Immunoglobulin for multifocal motor neuropathy

Abstract

Background

Multifocal motor neuropathy (MMN) is a rare, probably immune‐mediated disorder characterised by slowly progressive, asymmetric, distal weakness of one or more limbs with no objective loss of sensation. It may cause prolonged periods of disability. Treatment options for MMN are few. People with MMN do not usually respond to steroids or plasma exchange. Uncontrolled studies have suggested a beneficial effect of intravenous immunoglobulin (IVIg). This is an update of a Cochrane Review first published in 2005, with an amendment in 2007. We updated the review to incorporate new evidence.

Objectives

To assess the efficacy and safety of intravenous and subcutaneous immunoglobulin in people with MMN.

Search methods

We searched the following databases on 20 April 2021: the Cochrane Neuromuscular Specialised Register, CENTRAL, MEDLINE, Embase, ClinicalTrials.gov, and WHO ICTRP for randomised controlled trials (RCTs) and quasi‐RCTs, and checked the reference lists of included studies.

Selection criteria

We considered RCTs and quasi‐RCTs examining the effects of any dose of IVIg and subcutaneous immunoglobulin (SCIg) in people with definite or probable MMN for inclusion in the review. Eligible studies had to have measured at least one of the following outcomes: disability, muscle strength, or electrophysiological conduction block. We used studies that reported the frequency of adverse effects to assess safety.

Data collection and analysis

Two review authors independently reviewed the literature searches to identify potentially relevant trials, assessed risk of bias of included studies, and extracted data. We followed standard Cochrane methodology.

Main results

Six cross‐over RCTs including a total of 90 participants were suitable for inclusion in the review. Five RCTs compared IVIg to placebo, and one compared IVIg to SCIg. Four of the trials comparing IVIg versus placebo involved IVIg‐naive participants (induction treatment). In the other two trials, participants were known IVIg responders receiving maintencance IVIg at baseline and were then randomised to maintenance treatment with IVIg or placebo in one trial, and IVIg or SCIg in the other. Risk of bias was variable in the included studies, with three studies at high risk of bias in at least one risk of bias domain.

IVIg versus placebo (induction treatment): three RCTs including IVIg‐naive participants reported a disability measure. Disability improved in seven out of 18 (39%) participants after IVIg treatment and in two out of 18 (11%) participants after placebo (risk ratio (RR) 3.00, 95% confidence interval (CI) 0.89 to 10.12; 3 RCTs, 18 participants; low‐certainty evidence). The proportion of participants with an improvement in disability at 12 months was not reported. Strength improved in 21 out of 27 (78%) IVIg‐naive participants treated with IVIg and one out of 27 (4%) participants who received placebo (RR 11.00, 95% CI 2.86 to 42.25; 3 RCTs, 27 participants; low‐certainty evidence). IVIg treatment may increase the proportion of people with resolution of at least one conduction block; however, the results were also consistent with no effect (RR 7.00, 95% CI 0.95 to 51.70; 4 RCTs, 28 participants; low‐certainty evidence).

IVIg versus placebo (maintenance treatment): a trial that included participants on maintenance IVIg treatment reported an increase in disability in 17 out of 42 (40%) people switching to placebo and seven out of 42 (17%) remaining on IVIg (RR 2.43, 95% CI 1.13 to 5.24; 1 RCT, 42 participants; moderate‐certainty evidence) and a decrease in grip strength in 20 out of 42 (48%) participants after a switch to placebo treatment compared to four out of 42 (10%) remaining on IVIg (RR 0.20, 95% CI 0.07 to 0.54; 1 RCT, 42 participants; moderate‐certainty evidence).

Adverse events, IVIg versus placebo (induction or maintenance): four trials comparing IVIg and placebo reported adverse events, of which data from two studies could be meta‐analysed. Transient side effects were reported in 71% of IVIg‐treated participants versus 4.8% of placebo‐treated participants in these studies. The pooled RR for the development of side effects was 10.33 (95% CI 2.15 to 49.77; 2 RCTs, 21 participants; very low‐certainty evidence). There was only one serious side effect (pulmonary embolism) during IVIg treatment.

IVIg versus SCIg (maintenance treatment): the trial that compared continuation of IVIg maintenance versus SCIg maintenance did not measure disability. The evidence was very uncertain for muscle strength (standardised mean difference 0.08, 95% CI −0.84 to 1.00; 1 RCT, 9 participants; very low‐certainty evidence). The evidence was very uncertain for the number of people with side effects attributable to treatment (RR 0.50, 95% CI 0.18 to 1.40; 1 RCT, 9 participants; very low‐certainty evidence).

Authors' conclusions

Low‐certainty evidence from three small RCTs shows that IVIg may improve muscle strength in people with MMN, and low‐certainty evidence indicates that it may improve disability; the estimate of the magnitude of improvement of disability has wide CIs and needs further studies to secure its significance. Based on moderate‐certainty evidence, it is probable that most IVIg responders deteriorate in disability and muscle strength after IVIg withdrawal. SCIg might be an alternative treatment to IVIg, but the evidence is very uncertain. More research is needed to identify people in whom IVIg withdrawal is possible and to confirm efficacy of SCIg as an alternative maintenance treatment.

Plain language summary

Intravenous immunoglobulin for multifocal motor neuropathy

Review question

We reviewed the evidence about the effect of intravenous immunoglobulin (IVIg) and subcutaneous immunoglobulin (SCIg) in people with multifocal motor neuropathy (MMN). Immunoglobulin is a preparation of antibodies with no specific target purified from the blood and is usually given into a vein (intravenous immunoglobulin; IVIg); however, infusions under the skin (SCIg) might be an alternative.

Background

MMN is a rare condition that causes weakness of the limbs, which tends to worsen over time. The arms, especially the hands, are more often involved than the legs. The usual treatment for MMN is IVIg.

Study characteristics

We searched widely for studies on this topic and found six small trials involving a total of 90 people with MMN who received IVIg treatment. Three studies had limitations, and there was not enough information about the remaining studies to permit a full assessment. Five trials compared an infusion (slow injection) of IVIg with a dummy infusion. The remaining trial compared two different administration routes: IVIg infusion versus SCIg infusion. The effects were measured between 4 and 12 weeks after treatment. Two trials involved employees of an immunoglobulin manufacturer as authors. One of these studies, which was also sponsored by the manufacturer, underwent independent manuscript review by experts with full data access.

Key results and certainty of the evidence

IVIg may improve strength in people with MMN. IVIg may also improve disability, but we need further research to confirm or rule out an effect, as this evidence is less certain. The evidence for whether more people experience side effects on IVIg than with a dummy infusion is uncertain. There was only one serious side effect (pulmonary embolism) during IVIg treatment. Most people with MMN need ongoing immunoglobulin treatment, as disability and muscle strength probably deteriorate with immunoglobulin withdrawal. We need more evidence on whether SCIg might be an alternative to IVIg.

The evidence is current to April 2021.

Summary of findings

Summary of findings 1. IVIg compared to placebo for multifocal motor neuropathy.

IVIg compared to placebo for multifocal motor neuropathy
Patient or population: people with multifocal motor neuropathy Settings: variable: home, centre, or study site Intervention: intravenous immunoglobulin (IVIg) Comparison: placebo
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Placebo	IVIg
Induction
Proportion of participants with an improvement in disability scale used in the original study Follow‐up: 2 to 6 weeks after the last IVIg treatment	11 per 100	33 per 100 (10 to 112)	RR 3.00 (0.89 to 10.12)	18 (36 observations) (3 cross‐over RCTs)	⊕⊕⊝⊝ Lowa,b	IVIg treatment may improve disability, although the result is also consistent with no effect.
Proportion of participants with an improvement of muscle strength Follow‐up: 2 to 6 weeks after the last IVIg treatment	4 per 100	41 per 100 (11 to 156)	RR 11.00 (2.86 to 42.25)	27 (54 observations) (3 cross‐over RCTs)	⊕⊕⊝⊝ Lowa,b	IVIg treatment may improve muscle strength.
Proportion of participants with an improvement in disability at 12 months or later	Not measured
Proportion of participants in which at least 1 conduction block resolved after therapy Follow‐up: 2 to 6 weeks after the last IVIg treatment	0 per 100 (event rate 0 in 28)	Not calculable (event rate 6 in 28)	RR 7.00 (0.95 to 51.70)	28 (56 observations) (4 cross‐over RCTs)	⊕⊕⊝⊝ Lowa,c	IVIg treatment may increase the proportion of people with resolution of at least 1 conduction block; however, the data are also consistent with no effect. In general, spontaneous resolution of conduction block is considered rare, although axonal damage can lead to too low amplitudes to assess the presence of a conduction block.
Maintenance
Proportion of IVIg‐treated participants with an increase in disability during controlled IVIg withdrawal, as determined and defined by study authors Follow‐up: day 8 of 2‐week treatment cycle, day 15 of those treated every 3 or 4 weeks	17 per 100	41 per 100 (19 to 87)	RR 2.43 (1.13 to 5.24)	42 (84 observations) (1 cross‐over RCT)	⊕⊕⊕⊝ Moderatea,d	IVIg withdrawal probably leads to deterioration in disability.
Proportion of IVIg‐treated participants with a decrease in muscle strength during controlled IVIg withdrawal Follow‐up: day 8 of 2‐week treatment cycle, day 15 of those treated every 3 or 4 weeks	48 per 100	10 per 100 (3 to 26)	RR 0.2 (0.07 to 0.54)	42 (84 observations) (1 cross‐over RCT)	⊕⊕⊕⊝ Moderatea,d	IVIg withdrawal probably leads to deterioration in muscle strength.
Adverse effects (induction or maintenance)
Proportion of participants with side effects attributable to treatment Follow‐up: day 8 of 2‐week treatment cycle, day 15 of those treated every 3 or 4 weeks	5 per 100	49 per 100 (10 to 100)	RR 10.33 (2.15 to 49.77)	21 (42 observations) (2 cross‐over RCTs)	⊕⊝⊝⊝ Very lowe	The evidence is too uncertain to determine whether adverse effects attributable to treatment are more or less common with IVIg than with placebo. 1 trial report did not mention adverse effects. In 1 trial, minor adverse effects occurred but were not attributed to individual participants and therefore could not be included in the meta‐analysis. A third trial reported the number of infusions in which adverse events occurred and not the number of participants.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; IVIg: intravenous immunoglobulin; RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence   High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded once for imprecision: small sample size.
^bDowngraded once for study limitations: the studies provided insufficient detail for some risk of bias assessments, and two studies were at high risk of bias for blinding. Although the scales used to measure disability in the included studies were not tailored to multifocal motor neuropathy, this did not warrant further downgrading for indirectness.
^cWe downgraded a second time for imprecision, as only six events occurred. We did not further downgrade the evidence for study limitations, as diagnosis of a conduction block is unlikely to be affected by lack of blinding.
^dWe did not downgrade for study limitations. Although the trial was at unclear risk of other bias due to a short placebo period with possible incomplete IVIg wash‐out, this would tend to lead to an underestimation of withdrawal effects rather than an exaggeration.
^eWe downgraded the evidence a second time for imprecision, as the evidence involved two small trials and few events. We also downgraded the evidence because the studies provided insufficient detail for some risk of bias assessments, and one of the two trials was at high risk of bias.

Summary of findings 2. IVIg compared to SCIg for multifocal motor neuropathy.

IVIg compared to SCIg for multifocal motor neuropathy
Patient or population: people with multifocal motor neuropathy Settings: initial short hospital stay of 1 to 3 days, followed by self‐administration at home supervised by telephone calls from the study nurse Intervention: intravenous immunoglobulin (IVIg) Comparison: subcutaneous immunoglobulin (SCIg)
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	SCIg	IVIg
Induction
No trial studied the use of SCIg for induction.
Maintenance
Proportion of IVIg‐treated participants with an increase in disability during controlled IVIg withdrawal	Not reported
Change in mean muscle strength in participants on IVIg treatment at study entry Follow‐up: 10 weeks	See comment	The mean change in mean muscle strength in participants on IVIg treatment at study entry in the intervention group was 0.08 SD higher (−0.84 lower to 1.00 higher)	‐	9 (18 observations) (1 cross‐over RCT)	⊕⊝⊝⊝ Very lowa,b	The evidence is uncertain for the effects on muscle strength of maintenance treatment with SCIg vs IVIg.   As a rule of thumb, an effect size of 0.2 would be a small effect; 0.08 is unlikely to be important (Cohen 1988).
Adverse events
Proportion of participants with side effects attributable to treatment Follow‐up: baseline to final evaluation 105 days (45 to 140)	67 per 100	33 per 100 (12 to 93)	RR 0.50 (0.18 to 1.40)	9 (18 observations) (1 cross‐over RCT)	⊕⊝⊝⊝ Very lowa,b	The evidence is uncertain for adverse effects attributable to maintenance treatment with SCIg vs IVIg.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; IVIg: intravenous immunoglobulin; RCT: randomised controlled trial; RR: risk ratio; SCIg: subcutaneous immunoglobulin; SD: standard deviation
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded for study limitations: a single‐blinded RCT in which participants were not blinded for outcome. The study period per treatment arm varied between 54 and 168 days, which may have led to a considerable carry‐over effect in receiving SCIg, especially in participants with a shorter treatment period.
^bDowngraded twice for imprecision: a single trial with a small sample size (9 people, 18 observations). The event rate for adverse events was very low (3 of 9 participants on IVIg versus 6 of 9 participants on placebo).

Background

Description of the condition

Multifocal motor neuropathy (MMN) is characterised by slowly progressive, asymmetric, predominantly distal weakness of one or more limbs with no objective loss of sensation (Cats 2010; Lewis 1982; Nobile‐Orazio 2001; Nobile‐Orazio 2002). This weakness may be accompanied by muscular atrophy in later stages of the disease, and cramps and fasciculations are reported to occur in approximately 50% of people with MMN (Nobile‐Orazio 2001). The arms are usually more affected than the legs (Nobile‐Orazio 2001; Nobile‐Orazio 2002; Van den Berg‐Vos 2000a). Tendon reflexes are often decreased or absent in the affected limb, but can also be brisk in a minority of people with MMN (Cats 2010). Cranial nerve involvement and respiratory failure due to phrenic nerve palsy have occasionally been reported (Beydoun 2000; Cavaletti 1998; Kaji 1992; Pringle 1997). Recently, the extent of sensory signs and symptoms in MMN has been reconsidered, and the development of electrophysiological sensory changes with or without sensory signs and symptoms over the course of MMN have been described (Lambrecq 2009; Lievens 2009).

The hallmark of the disease is the presence of multifocal conduction block on electrophysiological testing outside the usual sites of nerve compression (Cornblath 1991; Kaji 1991; Parry 1992; Parry 1993; Van Asseldonk 2003). However, some people with otherwise typical MMN have no detectable conduction block, probably because these blocks are activity‐dependent (Nodera 2006), or are located in nerve segments which cannot be assessed by routine electrophysiological examination (Delmont 2006; Pakiam 1998). Conduction block is a reduction in the amplitude or area (or both) of the compound muscle action potential (CMAP) obtained by proximal versus distal stimulation of motor nerves in the absence of or with only focal abnormal temporal dispersion (Cornblath 1991; Kaji 2003; Nobile‐Orazio 2001). The extent of reduction of the CMAP amplitude or area necessary for conduction block, or both, is still a matter of debate. Definite conduction block is usually defined as an area reduction of 50% or more between proximal versus distal stimulation in a long nerve segment or an amplitude reduction of 30% or more over 2.5 cm (Franssen 1997; Rhee 1990; Van Asseldonk 2006). Probable conduction block is usually defined as an amplitude reduction of 30% or more between proximal versus distal stimulation in an arm nerve (Albers 1985; Oh 1994).

Three main lines of evidence suggest that MMN is an immune‐mediated disorder.

• Thirty per cent to 80% of people with MMN have serum immunoglobulin M (IgM) ganglioside‐monosialic acid (GM1) antibodies (Van Schaik 1995; Willison 2002). These antibodies are not specific to MMN and can be found in various other immune‐mediated neuropathies such as chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) and Guillain‐Barré syndrome (GBS). Furthermore, it has been shown that antibody titres do not correlate with clinical severity over time (Léger 2001; Van den Berg 1998).

• Increased signal intensities on T2‐weighted magnetic resonance images of the brachial plexus have been observed in people with MMN, suggesting an inflammatory process (Van Es 1997).

• Response to immune‐modulating treatment. The efficacy of human polyclonal immunoglobulin administered intravenously (IVIg) has been suggested in many open and uncontrolled studies. Elevated anti‐GM1 antibodies have been suggested to be a reliable predictor of a favourable response after IVIg treatment (Van den Berg‐Vos 2000a). However, a considerable number of participants lack these antibodies but do respond favourably to IVIg treatment (Azulay 1997; Bouche 1995; Chaudhry 1993; Katz 1997; Van den Berg 1995). A beneficial effect of cyclophosphamide (Chaudhry 1993; Feldman 1991; Krarup 1990; Meucci 1997; Pestronk 1988; Van Es 1997), interferon‐beta 1a (Martina 1999; Van den Berg‐Vos 2000b), ciclosporin (Nemni 2003), methotrexate (Nobile‐Orazio 2009), and azathioprine (Hausmanowa‐Petrusewicz 1991) in MMN has been suggested in several uncontrolled studies and reviewed in another Cochrane Review (Umapathi 2015).

Almost 80% of people with MMN are between 20 and 50 years of age at onset of the disease (Nobile‐Orazio 2001). Men are more frequently affected than women, with a ratio of 2.6:1 (Nobile‐Orazio 2001). The prevalence is estimated to be 1 to 2 per 100,000 (Nobile‐Orazio 2001). The diagnosis of MMN is based on clinical, laboratory, and electrophysiological characteristics (Hughes 2001; Nobile‐Orazio 2001; Parry 1992; Van den Berg‐Vos 2000a; Van Schaik 2010). A set of diagnostic criteria has been proposed that combines clinical, laboratory, and electrophysiological features of people with MMN, which may help to predict whether individuals will respond to treatment (Van den Berg‐Vos 2000a). The American Association of Electrodiagnostic Medicine has developed five criteria through a formal consensus process for diagnosing MMN with a high level of confidence (Olney 2003). These criteria for definite MMN are: weakness without objective sensory loss in the distribution of two or more nerves; definite conduction block in two or more nerves outside of common entrapment sites; normal sensory nerve conduction velocity across the same segments with demonstrated motor conduction block; normal results for sensory nerve conduction studies on all tested nerves, with a minimum of three nerves tested; and absence of upper motor neuron signs. The criteria for probable MMN are somewhat less strict for conduction block.

Description of the intervention

Immunoglobulins (Igs) are proteins produced by the immune system to neutralise pathogens. Ig is harvested from pooled blood (plasma) from healthy donors and used therapeutically after purification. Normal human Ig preparations consist of IgG (90% to 98%) and varying small amounts of IgM and IgA isotypes (Anonymous 2009). 

Ig can be administered through intramuscular injection, intravenously (IVIg), or subcutaneously (SCIg). The first IVIg treatment is administered at a dose of 2 g/kg body weight over two to five days. After this so‐called loading dose, most people require maintenance treatment at a lower dose and at varying intervals, usually every two to four weeks. When SCIg is chosen for maintenance treatment, patients are first treated with a loading dose of IVIg (NHS England 2018).

In most people with MMN, the effect of IVIg only lasts for a few weeks, and maintenance treatment with periodic IVIg infusions for extended periods of time is often indicated (Azulay 1997; Léger 2001; Meucci 1997; Van den Berg 1998; Van den Berg‐Vos 2002). Maintenance treatment regimens range from 0.4 g/kg once a week to 1 to 2 g/kg given over two or five days monthly, or are guided by individual response. If initial treatment with IVIg is effective, IVIg maintenance therapy appears to remain effective in most patients. However, the effectiveness usually declines during prolonged treatment, requiring increasing dose or higher frequency of administration, or both (Cats 2010; Terenghi 2004; Van den Berg‐Vos 2002). This long‐term decline in effectiveness is probably due to ongoing axonal degeneration, which is reflected in progressive reduction of CMAP amplitudes and decreased muscle strength (Terenghi 2004; Van den Berg 1998). In a published retrospective study, treatment with higher‐than‐normal maintenance doses of IVIg promoted reinnervation, decreased the number of conduction blocks, and prevented axonal degeneration in 10 people with MMN up to 12 years (Vucic 2004).

The need for ongoing IVIg treatment in most people with MMN is accompanied by the burden of frequent infusions, adverse effects in some people, and high healthcare costs. SCIg, which has the main advantage of self‐infusion not requiring admission to hospital, could be an alternative treatment to IVIg. This route of administration increases an individual's autonomy and possibly quality of life (Cocito 2014). In general, maintenance treatment with SCIg is administered once or twice per week, depending on the required dose, local adverse events, and the patient's preference. IVIg and SCIg dose and interval are titrated individually to find the lowest possible dose to prevent deterioration. Infusions can take place at hospitals, infusion centres, or at home when administered by specialised nurses.

Periodic IVIg and SCIg withdrawals are recommended to determine immunoglobulin dependency and to establish the lowest effective concentration. Whether and when IVIg or SCIg withdrawals should be attempted in people with MMN is unknown. In the largest retrospective study including 88 people with MMN, four (5%) IVIg responders did not require ongoing treatment (Cats 2010). In another study of 40 participants, IVIg withdrawal was possible in 22% (Léger 2008).

How the intervention might work

The ways by which immunoglobulin exerts its supposed beneficial effect in MMN are not clear, but various mechanisms have been suggested (Kazatchkine 2001; Van Schaik 1994; Yu 1999). Studies in other diseases treated with IVIg have demonstrated that IVIg may inhibit auto‐antibody production, neutralise pathogenic antibodies, decrease antibody‐dependent cellular cytotoxicity by blocking Fc‐receptors on macrophages, and prevent differentiation of B‐cells to IgG producing plasma cells by restoring expression of inhibitory FcRII receptors on B‐cells (Kazatchkine 2001). Furthermore, peripheral blood from people with MMN who have been treated with IVIg shows increased CD8‐positive suppressor T‐cell function (Delfraissy 1985; Leung 1987; Macey 1990). The most recent evidence suggests that IVIg inhibits complement activation mediated by anti‐GM1 IgM antibodies in a dose‐dependent fashion (Piepers 2007; Yuki 2011).

Why it is important to do this review

Treatment options for people with MMN are few. In contrast to the response in people with CIDP, people with MMN usually do not respond to steroids or plasma exchange, and may worsen when they receive these treatments (Nobile‐Orazio 2001). Previous studies have shown that adding mycophenolate mofetil, rituximab, and eculizumab did not lead to a significant reduction of the required maintenance IVIg dose (Chaudhry 2010; Fitzpatrick 2011; Piepers 2007). Cyclophosphamide has serious long‐term side effects (Meucci 1997), and interferon‐beta 1a has only been tested in very limited numbers of people with MMN (Van den Berg‐Vos 2000a). At the time of the first version of this Cochrane Review in 2005, four randomised controlled double‐blind trials of IVIg for treating MMN were known to have been performed (Azulay 1994; Federico 2000; Léger 2001; Van den Berg 1995), but no systematic review had been published. Two more eligible studies have subsequently been published. We have provided a systematic review of the randomised trials of immunoglobulin that have been undertaken in MMN.

Objectives

To assess the efficacy and safety of intravenous and subcutaneous immunoglobulin in people with multifocal motor neuropathy (MMN).

Methods

Criteria for considering studies for this review

Types of studies

We searched for all randomised controlled trials (RCTs) or quasi‐RCTs (i.e. trials that are not truly randomised but use alteration or another systematic method of allocation) examining the effects of IVIg and SCIg treatment in people with MMN. We considered quasi‐RCTs in spite of their higher risk of selection bias, because there is a paucity of evidence on this topic. We included two different types of studies dependent on the treatment status of the participants at study entry (baseline): a) studies with immunoglobulin‐naive participants (induction treatment) and b) studies with participants already on maintenance IVIg and SCIg treatment (i.e. immunoglobulin controlled‐withdrawal studies using placebo or alternative treatments to substitute immunoglobulin maintenance treatment). Randomised cross‐over trials were eligible.

We included studies regardless of publication status and language of publication.

Types of participants

Eligible studies had to include unselected participants with definite or probable MMN according to published criteria (Olney 2003; Van den Berg‐Vos 2000a; Van Schaik 2010). If fulfilment of published criteria were not stated, we defined MMN as a slowly or stepwise progressive asymmetric lower motor neuron syndrome with no bulbar or upper motor signs and evidence of definite or probable conduction block in motor nerves. Mild sensory symptoms at time of diagnosis were permitted as long as there were no sensory signs on examination and sensory nerve conduction studies were normal. People with upper motor neuron features or bulbar signs had to be excluded. Other related conditions, such as other neuropathies (diabetic, lead, porphyric, or vasculitic neuropathy, CIDP, Lyme neuroborreliosis, postradiation neuropathy, hereditary neuropathy with liability to pressure palsies, Charcot‐Marie‐Tooth disease, or paraproteinaemic neuropathies) and myopathies (facioscapulohumeral muscular dystrophy, inclusion body myositis) had to be excluded.

Types of interventions

We considered for inclusion studies of any dose of immunoglobulin compared with placebo or any other treatment. All brands of immunoglobulin were included, provided that the preparation was produced according to World Health Organization (WHO) guidelines (WHO 1982).

Types of outcome measures

In this update we added three secondary outcomes which permitted us to include studies focusing on the ongoing need for immunoglobulin maintenance treatment and studies exploring alternatives to maintenance intravenous immunoglobulin (IVIg) treatment, namely subcutaneous immunoglobulin (SCIg). See Differences between protocol and review.

Eligible studies had to have measured at least one of our efficacy outcomes, but in practice we did not exclude any studies on this basis.

Primary outcomes

• Proportion of participants with an improvement in disability between two and six weeks after the last treatment as determined and defined by the study authors and compared to baseline.

Only studies with immunoglobulin‐naive participants at baseline were eligible for the primary outcome. We defined the primary outcome measure as a proportion because the studies used different disability scales. Each study used the strictest available criteria to define improvement. As immunoglobulin is thought to induce and maintain improvement in the majority of people with MMN, but does not eradicate the disease, people with MMN have to be treated with periodic infusions for long periods. Outcome assessment between two and six weeks after the last IVIg treatment probably reflects the treatment responses of these people best. This period was also used for SCIg.

Secondary outcomes

Induction treatment

We assessed these secondary outcomes only for studies with IVIg‐naive participants at baseline (induction treatment).

• Proportion of participants with an improvement in muscle strength as determined and defined by the study authors, assessed between two and six weeks after the last treatment and compared to baseline.

• Mean change in muscle strength expressed as standardised mean difference (SMD), assessed between two and six weeks after the last treatment and compared to baseline in immunoglobulin‐naive participants at study entry.

• Proportion of participants with a sustained improvement in disability at 12 months or later as determined and defined by the study authors.

• Proportion of participants in whom at least one conduction block resolved after therapy, assessed between two and six weeks after the last treatment and compared to baseline.

Maintenance treatment

We assessed these secondary outcomes only for studies with participants already receiving IVIg/SCIg at baseline (maintenance treatment).

• Proportion of immunoglobulin‐treated participants with an increase in disability during controlled IVIg/SCIg withdrawal as determined and defined by the study authors.

• Proportion of immunoglobulin‐treated participants with a decrease in muscle strength during controlled IVIg/SCIg withdrawal as determined and defined by the study authors.

• Mean change in muscle strength expressed as SMD, assessed between two and six weeks after the last treatment and compared to baseline in participants on maintenance IVIg/SCIg treatment at study entry.

Induction or maintenance treatment

We assessed the frequency of adverse effects for all types of studies.

• Frequency of adverse effects attributable to treatment during the whole study period.

Search methods for identification of studies

Electronic searches

We searched the following databases on 20 April 2021.

• Cochrane Neuromuscular Specialised Register (in the Cochrane Register of Studies Web; Appendix 1)

• Cochrane Central Register of Controlled Trials (CENTRAL; Issue 4, 2021) (in the Cochrane Register of Studies Web; Appendix 2)

• MEDLINE (1946 to 19 April 2021; Appendix 3)

• Embase (1974 to 2021 week 15; Appendix 4)

• World Health Organization International Clinical Trials Registry Platform (https://trialsearch.who.int/ ; Appendix 5)

• US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov; Appendix 6)

Searching other resources

We searched the references listed in the published studies, reviews, textbooks, and relevant conference proceedings. We contacted investigators known as active in the field to identify unpublished or overlooked studies. We invite readers to suggest studies, particularly in other languages, that we should consider for inclusion when the review is updated.

Data collection and analysis

Selection of studies

Two review authors (IvS and LvdB for the first version of the review, and FE and IvS or FE and SK for the 2021 update) independently reviewed the titles and abstracts obtained from the literature searches for potential relevance. We obtained the full‐text reports of those studies deemed potentially relevant, and the review authors selected trials for inclusion in the review. Review authors were not blinded to author and source institution. Any disagreements were resolved by consensus. The review authors recorded the study selection process in sufficient detail to create a PRISMA flow diagram and Characteristics of excluded studies tables (Moher 2015).

Data extraction and management

Two review authors (SK and RB) independently extracted outcome data from newly identified trials using a data extraction tool. Both review authors entered data into Review Manager 5 (Review Manager 2020), with any disagreements resolved by consensus. 

For this update, one review author (SK) extracted data for the following characteristics: study design, eligibility criteria, date conducted, setting, participant characteristics, intervention detail, outcomes assessed, conflicts of interest amongst investigators, funding sources, and potentially relevant studies in references, and entered these data into Review Manager 5 (Review Manager 2020). There were two new studies at this update; piloting the data extraction form was not feasible. 

For studies requiring translation, we would extract data from a translation of the study. We would cross‐check numerical data from translated studies with the study report if possible.

Assessment of risk of bias in included studies

We completed a risk of bias table addressing the following domains: sequence generation, allocation concealment, blinding of participants and personnel (performance bias), blinding of outcome assessors (detection bias), incomplete outcome data, selective outcome reporting, and other sources of bias (e.g. biases arising from the cross‐over design of the included studies), according to the guidance in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Two review authors (RB and SK) independently completed risk of bias assessments, making a judgement of low, high, or unclear risk of bias for each domain. Any differences were resolved by discussion. The review authors considered risk of bias judgements for each outcome across domains when making GRADE assessments (see below).

Measures of treatment effect

For dichotomous data (the proportion of participants with an improvement in disability, proportion of participants with an improvement of muscle strength, proportion of participants with an improvement in disability at 12 months, proportion of participants in which at least one conduction block resolved after therapy, proportion of immunoglobulin‐treated participants with an increase in disability during controlled IVIg/SCIg withdrawal, proportion of immunoglobulin‐treated participants with a decrease in muscle strength during controlled IVIg/SCIg withdrawal, and frequency of adverse effects attributable to treatment during the whole study period), we calculated the risk ratio (RR) with corresponding 95% confidence intervals (CIs) for each study.

Regarding continuous data (changes in muscle strength in immunoglobulin‐naive participants and changes in muscle strength in participants on maintenance IVIg/SCIg treatment), we anticipated that different trials would use different scales to assess muscle strength. We calculated effect sizes for each study and pooled them using SMD, as muscle strength assessment was not sufficiently comparable between studies to allow for analysis with mean differences. The SMD is a dimensionless measure of effect for continuous data, which is useful for combining results of individual studies in which the same outcome is measured with different instruments or scales. The SMD is defined as the mean change in score of the placebo group minus mean change in score of the treatment group, divided by the pooled standard deviation (SD) of the change in scores of the two groups. We derived means, SDs, and standard errors by calculation or extraction from the available data. If reports provided only CIs, we calculated SDs using values specific to the distribution given the expected small sample sizes of the study (Deeks 2019).

Unit of analysis issues

If there were insufficient data to allow paired analysis in cross‐over trials, we considered cross‐over trials as single parallel‐design trials, assuming that the groups from both parts of the cross‐over trial were independent, rather than as a single group of participants, and assuming that no carry‐over effect had occurred. We intended to include data from participants who did not cross over to the second treatment for the first treatment only. If multiple trial arms were reported in a single trial, we would include only the treatment arms relevant to the review topic. If two intervention groups in the same meta‐analysis had a shared comparison group (e.g. drug A versus placebo and drug B versus placebo), we would follow the guidance in Chapter 23 of the Cochrane Handbook for Systematic Reviews of Interventions to avoid double‐counting (Higgins 2019).

Dealing with missing data

We obtained individual participant data for one study (Van den Berg 1995). There were no other missing data.

Assessment of heterogeneity

We used the I² statistic to measure heterogeneity amongst the trials in each analysis (Higgins 2003). If we identified substantial unexplained heterogeneity, we reported it and examined trial characteristics for likely causes. We used the rough guide to interpretation of I² as outlined in Chapter 11 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2019), as follows:

• 0% to 40%: might not be important;

• 30% to 60%: may represent moderate heterogeneity;

• 50% to 90%: may represent substantial heterogeneity;

• 75% to 100%: considerable heterogeneity.

We avoided the use of absolute cut‐off values, instead interpreting I² in relation to the size and direction of effects and strength of evidence for heterogeneity (e.g. P value from the Chi² test, or CI for I²) (Deeks 2019).

Assessment of reporting biases

We searched clinical trials registries for ongoing or completed but unpublished trials to limit publication, citation, location, and outcome reporting bias. There was an insufficient number of studies (at least 10 are required) to create a funnel plot to assess small‐study effects (Page 2019).

Data synthesis

When in interpreting the I², the size and direction of effects and strength of evidence for heterogeneity showed data to be heterogeneous, we used the random‐effects model of DerSimonian and Laird (Ioannidis 1995), employing Review Manager 5 (Review Manager 2020). If no heterogeneity could be demonstrated, we used a fixed‐effect model (Mantel‐Haenszel RR method) (Rothman 1986). To assess overall efficacy on proportional outcomes from all the studies, we calculated pooled RR estimates. We obtained absolute risk differences using Review Manager 5 analyses. For continuous data, we pooled individual SMDs with the generic inverse variance method available in Review Manager 5. In this method, 1/variance is used as weights. We expressed statistical uncertainty with 95% CI. We analysed data from studies exploring alternative treatments to maintenance IVIg treatment in a separate comparison (IVIg versus SCIg) and analysed induction studies and maintenance studies separately, other than for adverse events, for which we pooled the data from both types of study.

Subgroup analysis and investigation of heterogeneity

If there was heterogeneity, we investigated possible sources by repeating the analysis after elimination of trials at high or unclear risk of bias, paying particular attention to allocation concealment.

We did not plan subgroup analyses, given the likely size of included studies in this rare disease.

Sensitivity analysis

In the original protocol, we stated that sensitivity analysis was to be performed only in the presence of heterogeneity, in which case the source was to be investigated by repeating the analysis after elimination of trials that scored less than A (adequate) on each of the indices of quality then in use, paying particular attention to allocation concealment. For the review update, we planned the following sensitivity analyses:

• repeat the analysis excluding unpublished studies;

• repeat the analysis excluding studies at high risk of bias in the event of heterogeneity.

The protocol also specified an intention to assess the effect of including cross‐over studies; however, all of the included studies had a cross‐over design.

We performed a sensitivity analysis excluding Van den Berg 1995, as the validity of the trial design was debatable (see Included studies for details). This was the only sensitivity analysis, as there was no statistical heterogeneity, and all studies were published. 

We reviewed open and uncontrolled studies and compared them to the results of the included studies in the Discussion. We discussed adverse events of IVIg from the non‐randomised literature in relation to the adverse events found in this review in the Discussion.

Summary of findings and assessment of the certainty of the evidence

We created summary of findings tables using GRADEpro GDT software (GRADEpro GDT), and presented the following outcomes:

For the comparison IVIg/SCIg compared to placebo for MMN

Primary outcome

• Proportion of participants with an improvement in disability as determined and defined by the study authors, assessed between two and six weeks after the last treatment, compared to baseline.

Secondary outcomes

Induction treatment

• Proportion of participants with an improvement of muscle strength as determined and defined by the study authors, assessed between two and six weeks after the last treatment and compared to baseline.

• Proportion of participants with an improvement in disability at 12 months or later as determined and defined by the study authors.

• Proportion of participants in whom at least one conduction block resolved after therapy, assessed between two and six weeks after the last treatment and compared to baseline.

Maintenance treatment

• Proportion of IVIg/SCIg‐treated participants with an increase in disability during controlled immunoglobulin withdrawal as determined and defined by the study authors.

• Proportion of IVIg/SCIg‐treated participants with a decrease in muscle strength during controlled immunoglobulin withdrawal as determined and defined by the study authors.

Induction or maintenance treatment

• Frequency of adverse effects attributable to treatment during the whole study period.

For the comparison IVIg compared to SCIg for MMN

This comparison was not applicable to induction treatment.

Outcomes

• Proportion of IVIg/SCIg‐treated participants with an increase in disability during controlled immunoglobulin withdrawal as determined and defined by the study authors.

• Proportion of IVIg/SCIg‐treated participants with a decrease in muscle strength during controlled immunoglobulin withdrawal as determined and defined by the study authors.

• Frequency of adverse effects attributable to treatment during the whole study period.

Two review authors (FE and IvS) used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias) to independently assess the certainty of a body of evidence (i.e. studies that contribute data for the prespecified outcomes). We used the methods and recommendations described in Chapters 14 and 15 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2019a; Schünemann 2019b). Any disagreements were resolved by discussion or by involving another review author (LvdB). We assessed trial quality according to the GRADE criteria. We considered RCTs as high‐certainty evidence if the above five factors were not present to any serious degree, downgrading the certainty to moderate, low, or very low as required. We downgraded the certainty of the evidence once if a GRADE consideration was serious and twice if very serious. We justified all decisions to down‐ or upgrade the certainty of the evidence using footnotes and made comments to aid the reader's understanding of the review where necessary.

We based our conclusions only on findings from the quantitative or narrative synthesis of included studies for this review. The recommendations for practice and our implications for research suggest priorities for future research and outline the remaining uncertainties in the area. We assessed the certainty of the evidence and stated uncertainties.

Results

Description of studies

Results of the search

A flow chart illustrating the study selection process is provided in Figure 1.

1.

Study selection flow chart.

The previous version of the review included four studies (Van Schaik 2005).

For this update, our database searches retrieved 451 references. After removal of 149 duplicates, we screened 302 records. From these, we identified 28 reports, from which we selected two new included studies (eight associated references) and excluded 11 studies (20 associated references). We identified no ongoing studies or studies awaiting classification.

In total, six RCTs (13 reports) from the previous version of the review and this update were eligible for inclusion (Azulay 1994; Federico 2000; Hahn 2013; Harbo 2009; Léger 2001; Van den Berg 1995).

We excluded 23 studies (Al‐Zuhairy 2019; Auer 1994; Baumann 2009; Burrell 2011; Cats 2008; Cats 2010; Charles 1992; Eftimov 2009; Herraets 2019; Jafari 2000; Kelly 1992; Kermode 1992; Köller 2006; Komiyama 1998; Kubori 1999; Kuwabara 2018; Léger 2008; Léger 2019; Pestronk 1995; Tan 1994; Turnbull 2001; Van den Berg‐Vos 2002; Yuki 1993). See Excluded studies.

Included studies

Participants

Six RCTs including 90 randomised participants with MMN were eligible (Azulay 1994; Federico 2000; Hahn 2013; Harbo 2009; Léger 2001; Van den Berg 1995).

Four studies included immunoglobulin‐naive participants at study entry and used IVIg as study treatment; in two of these studies all participants were treatment naive (Azulay 1994; Federico 2000). One study included both treatment‐naive participants and participants previously treated with IVIg (Léger 2001), and one study was designed as a proof‐of‐principle study (Van den Berg 1995), in which all participants had previously been treated with IVIg. Participants from Léger 2001 who were previously treated with IVIg were not included in our analysis. Although participants in Van den Berg 1995 were selected on the basis of IVIg response in an open trial before entering a double‐blind, placebo‐controlled, randomised, cross‐over trial, we decided to include this study because all participants were unselected when entering the open phase of the trial, and all of these participants responded to IVIg, therefore all participants entered the second phase. As the validity of this trial design is debatable, we presented results with and without this study.

Two studies included known IVIg responders receiving maintenance IVIg treatment at baseline: one study compared continuation of IVIg with placebo (Hahn 2013), and the other continuation of IVIg with SCIg (Harbo 2009).

Interventions

The included studies used different treatment regimens of total infusion dose and interval between infusions. Three studies used a total of 2 g/kg body weight of IVIg administered over five days (Azulay 1994; Federico 2000; Van den Berg 1995), whilst one study used 2.5 g/kg body weight of IVIg over five days (Léger 2001). In two studies, IVIg dose and intervals were equal to the individual pre‐study regimen. In one of these studies, mean monthly dose was 1.2 g/kg, with intervals ranging between two and four weeks (Hahn 2013); in the other study, IVIg dose ranged between 0.6 g/kg and 2.1 g/kg per month, with intervals ranging between 18 and 56 days (Harbo 2009). SCIg dose was equivalent to IVIg dose in individual participants and ranged between 0.2 g/kg and 0.5 g/kg per week (Harbo 2009). The total SCIg amount was administered in two or three doses over the week.

Study design

All six RCTs had a cross‐over design. Cross‐over took place eight weeks after the first treatment in one study (Azulay 1994). In another study, participants crossed over into the second arm at day 28 if they remained unchanged or deteriorated (Federico 2000). Participants who had improved at day 28 did not enter the second arm until they were back at baseline levels. Léger and colleagues treated participants each month for three months (Léger 2001). After three treatments, all participants who were responders remained on the same treatment for three months, and the non‐responders crossed over for the following three months. Van den Berg and colleagues treated participants four times with IVIg or placebo in random order (Van den Berg 1995). The interval between each treatment was determined by the time it took for the participant to get back to baseline levels, with a minimum of one month. Our prespecified outcomes were assessed at one month after the last IVIg infusion in three studies (Azulay 1994; Federico 2000; Léger 2001), and at variable time points ranging between two and four weeks in one study (Van den Berg 1995). In Hahn and colleagues, participants received pre‐treatment with IVIg for 12 weeks before they entered the first study treatment period of 12 weeks. After the first study period, all participants received open‐label IVIg for 12 weeks ('wash‐out') before entering the second study period of 12 weeks. Participants who deteriorated during the first period were restabilised during the open‐label IVIg period and crossed to the second treatment ('accelerated switch') (Hahn 2013). The study comparing SCIg with IVIg was preceded by a prolonged treatment‐free interval of a maximum of 10 weeks to ensure that participants were responsive to immunoglobulins and a restabilisation phase of IVIg infusion (Harbo 2009). Interventions were separated by a variable wash‐out period during which participants received two regular doses of IVIg.

Outcomes

Each study used different outcome measures. For details, see Characteristics of included studies. Only individual participant data were obtained from one study (Van den Berg 1995).

Settings

Two studies were conducted in France (Azulay 1994; Léger 2001), one in neuromuscular clinics in Canada (Federico 2000), one multicentre study in 17 centres in the USA, Canada, and Denmark (with the option of home infusion in the USA and Canada) (Hahn 2013), one in Denmark (hospital and thereafter home self‐administration) (Harbo 2009), and one in hospitals in the Netherlands (Van den Berg 1995).

Funding and conflicts of interest

The trial comparing IVIg versus placebo maintenance treatment was sponsored by an IVIg manufacturer; employees were also involved in the study. However, the manuscript was reviewed by independent experts to whom the trial data were made available (Hahn 2013). An additional trial involved authors with conflicts of interest through employment (Léger 2001). The remaining trials had no reported commercial funding and were without conflicts of interest, or did not provide this information.

Excluded studies

We listed 23 excluded reports or studies. The reasons for exclusion were as follows:

• four letters to editor (Auer 1994; Kelly 1992; Pestronk 1995; Turnbull 2001);

• nine case series or case reports (Baumann 2009; Cats 2010; Charles 1992; Jafari 2000; Kermode 1992; Köller 2006; Komiyama 1998; Léger 2008; Yuki 1993);

• seven non‐randomised prospective studies (Burrell 2011; Cats 2008; Eftimov 2009; Herraets 2019; Kubori 1999; Kuwabara 2018; Van den Berg‐Vos 2002);

• one review (Tan 1994);

• two non‐inferiority RCTs comparing IVIg brands (Al‐Zuhairy 2019; Léger 2019).

See Characteristics of excluded studies tables. We listed all studies excluded at full‐text review, as these papers may be of interest given the paucity of evidence on this topic.

Risk of bias in included studies

Risk of bias assessments for each trial are presented in Figure 2.

2.

Methodological quality summary: review authors' judgements about each methodological quality item for each included study. Key: red (‐) = high risk of bias, yellow (?) = unclear risk of bias, green (+) = low risk of bias.

None of the included trials was at low risk of bias in all domains. Three trials were at high risk of bias in at least one domain (Azulay 1994; Hahn 2013; Harbo 2009). The remaining trials were at unclear risk of bias in at least two domains (Federico 2000; Léger 2001; Van den Berg 1995).

Allocation

Federico 2000 used a random number generator, and the Hahn 2013 protocol indicates use of "computer generated pseudo‐random numbers". Harbo 2009 described block randomisation by the pharmacist. We judged all three trials as at low risk of bias for random sequence generation. We assessed Azulay 1994, Léger 2001, and Van den Berg 1995 as at unclear risk of bias, as insufficient detail was provided on the method of randomisation. 

Harbo 2009 and Hahn 2013 described their method of allocation concealment and was therefore assessed as at low risk of bias. Azulay 1994, Léger 2001, and Van den Berg 1995 did not describe allocation concealment and were therefore assessed as at unclear risk of bias for this domain. In Federico 2000, a central pharmacist distributed opaque infusion bags, but measures to conceal allocation were not clear.

Blinding

Only Federico 2000 and Hahn 2013 were at low risk of both performance and detection bias, as participants, personnel, and outcome assessors were all blinded, and the reports describe methods to maintain blinding. In Léger 2001, participants and assessors were reported as blinded, but as the trial authors did not describe measures to maintain blinding, our judgement was unclear risk for both blinding domains. In Van den Berg 1995, blinding was attempted, but blinding of participants was not maintained, and all participants could correctly name the treatment sequence. We judged this trial as at high risk of bias for blinding of participants and personnel, and unclear risk of bias for blinding of outcome assessment. In Azulay 1994, treatment providers were not blinded, but did not assess participants; we judged the risk of bias as high for blinding of participants and personnel and unclear for blinding of outcome assessment. was at high risk of bias for both performance and detection bias, as participants and personnel were not blinded, and although assessors were blinded, treatment preference could potentially have influenced outcome measures (e.g. muscle strength).

Incomplete outcome data

All of the included studies addressed the occurrence and reasons for incomplete outcome data. We assessed all studies as at low risk of attrition bias, except for Harbo 2009, which was judged unclear.

Selective reporting

All of the included studies reported all outcomes that had been prespecified in the methods section in the results section. Protocols were not available for most of the studies. We judged all six studies as at low risk of reporting bias.

Other potential sources of bias

All of the included studies were at unclear risk of other bias, except for Harbo 2009, which was at high risk. In two studies, all participants had to be immunoglobulin responders (; ). As the main goal of these trials was to determine whether IVIg maintenance was needed (in ) and whether SCIg is as effective and safe as IVIg maintenance treatment (in ), this recruitment bias seems justified. In Harbo 2009, participants were treated with IVIg twice during the 'wash‐out' period, which may have led to a considerable carry‐over effect into the SCIg arm, which lasted only three IVIg cycles; this resulted in a judgement of high risk of bias. In Hahn 2013, the blinded period lasted 12 weeks, which is relatively short and could have resulted in a carry‐over effect in the placebo arm, as participants had been treated previously with IVIg (wash‐out periods were scheduled prior to study entry and between study periods). However, in clinical practice most people receive IVIg on a monthly basis because of a four‐week duration of effect, so we judged the risk of bias as unclear rather than high.

In one trial it was unclear whether participants with MMN had been treated with immunoglobulins in the past (), which may have led to recruitment bias. The wash‐out period was eight weeks in Azulay 1994, possibly leading to a carry‐over effect in participants who switched from active treatment to placebo treatment. In one trial a variable wash‐out period was defined that lasted between 38 and 120 days and may have led to a variable carry‐over effect between participants (). In this study additional (immunomodulatory) treatments were not addressed. One trial used a response‐conditional cross‐over design, meaning that not all participants received both treatments (Léger 2001); however, we used first‐period data from this trial. It was unclear from the Van den Berg 1995 report whether additional treatments were allowed. We judged these four trials as at unclear risk of bias.

Effects of interventions

See: Table 1; Table 2

Immunoglobulins versus placebo

See Table 1.

Primary outcome

Proportion of participants with an improvement in disability between two and six weeks after the last treatment as determined and defined by the study authors and compared to baseline

Three RCTs summarising the results of 18 participants were suitable for this analysis (Azulay 1994; Léger 2001; Van den Berg 1995). All of these trials included IVIg as study treatment, and none used SCIg as study treatment. 

Different disability scales were used: none of five participants improved on the modified Norris scale after IVIg or placebo (Azulay 1994); four out of seven participants improved after IVIg and two out of seven after placebo on a self‐evaluation scale scoring five motor activities of daily life, chosen for each participant together with the examiner at baseline (Léger 2001); and three out of six participants improved after IVIg on the modified Rankin scale and none after placebo (Van den Berg 1995). 

Outcomes were assessed at 28 days after onset of treatment (Azulay 1994), every month up to four or seven months (Léger 2001), and every week up to 28 weeks (Van den Berg 1995). Intravenous immunoglobulin or placebo was each given 18 times. An improvement was reported in seven out of 18 (39%) IVIg treatments and two out of 18 (11%) placebo treatments. A higher proportion of participants improved after IVIg therapy as compared with placebo, with a pooled risk ratio (RR) of 3.00, but confidence intervals (CIs) did not exclude the possibility of no effect (95% CI 0.89 to 10.12; 3 RCTs, 18 participants; low‐certainty evidence). We used a fixed‐effect model, as the studies were of a similar small size, and the study results did not show evidence of heterogeneity (Figure 3; Analysis 1.1).

3.

IVIg versus placebo. Proportion of participants with an improvement in disability as determined and defined by the study authors. In all forest plots, the 'Total' columns show the number of observations.

1.1. Analysis.

Comparison 1: IVIg versus placebo, Outcome 1: Proportion of participants with an improvement in disability as determined and defined by the study authors

As Van den Berg 1995 included participants that were already treated, we performed a sensitivity analysis to determine to what degree it affected the results. When analysing this outcome without the study of Van den Berg and colleagues, an improvement was reported in four out of 12 IVIg treatments and spontaneous improvements in two out of 12 placebo treatments. A higher proportion of participants improved after IVIg therapy compared with placebo, with a pooled RR of 2.00, but CIs did not exclude the possibility of an effect in favour of placebo (95% CI 0.53 to 7.60; 2 RCTs, 12 participants).

Both comparisons showed wide CIs. Given the small numbers of participants, a clinically relevant effect cannot be confirmed or excluded.

Secondary outcomes

Secondary outcomes were assessed at 28 days after the start of treatment (Azulay 1994; Federico 2000), every month up to four or seven months (Léger 2001), every week up to 28 weeks (Van den Berg 1995), and during the last treatment cycle of the blinded 12‐week period, or earlier in case of deterioration (Hahn 2013).

Induction treatment

Proportion of participants with an improvement in muscle strength as determined and defined by the study authors, assessed between two and six weeks after the last treatment and compared to baseline

Three RCTs involving 27 participants provided data for this outcome and were included in meta‐analysis (Azulay 1994; Federico 2000; Van den Berg 1995). All three trials used only IVIg as the study treatment.

Azulay and colleagues obtained a muscle score by summing the strength assessed in two muscles that had to be impaired and were selected at the beginning of the study. An improvement was defined as an increased muscle score of more than 50% compared to baseline (Azulay 1994). In this study, five participants improved in muscle strength after IVIg and none after placebo. Van den Berg and colleagues assessed 11 different muscles using a handheld dynamometer. An improvement was defined as an increase of 50% or more in at least two muscles, without a decrease of at least 25% in more than one other muscle (Van den Berg 1995). In this study, five out of six participants improved after IVIg and one after placebo. Federico and colleagues asked participants to subjectively rate their strength at the end of a treatment period compared to the beginning (Federico 2000). In this study, 11 out of 16 participants rated their strength as improved after treatment, whereas none of the participants did so after placebo.

Overall, an improvement of muscle strength was reported in 21 out of 27 (78%) IVIg treatments and one out of 27 (4%) placebo treatments. A higher proportion of participants improved after IVIg therapy compared with placebo, with a pooled RR of 11.00 (95% CI 2.86 to 42.25; 3 RCTs, 27 participants; low‐certainty evidence; Analysis 1.2). Analysing these outcome data without Van den Berg 1995 resulted in an RR of 17.00 (95% CI 2.48 to 116.59; 2 RCTs, 21 participants; Figure 4).

1.2. Analysis.

Comparison 1: IVIg versus placebo, Outcome 2: Proportion of participants with an improvement in muscle strength

4.

IVIg versus placebo. Proportion of participants with an improvement in muscle strength. 

Mean change in muscle strength expressed as standardised mean difference (SMD), assessed between two and six weeks after the last treatment and compared to baseline

Mean muscle strength scores were available in three trials assessing 29 participants (Azulay 1994; Federico 2000; Léger 2001). All three trials used only IVIg as the study treatment.

Azulay and colleagues obtained a muscle score by summing the strength expressed in Newtons (N) assessed in two selected muscles that had to be impaired and were selected at the beginning of the study (Azulay 1994). On average, muscle strength improved by 280 N on IVIg and decreased by 31 N on placebo. Federico and colleagues assessed grip strength with a handheld dynamometer (Federico 2000). Grip strength improved by 6.4 kg with IVIg treatment and worsened by 1.0 kg with placebo. The Léger 2001 trial measured the Medical Research Council (MRC) sum score of 28 muscles (maximum score 140). The mean improvement after IVIg treatment was 3 points, whereas participants treated with placebo improved 1 point. The pooled SMD for all studies was 1.13 (95% CI −0.70 to 2.96; 3 RCTs, 29 participants). This indicated that the mean change in muscle strength on IVIg was approximately one standard deviation higher than the mean on placebo, but CI allowed for effects in either direction (Figure 5; Analysis 1.3).

5.

IVIg versus placebo. Change in mean muscle strength in participants without IVIg treatment at study entry.

1.3. Analysis.

Comparison 1: IVIg versus placebo, Outcome 3: Change in mean muscle strength in participants without IVIg treatment at study entry

Proportion of participants with a sustained improvement in disability at 12 months or later as determined and defined by the study authors

None of the trials assessed or reported this outcome.

Proportion of participants in whom at least one conduction block resolved after therapy, assessed between two and six weeks after the last treatment and compared to baseline

All four trials in immunoglobulin‐naive participants (n = 28) assessed and reported this outcome (Azulay 1994; Federico 2000; Léger 2001; Van den Berg 1995). All of these trials used only IVIg as the study treatment.

Resolution of at least one conduction block occurred six times after IVIg and never after placebo (RR 7.00, 95% CI 0.95 to 51.70; 4 RCTs, 28 participants; low‐certainty evidence). Analysing these outcome data without Van den Berg 1995 resulted in an RR of 11.00 (95% CI 0.69 to 175.86; 3 RCTs, 22 participants; Figure 6; Analysis 1.4).

6.

IVIg versus placebo. Proportion of participants in whom at least 1 conduction block resolved after therapy. 

1.4. Analysis.

Comparison 1: IVIg versus placebo, Outcome 4: Proportion of participants in whom at least 1 conduction block resolved after therapy

Maintenance treatment

For the secondary outcomes for studies with participants using immunoglobulin maintenance treatment, there was only one eligible study, involving 44 randomised participants, in which IVIg was used as the only study treatment (Hahn 2013).

Proportion of immunoglobulin‐treated participants with an increase in disability during controlled IVIg/SCIg withdrawal as determined and defined by study authors

Disability was measured using the upper limb portion of the Guy’s Neurological Disability Score. An increase in disability was reported in 17 of 42 evaluable participants (40%) after placebo and seven of 42 participants (17%) after IVIg. A higher proportion of participants deteriorated after switching from IVIg to placebo therapy as compared to those continuing on IVIg, with an RR of 2.43 (95% CI 1.13 to 5.24; 1 RCT, 42 participants; moderate‐certainty evidence; Analysis 1.5).

1.5. Analysis.

Comparison 1: IVIg versus placebo, Outcome 5: Proportion of IVIg‐treated participants with an increase in disability during controlled IVIg withdrawal, as defined and determined by the study authors

Proportion of immunoglobulin‐treated participants with a decrease in muscle strength during controlled IVIg/SCIg withdrawal as determined and defined by study authors

In 42 evaluable participants, muscle strength was measured using the maximal grip strength of the more affected hand (Hahn 2013). The study used a cut‐off of 30% decrease of grip strength in the more affected hand to define a decline, which was a secondary outcome in the trial. Compared to baseline there was a decrease of grip strength in the most affected hand of 30% or more in four of 42 participants (10%) treated with IVIg compared to 20 of 42 participants (48%) treated with placebo. The proportion of participants with a decrease in muscle strength was lower in those receiving IVIg compared to those receiving placebo (RR 0.20, 95% CI 0.07 to 0.54; 1 RCT, 42 participants; moderate‐certainty evidence; Analysis 1.6).

1.6. Analysis.

Comparison 1: IVIg versus placebo, Outcome 6: Proportion of IVIg‐treated participants with a decrease in muscle strength during controlled IVIg withdrawal, as defined and determined by the study authors

Mean change in muscle strength expressed as SMD, assessed between two and six weeks after the last treatment and compared to baseline in participants on maintenance IVIg/SCIg treatment at study entry

Maximal grip strength of the more affected hand increased by 3.75% after IVIg treatment and decreased by 31.38% after placebo treatment. There was an increase in muscle strength, with an SMD of 3.78 in favour of IVIg (95% CI 3.06 to 4.51; 1 RCT, 42 participants; Analysis 1.7).

1.7. Analysis.

Comparison 1: IVIg versus placebo, Outcome 7: Change in mean muscle strength in participants with IVIg treatment at study entry

Induction or maintenance treatment

Frequency of adverse effects attributable to treatment during the whole study period

In one trial side effects were not mentioned (Van den Berg 1995). During the double‐blind periods of the largest study, the rate of adverse events that were possibly caused by study treatment was 10% for IVIg (total of 242 infusions) and 12% for placebo (total of 129 infusions) (Hahn 2013). The report gave the number of participants with serious adverse events, but not adverse events; these data could not be included in the meta‐analysis. Two of 43 participants (4.7%) in this study reported one or more moderate or severe adverse effects after IVIg infusion, compared to 10 of 43 participants (23%) after placebo, regardless of causality. One serious adverse effect (pulmonary embolism) and 100 non‐serious adverse effects were encountered throughout four 12‐week IVIg treatment periods (three open‐label and one double‐blind). The other three studies reported only minor clinical side effects. Cutaneous rash and transient fever were seen in two participants out of five treated with IVIg in Azulay 1994. Federico and colleagues noted minor side effects in 13 of 16 participants treated with IVIg: headache (5), headache and rash (3), rash alone (2), headache and malaise (1), anorexia, chills and fever (1), transient hypertension (1); and in one participant after placebo treatment: headache, fever and chills (Federico 2000). Minor side effects were reported in Léger 2001, but were not attributable to individual participants and could not be included in the analysis. Participants in this study treated with IVIg complained of headache three times, flushing once, shivering twice, fever once, blurred vision twice and eczema once; the only reported side effect after placebo treatment was cold feet, in one participant. In the two meta‐analysed studies, adverse events occurred in 71% of people receiving IVIg and 4.8% of people receiving placebo, with a pooled RR for the development of side effects of 10.33 (95% CI 2.15 to 49.77; 2 RCTs, 21 participants; very low‐certainty evidence) (Analysis 1.8).

1.8. Analysis.

Comparison 1: IVIg versus placebo, Outcome 8: Proportion of participants with side effects attributable to treatment

IVIg versus SCIg

One study compared the continuation of IVIg with a switch to SCIg in nine participants on IVIg maintenance treatment (Harbo 2009). Dynanometric muscle strength was the primary outcome measure; this study did not assess disability. See Table 2.

Primary outcome

Proportion of participants with an improvement in disability between two and six weeks after the last treatment as determined and defined by the study authors and compared to baseline

The primary outcome could not be assessed for this study, as all participants were treated with IVIg at study entry.

Secondary outcomes

Induction treatment

As all participants were known IVIg responders and received maintenance IVIg treatment, it was not possible to assess the secondary outcomes for induction treatment.

Maintenance treatment 

The only secondary outcomes for maintenance treatment available for analysis were mean change in muscle strength and the frequency of adverse effects attributable to treatment during the whole study period. 

Mean change in muscle strength expressed as SMD, assessed between two and six weeks after the last treatment and compared to baseline in participants on maintenance IVIg/SCIg treatment at study entry

Muscle strength was expressed as a combined dynamometric strength score assessed halfway between the two previous regular doses of IVIg for each participant. The strength of five to six affected muscle groups at three joints and at handgrip were taken and expressed as a percentage of normal strength. Selected muscles had to be impaired and were selected at the beginning of the study. The combined dynamometric strength score at baseline was similar in the two treatment arms: 69.5 ± 14% of normal strength in SCIg‐treated participants and 69.9 ± 18% of normal strength in IVIg‐treated participants. There was an increase of 3.6% (95% CI −3.6% to 10.9%) during SCIg treatment and an increase of 4.3% (95% CI −1.3% to 10.0%) during IVIg treatment. The evidence on muscle strength after SCIg compared to IVIg maintenance treatment was too uncertain for any conclusions to be drawn (SMD 0.08, 95% CI −0.84 to 1.00; 1 RCT, 9 participants; very low‐certainty evidence; Analysis 2.1).

2.1. Analysis.

Comparison 2: IVIg versus SCIg maintenance, Outcome 1: Change in mean muscle strength in participants with IVIg treatment at study entry

Frequency of adverse effects attributable to treatment during the whole study period

During SCIg treatment (84 days, range 36 to 112), six of nine participants reported adverse effects, all of which were reported to be mild and transient. One participant had sustained erythema and oedema at the injection sites, which required dose adjustments. Three of nine IVIg‐treated participants had a skin rash, phlebitis, or an infected implantable port. Evidence about adverse effects was too uncertain for any conclusions to be drawn (RR 0.50, 95% CI 0.18 to 1.40; 1 RCT, 9 participants; very low‐certainty evidence; Analysis 2.2).

2.2. Analysis.

Comparison 2: IVIg versus SCIg maintenance, Outcome 2: Proportion of participants with side effects attributable to treatment

Discussion

Summary of main results

The review included six RCTs involving a total of 90 randomised participants with MMN.

Induction treatment with immunoglobulins 

In immunoglobulin‐naive participants, low‐certainty evidence indicates that more people may experience an improvement in disability with initiation of IVIg treatment than with placebo; however, this finding is based on a small number of participants, and the estimated CIs are wide and also include the possibility of no effect. There is no evidence to support the use of SCIg as induction treatment in MMN. 

Muscle strength improved spontaneously in one out of 27 participants (4%), but treatment with IVIg increased this chance of improvement to 78% (21 out of 27 participants). This absolute risk difference of 74% gives a number needed to treat for an additional beneficial outcome of 1.4. The percentage of participants with an improvement in disability was 39% with IVIg treatment and 11% with placebo. This difference in effect when outcome was measured at the disability level or at the impairment level is not surprising, since only 34 participants were studied. The power of the included studies together to detect a significant proportional difference at disability level between treated and untreated participants is low. For this reason, we graded the certainty of evidence for both disability and muscle strength as low.

The mean muscle strength showed no important change after IVIg treatment compared to placebo, but the result was imprecise and allowed for effects in either direction.

None of the trials reported or assessed the number of participants with a sustained improvement in disability (at 12 months). Low‐certainty evidence suggests that a higher proportion of participants had at least one conduction block resolved after IVIg treatment compared with placebo; the CIs also allowed for the possibility of no effect.

Maintenance treatment with immunoglobulins

The need for ongoing treatment in most people with MMN was illustrated by Hahn and colleagues, as 40% of placebo participants deteriorated in disability, and 48% deteriorated in muscle strength (grip strength) by 30% or more in the more affected hand, based on moderate‐certainty evidence (Hahn 2013). When decline in participants' ability to perform activities of daily living was also considered, the proportion of untreated participants on placebo who deteriorated was 86%. This emphasises that the disability scale used (Guy's Neurological Disability Score) and grip strength do not capture all changes that can interfere with daily living. Nevertheless, it is also important to note that 14% of participants on placebo did not experience deterioration, although the 12‐week placebo period might have been too short for this to be experienced by some participants.

SCIg might be an alternative to IVIg maintenance treatment. A single trial assessed the feasibility of switching participants from IVIg to SCIg. The trial did not address disability. Muscle strength remained stable in participants switching from IVIg maintenance treatment to SCIg, but due to the very low certainty of the evidence we cannot draw any conclusions (Harbo 2009).

Adverse events

The evidence for all adverse events estimates was of very low certainty. Based on reported data of 21 participants treated with IVIg that could be meta‐analysed, 71% of participants experienced any adverse events during the trial, compared to 4.8% receiving placebo. One participant treated with IVIg had a pulmonary embolism, which was regarded as a serious adverse event.

In the trial comparing IVIg with SCIg, mild transient side effects were reported in 33% of participants treated with IVIg and 66% of those treated with SCIg. SCig side effects were all related to injection‐site reactions, with erythema, itch, and tenderness.

Cost

Cost‐utility analyses of immunoglobulins used in MMN have not been published. A cost‐utility study has been done in chronic inflammatory demyelinating polyradiculoneuropathy (CIDP), which showed that compared with prednisolone, IVIg is likely to be cost‐effective in this group (McCrone 2003). However, the analyses also showed that this cost‐effectiveness was greatly dependent on the price and the amount of IVIg administered. Furthermore, corticosteroids are not used in MMN. Given these factors, it is uncertain whether the outcome of this study can be extrapolated to the MMN population.

Overall completeness and applicability of evidence

Not all outcomes on efficacy were addressed in this review. One trial documented patient global assessment of change, for which there was no change between assessments in those receiving IVIg compared to reports of minimal worsening to much worse in those receiving placebo (Hahn 2013). When comparing IVIg to SCIg treatment (Harbo 2009), four out of nine cases preferred SCIg due to lack of end‐of‐dose deterioration and the ability to receive treatment at home, compared to two participants who preferred IVIg due to the avoidance of several treatments required in one week. The use of hyaluronidase in combination with SCIg has been reported to enable a reduction of SCIg dosing frequency, which may improve patient preference (Herraets 2019), but more trial data are required to define which therapy is considered superior.

No long‐term outcomes were reported, and data from small RCTs do not allow adequate assessment of adverse events.

Certainty of the evidence

Induction treatment with immunoglobulins

We downgraded the certainty of the evidence for all outcomes to low because of imprecision due to study limitations, small sample sizes, and wide CI. The CI for disability between two and six weeks after the last treatment encompassed both a clinically relevant effect in favour of IVIg and little or no effect. The power of the included studies to detect a significant proportional difference at disability level between treated and untreated participants was low.

In clinical phase III trials, the primary outcome should be disability and not impairment, as the primary question to be answered is whether a person benefits from a particular treatment. From this point of view, our evidence is less certain that people with MMN will benefit from IVIg treatment. However, since weakness is the only determinant of disability in people with MMN, it is to be expected that in those whose muscle strength improves after IVIg treatment, disability will improve as well.

Another reason for the divergent results between disability and muscle strength could be the fact that the disability scales used in the three trials are not reliable and tailored for use in people with MMN and are unresponsive to change. Three different disability scales were used: the modified Norris scale (Azulay 1994), the modified Rankin scale (Van den Berg 1995), and a self‐evaluation scale scored from zero (normal) to five (impossible) for five motor activities of daily life, chosen for each participant together with the examiner at baseline (Léger 2001). The Norris scale, which evaluates limb function, was designed for amyotrophic lateral sclerosis (ALS) assessment and is a combined disability and impairment scale (Norris 1974). The combination of disability and impairment in one scale potentially confuses the clinical interpretation of changes on such a scale and should be avoided. Furthermore, this scale has not been evaluated for use in people with MMN. The modified Rankin score measures disability, and has been shown to be reliable and valid in assessing people with neuropathy (de Haan 1993; Molenaar 1998; Molenaar 1999); however, whether this is also true for people with MMN with predominantly upper limb and focal involvement has not been investigated. The disability scale used by Léger and colleagues seems to be more tailored for use in people with MMN. However, this scale has not been formally evaluated for reliability, validity, responsiveness, and inter‐ and intrarater variability in people with MMN (Léger 2001).

The results from the other outcomes (i.e. mean change in muscle strength, proportion of participants with an improvement in disability at 12 months or later, and proportion of participants in which at least one conduction block has resolved after therapy) all supported the efficacy of IVIg in MMN for induction treatment, but had the same limitations, that is small numbers of participants or events and wide CIs. 

As previously stated, there is no evidence to support the use of SCIg as induction treatment in MMN. 

Maintenance treatment with immunoglobulins

The Hahn 2013 trial, which examined the need for ongoing IVIg treatment in people with MMN, was the largest study in the review, with 44 randomised participants. The data, whilst imprecise, showed clear effects on disability and strength, and we considered the certainty of evidence for these effects to be moderate. We considered the 12‐week wash‐out period to be short, but did not consider this to be a serious study limitation.

Muscle strength remained stable in participants switching from IVIg maintenance treatment to SCIg in the one RCT evaluating SCIg in MMN (Harbo 2009). The major limitation of this study was the short study period, in which an IVIg carry‐over effect could have occurred. Harbo 2009 included only nine participants, and we downgraded the certainty of the evidence twice for very serious imprecision.

Adverse events

Although adverse events were reported in the included studies, the estimate of the occurrence and severity of adverse events with IVIg and SCIg is better made based on large observational studies with long‐term follow up rather than small trials, including studies on more prevalent diseases such as CIDP. Non‐RCT evidence on adverse events is presented in the Agreements and disagreements with other studies or reviews section of the review.

The certainty of the evidence is mostly limited due to the small sample size of the included trials. However, the trials provided mostly homogenous results, in which all relevant outcomes suggested efficacy of IVIg in MMN, albeit with low certainty.

Potential biases in the review process

Due to the rarity of MMN, the number of participants included in each trial was low. All of the included trials used a cross‐over design. Such a trial design is suitable for evaluating interventions with a temporary effect in the treatment of stable, chronic conditions, for which MMN would be applicable. Carry‐over of treatment effect is a potential bias of such studies; this was mitigated against in each trial through the inclusion of wash‐out periods. Reporting of the cross‐over trials was variable, as some studies did not provide individual data or sufficient details on used statistical methods to confirm that appropriate paired analysis was performed, especially for the proportional outcomes. For this reason we were not able to perform a paired analysis for meta‐analysis, thus we analysed results from treatment versus placebo groups as if the trial were a parallel‐group trial comparing the two. This approach gives rise to a unit of analysis error, which resulted in relatively broad CIs. However, the analysis is conservative, and thus the statistically significant findings discussed in the results are under‐weighted.

Agreements and disagreements with other studies or reviews

Induction treatment with immunoglobulins

There are many open and uncontrolled studies of IVIg initiation in MMN. The reported response to IVIg in larger retrospective studies varies between 67% and 94% (Azulay 1997; Bouche 1995; Cats 2010; Léger 2008; Nobile‐Orazio 2002; Terenghi 2004), which is comparable with the rate of muscle strength improvement found in this review.

Maintenance treatment with immunoglobulins 

The need for ongoing treatment in most participants with MMN was illustrated by Hahn 2013. This is in line with the largest retrospective study, including 88 people with MMN, in which four IVIg responders (5%) did not require ongoing treatment (Cats 2010). In another study involving 40 participants, IVIg withdrawal was possible in 22% of participants (Léger 2008). Nevertheless, it is also important to note that 14% of participants did not experience deterioration, although the 12‐week placebo period might have been too short for some participants to experience deterioration. In other chronic inflammatory neuropathies, such as CIDP, periodic immunoglobulin withdrawals are recommended to determine immunoglobulin dependency. Whether and when IVIg withdrawals should be attempted in people with MMN is unknown.

Muscle strength remained stable in participants switching from IVIg maintenance treatment to SCIg in the included RCT of SCIg maintenance treatment in MMN (Harbo 2009). In an open‐label, prospective study, four of five participants remained stable after switching to SCIg when SCIg was administered in equal monthly doses to the individual pre‐study IVIg dose (Eftimov 2009). Participants who received lower SCIg dosages compared to pre‐study IVIg dose deteriorated, suggesting that at least equal SCIg doses are needed to maintain clinical efficacy of immunoglobulin therapy. In a 2014 prospective observational study, 21 of 22 people (95%) with MMN and wearing‐off signs during maintenance IVIg treatment were switched to SCIg and remained stable in disability (Cocito 2014).

Adverse events

Only one serious side effect was reported in this review, but mild, transient side effects were reported in 71% of those treated with IVIg, which is more than has been reported in the literature on non‐randomised studies (Duhem 1994; Martin 2000; Stiehm 1996).

Most side effects of IVIg are mild and transient, and are estimated to occur in 1% to 15% of infusions (Dalakas 2003; Duhem 1994; Eijkhout 2002; Stiehm 1996; Wittstock 2003). Rashes, chills, fever, mild hypo‐ or hypertension, nausea, malaise, headache, and mild arthralgias seem to occur most frequently. Occasionally, a short‐lasting aseptic meningitis may occur without or with only minimal long‐term sequelae. The most worrisome, albeit rare, complications are severe and potentially fatal anaphylactic shock, thromboembolic events, and temporary renal impairment, which is relatively common in individuals with pre‐existing renal disease (Dalakas 2003; Pierce 2003). The exact incidence of these severe side effects is unknown (Dalakas 2003), although one study estimated the risk of thromboembolic events at 1% of infusions (Rajabally 2011). A more recent study demonstrated a thromboembolic event odds ratio of 53.6 in people on IVIg, compared to 7.6 in the general population (Kapoor 2020). In an ongoing postmarketing clinical pharmacovigilance study, an adverse reaction rate of less than 0.5% for more than 26,000 infusions and less than 4% in 2554 people has been reported (Martin 2000).

Alternatively, SCIg more often leads to local adverse reaction, with a reported frequency of adverse events of 17% in a large study of people with primary immunodeficiencies (Gardulf 1995). These local adverse reactions are usually minor, generally considered acceptable, and present especially when initiating treatment (Cocito 2014; Eftimov 2009; Harbo 2009). However, systemic adverse events are less frequent compared to IVIg. Based on these recent small studies, there is limited evidence that SCIg can be used as a substitute for maintenance IVIg treatment. People with poor venous access and recurring systemic adverse reaction after IVIg might be particularly good candidates for this alternative, but more evidence is needed to confirm the efficacy of SCIg.

Authors' conclusions

Implications for practice.

Evidence from small randomised controlled trials (RCTs) shows that there may be an improvement in disability in people with multifocal motor neuropathy (MMN) after intravenous immunoglobulin (IVIg) compared with placebo; however, our confidence in the estimate will be increased with further studies. There may be an improvement in muscle strength. As weakness is the only determinant of disability in people with MMN, increase in muscle strength, and especially grip strength, can be used to assess individual response to IVIg. However, improvement in disability is the primary goal of treatment and the main parameter on which to decide whether to continue treatment. People whose complaints do not interfere with daily life are often closely followed without treatment, given the minimal expected benefit of treatment, the burden of regular infusions, and high treatment costs.

In most, but not all, people with MMN receiving maintenance IVIg treatment, IVIg withdrawal leads to deterioration in disability and muscle strength. There is very limited evidence from a single, small RCT on the use of subcutaneous immunoglobulin (SCIg) as an alternative to IVIg maintenance treatment. The evidence on adverse effects with IVIg compared to placebo or SCIg is very limited.

Implications for research.

In addition to the limited evidence of IVIg efficacy in MMN from RCTs, increase of muscle strength has been confirmed by larger retrospective cohort studies. RCTs with large numbers of newly diagnosed patients to confirm improvement in disability are unlikely given the rarity of disease and the current evidence of IVIg efficacy in increasing muscle strength. In general, for cross‐over trials in rare diseases such as MMN, it is recommended that sufficient data be provided and that analysis be performed in such a way that allows future pooling of paired data, which was not possible in this meta‐analysis. Prospective international registries could be useful in providing evidence based on larger numbers of people, focusing on improvement in disability. Newer and more responsive disability scales such as the Rasch Overall Disability Scale (RODS) might be better than the disability scales used in previous RCTs. More importantly, a prospective registry should identify predictors of ongoing disease activity and provide guidance on how to individualise maintenance treatment.

SCIg might be an alternative to IVIg maintenance treatment, but confirmation from RCTs with larger numbers is needed.

What's new

Date	Event	Description
20 April 2021	New citation required and conclusions have changed	Searches identified two new randomised controlled trials for inclusion: a placebo‐controlled study of intravenous immunoglobulin withdrawal and a comparison of subcutaneous immunoglobulin and intravenous immunoglobulin as maintenance treatment.
20 April 2021	New search has been performed	The strategies and randomised controlled trial filters have been updated. The title was changed from 'Intravenous immunoglobulin for multifocal motor neuropathy', since subcutaneous administration has now been evaluated in trials. Two new authors, Stephen Keddie and Ruth Brassington, were added. The Methods were updated. We included summary of findings tables and assessed the certainty of evidence using the GRADE approach.

History

Protocol first published: Issue 4, 2003
Review first published: Issue 2, 2005

Date	Event	Description
28 October 2008	Amended	Converted to new review format
16 March 2007	New search has been performed	The searches were updated in March 2007, but no new trials were identified. There was also no new non‐randomised evidence.
27 January 2005	New citation required and conclusions have changed	Substantive amendment

Appendices

Appendix 1. Cochrane Neuromuscular Disease Group Specialised Register (CRS) search strategy

1 multifocal NEAR5 "motor neuropath*" AND INREGISTER 39

2 "multifocal motor" NEAR5 neuropath* AND INREGISTER 38

3 "motor neuron* syndrome*" AND INREGISTER 5

4 "anti GM1" and (neuropath* or polyneuropath* or polyradiculoneuropath*) AND INREGISTER 8

5 #1 or #2 or #3 or #4 44

6 ((intravenous OR IV OR subcutaneous) NEAR3 immunoglobulin*) AND INREGISTER 266

7 ((intravenous OR IV OR subcutaneous) NEAR3 "immune globulin*") AND INREGISTER 27

8 ivig AND INREGISTER 204

9 #6 or #7 or #8 303

10 #5 and #9 37

Appendix 2. CENTRAL (CRS Web) search strategy

1 multifocal NEAR5 "motor neuropath*" AND CENTRAL:TARGET 64

2 "multifocal motor" NEAR5 neuropath* AND CENTRAL:TARGET 62

3 "motor neuron* syndrome*" AND CENTRAL:TARGET 19

4 "anti GM1" and (neuropath* or polyneuropath* or polyradiculoneuropath*) AND CENTRAL:TARGET 13

5 #1 or #2 or #3 or #4 85

6 ((intravenous OR IV OR subcutaneous) NEAR3 immunoglobulin*) AND CENTRAL:TARGET 2272

7 ((intravenous OR IV OR subcutaneous) NEAR3 "immune globulin*") AND CENTRAL:TARGET 286

8 ivig AND CENTRAL:TARGET 1417

9 #6 or #7 or #8 2627

10 #5 and #9 57

Appendix 3. MEDLINE (OvidSP) search strategy

Ovid MEDLINE(R) and Epub Ahead of Print, In‐Process, In‐Data‐Review & Other Non‐Indexed Citations and Daily <1946 to April 19, 2021>

1 ((randomized controlled trial or controlled clinical trial).pt. or (randomi?ed or placebo or randomly or trial or groups).ab. or drug therapy.fs.) not (exp animals/ not humans.sh.) (4338322)

2 (((multifocal adj5 motor neuropath$) or motor neuron$ syndrome$ or multifocal motor) adj5 neuropath$).tw. or (anti‐GM1 and (neuropath$ or polyneuropath$ or polyradiculoneuropath$)).mp. (1136)

3 Immunoglobulins, Intravenous/ or (((intravenous or IV or subcutaneous) adj3 (ig or immunoglobulin$)) or ivig).mp. (22931)

4 1 and 2 and 3 (230)

Appendix 4. Embase (OvidSP) search strategy

Embase <1974 to 2021 Week 15>

1 (crossover‐procedure or double‐blind procedure or single‐blind procedure or randomized controlled trial).sh. or (random$ or crossover$ or cross over$ or placebo$ or (doubl$ adj blind$) or allocat$).tw,ot. or trial.ti. (2125563)

2 (exp animal/ or exp invertebrate/ or animal.hw. or non human/ or nonhuman/) not (human/ or human cell/ or human tissue/ or normal human/) (6711823)

3 1 not 2 (1880593)

4 limit 3 to (conference abstracts or embase) (1565008)

5 ((multifocal adj5 motor neuropathy$) or (multifocal motor adj5 neuropath$) or (anti‐GM1 and (neuropath$ or polyneuropath$ or polyradiculoneuropath$))).mp. or motor neuron$ syndrom$.tw. (2384)

6 immunoglobulin/iv, sc (32730)

7 immunoglobulin/ and (intravenous drug administration/ or subcutaneous drug administration/) (6459)

8 (((intravenous or IV or subcutaneous) adj3 (ig or immunoglobulin$)) or ivig).mp. (34402)

9 6 or 7 or 8 (61589)

10 4 and 5 and 9 (109)

Appendix 5. ClinicalTrials.gov search strategy

Advanced Search

Condition or disease: Multifocal Motor Neuropathy

Study type: Interventional Studies (Clinical Trials)

Intervention/treatment: Immunoglobulin* OR Immune Globulin*

9 Studies

Appendix 6. WHO ICTRP search strategy

Advanced Search

Multifocal Motor Neuropathy in the Condition

Immunoglobulin* OR Immune Globulin* in the Intervention

Recruitment Status is ALL

16 records for 9 trials found!

Data and analyses

Comparison 1. IVIg versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Proportion of participants with an improvement in disability as determined and defined by the study authors	3	36	Risk Ratio (M‐H, Fixed, 95% CI)	3.00 [0.89, 10.12]
1.1.1 Without study van den Berg 1995	2	24	Risk Ratio (M‐H, Fixed, 95% CI)	2.00 [0.53, 7.60]
1.1.2 Study van den Berg 1995	1	12	Risk Ratio (M‐H, Fixed, 95% CI)	7.00 [0.44, 111.91]
1.2 Proportion of participants with an improvement in muscle strength	3	54	Risk Ratio (M‐H, Fixed, 95% CI)	11.00 [2.86, 42.25]
1.2.1 Without study van den Berg 1995	2	42	Risk Ratio (M‐H, Fixed, 95% CI)	17.00 [2.48, 116.59]
1.2.2 Study van den Berg 1995	1	12	Risk Ratio (M‐H, Fixed, 95% CI)	5.00 [0.81, 31.00]
1.3 Change in mean muscle strength in participants without IVIg treatment at study entry	3	58	Std. Mean Difference (IV, Fixed, 95% CI)	1.13 [‐0.70, 2.96]
1.4 Proportion of participants in whom at least 1 conduction block resolved after therapy	4	56	Risk Ratio (M‐H, Fixed, 95% CI)	7.00 [0.95, 51.70]
1.4.1 Without study van den Berg 1995	3	44	Risk Ratio (M‐H, Fixed, 95% CI)	11.00 [0.69, 175.86]
1.4.2 Study van den Berg 1995	1	12	Risk Ratio (M‐H, Fixed, 95% CI)	3.00 [0.15, 61.74]
1.5 Proportion of IVIg‐treated participants with an increase in disability during controlled IVIg withdrawal, as defined and determined by the study authors	1	84	Risk Ratio (M‐H, Fixed, 95% CI)	2.43 [1.13, 5.24]
1.6 Proportion of IVIg‐treated participants with a decrease in muscle strength during controlled IVIg withdrawal, as defined and determined by the study authors	1	84	Risk Ratio (M‐H, Fixed, 95% CI)	0.20 [0.07, 0.54]
1.7 Change in mean muscle strength in participants with IVIg treatment at study entry	1	84	Std. Mean Difference (IV, Fixed, 95% CI)	3.78 [3.06, 4.51]
1.8 Proportion of participants with side effects attributable to treatment	2	42	Risk Ratio (M‐H, Fixed, 95% CI)	10.33 [2.15, 49.77]

Comparison 2. IVIg versus SCIg maintenance.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Change in mean muscle strength in participants with IVIg treatment at study entry	1	18	Std. Mean Difference (IV, Fixed, 95% CI)	0.08 [‐0.84, 1.00]
2.2 Proportion of participants with side effects attributable to treatment	1	18	Risk Ratio (M‐H, Fixed, 95% CI)	0.50 [0.18, 1.40]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Azulay 1994.

Study characteristics
Methods	Double‐blind, cross‐over design
Participants	Inclusion criteria: study included 12 people with "motor neuron syndromes" and high titre GM1 antibodies, of whom a subgroup of 5 with electrophysiological evidence of motor conduction block were eligible for this review Study exclusion criteria: evidence of motor neuron disease No immunosuppressive drug was administered during the trial or in the 3 months before Number of participants assessed for eligibility: not stated Number (with MMN) randomised: 5; 5 IVIg phase, 5 placebo phase Sex: 4 men, 1 woman Age (years): range 40 to 63, mean 48.4 Disease duration: 16 months to 10 years Recruitment: not described Setting: France (Marseilles)
Interventions	IVIg 0.4 g/kg/day for 5 consecutive days versus placebo (saline). The 2nd treatment was administered 8 weeks after the first treatment. Source IVIg: National Center of Blood Transfusion France
Outcomes	Muscle strength (tested by a computerised analyser (Myocomp, Meditronic Instrument)). Average strength (in N) of a maximal isometric contraction lasting 10 seconds for 2 selected muscles which were deemed clinically impaired but with a degree of preserved contraction. The measurement was always performed by the same examiner. A muscle score was obtained by summing the strength values. A participant was considered improved if the muscle score increased more than 50% of the initial value. Norris scale (disability) (score from 63 to 0) ('improved' not defined) Neurophysiological measurements GM1 antibodies Time points of assessment during each phase: before drug administration and 5, 28, and 56 days after Side effects (assessed throughout the session by clinical examination and the monitoring of blood pressure, heart rate, and routine blood tests)
Funding	Not reported
Potential conflicts of interest	None stated.
Notes	Study dates: not stated Wash‐out period: 8 weeks
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Each subject was given each of the two treatments, which were assigned to the 12 according to a crossover design under double‐blind conditions" Comment: randomisation procedure not described; it is unclear whether the order of treatments was randomised
Allocation concealment (selection bias)	Unclear risk	Quote: "Each subject was given each of the two treatments, which were assigned to the 12 according to a crossover design under double‐blind conditions" Comment: measures taken to ensure concealment of allocation are not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "Each subject was given each of the two treatments, which were assigned to the 12 according to a crossover design under double‐blind conditions" and "IVIg of saline was administered intravenously by a nurse who was independent from the study. IVIg was produced by the National Centre of Blood Transfusion and the placebo by the hospital pharmacy" and "Only minor clinical side effects occurred. Cutaneous rash and transient fever were noted in two. No biological side effects occurred. Thus, blindness of the study was not jeopardized." Comment: it was stated that participants were blinded to treatment. Saline was produced by the hospital pharmacy and IVIg by a national blood transfusion centre. They may have looked different. There are no details on whether the administered treatment was properly concealed. Although treatment was administered by an independent nurse, the nurse does not appear to have been blinded in this study, which could unblind participants.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "Each subject was given each of the two treatments, which were assigned to the 12 according to a crossover design under double‐blind conditions" Comment: assessors were reported as blinded; however, the procedure was not described. Treatment providers did not perform assessments.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All outcomes prespecified in the methods section were available for all participants.
Selective reporting (reporting bias)	Low risk	All outcomes prespecified in the methods section were reported in the results section for all participants.
Other bias	Unclear risk	Selection of participants is not clearly described. For example, whether MMN patients had been treated with IVIg in the past is not clear. This may have led to recruitment bias. Cross‐over study with 8‐week wash‐out. Carry‐over effects possible. Comment: unclear whether 8‐week wash‐out period is sufficient for all patients, although in clinical practice most receive IVIg on a monthly basis due to a 4‐week duration of effect.

Federico 2000.

Study characteristics
Methods	Double‐blind, cross‐over design
Participants	Inclusion criteria: MMN according to well‐defined criteria (asymmetric lower motor neuron syndrome without sensory or bulbar signs and evidence of partial conduction block (> 30%) in motor nerves with normal sensory responses) Exclusion criteria: ALS, paraproteinaemic neuropathy, CIDP Number of participants assessed for eligibility: 18 Number randomised: 16; 16 IVIg phase, 16 placebo phase Sex: 15 men, 1 woman Age (years): range 26 to 68, mean 38.9 ± 2.8 years Duration of symptoms (years) 0.5 to 15 (mean 5.3 ± 1.2) Recruitment: prospective recruitment from neuromuscular clinic Setting: neuromuscular clinics of the University of Calgary, Alberta, and the University of Western Ontario, London
Interventions	IVIg 0.4 g/kg/day for 5 days versus placebo (5% dextrose or 0.9% saline i.v.) Source IVIg: Bayer Inc, Toronto, Canada After the 1st phase of treatment, participants were crossed over if unchanged or deteriorated by day 28; if not, participants did not enter the 2nd arm until outcomes were back to baseline. Additional allowed therapies not mentioned.
Outcomes	Subjective assessment of muscle strength before treatment and after 28 days Change in NDS modified by Dick and colleagues (summed score of strength in 26 muscle groups (0 normal, 1 mildly weak, 2 moderately weak, 3 severely weak, 4 paralysis); summed score of sensation in all 4 limbs and assessment of tendon reflexes (0 normal, 1 reduced, 2 absent) and tremor (0 absent, 1 present)) Grip strength in both hands Neurophysiological measurements (change in conduction block in affected nerves) Time points for neurological examination, NDS, and grip strength on days 1, 5, and 28
Funding	Supported by the Bayer‐Canadian Red Cross Research and Development Fund and the Muscular Dystrophy Association of Canada
Potential conflicts of interest	1 of the authors (DWZ) was a Medical Scholar of the Alberta Heritage Foundation for Medical Research.
Notes	Study dates: 1992 to 1998 Predefined set of rules to determine length of wash‐out
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Treatment order was assigned to each patient individually using a random number generator."
Allocation concealment (selection bias)	Unclear risk	Comment: allocation concealment not described
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The participating, evaluating neurologist, electromyographer, and treating nurses were all masked to the actual treatment protocol. Only the study coordinator and blood bank personnel were nonblinded. These nonblinded individuals had no role in the patient evaluation or data analysis." and "Treatment blinding was ensured by having the infusions delivered from the transfusion services in identical opaque transfer packs"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "The participating, evaluating neurologist, electromyographer, and treating nurses were all masked to the actual treatment protocol. Only the study coordinator and blood bank personnel were nonblinded. These nonblinded individuals had no role in the patient evaluation or data analysis." and "Treatment blinding was ensured by having the infusions delivered from the transfusion services in identical opaque transfer packs" and "neurologic examination in each patient was performed by the same blinded observer on days 1, 5, and 28" Quote: “…the adverse effects produced by IVIg in this study were not unexpected and may have led to unmasking of treatment order in some patients….To eliminate bias, evaluating neurologists did not inquire about treatment‐related side effects. Thus, measures of grip strength and NDS as well as electrophysiologic studies carried out examiners blinded to these side effects were unlikely to be affected by observer bias.”
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "All completed both treatment arms; however, follow‐up data were incomplete as follows: NDS (14/16), grip strength (15/16), electrophysiologic measures (10/16). The reason for the incomplete follow‐up data was that examining neurologists did not complete all measurements on a few occasions." Authors state that this was “not related to patient profile, order of treatment, response to treatment, adverse effects, examiner unblinding, or other sources of bias”.
Selective reporting (reporting bias)	Low risk	All outcomes prespecified in the methods section were reported in the results section.
Other bias	Unclear risk	A predefined set of rules to determine the length of wash‐out was used. This led to a variable wash‐out period (from 38 to 120 days), which in turn may have resulted in a variable carry‐over effect. Additional other (immunomodulatory) treatment was not addressed.

Hahn 2013.

Study characteristics
Methods	Randomised withdrawal double‐blind, placebo‐controlled, cross‐over study
Participants	Inclusion criteria: MMN according to American Association of Electrodiagnostic Medicine criteria. Participants were on a stable regimen of IVIg for over 3 months at 0.4 g/kg to 2.0 g/kg body weight every 2 to 5 weeks. Exclusion criteria: not stated Number of participants assessed for eligibility: 50 Participants randomised: 44 (42 completed both phases) Number of participants analysed: 44 Age (years): range 31 to 72, mean 51.64 (SD 10.25) Sex: 30 men, 12 women Disease duration: participants had been receiving IVIg for a median of 6 months before enrolment (range 2.8 to 184.6 months) Setting: 17 medical centres in the USA, Canada, and Denmark. Participants in the USA or Canada had the option of home infusion given by nurse.
Interventions	After 12‐week open‐label IVIg, each participant was randomised to IVIg or placebo for 12 weeks, then the reverse. IVIg in pre‐study dose and interval: AMMAGARD LIQUID/KIOVIG (Baxter Healthcare Corporation, Westlake Village, CA, USA), a 10% liquid IgG preparation. Pre‐IVIg dosing levels were maintained. Those treated every 5 weeks pre‐study were changed to 4 weeks at equivalent dose per kg/week, given over ≤ 5 days at rates of up to a maximum of 5 mL/kg body weight/hour. Placebo: 0.25% human albumin prepared with normal saline from BUMINATE 25%, or human albumin 200 g/L Baxter Solution for Infusion
Outcomes	Guy's Neurological Disability Score (upper limb portion, which ranges from 0 (no problem) to 5 (unable to use for purposeful movements) Grip strength (DnEx dynamometer) Overall Disability Sum Score Patient global impression of change score Visual analogue scale of disability Side effects "Assessments took place midway through the last treatment cycle in each study period: on day 8 (±1 day) of a 2‐week treatment interval, and on day 15 (±2 days) for those treated every 3 or 4 weeks."
Funding	The study was sponsored by the Baxter Healthcare Corporation.
Potential conflicts of interest	3 of the authors were employees of the study sponsor. An independent review of the manuscript was performed by experts in the field who had full access to study data.
Notes	Study dates: August 2008 to August 2011 ClinicalTrials.gov id: NCT00666263 "Washout" period consisted of open‐label IVIg infusions. The study used "accelerated" switch in case of deterioration interfering with daily activities or if grip strength declined by 50% or more in the more affected hand.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: “On completion of the initial open‐label period of IVIG, each study subject was assigned a randomization code by the sponsor. An unblinded central pharmacist dispensed blinded investigational product in opaque infusion bags.” (Full details of randomisation procedure available in the study protocol in appendix to the paper.) “Randomisation codes assigned in chronological order.”
Allocation concealment (selection bias)	Low risk	Quote: “On completion of the initial open‐label period of IVIG, each study subject was assigned a randomization code by the sponsor. An unblinded central pharmacist dispensed blinded investigational product in opaque infusion bags.” (Full details of randomisation procedure available in the study protocol in appendix to the paper.) “Randomisation codes assigned in chronological order.”
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: “All subjects, investigators and the sponsor remained blinded during the entire study.”
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: “All subjects, investigators and the sponsor remained blinded during the entire study.”
Incomplete outcome data (attrition bias) All outcomes	Low risk	2 withdrawals due to adverse events on IVIg, 1 withdrawal due to relocation after the end of both blinded phases. All randomised participants were included in the intention‐to‐treat analyses. Quote: "This was repeated as a sensitivity analysis using the worst imputation technique to account for missing data."
Selective reporting (reporting bias)	Low risk	All outcomes prespecified in the methods section were reported in the results section for all participants. No apparent selective reporting based on comparison with ClinicalTrials.gov entry and protocol (available in an appendix)
Other bias	Unclear risk	Quote: “trial design allowed a switch to open‐label IVIG when substantial functional deterioration occurred during blinded treatment. By this design, the period of blinded placebo was shortened in the subjects (69%) who prematurely switched to open‐label IVIG treatment and lasted a median 28 days (range: 7–86) compared to a median 84 days (range: 13–91) of blinded IVIG.” The main goal of this withdrawal trial was to determine whether there was an ongoing need for IVIg treatment. All participants were on a stable IVIg regimen at study entry, and all received open‐label IVIg infusions during the "washout" period prior to switching to the next intervention. The blinded placebo period lasted 12 weeks, which is short and could have resulted in a carry‐over effect from IVIg.

Harbo 2009.

Study characteristics
Methods	Randomised (block), single‐blind (physician blinding), cross‐over design
Participants	Inclusion criteria: definite or probable MMN according to EFNS/PNS criteria. On IVIg maintenance therapy Exclusion criteria: systemic or malignant disease, allergic reaction to IVIg, pregnancy, breastfeeding, age under 18 or above 80, other immune therapies within the last 6 months, and conditions associated with prolonged coagulation times Number of participants assessed for eligibility: 14 (4 refused participation. 1 person did not respond to initial IVIg, so was excluded.) Number randomised and completing: 9; 9 IVIg phase, 9 SCIg phase Sex: 4 men, 5 women Age (years): range 28 to 63, mean 49 Symptom duration (years): between 0.5 and 22.3 (mean 12.3) Setting: Denmark (Aarhus). Participants had a first treatment in a hospital stay of 1 to 3 days. Afterwards they self‐administered SCIg at home supervised by telephone calls from the study nurse.
Interventions	SCIg 0.2 g/kg to 0.5 g/kg per week versus IVIg 0.6 g/kg to 2.1 g/kg per month. IVIg doses were individually tailored in a pre‐study period. SCIg dose was equivalent to IVIg dose in the individual. The 1st treatment period of IVIg or SCIg was equal to 3 IVIg treatment intervals of 18 days to 56 days; a wash‐out period preceded the 2nd treatment period. Source of SCIg: Subcuvia, 160 mg/mL, Baxter A/S, Vienna, Austria; source of IVIg: Endobulin, 50 mg/mL, Baxter A/S, Vienna, Austria.
Outcomes	Combined dynamometric strength score in 5 to 6 affected muscle groups Grip strength (Jamar dynamometer) MRC sum score 9‐hole peg test 10‐metre walking test GM1 antibodies Neurophysiological measurements 36‐Item Short Form Health Survey (SF‐36) Outcomes measured at baseline and at the end of each treatment period. "The period of SCIG infusions was 84 days (range 36–112) and the period from baseline to final evaluation 105 days (45 to 140)"
Funding	Study sponsor was University of Aarhus (ClinicalTrials.gov).
Potential conflicts of interest	None stated.
Notes	Study dates: recruitment 1 June 2005 to 1 March 2007 ClinicalTrials.gov id: NCT00268788 The study was preceded by a prolonged treatment‐free interval of a maximum of 10 weeks to ensure that participants were responsive to immunoglobulins; responsiveness was defined as a decline in strength of at least 10% during this period. Interventions were separated by a wash‐out period during which participants received 2 doses of IVIg.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: “randomized single‐blinded cross‐over study” “the hospital pharmacy assigned the order of treatment using block randomization with a block size of four and delivered the medication in numbered containers.”
Allocation concealment (selection bias)	Low risk	Quote: "The principle investigator consecutively enrolled the patients, whereas the hospital pharmacy assigned the order of treatment using block randomization with a block size of four and delivered the medication in numbered containers. The evaluating physician and clinical neurophysiologist were masked to the actual treatment." Comment: measures taken to ensure concealment of allocation appear to be adequate
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blinded to treatment. Single‐blind; “The evaluating physician and clinical neurophysiologist were masked to the actual treatment”
Blinding of outcome assessment (detection bias) All outcomes	High risk	"The evaluating physician and clinical neurophysiologist were masked to the actual treatment." and "A detailed description of preference for therapy was given by the patients and the study nurse continuously registered adverse effects" Comment: assessors were reported as blinded; however, the procedure, especially whether assessors were blinded for local and systemic adverse events, was not described. The primary outcome was dynamometric muscle strength. Despite the fact that dynamometry is considered to be a fairly objective outcome measure, the single‐blinded design could have introduced potential assessment bias when participants had a treatment preference.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Quote: "Table 1 shows clinical data for the nine randomized either to SCIG followed by IVIG (n, 4) or to IVIG followed by SCIG (n, 5). All completed both the treatments and all evaluations, except one patient (#4) who refused participation in the electrophysiological study. One patient (# 6) did not comply with the scheduled SCIG treatment and infused 1040 ml only of the planned 1600 ml." and "All randomized were included for calculation of the primary end‐point according to the principle of intention to treat, whereas the patient (#6) who did not comply with the SCIG treatment schedule was excluded from the exploratory end‐point analysis."
Selective reporting (reporting bias)	Low risk	All outcomes prespecified in the methods section were reported in the results section.
Other bias	High risk	All participants had to be immunoglobulin responders. As the main goal of this trial was to determine whether SCIg is as effective and safe as IVIg, this recruitment bias seems justified.  A high risk of bias due to the study design. The "washout" period consisted of stabilisation using 2 treatments of IVIg. The subsequent SCIg treatment period would be insufficient to offset effects of the preceding IVIg.

Léger 2001.

Study characteristics
Methods	Double‐blind, partial cross‐over design. At month 4 (after 3 treatments with IVIg or placebo, randomly assigned), all participants had a double‐blind clinical evaluation. Responders remained on therapy for 3 more months. Non‐responders switched to alternative for the following 3 months.
Participants	Inclusion criteria: MMN according to well‐defined criteria (clinically and electrophysiologically as per criteria described in Léger 1994). If newly diagnosed, progressive neuropathy for > 2 months. If established, static, or showing recent deterioration Exclusion criteria: severe concomitant medical problem that might cause neuropathy or interfere with treatment; pregnant; or under 18 No immunosuppressants in the last 2 months or IVIg in last 3 months Number of participants assessed for eligibility: "19 screened" Number randomised: 19; 10 IVIg‐naive participants were eligible for this review: 5 initially received placebo, and 5 received IVIg (1 of whom was lost to follow‐up) In all randomised participants (characteristics not reported separately for IVIg‐naive group): Sex: 13 men, 6 women Age (years): range 40 to 65, mean 54.6 Disease duration (years ± SD): placebo group (n = 9) 8.2 ± 5.6; IVIg group (n = 10) 9.8 ± 8.7 Recruitment: patients selected from clinics Setting: France (Paris)
Interventions	IVIg 0.5 g/kg/day for 5 consecutive days versus placebo (1% human albumin); source IVIg (Endobulin; Baxter, Vienna, Austria) At month 4, amongst the 9 IVIg‐naive participants, 3/5 in the placebo group and 2/4 in the IVIg group were non‐responders, and were crossed over to receive IVIg and placebo, respectively, for the 2nd phase. (Response defined as at least 1 more MRC point in 2 affected muscles plus 1 point less in 2 activities of daily life compared with baseline.) The other participants continued on 1st‐phase interventions.
Outcomes	MRC sum score in 28 muscles Self‐evaluation scale (0 normal to 5 impossible) on 5 activities of daily living Neurophysiological measurements GM1 antibodies Side effects Neurological and GM1 antibody assessment at baseline and every month for 7 months. Primary endpoint was difference in MRC score between baseline and 4 months.
Funding	This study was supported by Assistance Publique‐Hôpitaux de Paris, Baxter AG (Vienna) for logistic and financial support, and by a BIOMED contract (INCAT) of the European Community, Brussels.
Potential conflicts of interest	Employees of Baxter (the pharmaceutical company that provided the study product) assisted in study management and safety assessments, documenting samples, and preparation of study report.
Notes	Study dates: December 1995 to October 1997 Predefined set of rules to decide whether to cross over
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "We conducted a double‐blind, randomized, placebo‐controlled study. The were randomly assigned to receive either IVIg (Endoglobulin, 5% lyophilized; Baxter, Vienna, Austria) at 500 mg/kg/day for 5 consecutive days, or placebo (1% human albumin, lyophilized) once a months for 3 months." Comment: randomisation procedure not described
Allocation concealment (selection bias)	Unclear risk	Quote: "We conducted a double‐blind, randomized, placebo‐controlled study. The were randomly assigned to receive either IVIg (Endoglobulin, 5% lyophilized; Baxter, Vienna, Austria) at 500 mg/kg/day for 5 consecutive days, or placebo (1% human albumin, lyophilized) once a months for 3 months." Comment: measures taken to ensure concealment of allocation are not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "At month 4 (after three treatment periods), all had a double‐blind clinical evaluation." Comment: participants were reported as blinded; however, the procedure was not described in sufficient detail. There were no details on treatment administration and blinding of treatment providers.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "At month 4 (after three treatment periods), all had a double‐blind clinical evaluation." and "All were assessed by the same blinded observers." Comment: assessors were reported as blinded; however, the procedure was not described in sufficient detail
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: “Nine of the 10 patients in Group 1 [not previously treated with IVIg] completed the study, one patient being lost to follow‐up for personal reasons.” A response‐conditional cross‐over design was used, which means that not all participants received both treatments. Data reported after the first period. An intention‐to‐treat analysis is not possible.
Selective reporting (reporting bias)	Low risk	All outcomes prespecified in the methods section were reported in the results section.
Other bias	Unclear risk	Quote: "Those who were found to be responders (according to the criteria defined below) remained on the same treatment for 3 months, and those who were considered to be non‐responders were switched to the alternative study drug for the 3 following months." and "At 4 months, were considered as responders if they had at least 1 more point in two affected muscles plus 1 point less in two activities of daily life compared with baseline." Comment: a response‐conditional, cross‐over design was used, which means that not all participants received both treatments. However, in the review we report results from the first period.

Van den Berg 1995.

Study characteristics
Methods	Double‐blind, placebo‐controlled, cross‐over design
Participants	Inclusion criteria: MMN according to well‐defined criteria (progressive asymmetrical weakness and atrophy without sensory loss, with electrophysiological evidence of motor conduction block) Exclusion criteria: upper motor neuron findings or CIDP Number of participants assessed for eligibility: not stated Number randomised: 6 Sex: 4 men, 2 women Age (years): range 33 to 64, mean 46 Symptom duration: (years) 2 to 8 (mean 6) Recruitment: not documented Setting: the Netherlands (Utrecht). Participants were admitted to hospital for 6 consecutive days during each treatment course.
Interventions	IVIg 0.4 g/kg/day for 5 consecutive days versus placebo (pasteurised plasma solution); source IVIg: Central Laboratory of Blood Transfusion, Amsterdam, the Netherlands 4 participants received 2 IVIg treatments (0.4 g/kg for 5 consecutive days) and 2 placebo treatments in a randomised order. 2 participants received only 1 IVIg treatment and 1 placebo treatment for practical reasons.
Outcomes	Muscle strength (MRC scale) with handheld dynamometer (improvement defined as an increase of 50% or more in at least 2 muscles or muscle groups, without a decrease of at least 25% in more than 1 other muscle or muscle group) Modified Rankin scale Neurophysiological measurements before treatment and on days 6 and 14 after each treatment
Funding	This work was supported by the Dammers Foundation for research into motor neuron disease and the Netherlands Organisation for Scientific Research.
Potential conflicts of interest	None stated.
Notes	Study dates: not given Predefined set of rules to determine length of wash‐out
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The effect of IVIg treatment was studied for each patient in a single patient randomised trial. Four (1‐4) received two IVIg treatments (0 4 g/kg for five consecutive days) and two placebo treatments (pasteurised plasma solution for five consecutive days) in a randomised order." Comment: randomisation procedure not described
Allocation concealment (selection bias)	Unclear risk	Quote: "Treatments were blinded for the patient and the physician by the participating pharmacist." and "The code was revealed by the participating pharmacist after all treatment courses had been completed and all measurements evaluated." Comment: measures taken to ensure concealment of allocation are not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "Treatments were blinded for the patient and the physician by the participating pharmacist." and "The code was revealed by the participating pharmacist after all treatment courses had been completed and all measurements evaluated." Blinding was attempted, but the study design meant that participants were highly likely to know which treatment they were receiving. Quote: “At the end of the four treatment courses, all five patients succeeded in naming the correct order of IVIg or placebo treatments before the code was broken.”
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "Treatments were blinded for the patient and the physician by the participating pharmacist." and "The code was revealed by the participating pharmacist after all treatment courses had been completed and all measurements evaluated." Comment: assessors were reported as blinded, but the procedure was not described. Participant unblinding risks assessor unblinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All outcomes prespecified in the methods section were available for all participants.
Selective reporting (reporting bias)	Low risk	All outcomes prespecified in the methods section were reported in the results section.
Other bias	Unclear risk	Quote: "The interval between each treatment was determined by the time it took for the patient to return clinically to the pretreatment state. To prevent cumulative dose effects, the minimal time interval between two treatment courses was kept at one month." Comment: this predefined rule precludes the possibility of a carry‐over effect Quote: "Open trial. were treated with IVIg (Central Laboratory Blood Transfusion, Amsterdam; 0 4 g/kg for five consecutive days). who responded as defined later entered the double blind placebo controlled trial, which was started when the patient had returned clinically to the pretreatment state." Comment: entered an open phase before the actual blinded trial and had to respond during this open phase. This may have led to recruitment bias; however, all participants entering the trial were also included in the blinded study phase. The published report did not say whether additional other (immunomodulatory) treatments were permitted.

ALS: amyotrophic lateral sclerosis
CIDP: chronic inflammatory demyelinating polyradiculoneuropathy
EFNS/PNS: European Federation of Neurological Societies/Peripheral Nerve Society
GM1: ganglioside‐monosialic acid
IgG: immunoglobulin G
i.v.: intravenously
IVIg: intravenous immunoglobulin
MMN: multifocal motor neuropathy
MRC sum score: Medical Research Council sum score
NDS: Neurological Disability Scale
SCIg: subcutaneous immunoglobulin
SD: standard deviation
 

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Al‐Zuhairy 2019	Non‐inferiority cross‐over RCT comparing IVIg brands ‐ comparison not eligible
Auer 1994	Letter to editor with comment, no IVIg treatment
Baumann 2009	A non‐randomised, non‐controlled case series
Burrell 2011	Non‐randomised prospective study
Cats 2008	Open‐label, non‐controlled, non‐randomised prospective study
Cats 2010	Non‐randomised case series
Charles 1992	Case report, no randomisation procedure
Eftimov 2009	Open‐label, non‐controlled, non‐randomised pilot study
Herraets 2019	Non‐randomised, non‐controlled open‐label study in which all patients with IVIg are switched to facilitated SCIg
Jafari 2000	Case report, no randomisation procedure
Kelly 1992	Letter to editor with comment, no IVIg treatment
Kermode 1992	Case series (n = 5), no randomisation procedure
Köller 2006	Case report, no randomisation procedure
Komiyama 1998	Case report, no randomisation procedure, no IVIg treatment
Kubori 1999	Not properly randomised: lacking a placebo/control group. Participants were divided depending on clinical severity into 3 groups and treated with different dosages of IVIg.
Kuwabara 2018	Phase III clinical trial, lacking a control group
Léger 2008	Retrospective case series, non‐controlled, non‐randomised
Léger 2019	Non‐inferiority cross‐over RCT comparing IVIg brands ‐ comparison not eligible
Pestronk 1995	Letter to editor with comment, no IVIg treatment
Tan 1994	No randomisation procedure, no IVIg treatment
Turnbull 2001	Letter to editor with comment, no IVIg treatment
Van den Berg‐Vos 2002	No randomisation procedure
Yuki 1993	Case report, no randomisation procedure

IVIg: intravenous immunoglobulin; RCT: randomised controlled trial; SCIg: subcutaneous immunoglobulin

Differences between protocol and review

In the protocol, Van Schaik 2003, and first meta‐analysis, Van Schaik 2005, we focused on improvement in unselected participants starting with intravenous immunoglobulin (IVIg) treatment.

In this update we included both IVIg as well as subcutaneous immunoglobulin (SCIg) treatment, and revised the search strategy accordingly. We made a distinction between studies including participants in whom IVIg or SCIg treatment was being initiated (immunoglobulin‐naive participants) and studies including participants who were receiving maintenance immunoglobulin therapy at study entry (IVIg or SCIg responders). These studies have different objectives. Studies that initiate IVIg/SCIg treatment aim to achieve improvement in disability and muscle strength. Alternatively, studies including participants on maintenance treatment explore the ongoing need for IVIg/SCIg treatment or alternative treatments to substitute maintenance IVIg/SCIg treatment in preserving functional status and muscle strength. As the latter objective is equally important in a chronic disease such as multifocal motor neuropathy, we added secondary outcomes to enable data analysis from recent and future trials. We added three secondary outcomes for studies of ongoing immunoglobulin maintenance treatment and studies exploring alternative treatments for maintenance immunoglobulin treatment.

In the current update we used the Cochrane risk of bias tool (Higgins 2011), rather than the quality assessment procedure described in the protocol (Van Schaik 2003).

We included summary of findings tables in the current update.

We changed the time to respond for the primary outcome from between two and four weeks to between two and six weeks.

Previously outcomes were defined as 'significant' improvement in disability and muscle strength as defined by trial authors. We removed 'significant' from the outcome definition, as the term was not generally used in defining outcomes in trials and is liable to misinterpretation.

To comply with current Cochrane MECIR standards, we also (Higgins 2020):

• reported the role of outcomes in study selection;

• searched two trials registries;

• included a PRISMA flow chart to illustrate study selection;

• documented the process for data entry and checking and a planned approach for studies requiring translation;

• extracted additional study characteristics data, e.g. funding sources and conflicts of interest for the included studies;

• described a planned approach to trials with multiple treatment arms;

• described methods for measurement of heterogeneity.

Contributions of authors

All original authors were involved in all aspects of the development of the protocol. INvS and LvdB assessed the trials for the original review and extracted data. RdH, INvS, and FE calculated the effect sizes and performed the meta‐analysis.

INvS wrote the original review.

SK and RB extracted outcome data from trials for this update.

SK extracted additional characteristics of studies data, which RB spot checked and entered into Review Manager 5.

FE wrote the update, with additional drafting by RB and SK.

All authors criticised and amended the draft.

Sources of support

Internal sources

• Academic Medical Center, University of Amsterdam, Department of Neurology, Netherlands

  FE, INvS

• Academic Medical Center, University of Amsterdam, Department of Biostatistics and Clinical Epidemiology, Netherlands

  RdH

• University Medical Center Utrecht, Department of Neurology, Utrecht, Netherlands

  LvdB

External sources

• NIHR Infrastructure award to Cochrane Neuromuscular, UK

  RB

Declarations of interest

Stephen Keddie: is employed by Barts and the Royal London NHS Trust and University College London Hospitals NHS trust bank for clinical neurology work. Dr Keddie's research is funded by a clinical training research fellowship awarded by the Association of British Neurologists and Guarantors of Brain. Dr Keddie is a neurology specialist registrar and manages patients with multifocal motor neuropathy.

Filip Eftimov: Dr Eftimov received support for printing of his thesis from Sanquin Bloedvoorziening (Dutch blood bank and IVIg manufacturer) and CSL‐Behring in 2014. He reports lecture fees from CSL‐Behring, Kedrion, and Grifols. Outside of the submitted work, as principal investigator of INCbase, he also reports investigator‐initiated grants from Kedrion, Terumo BCT, CSL‐Behring, and Takeda Pharmaceutical Company, and grants from ZonMw (Dutch governmental agency) and Prinses Beatrix Spierfonds (a Dutch charity). In addition, his institution has received fees from UCB Pharma, CSL‐Behring, and Takeda for advisory board membership. All grants and fees were paid to his institution. He is a member of the Cochrane Neuromuscular Editorial Board.

Leonard H van den Berg: has co‐ordinated one randomised controlled trial of IVIg in multifocal motor neuropathy, which is included in this review. He received remuneration for sitting on scientific advisory boards of Baxter and Biogen Idec. His institution has received financial support from Baxter for organisation of an multifocal motor neuropathy masterclass.

Ruth Brassington: none known. She is Managing Editor of Cochrane Neuromuscular. She relinquished editorial responsibility for the review after joining the author team.

Rob J de Haan: none known.

Ivo N van Schaik: chaired a steering committee for a CSL‐Behring study investigating the safety and efficacy of SCIg in chronic inflammatory demyelinating polyradiculoneuropathy and received departmental honoraria for serving on scientific advisory boards for CSL‐Behring and Kedrion. He received departmental research support from the Netherlands Organisation for Scientific Research and the Dutch Prinses Beatrix Fonds. All lecturing and consulting fees for INvS were donated to the Stichting Klinische Neurologie, a local foundation that supports research in the field of neurological disorders. INvS served on the editorial board of Cochrane Neuromuscular, was a member of the organising committee of the Inflammatory Neuropathy Consortium (INC), a standing committee of the Peripheral Nerve Society, and was a member of the Scientific Board of the Kreuth III meeting on the optimal use of plasma‐derived medicinal products, especially coagulation factors and normal immunoglobulins, organised under the auspices of the European Directorate for the Quality of Medicines & HealthCare (EDQM).

This review does not comply with the relevant Cochrane Conflicts of Interest policy, revised in 2014, which requires the majority of review authors to have no relevant conflicts of interest. The review will be updated a year after publication, when the review will comply with the policy.

New search for studies and content updated (conclusions changed)