Article Text

Download PDFPDF

Intravenous immunoglobulin as short- and long-term therapy of multifocal motor neuropathy: a retrospective study of response to IVIg and of its predictive criteria in 40 patients
  1. J-M Léger1,
  2. K Viala1,
  3. F Cancalon2,
  4. T Maisonobe1,
  5. B Gruwez3,
  6. T Waegemans2,
  7. P Bouche1
  1. 1
    Department of Neurology, Hôpital de la Salpêtriere, Paris, France
  2. 2
    Laboratoire Français du Fractionnement et des Biotechnologies (LFB), Les Ulis, France
  3. 3
    Department of Pharmacology, Hôpital Cochin, Paris, France
  1. Prof Jean-Marc Léger, Centre de Référence Maladies Neuromusculaires rares, Bâtiment Babinski, Hôpital de la Salpêtriere, 47 boulevard de l’Hôpital, 75651 Paris Cedex 13, France; jean-marc.leger{at}


Objective: To document short-term and long-term responses to a single type of intravenous immunoglobulin (IVIg) in a large cohort of patients with multifocal motor neuropathy (MMN).

Methods: A retrospective study was conducted in 40 patients with MMN included on ENMC Workshop criteria, and treated with periodic IVIg infusions between 1995 and 2003. The short-term response was defined as improvement of at least 1 point on the MRC score in at least two affected muscles at 6 months. The population comprised 22 treatment-naïve patients (who had never received IVIg before inclusion), and 18 previously treated patients. For the long-term evaluation (>6 months), the patients were classified into three groups according to the dependency or not on periodic IVIg. In addition, changes in conduction block (CB) and predictive criteria for response to IVIg were explored.

Results: The MRC score significantly improved (p<0.0001) in 14 (70%; 95% CI 0.46 to 0.88) of the 20 treatment-naïve patients (missing data for 2 patients). None of the predictive criteria studied were found to be significant. At the end of follow-up (mean of 2.2±2.0 years), only 8 of the 40 patients (22%) had significant remission, whereas 25 patients (68%) were dependent on periodic IVIg infusions. The number of CBs decreased or remained unchanged in 12 treatment-naïve patients and increased in 2 such patients.

Conclusions: This study confirmed a significantly high short-term response to IVIg of patients with MMN, but showed contrasted results in long-term follow-up. No predictive factors for response to IVIg were found.

  • multifocal motor neuropathy
  • conduction block
  • intravenous immunoglobulin
  • treatment naïve
  • clinical study

Statistics from

Multifocal motor neuropathy (MMN) was differentiated from other motor neuropathies 20 years ago,1 2 thus becoming a well-defined condition.3 4 In the past 4 years, diagnostic criteria and guidelines on the management of MMN have been edited by European and American experts.57 Several uncontrolled810 and randomised controlled trials (RCTs)1115 have demonstrated the effectiveness of high-dose intravenous immunoglobulins (IVIg) in MMN. Since these reports, selected outcome measures have been proposed16 and evaluated.17 In addition, these RCTs did not take into account the marked variation of the response to treatment or the clinical progression of the disease over time.18 Only a few studies have addressed the issue of long-term effectiveness of IVIg in MMN, the results of which are controversial.19 Moreover, exploration of potential factors predicting a good response to IVIg has rarely been performed.20



We conducted a retrospective study in patients with defined MMN and treated at least once with one type of IVIg (Tégéline®; LFB, Les Ulis, France) between 1 January 1995 and 15 February 2003. Patients were identified through our database, and informed by mail that their data would be collected and of their right to refuse. We used a slightly modified version of the ENMC diagnostic criteria5 for MMN (except for upper age limit: 80 years and CSF analysis), to include the patients. All patients were treated with monthly IVIg at the dose of 2 g/kg over 3–5 days for at least 6 months, followed by maintenance infusions (at the same dose), if needed by recurrent clinical worsening.

Short-term assessment

To test for bias due to the effects of previous treatment, we divided the 40 patients into 22 patients who had not previously been treated, and 18 patients who had previously received IVIg and/or another immunomodulator, before the studied period. Treatment response at 6 months among the 22 treatment-naïve patients was defined as an improvement of at least one point on the MRC score (measured in the eight most affected muscles), compared with the baseline score, in at least two muscles corresponding to affected nerves. For comparison purposes, we used data from a randomised, placebo-controlled trial to form a historical placebo group.14 In a second analysis, patients in whom the number of muscles with deterioration (defined as loss of at least 1 point on the MRC score) exceeded those with improvement, were considered to be non-responders. We also compared the overall number of conduction blocks (CBs) before treatment and at the last available electromyographic examination under therapy.

Predictive factors

Age at diagnosis, gender, prominence of clinical involvement in the upper/lower limbs, normal/abolished tendon reflexes, initial MRC sum-score, serum anti-GM1 IgM antibody titres, delay between first symptoms and first course of IVIg, and electrophysiological features were analysed as predictive factors for clinical response in the only treatment-naïve population.

Long-term assessment

To evaluate the long-term efficacy (>6 months), all 40 patients were studied and divided into three groups at the end of clinical follow-up. The groups were the following: group 1  =  remission defined as lasting stabilisation of clinical improvement (>6 months), without further treatment, after initial IVIg therapy during at least 6 months; group 2  =  stabilisation of clinical improvement dependent on maintenance IVIg infusions (2a  =  without additional immunosuppressive agent; 2b  =  with immunosuppressive agent); group 3  =  non-responders.

Statistical analysis

Data are reported as numbers and percentages, median (range) or mean (SD) where appropriate. The responder rate at 6 months was calculated using the last-observation-carried-forward (LOCF) method for all patients who did not complete the six treatment courses. Response rates compared with the 20% historical placebo rate were expressed as medians with 95% confidence intervals (CIs). A CI with a lower limit greater than 20% was considered to be therapeutically significant. To assess changes in the MRC score between baseline and 6 months, the absolute difference in the mean continuous MRC score was compared using the null test. A p-value <0.05 was considered statistically significant. For analysis of predictive criteria, a logistic regression using backward elimination was used. Analyses were performed with SAS 8.2 software (SAS Institute, Cary, NC, USA).


Patient characteristics at baseline in the 40 patients studied are shown in table 1.

Table 1 Baseline patient characteristics

For the primary analysis, among the 22 treatment-naïve patients, a comparison between MRC scores at inclusion and at 6 months was available for 20 patients (data missing for 2 patients). Fourteen of these 20 patients (70%; 95% CI 0.46 to 0.88) were responders at 6 months and 6 (30%) were non-responders. The 46% lower CI for treatment response was above the 20% responder rate in the historical placebo group (p<0.0001). In a further analysis, in which 1 out of 22 patients had deterioration in three muscles and improvement in two muscles, and was therefore considered to be a non-responder, the response rate was 59.1% (95% CI 0.36 to 0.79).

Among the 22 treatment-naïve patients, with a mean follow-up of 7.1 months, the number of CBs decreased for 8 patients, with complete disappearance of CB for 2 patients, remained stable for 4 patients and increased for 2 patients. This measure was not assessable for the last 8 patients.

Analysis of predictive criteria revealed that the best two predictive factors (although not significant) were female gender (p = 0.08) and lower MRC score at inclusion (p = 0.07).

At the end of follow-up (mean of 2.2±2.0 years) and among the entire population (40 patients), 8 patients (22%) were in group 1 and 25 patients (68%) were in group 2—that is, requiring periodic IVIg infusions to maintain good clinical condition. Among these 25 patients, 8 (46%) were given additional immunosuppressive agents during various periods. Four patients (11%) were in group 3. Data were missing for 3 patients.

The variation of the MRC score in 6 treatment-naïve patients, for whom data were available over a period greater than 12 months, is shown in figure 1. All were responders, but the MRC score at last-follow-up was lower than after the first IVIg infusion in 4 patients and higher in patients 4 and 13.

Figure 1 Validation of the MRC score in six treatment-naive patients over a period greater than 12 months.

Adverse effects

In total, there were 150 adverse effects observed in 32 patients. They were in line with those previously described14 21 and only caused minor inconvenience.


We first found that the MRC score improved in 70% of the 20 treatment-naïve patients with MMN in our cohort, retrospectively studied over a 6-month period. These results are in line with previous reports of 60–80% short-term efficacy of IVIg,15 and those of a comparable retrospective study conducted in 34 patients treated with a single course of IVIg.22 However, our patients additionally had criteria established by a consensus of experts,5 which have been recently both revised and validated,7 and a longer follow-up period. The first point seems to be crucial, as some authors20 have clearly shown that the response rate was different depending on the set of criteria selected.

A well-recognised impairment outcome measure, the MRC score, was used for short-term assessment. Different outcome measures have recently been proposed for peripheral neuropathy clinical trials,16 but only the INCAT Overall Disability Sum Score (ODSS) has been shown to correlate well with patients’ own clinical judgment in immune-mediated neuropathies.17 Moreover, as MMN has an asymmetrical clinical pattern, an alternative sum score should be proposed covering at least 8 more affected muscles.16 Nevertheless, until there is a consensus on a simple, reliable, validated functional scale to assess disability in patients with MMN, comparison between studies will remain difficult. Previous studies have documented variable effects on CBs, as in our study. Some reported resolution and improvement in CBs, whereas others report no significant improvement.814 Consequently, the measure of CBs does not seem to be highly relevant in the assessment of patients with MMN for therapeutic trials.16

Only one retrospective study disclosed significant results for predictive criteria of response to IVIg, mainly age at onset and number of affected limb regions.20 Our results do not confirm these data, and it therefore seems difficult to predict a good response to IVIg in patients with MMN.

The data concerning the long-term response to IVIg showed marked variations in terms of MRC score, disease duration and IVIg infusion regimens, as previously described.18 Striking differences in the IVIg regimens used in the few published studies make inter-study comparisons difficult. Two comparable retrospective studies found slight motor deterioration on, respectively, 4–8 years,21 and mean 8.2 years follow-up,23 in patients treated with periodic infusions at clinical worsening, whereas another study demonstrated sustained clinical improvement in mean 7.25-years’ follow-up in patients treated with monthly infusions.24 Consequently, it seemed to us more relevant to address the question of dependency on long-term IVIg. Only 22% of patients in our cohort were able to stop IVIg therapy after a few months or years without clinical worsening, and the majority needed periodic IVIg infusions to maintain their clinical motor condition. Consequently, MMN should merit further prospective studies for better knowledge of its natural history and of the potential beneficial effects of additional long-term immunomodulating therapy.


The authors would like to thank S Mellac and D Simon for having monitored this study, and D Magrez and E Guemas for the statistical analyses.


View Abstract


  • Competing interests: François Cancalon and Tony Waegemans work in the Research and Development Department of LFB.

Request Permissions

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