the conclusions differ we will explore the reasons for any difference in relation to the AMSTAR scores of the included reviews.

Where comparable data and outcomes exist for different treatments, we will perform indirect comparisons between them using a network analysis and multiple treatments meta-analysis if appropriate (White 2011). We will judge formal multiple-treatments meta-analysis to be appropriate if the same outcome has been used in the meta-analyses included and if the trial populations are broadly comparable in age, gender, diagnostic criteria, disease duration and disease severity. We will accept limited variation in these parameters, provided that there is some overlap. For example if all participants in one meta-analysis had 'mild' disease and those in another all moderate', then we would not combine them formally. However if the mixtures were something like 30:70 in one meta-analysis and 70:30 in another, then we would combine them but express the need for caution in interpretation. To be considered comparable the outcomes should have been measured in clinically equivalent ways in each meta-analysis over clinically similar periods of follow-up.  

The method of White 2011 places multiple-treatment meta-analysis within the wider context of multivariate meta-analysis. Our approach will parallel that described by White for his analysis combining data from 24 randomised controlled trials assessing 4 different interventions aiming to promote smoking cessation.  We will first fit so-called 'consistency' models that assume no design by treatment interactions and use these to estimate pairwise treatment differences (with 95% confidence intervals) and between-studies heterogeneity. We will then fit 'inconsistency models' to assess the level of evidence for design by treatment interactions. If there is evidence of such interactions then we will seek to explain these using any relevant factors that differ between studies. As an example, consider three treatments A, B and C for which trials that directly compare A with B favour A, and trials that directly compare A with C favour C, but trials that compare B with C favour B. The data are contradictory unless one can identify a factor that explains the apparent discrepancy. For instance, C might be the best treatment in 'severe' cases, but the worst in 'mild, cases, and the comparisons of A with C were done predominantly in severe cases, but those comparing B with C have been done predominantly in mild cases. If there are unexplainable design by treatment interactions, then interpretation will be suitably cautious.

Such an analysis comparing corticosteroids and IVIg will be of special interest and is likely to be possible.


We thank the GBS Support Group UK for financial support, funded using money raised in memory of Candice Marie Roberts. The Cochrane Neuromuscular Centre is supported by the MRC Centre for Neuromuscular Diseases.


Appendix 1. MEDLINE (OvidSP) search strategy

Database: Ovid MEDLINE(R) <1946 to October Week 4 2012>
Search Strategy:
1 randomized controlled (340101)
2 controlled clinical (85462)
3 randomized.ab. (242742)
4 placebo.ab. (135753)
5 drug therapy.fs. (1580100)
6 randomly.ab. (174151)
7 trial.ab. (251434)
8 groups.ab. (1138496)
9 or/1-8 (2943095)
10 exp animals/ not (3799729)
11 9 not 10 (2500234)
12 Polyradiculoneuropathy, Chronic Inflammatory Demyelinating/ (822)
13 ((chronic adj3 inflammatory adj3 demyelinating adj3 polyradiculoneuropathy) or (chronic adj3 inflammatory adj3 demyelinating adj3 polyneuropathy) or cidp).mp. (1696)
14 inflammatory (2804)
15 (polyradiculoneuropath$3 or polyneuropath$3).tw. (10441)
16 (polyneuritis or polyradiculoneuritis).tw. (1648)
17 polyneuropathies/ or Polyradiculoneuropathy/ (7507)
18 or/15-17 (16077)
19 chronic (221872)
20 14 and 18 and 19 (324)
21 or/12-13,20 (1709)
22 11 and 21 (501)
23 exp pain/ or (pain or painful).mp. (496384)
24 23 and (neuropath$.mp. or exp peripheral nervous system diseases/) (33150)
25 (57294)
26 24 or 25 (90119)
27 22 and 26 (35)
28 meta-analysis/ (37222)
29 (37222)
30 (meta analy$ or metaanaly$ or meta?analy$).tw. (44044)
31 ((health technology adj5 assessment) or hta).tw. (1902)
32 (systematic adj3 review$).mp. (36881)
33 (systematic adj3 overview$).mp. (579)
34 consensus development (8270)
35 practice (17335)
36 or/28-35 (105127)
37 21 and 36 (25)
38 27 or 37 (60)

Contributions of authors

RACH prepared the first draft of the protocol. All authors commented on and edited the first and subsequent drafts, and approved the final version.

Declarations of interest

RACH was one of the authors of two of the reviews likely to be included in this overview, and of two of the trials in those reviews. RACH has been paid for consultancies by Baxter, CSL Behring, Grifols/Talecris, LFB, Octapharma, companies which manufacture human immunoglobulin (an intervention included in this overview) and also BMS and Novartis. RACH is also a Member of the Medical Advisory Board of GBS CIDP Foundation International and Medical Patron of GBS Support Group UK.

IVS was one of the authors of a review likely to be included in this overview and two of the trials in those reviews. IVS has received departmental honoraria for serving on scientific advisory boards and a steering committee for CSL Behring.

MPTL has been paid for consultancies by Baxter, CSL Behring and LFB and has received travel support from Grifols, all manufacturers of IVIg.

CF has been paid for service on a Data Monitoring Committee by CSL-Behring.

Sources of support

Internal sources

  • National Institute for Health Research, Cochrane Review Group Infrastructure Award, UK.

    Financial support

External sources

  • GBS Support Group, UK.

    Financial support